JackSmith077 commited on
Commit
d444338
·
verified ·
1 Parent(s): fbdc93e

Limit G055_inner_court_investigation PR to assets/3d_object

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +0 -16
  2. outputs/G055_inner_court_investigation/.generation/3d_object/meshy/obj200_meshy_tripo_20261015_v2/asset_task_index.json +0 -111
  3. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/2fe8790006de37b67087226226b42b450fd22be3f3c30f7b8d0ddbaf50246d62.meta.json +0 -33
  4. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/440b4733d80d11bc6c02e030f10f3e14a40728326f88b899eb77291c6944107d.meta.json +0 -32
  5. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/8fc448662209e1e2fdf9c5ebb0e8e8a012c7533282406ec50f103a9b5b99f7b6.meta.json +0 -33
  6. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/c4784de99ad02419f6e080c08b3388e1e8700ed26b6d9e46e7dbb72cafa7311f.meta.json +0 -0
  7. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/072537beaa31697bd206f8b13b746d96ac409e06dffd39e613fddaa8e2f755f0.meta.json +0 -30
  8. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/2ce49f964f8b611e965e5f39d37edc4b84a43acb9238f0caaabd9215699fbabf.meta.json +0 -30
  9. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/c0f6d2616b8bb8de62f37adee4acaa3039af81bbe0a507b57e9e150aaa700dad.meta.json +0 -30
  10. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/ecb7e508f9dc367e4d5c4b91e9c7b37b68dbb47a2d9173ce1b5c30f01813d8a9.meta.json +0 -30
  11. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/frozen_game_inputs.json +0 -222
  12. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/general_requirement.txt +0 -821
  13. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/method_selection.md +0 -152
  14. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/object_requirement.txt +0 -9
  15. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/localization_integrity.json +0 -211
  16. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/frozen_parent_plan_original.zip +0 -3
  17. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/records_before_localization_35f1ef02271bc9d9.zip +0 -3
  18. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/records_before_localization_967e45174af30b60.zip +0 -3
  19. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_001.json +0 -56
  20. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_002.json +0 -56
  21. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_003.json +0 -88
  22. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_004.json +0 -88
  23. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_001.json +0 -72
  24. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_002.json +0 -72
  25. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_003.json +0 -54
  26. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_004.json +0 -54
  27. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png +0 -3
  28. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png +0 -3
  29. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/cloud_api.py +0 -557
  30. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/glb_utils.py +0 -578
  31. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/meshy_model.py +0 -463
  32. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/tripo_model.py +0 -532
  33. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/operators/gen_3d_object/operator.py +0 -596
  34. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/paths.py +0 -444
  35. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/record_paths.py +0 -77
  36. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/prompts/stage2_assets.md +0 -95
  37. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/cloud_journal.py +0 -178
  38. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/generate_3d_comparison.py +0 -603
  39. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/package-lock.json +0 -763
  40. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/process_mesh.mjs +0 -77
  41. outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/sources_index.json +0 -178
  42. outputs/G055_inner_court_investigation/.generation/3d_object/tripo/obj200_meshy_tripo_20261015_v2/asset_task_index.json +0 -111
  43. outputs/G055_inner_court_investigation/assets/3d_object/.gitattributes +16 -0
  44. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/contact_sheet.jpg +0 -3
  45. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/iso.png +0 -3
  46. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/neg_x.png +0 -3
  47. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/neg_z.png +0 -3
  48. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/pos_x.png +0 -3
  49. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/pos_y.png +0 -3
  50. outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/pos_z.png +0 -3
.gitattributes CHANGED
@@ -58,19 +58,3 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_001/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
62
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_001/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
63
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_001/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
64
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_001/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
65
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_002/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
66
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_002/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
67
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_002/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
68
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_002/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
69
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_003/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
70
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_003/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
71
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_003/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
72
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_003/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
73
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_004/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
74
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_004/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
75
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_004/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
76
- outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_004/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
 
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/meshy/obj200_meshy_tripo_20261015_v2/asset_task_index.json DELETED
@@ -1,111 +0,0 @@
1
- {
2
- "plan_sha256": "5c287b8399f00fbb0dddf94c9e3f5ffcbcc2e679ff76882bc8fc369d6eae840d",
3
- "game_id": "G055_inner_court_investigation",
4
- "provider": "meshy",
5
- "scope": "3d_object_only_not_complete_stage2",
6
- "tasks": [
7
- {
8
- "task": {
9
- "game_id": "G055_inner_court_investigation",
10
- "task_id": "G055_inner_court_investigation_obj_001",
11
- "asset_name": "shen_zhiwei",
12
- "name_zh": "沈知微",
13
- "role": "player",
14
- "input_mode": "image_to_3d",
15
- "prompt": "Original young court investigator Shen Zhiwei: understated dark-teal commoner robe with faint embroidery, hair in a plain bun with wooden pin, cloth shoes, sharp calm gaze, stylized semi-realistic ancient court style, vermilion-and-indigo palette, T-pose, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
16
- "seed": 5501,
17
- "decimation_target": 8000,
18
- "texture_size": 2048,
19
- "rig": "humanoid",
20
- "animation_strategy": "skeletal",
21
- "spec_ref": "P01,W06",
22
- "image_path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
23
- "poly_policy": "low_poly_priority",
24
- "l_tier_repair_target": true
25
- },
26
- "shared_prompt": null,
27
- "reference": {
28
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
29
- "sha256": "ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55",
30
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png"
31
- },
32
- "subject_group": "character",
33
- "context_condition": "reference_only"
34
- },
35
- {
36
- "task": {
37
- "game_id": "G055_inner_court_investigation",
38
- "task_id": "G055_inner_court_investigation_obj_002",
39
- "asset_name": "evidence_pouch_ash",
40
- "name_zh": "旧香囊残灰",
41
- "role": "evidence",
42
- "input_mode": "image_to_3d",
43
- "prompt": "Small scorched silk incense pouch with charred rim and faint residual ash, tied with a cut thread, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
44
- "seed": 1552,
45
- "decimation_target": 4000,
46
- "texture_size": 1024,
47
- "rig": "none",
48
- "animation_strategy": "procedural",
49
- "spec_ref": "P20.E01",
50
- "image_path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
51
- "reference_image_role": "single_object_visual_anchor",
52
- "poly_policy": "low_poly_priority",
53
- "l_tier_repair_target": true
54
- },
55
- "shared_prompt": null,
56
- "reference": {
57
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
58
- "sha256": "b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d",
59
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png"
60
- },
61
- "subject_group": "rigid",
62
- "context_condition": "reference_only"
63
- },
64
- {
65
- "task": {
66
- "game_id": "G055_inner_court_investigation",
67
- "task_id": "G055_inner_court_investigation_obj_003",
68
- "asset_name": "rota_manuscript",
69
- "name_zh": "排班抄本",
70
- "role": "evidence",
71
- "input_mode": "text_to_3d",
72
- "prompt": "Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
73
- "seed": 1553,
74
- "decimation_target": 5000,
75
- "texture_size": 1024,
76
- "rig": "none",
77
- "animation_strategy": "procedural",
78
- "spec_ref": "P20.E03",
79
- "poly_policy": "low_poly_priority",
80
- "l_tier_repair_target": true
81
- },
82
- "shared_prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、��令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
83
- "reference": null,
84
- "subject_group": "rigid",
85
- "context_condition": "shared_short_context"
86
- },
87
- {
88
- "task": {
89
- "game_id": "G055_inner_court_investigation",
90
- "task_id": "G055_inner_court_investigation_obj_004",
91
- "asset_name": "seal_wax_imprint",
92
- "name_zh": "火漆拓印",
93
- "role": "evidence",
94
- "input_mode": "text_to_3d",
95
- "prompt": "Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
96
- "seed": 1554,
97
- "decimation_target": 4000,
98
- "texture_size": 1024,
99
- "rig": "none",
100
- "animation_strategy": "procedural",
101
- "spec_ref": "P20.E06",
102
- "poly_policy": "low_poly_priority",
103
- "l_tier_repair_target": true
104
- },
105
- "shared_prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
106
- "reference": null,
107
- "subject_group": "rigid",
108
- "context_condition": "shared_short_context"
109
- }
110
- ]
111
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/2fe8790006de37b67087226226b42b450fd22be3f3c30f7b8d0ddbaf50246d62.meta.json DELETED
@@ -1,33 +0,0 @@
1
- {
2
- "provider": "meshy",
3
- "payload": {
4
- "ai_model": "meshy-6",
5
- "model_type": "standard",
6
- "topology": "triangle",
7
- "target_polycount": 5000,
8
- "should_remesh": true,
9
- "mode": "preview",
10
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
11
- "seed": 1553
12
- },
13
- "model": "meshy-6",
14
- "task_id": "01a1430a-f6b4-747f-a843-36a9fb07347c",
15
- "task_ids": [
16
- "01a14309-c130-7684-ae32-797c57e1dd4b",
17
- "01a1430a-f6b4-747f-a843-36a9fb07347c"
18
- ],
19
- "elapsed_sec": 185.3,
20
- "credits": null,
21
- "cached": false,
22
- "format": "glb",
23
- "preview_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/preview.png?Expires=4945708800&Signature=Q5TlMTiuz4nC9lIV5eo3Iw7Ynk3ep0au8qg-GPkaYvxx378pwu5I8z1ofbJoGjnBycSeBytb-zPselBJcl9A4~dRDDP9wprNW96EIp~xd-dwaoBJiVvv4slTXz2DkiCv-T0NY1CefkTsor3Exv5uJ59WOLHQ8AJQs4pImAc5nRlTUjsPQh1ZcaUET4a4xGX4EnSgGRBafvzIVrVGmvXJAZInML4Y4BnwQlYsHrUAifFE2f6goNfNajqLDuHkOWGVCFJf8snYTVXJ-oARhDJlrUuO-xnfELJKSYPOrbw0MJLf3r9k6z2rPupzJpqeFsKxrn4v~l5YZPkJwnowuAZEoQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
24
- "bytes": 9093236,
25
- "triangles": 4584,
26
- "meshes": 1,
27
- "materials": 1,
28
- "textures": 4,
29
- "images": 4,
30
- "nodes": 1,
31
- "generator": "pygltflib@v1.16.5",
32
- "cache_key": "2fe8790006de37b67087226226b42b450fd22be3f3c30f7b8d0ddbaf50246d62"
33
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/440b4733d80d11bc6c02e030f10f3e14a40728326f88b899eb77291c6944107d.meta.json DELETED
@@ -1,32 +0,0 @@
1
- {
2
- "provider": "meshy",
3
- "payload": {
4
- "ai_model": "meshy-6",
5
- "model_type": "standard",
6
- "topology": "triangle",
7
- "target_polycount": 4000,
8
- "should_remesh": true,
9
- "should_texture": true,
10
- "enable_pbr": true,
11
- "image_url": "data:image/png;base64,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"
12
- },
13
- "model": "meshy-6",
14
- "task_id": "01a14309-545d-72a9-bc79-327f44ab8d13",
15
- "task_ids": [
16
- "01a14309-545d-72a9-bc79-327f44ab8d13"
17
- ],
18
- "elapsed_sec": 169.77,
19
- "credits": null,
20
- "cached": false,
21
- "format": "glb",
22
- "preview_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/preview.png?Expires=1792385413&Signature=jMWVrIkhbtIEjx~kBiuMo2YxpqdR91XA1glI~3O95za1sYO7VUOVfUbULMHfJlcU18Krtykton8wYOFbKUN7TcH-32RkA7VWyF2Sg-mv7bm2iUYPNkRtw6rXu0TVxrhZ8Y3hnrsLS7h2tUDH8gIwQGQam3iv9LBW1tsYRGTKnbbW6CWf13rXJPicYJqI9n7NeYCoIreCrXcXzHec1WNNSOXmmeMPhRL1oQukn7mGQtJ8GohFLvinIFWxIuPphNWVxZOv-YXGuzsjThRATPVyCICTnPiolTNCZF5Xra245ssmtblQGQKslpQg7LkACoAgqdotacCtgJ5gyx6YybJitA__&Key-Pair-Id=K1VGYTHIYLM9UM",
23
- "bytes": 7855852,
24
- "triangles": 3100,
25
- "meshes": 1,
26
- "materials": 1,
27
- "textures": 4,
28
- "images": 4,
29
- "nodes": 1,
30
- "generator": "Khronos glTF Blender I/O v4.4.55",
31
- "cache_key": "440b4733d80d11bc6c02e030f10f3e14a40728326f88b899eb77291c6944107d"
32
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/8fc448662209e1e2fdf9c5ebb0e8e8a012c7533282406ec50f103a9b5b99f7b6.meta.json DELETED
@@ -1,33 +0,0 @@
1
- {
2
- "provider": "meshy",
3
- "payload": {
4
- "ai_model": "meshy-6",
5
- "model_type": "standard",
6
- "topology": "triangle",
7
- "target_polycount": 4000,
8
- "should_remesh": true,
9
- "mode": "preview",
10
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
11
- "seed": 1554
12
- },
13
- "model": "meshy-6",
14
- "task_id": "01a1430b-df46-71b4-9dbf-41a42f98f5be",
15
- "task_ids": [
16
- "01a1430a-442d-768f-90d5-a63534dfd811",
17
- "01a1430b-df46-71b4-9dbf-41a42f98f5be"
18
- ],
19
- "elapsed_sec": 223.22,
20
- "credits": null,
21
- "cached": false,
22
- "format": "glb",
23
- "preview_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/preview.png?Expires=4945708800&Signature=V7dBr~D1mpC9OpiQk-3bdFqBSVPrBNRo7AIzc11fHefYdYEz57nOpRmSVZL-04znHaNaQmyGxDBeNwdCWHHcneuB0iorcOSQZeX1BAAcLFIlM7~AZ7ixbCzUbIVaGeRwHAz~AWl0JzjnPUqvY-1UN1r6wTGzQSOcfLromVilsfRPiYBvKha3RVAL7yYwNgLGbhgoJe54UC4gAJDBhTQcZWYdS8j3Wxg~d-OOin-AwxBQ386aKGLqqlDNbnakUGVR6gMNlMVQ1JVRXTsIdfCL-wqy-SyZAGPXIAZrEUrFhg39jnLPppzBagkzit-Qwo2NHAcM7sXIMRGTYzGEI9c-aQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
24
- "bytes": 7683864,
25
- "triangles": 4007,
26
- "meshes": 1,
27
- "materials": 1,
28
- "textures": 4,
29
- "images": 4,
30
- "nodes": 1,
31
- "generator": "pygltflib@v1.16.5",
32
- "cache_key": "8fc448662209e1e2fdf9c5ebb0e8e8a012c7533282406ec50f103a9b5b99f7b6"
33
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/c4784de99ad02419f6e080c08b3388e1e8700ed26b6d9e46e7dbb72cafa7311f.meta.json DELETED
The diff for this file is too large to render. See raw diff
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/072537beaa31697bd206f8b13b746d96ac409e06dffd39e613fddaa8e2f755f0.meta.json DELETED
@@ -1,30 +0,0 @@
1
- {
2
- "provider": "tripo",
3
- "payload": {
4
- "model": "v3.1-20260211",
5
- "texture": true,
6
- "pbr": true,
7
- "texture_quality": "standard",
8
- "geometry_quality": "standard",
9
- "smart_low_poly": true,
10
- "face_limit": 4000,
11
- "model_seed": 1552,
12
- "texture_seed": 1552,
13
- "input": "file_c4378b18-bced-45a0-a2b6-f186bc0578e7"
14
- },
15
- "model": "v3.1-20260211",
16
- "task_id": "51fc3810-e559-4357-8674-9f24c13383dc",
17
- "elapsed_sec": 218.19,
18
- "credits": 40.0,
19
- "cached": false,
20
- "preview_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/51fc3810-e559-4357-8674-9f24c13383dc/legacy_mesh.webp?auth_key=1792227420-bdngxogf-0-f36c118791ad76fe1084b93cc391f5e3",
21
- "bytes": 2113876,
22
- "triangles": 7770,
23
- "meshes": 1,
24
- "materials": 1,
25
- "textures": 3,
26
- "images": 3,
27
- "nodes": 1,
28
- "generator": "https://tripo3d.ai",
29
- "cache_key": "072537beaa31697bd206f8b13b746d96ac409e06dffd39e613fddaa8e2f755f0"
30
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/2ce49f964f8b611e965e5f39d37edc4b84a43acb9238f0caaabd9215699fbabf.meta.json DELETED
@@ -1,30 +0,0 @@
1
- {
2
- "provider": "tripo",
3
- "payload": {
4
- "model": "v3.1-20260211",
5
- "texture": true,
6
- "pbr": true,
7
- "texture_quality": "standard",
8
- "geometry_quality": "standard",
9
- "smart_low_poly": true,
10
- "face_limit": 4000,
11
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
12
- "model_seed": 1554,
13
- "texture_seed": 1554
14
- },
15
- "model": "v3.1-20260211",
16
- "task_id": "f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281",
17
- "elapsed_sec": 204.89,
18
- "credits": 30.0,
19
- "cached": false,
20
- "preview_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281/legacy_mesh.webp?auth_key=1792227624-toqzmsks-0-e5c3ed58fcb6bc157d98452f5c922142",
21
- "bytes": 2980112,
22
- "triangles": 5727,
23
- "meshes": 1,
24
- "materials": 1,
25
- "textures": 3,
26
- "images": 3,
27
- "nodes": 1,
28
- "generator": "https://tripo3d.ai",
29
- "cache_key": "2ce49f964f8b611e965e5f39d37edc4b84a43acb9238f0caaabd9215699fbabf"
30
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/c0f6d2616b8bb8de62f37adee4acaa3039af81bbe0a507b57e9e150aaa700dad.meta.json DELETED
@@ -1,30 +0,0 @@
1
- {
2
- "provider": "tripo",
3
- "payload": {
4
- "model": "v3.1-20260211",
5
- "texture": true,
6
- "pbr": true,
7
- "texture_quality": "standard",
8
- "geometry_quality": "standard",
9
- "smart_low_poly": true,
10
- "face_limit": 5000,
11
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
12
- "model_seed": 1553,
13
- "texture_seed": 1553
14
- },
15
- "model": "v3.1-20260211",
16
- "task_id": "e4c923ad-3363-4403-9067-bcb9ebb5b3f0",
17
- "elapsed_sec": 209.18,
18
- "credits": 30.0,
19
- "cached": false,
20
- "preview_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/e4c923ad-3363-4403-9067-bcb9ebb5b3f0/legacy_mesh.webp?auth_key=1792227597-ugaktvjq-0-c50629aab104ee767233ce8dd04c1651",
21
- "bytes": 3300940,
22
- "triangles": 6211,
23
- "meshes": 1,
24
- "materials": 1,
25
- "textures": 3,
26
- "images": 3,
27
- "nodes": 1,
28
- "generator": "https://tripo3d.ai",
29
- "cache_key": "c0f6d2616b8bb8de62f37adee4acaa3039af81bbe0a507b57e9e150aaa700dad"
30
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/ecb7e508f9dc367e4d5c4b91e9c7b37b68dbb47a2d9173ce1b5c30f01813d8a9.meta.json DELETED
@@ -1,30 +0,0 @@
1
- {
2
- "provider": "tripo",
3
- "payload": {
4
- "model": "v3.1-20260211",
5
- "texture": true,
6
- "pbr": true,
7
- "texture_quality": "standard",
8
- "geometry_quality": "standard",
9
- "smart_low_poly": true,
10
- "face_limit": 8000,
11
- "model_seed": 5501,
12
- "texture_seed": 5501,
13
- "input": "file_00cd8007-e2d1-4126-9443-5eb00661da5a"
14
- },
15
- "model": "v3.1-20260211",
16
- "task_id": "735f2f70-804b-4563-983e-7ce17bfa0be8",
17
- "elapsed_sec": 238.82,
18
- "credits": 40.0,
19
- "cached": false,
20
- "preview_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/735f2f70-804b-4563-983e-7ce17bfa0be8/legacy_mesh.webp?auth_key=1792227387-tddtsndb-0-a3b81fd0aaa003c2df1532eedb83315e",
21
- "bytes": 2459796,
22
- "triangles": 9597,
23
- "meshes": 1,
24
- "materials": 1,
25
- "textures": 3,
26
- "images": 3,
27
- "nodes": 1,
28
- "generator": "https://tripo3d.ai",
29
- "cache_key": "ecb7e508f9dc367e4d5c4b91e9c7b37b68dbb47a2d9173ce1b5c30f01813d8a9"
30
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/frozen_game_inputs.json DELETED
@@ -1,222 +0,0 @@
1
- {
2
- "schema_version": "game_delivery_view_not_execution_plan_v1",
3
- "batch_id": "obj200_meshy_tripo_20261015_v2",
4
- "game_id": "G055_inner_court_investigation",
5
- "parent_frozen_plan_sha256": "5c287b8399f00fbb0dddf94c9e3f5ffcbcc2e679ff76882bc8fc369d6eae840d",
6
- "parent_frozen_plan_file_sha256": "3c4798c34345ab6f7fc07f850f25deca44051565185a1c8a80058819e1bbd685",
7
- "task_count": 4,
8
- "parent_plan_original": {
9
- "archive": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/frozen_parent_plan_original.zip",
10
- "archive_member": "bench/test_data/outputs/_asset_batches/obj200_meshy_tripo_20261015_v2/plan.json",
11
- "original_file_sha256": "3c4798c34345ab6f7fc07f850f25deca44051565185a1c8a80058819e1bbd685",
12
- "plan_sha256": "5c287b8399f00fbb0dddf94c9e3f5ffcbcc2e679ff76882bc8fc369d6eae840d"
13
- },
14
- "scope": "3d_object_only_not_complete_stage2",
15
- "providers": {
16
- "meshy": {
17
- "model": "meshy-6",
18
- "api_base": "https://api.meshy.ai",
19
- "options": {
20
- "low_poly": false,
21
- "texture": true,
22
- "enable_pbr": true,
23
- "should_remesh": true,
24
- "topology": "triangle"
25
- },
26
- "credits": {
27
- "text_preview": 20,
28
- "text_refine": 10,
29
- "image": 30
30
- }
31
- },
32
- "tripo": {
33
- "model": "v3.1-20260211",
34
- "api_base": "https://openapi.tripo3d.com/v3",
35
- "options": {
36
- "low_poly": true,
37
- "texture": true,
38
- "pbr": true,
39
- "texture_quality": "standard",
40
- "geometry_quality": "standard"
41
- },
42
- "credits": {
43
- "text": 30,
44
- "image": 40
45
- }
46
- }
47
- },
48
- "postprocess": {
49
- "tool": "gltf-transform-4.2.1_meshoptimizer-0.24.0",
50
- "simplify_error": 0.01,
51
- "simplify_passes": 1,
52
- "triangle_budget": "per_task_decimation_target_at_most_8000",
53
- "texture_budget": "per_task_texture_size_downsize_only",
54
- "preserve_native_output": true
55
- },
56
- "runtime_sources": {
57
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/generate_3d_comparison.py": "70f8d79c8596908f7edcd7ff68120f558cea6881f23dc2f0e29ef62f2bab75c4",
58
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/cloud_journal.py": "92d075ef9308e0d6496a901d85b7847b8f602f88de8e19caeaf85ac893081253",
59
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/record_paths.py": "93c42e454d60d03c30b12e58e8e75453c7caa8fc92ff30bbc20aa645b3fc23dd",
60
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/paths.py": "ed1cf1105f95c76b4893e98a84b3e37ef7d1f87c873c93c8e58413457104d28b",
61
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/prompts/stage2_assets.md": "172339b9f6c6d834ef0d782c0d3dd779b62a12bd34acddc746b80d7cd99f63c2",
62
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/operators/gen_3d_object/operator.py": "5f88c457ee58f73de5ddbd4d96fa91e474c29bd3ecefbb02f3faf18b93c4365a",
63
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/meshy_model.py": "ef437ff79554469390c0d6c71f11788431fe4a1d1e51e92feae3adf296cebaec",
64
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/tripo_model.py": "29ab3944260d62585a991f384eee4917368f97c4476abc3eda85299bc1a8089e",
65
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/cloud_api.py": "a3c363a3439b3864e05e782340eb3059380c4b1c28eccb2524e9cbf9131fe996",
66
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/glb_utils.py": "6f3ee0147b345415cb6d8b1b1433fa557d0d543bb92276840dc85b43cd7a66ad",
67
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/process_mesh.mjs": "b1b6ada273a576083bb167772ec03fee6fe3275470180563722e56687688995c",
68
- "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/package-lock.json": "0d9c3ad5898a78a0db22f361caa7242b48ff86586e898f5ae5c1bb4a7134631a"
69
- },
70
- "mesh_runtime": {
71
- "status": "ready",
72
- "node": "20.17.0",
73
- "dependencies": {
74
- "@gltf-transform/core": "4.2.1",
75
- "@gltf-transform/extensions": "4.2.1",
76
- "@gltf-transform/functions": "4.2.1",
77
- "draco3dgltf": "1.5.7",
78
- "meshoptimizer": "0.24.0",
79
- "sharp": "0.35.5"
80
- }
81
- },
82
- "documents": {
83
- "general_requirement": {
84
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/general_requirement.txt",
85
- "sha256": "285a181c4ed0275c4c96a9cd71c6e22abfd90212018ef6c284f55315037589a3",
86
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/general_requirement.txt"
87
- },
88
- "object_requirement": {
89
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/object_requirement.txt",
90
- "sha256": "8b8e455ca2bb791b3718602c902186b576bfd8a7e172262e9e4ff3ba25c51173",
91
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/object_requirement.txt"
92
- },
93
- "context_excerpt": "宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、",
94
- "context_method": "deterministic title/world/core-loop excerpts; full brief retained locally"
95
- },
96
- "tasks": [
97
- {
98
- "task": {
99
- "game_id": "G055_inner_court_investigation",
100
- "task_id": "G055_inner_court_investigation_obj_001",
101
- "asset_name": "shen_zhiwei",
102
- "name_zh": "沈知微",
103
- "role": "player",
104
- "input_mode": "image_to_3d",
105
- "prompt": "Original young court investigator Shen Zhiwei: understated dark-teal commoner robe with faint embroidery, hair in a plain bun with wooden pin, cloth shoes, sharp calm gaze, stylized semi-realistic ancient court style, vermilion-and-indigo palette, T-pose, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
106
- "seed": 5501,
107
- "decimation_target": 8000,
108
- "texture_size": 2048,
109
- "rig": "humanoid",
110
- "animation_strategy": "skeletal",
111
- "spec_ref": "P01,W06",
112
- "image_path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
113
- "poly_policy": "low_poly_priority",
114
- "l_tier_repair_target": true
115
- },
116
- "shared_prompt": null,
117
- "reference": {
118
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
119
- "sha256": "ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55",
120
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png"
121
- },
122
- "subject_group": "character",
123
- "context_condition": "reference_only"
124
- },
125
- {
126
- "task": {
127
- "game_id": "G055_inner_court_investigation",
128
- "task_id": "G055_inner_court_investigation_obj_002",
129
- "asset_name": "evidence_pouch_ash",
130
- "name_zh": "旧香囊残灰",
131
- "role": "evidence",
132
- "input_mode": "image_to_3d",
133
- "prompt": "Small scorched silk incense pouch with charred rim and faint residual ash, tied with a cut thread, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
134
- "seed": 1552,
135
- "decimation_target": 4000,
136
- "texture_size": 1024,
137
- "rig": "none",
138
- "animation_strategy": "procedural",
139
- "spec_ref": "P20.E01",
140
- "image_path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
141
- "reference_image_role": "single_object_visual_anchor",
142
- "poly_policy": "low_poly_priority",
143
- "l_tier_repair_target": true
144
- },
145
- "shared_prompt": null,
146
- "reference": {
147
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
148
- "sha256": "b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d",
149
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png"
150
- },
151
- "subject_group": "rigid",
152
- "context_condition": "reference_only"
153
- },
154
- {
155
- "task": {
156
- "game_id": "G055_inner_court_investigation",
157
- "task_id": "G055_inner_court_investigation_obj_003",
158
- "asset_name": "rota_manuscript",
159
- "name_zh": "排班抄本",
160
- "role": "evidence",
161
- "input_mode": "text_to_3d",
162
- "prompt": "Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
163
- "seed": 1553,
164
- "decimation_target": 5000,
165
- "texture_size": 1024,
166
- "rig": "none",
167
- "animation_strategy": "procedural",
168
- "spec_ref": "P20.E03",
169
- "poly_policy": "low_poly_priority",
170
- "l_tier_repair_target": true
171
- },
172
- "shared_prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
173
- "reference": null,
174
- "subject_group": "rigid",
175
- "context_condition": "shared_short_context"
176
- },
177
- {
178
- "task": {
179
- "game_id": "G055_inner_court_investigation",
180
- "task_id": "G055_inner_court_investigation_obj_004",
181
- "asset_name": "seal_wax_imprint",
182
- "name_zh": "火漆拓印",
183
- "role": "evidence",
184
- "input_mode": "text_to_3d",
185
- "prompt": "Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
186
- "seed": 1554,
187
- "decimation_target": 4000,
188
- "texture_size": 1024,
189
- "rig": "none",
190
- "animation_strategy": "procedural",
191
- "spec_ref": "P20.E06",
192
- "poly_policy": "low_poly_priority",
193
- "l_tier_repair_target": true
194
- },
195
- "shared_prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
196
- "reference": null,
197
- "subject_group": "rigid",
198
- "context_condition": "shared_short_context"
199
- }
200
- ],
201
- "input_policy": "shared_short_context_and_reference_only",
202
- "context_limitations": "Text APIs received shared short context plus asset prompt; image APIs received reference only. Full requirements retained as evidence, not sent to APIs.",
203
- "benchmark_metrics": {
204
- "C": {
205
- "status": "UNKNOWN",
206
- "value": null
207
- },
208
- "Q1": {
209
- "status": "UNKNOWN",
210
- "value": null
211
- },
212
- "Q2": {
213
- "status": "UNKNOWN",
214
- "value": null
215
- },
216
- "Q3": {
217
- "status": "UNKNOWN",
218
- "value": null
219
- }
220
- },
221
- "view_sha256": "5fd977b003009784fb0799b8eb93c01d03f9381a5638de8404d3da2b6ee63722"
222
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/general_requirement.txt DELETED
@@ -1,821 +0,0 @@
1
- # 《宫墙影语》|宫廷潜行解谜机制输入规范
2
-
3
- ```yaml
4
- case_id: G055
5
- base_case_id: G055
6
- title: 宫墙影语
7
- category: RPG / Narrative / Social
8
- tier: L-Regular
9
- engine: ue5
10
- report_track: Standard
11
- paired_conditions: [R, N]
12
- required_real_clients: 1
13
- fixed_seed: 1055
14
- spec_version: v2
15
- status: DRAFT_LOCAL
16
- mechanism_source: null
17
- task_spec_version: null
18
- task_spec_hash: null
19
- frozen_by: null
20
- frozen_at: null
21
- dependency_hashes: null
22
- ```
23
-
24
- - 版本:mechanics-v2(2026-10-11 完整性修订:冻结运行引擎为 Unreal Engine 5.x;新增演出反馈补充 §29、过场视频 §30、阶段重开与历史状态保留 §31,并将 §22 创作边界正向化;全部 G01–G12、P01–P22 规则编号与冻结数值原样保留);运行引擎:**Unreal Engine 5.x**(spec、rubric 与 asset 任务 JSONL 引擎字段保持一致,引擎唯一冻结);输入方式:键盘与鼠标。
25
- - 整理声明:本文件由团队包 `inner_court_investigation_package/game_spec.md`(设计文档版本 1.0)忠实移植为 G055 的 game_spec.md,全部规则编号(G01–G12、P01–P22)与冻结数值原样保留;团队包 `rubric.txt` 作为 G055 rubric 的素材基线另行重核。fixed_seed=1055 用于评测注入(§4.3 概率判定的抽样复现),不改变正文规则。
26
- - 独立 oracle:§7.2.3「旧香囊残灰」算分示例(30→36→44→54→61→71→81)与 §7.2.1 证据强度公式、§4.3 潜行/交涉判定公式为冻结对账口径;任何偏离即 FAIL。
27
- - 参考图:本测例无绑定参考图,视觉要求仅由本文件原创描述冻结(不新增、不伪造任何图片绑定)。
28
-
29
- 【使用说明】
30
- 本文件用于游戏生成模型,目标是输出可操作的第三人称宫廷潜行解谜游戏。请按本文件实现,不把故事介绍、静态页面或开发者说明当作已完成玩法。角色使用原创中文姓名;稳定ID仅用于需求和测试指标,不代替人物姓名。所有人物、对白、事件和美术都必须原创,禁止加入其他作品的商标、具名角色、徽记、造型、专有台词或剧情复刻。
31
- 本作品不是即时战略或回合战棋。没有RTS式资源基地和大规模军队;玩家操控沈知微探索宫廷、潜行、推理并作出关系选择。键盘与鼠标是唯一操作设备,不生成手柄操作说明。
32
- 规则优先级:本文件全局数值和规则 > 机制章节 > 关卡实例 > 生成器自由补充。未指定的数值可作为明确标注的配置默认值,不得伪装成本文既定要求。任何任务关键线索都不得被随机失败永久删除。
33
-
34
- ## AI 游戏生成平台总设计提示词|剧情潜行解谜互动游戏
35
-
36
- > 请依据本文档生成一款第三人称、单人、剧情驱动的宫廷潜行与环境解谜游戏。核心体验由空间探索、环境机关、身份伪装、观察路线和多方案潜入构成;故事、人物与美术均为原创设定。作品采用完全原创的王朝、人物岗位、事件、美术与对白;角色使用功能称谓和稳定ID,不使用现实人物或既有影视作品中的具名角色、标志与台词。
37
-
38
- ## 使用规范 [G01–G04]
39
-
40
- - 将下述设定视为统一的设计约束;若平台能力有限,优先保留主线、潜行、线索板、警戒与多结局。
41
- - 生成可游玩的第三人称单机冒险,主线时长约 12–16 小时,支线 5–8 小时。
42
- - 目标平台:PC;全程仅使用键盘与鼠标;支持自动存档与手动存档。
43
- - 画面:写实风格化架空古代宫廷美术,避免历史纪录片式写实;夜景需保证可读性。
44
- - 默认难度“宫规”,提供“初入宫门”“宫规”“深宫夜行”三档。
45
- - 所有数字均为初始平衡值,可在内部测试后整体微调 10%–20%,不得改变核心关系。
46
- - 不允许把“解谜”退化成单一密码锁;同一目标至少提供两种解法或两条路径。
47
- - 不允许把女性角色写成单一善恶标签;每位关键人物都要有利益、恐惧、底线与可验证动机。
48
- - 战斗是高风险备选,不是主要解题方式;击杀关键人物不能成为默认最优解。
49
- - 生成对话时避免挪用任何既有影视台词,采用原创表达与新情境。
50
-
51
- ## 1. 一句话概念与体验承诺
52
-
53
- 玩家扮演沈知微,在一座以礼法、流言、门禁和人情构成的“活迷宫”中调查一份被篡改的宫廷名册;通过观察、伪装、潜行、机关、证词交叉验证与关系经营,在保全自己及同伴的前提下决定真相如何被揭开。
54
-
55
- ### 1.1 体验支柱
56
-
57
- - 读懂空间:宫殿不是走廊集合,而是由视线、作息、门禁、值守和礼仪共同组成的可推理地图。
58
- - 先观察再行动:听脚步、看巡逻、辨衣饰等级、记录谁能进入哪扇门。
59
- - 以身份作工具:借衣、借差事、借人情,不同身份开启不同通路,也带来不同风险。
60
- - 让线索彼此说话:物证、证���、时间表、账簿和关系网互相验证。
61
- - 选择有回声:短期安全、长期信任、名声、证据完整度影响后续章节与结局。
62
- - 失败仍推进:被发现后可能引发盘问、搜查、禁足或关系损失,不用频繁读档重来。
63
-
64
- ### 1.2 目标受众与内容边界
65
-
66
- - 适合喜爱叙事冒险、潜行、推理与角色关系经营的玩家。
67
- - 建议年龄分级:青少年以上;不描写酷刑细节、血腥处决或性暴力。
68
- - 宫斗冲突以权力、名誉、资源、信任和制度压力呈现。
69
- - 玩家可选择克制、合作、欺瞒或揭发;不强迫玩家采用单一“正确”道德路线。
70
-
71
- ## 2. 世界背景与故事前提
72
-
73
- 架空大周王朝,年号景和。宫中准备修订《内廷岁档》,它记录各宫人员调动、赏赐、药材领用、出入时辰与旧案批注。档案中一页“冬至夜值房名册”被人调包,牵涉多年前的一场火灾与一位被迫离宫的女官。
74
- 沈知微以新人身份入宫,最初只想安稳度日。她在锦瑟院发现一只不属于自己的旧香囊,内藏半张值房轮值签。顾怀青提醒她:香囊上的药香配方不可能来自御药房常规领用。许令仪则发现某名册页码与当值太监的口供对不上。玩家由此开始查案。
75
- 案件并非简单的“找出坏人”:名册被改动可能是为掩盖救人、栽赃、替人顶罪或换取权力。玩家需确认谁改、为何改、谁受益、何时能安全公开。
76
-
77
- ### 2.1 章节结构
78
-
79
- - **序章:初入宫门**:熟悉操作、礼法、第一条巡逻路线;完成一次不伤人的潜入。 目标:找回遗失的引荐牌,并确认锦瑟院并非随机分配。
80
- - **第一章:锦瑟院夜香**:调查香囊、药材与夜间值房;建立线索板。 目标:证明香囊来自宫中旧库,而非玩家角色自带。
81
- - **第二章:花影与名册**:潜入尚仪局、花房与档房,拼出轮值表。 目标:确认名册被改的时间窗口及三名可接触档案的人。
82
- - **第三章:宫宴无声**:大型社交潜行关;礼仪、视线、座次与菜品成为谜题。 目标:在不引发公开冲突下取得宴前药膳记录。
83
- - **第四章:雨夜太医院**:纵向探索太医院屋脊、回廊、药库;追查药材批次。 目标:将两份证词与实物批号对应,决定是否保护证人。
84
- - **第五章:冷宫旧档**:限时搜证与情感抉择;老宫人提供互相矛盾的记忆。 目标:找到火灾当夜的原始抄录,确认被隐去的救援路线。
85
- - **第六章:御前校档**:终局潜入;通过多条路线进入校档房,安排证据呈递顺序。 目标:玩家决定证据、对象、时机与承担后果的人。
86
- - **尾声:墙外春信**:依据前述变量呈现不同结局与人物关系收束。 目标:展示选择的长期影响,不以单一分数评价玩家。
87
-
88
- ## 3. 核心循环
89
-
90
- 1. 抵达区域:查看区域平面、门禁和当日作息摘要。
91
- 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。
92
- 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。
93
- 4. 选定身份与路线:制服、差事、口令、关系或环境机关。
94
- 5. 执行行动:潜行、交谈、取证、解锁、调度 NPC 或制造转移注意。
95
- 6. 处理后果:警戒、怀疑、信任、证据完整度与宫中流言变化。
96
- 7. 回到安全点:整理证据,解锁新的推理问题和下一章入口。
97
-
98
- ## 4. 数值口径与全局规则 [G05–G12]
99
-
100
- ### 4.1 统一刻度
101
-
102
- | 指标 | 范围 | 解释 |
103
- |---|---:|---|
104
- | 角色属性 | 1–10 | 能力基线;每点约影响 5%–8% 相关判定。 |
105
- | 技能熟练 | 0–100 | 行为熟练度;每 20 点解锁一个稳定优势。 |
106
- | 区域警戒 | 0–100 | 该区域对玩家的即时追查强度。 |
107
- | 个人怀疑 | 0–100 | NPC 对玩家身份或说辞的怀疑。 |
108
- | 关系信任 | -100–100 | 单个人物对沈知微的信任,0 为中性。 |
109
- | 名声 | -100–100 | 宫中对沈知微总体印象,分谨慎、贤名、锋芒等标签。 |
110
- | 证据强度 | 0–100 | 证据可验证性、独立来源和保全状态综合值。 |
111
- | 资源 | 整数 | 银钱、香料、药材、通行凭据等可消耗物。 |
112
- | 时间 | 游戏内分钟 | 用于巡逻、仪式、差事和限时窗口。 |
113
-
114
- ### 4.2 难度系数
115
-
116
- | 难度 | 侦测倍率 | 线索提示 | 盘问宽限 | 解谜辅助 |
117
- |---|---:|---|---:|---|
118
- | 初入宫门 | 0.70 | 关键线索自动标注一次 | 2 次失误 | 开启机关轮廓与路线提示 |
119
- | 宫规 | 1.00 | 需主动观察 | 1 次失误 | 可选轻提示,不给答案 |
120
- | 深宫夜行 | 1.25 | 不自动标注 | 0 次失误 | 无方向提示,仍保留无障碍选项 |
121
-
122
- - 难度只影响侦测速度、提示和容错,不改变剧情可达性。
123
- - 重要证据不得因随机骰点彻底丢失;失败应改变取得成本或证据质量。
124
- - 存档点设在安全��、章节转换点和重要社交场景前。
125
-
126
- ### 4.3 判定公式
127
-
128
- 潜行成功率 = 基础值 55 + (潜行熟练 - 50) × 0.55 + 环境修正 + 身份匹配修正 - 警戒惩罚。
129
- 交涉成功率 = 50 + (礼仪或洞察 - 5) × 6 + 信任修正 + 证据支撑 - 对方戒心。
130
- 解锁成功不采用纯随机:工具、观察到的结构线索与操作节奏决定结果;失误增加噪声或耗时。
131
- 任何概率判定需显示“稳妥 / 有风险 / 高风险”三档提示,不强制显示精确百分数。
132
- 环境遮蔽修正:屏风 +20,暗处 +15,拥挤人群 +10,空旷明处 -20,逆光窗前 -10。
133
- 身份匹配修正:同差事 +20,临时借衣 +10,衣饰等级不符 -15,口令过期 -25。
134
- 证据支撑修正:一项独立物证 +10,两项相互独立证据 +20,单一传闻 +0。
135
-
136
- ## 5. 玩家角色:沈知微 [P01]
137
-
138
- | 项目 | 数值 / 设定 |
139
- |---|---|
140
- | 年龄 | 17(剧情开始时) |
141
- | 身高 | 162 cm |
142
- | 体重 | 48 kg(仅视觉设定,不参与属性判定) |
143
- | 体力 | 6/10 |
144
- | 敏捷 | 7/10 |
145
- | 洞察 | 8/10 |
146
- | 礼仪 | 7/10 |
147
- | 记忆 | 9/10 |
148
- | 交涉 | 7/10 |
149
- | 潜行 | 6/10 |
150
- | 初始名声 | 0 |
151
- | 初始银钱 | 18 两 |
152
- | 初始关系 | 许令仪 +20;顾怀青 +15;虞蘅 +5 |
153
-
154
- ### 5.1 能力成长
155
-
156
- - 技能点每章 1 点,支线与无伤关卡可额外获得,总上限 12 点。
157
- - 体力、敏捷、洞察、礼仪、记忆、交涉、潜行均可升级;单项上限 10。
158
- - 专精不要求玩家全能;低属性通过道具、同伴帮助和替代路线补足。
159
- - 沈知微不具备超自然能力;“感知”表现为专注观察与推断,不是透视。
160
-
161
- ### 5.2 玩家弱点与约束
162
-
163
- - 新人身份:资历不足,主动命令太监或宫女会增加怀疑。
164
- - 行动窗口短:宵禁后长时间停留会逐步抬高区域警戒。
165
- - 非战斗专家:正面对抗最多制服普通看守,无法击败精锐护卫。
166
- - 关系有代价:替某人隐瞒一次可能让另一条线索延迟出现。
167
- - 证据需保全:湿纸、易挥发香灰、被污染器具会降低证据强度。
168
- - 礼法能保护也能限制:公开质疑高位人物会迅速改变名声。
169
-
170
- ## 6. 潜行与身份扮演 [P02]
171
-
172
- 潜行的核心不是蹲伏,而是“看起来有理由出现在这里”。每个区域都标注可进入身份、合理行动、禁区与盘问触发点。
173
-
174
- ### 6.1 身份类型
175
-
176
- | 身份 | 可进入区域 | 优势 | 风险 |
177
- |---|---|---|---|
178
- | 嫔妃随侍 | 外廷偏厅、寝殿外间 | 可随队通行、解释简单 | 被问主位姓名或差事时需答对。 |
179
- | 尚仪局女史 | 礼仪房、典册架外围 | 可查看排班与仪式用具 | 衣饰细节严格,错用称谓易露馅。 |
180
- | 医女 | 太医院药房、诊候廊 | 可接触药材与病历摘要 | 需辨认药名,错误配伍会暴露。 |
181
- | 花房杂役 | 花房、内苑小径 | 可携带花木与剪具 | 不适合进入档案重地。 |
182
- | 传膳宫人 | 膳房、宴席后廊 | 可在宴前后穿行 | 托盘被检查时需有真实单据。 |
183
- | 内侍书办 | 档房外间、值房 | 可询问纸张与印泥 | 身形、口音或腰牌不符会遭盘问。 |
184
- | 无身份便装 | 公共区域、暗道 | 低期待值、无制服限制 | 所有正式门禁都需绕行。 |
185
-
186
- ### 6.2 伪装规则
187
-
188
- - 服饰分主色、纹样、腰牌、鞋履、配饰五项;至少四项匹配才算稳定伪装。
189
- - 衣物可借、换、污损或被认出;重要角色记住玩家常用装束。
190
- - 同一身份连续使用超过 3 个区域,会触发“熟面孔”检查。
191
- - 伪装不等于隐身:同僚会提出符合身份的问题,回答可由线索或记忆准备。
192
- - 禁止在高位人物面前使用低阶身份;尊卑差错增加怀疑 25 点。
193
- - 每套伪装有“合理停留时间”,通常为 4–8 分钟,真实差事可延长。
194
-
195
- ### 6.3 侦测模型
196
-
197
- - NPC 侦测由视线、距离、光照、噪声、身份、行为异常六项共同决定。
198
- - 视线锥分近视 0–5 m、中视 5–12 m、远视 12–20 m;屏风和门框会遮挡。
199
- - 脚步噪声:慢走 5、快走 15、奔跑 35、碰撞物件 45(0–100)。
200
- - 对话噪声:轻声 5、普通交谈 12、争执 35、呼喊 60。
201
- - 发现异常先进入“留意”2 秒,再进入“确认”5 秒,最后进入“追查”。
202
- - NPC 可看到玩家进入掩体;掩体不能抹除已有目击。
203
- - 警戒低于 25:闲谈;25–49:巡逻改变;50–74:临时盘查;75–99:搜查封门;100:抓获或章节级后果。
204
-
205
- ## 7. 证据推理与线索板 [P03]
206
-
207
- 线索板是玩家的工作台,不是自动给出答案的剧情百科。每条线索记录来源、时间、可信度、可验证点和潜在偏见。
208
-
209
- ### 7.1 证据类别
210
-
211
- - 实物证据:纸张纤维、封泥、香灰、药包、衣料纤维、锁具磨损。
212
- - 记录证据:轮值册、赏赐账、药材领用单、宴席座次图、门房簿。
213
- - 口述证据:证人回忆、闲谈、盘问口供;需考虑恐惧、利益与记忆偏差。
214
- - 空间证据:脚印方向、可达路线、视线死角、门窗开合痕迹。
215
- - 关系证据:谁替谁传话、谁有权限、谁在关键时辰缺席。
216
-
217
- ### 7.2 线索质量
218
-
219
- | 质量 | 分值 | 表现 |
220
- |---|---:|---|
221
- | 流言 | 10–24 | 来源模糊,只用于提出假设。 |
222
- | 线索 | 25–49 | 存在事实指向,但可能有替代解释。 |
223
- | 佐证 | 50–74 | 有独立来源或可复验细节。 |
224
- | 强证据 | 75–89 | 能排除多数替代解释,来源完整。 |
225
- | 决定性证据 | 90–100 | 能重建事件链,且保全与来源链可靠。 |
226
-
227
- #### 7.2.1 证据强度的计算口径
228
-
229
- 证据强度不是主观印象值。每条证据建立时先计算基础分,再按验证、独立性、来源链和保全状态加减分;系统将结果限制在 0–100。
230
- `证据强度 = clamp(基础可靠度 + 可复验分 + 独立佐证分 + 来源链分 + 保全分 − 污染/胁迫/矛盾扣分, 0, 100)`
231
-
232
- | 组成项 | 范围 | 计算口径 |
233
- |---|---:|---|
234
- | 基础可靠度 | 0–40 | 证据类型与原始完整程度;原件高于转述件,直接观察高于传闻。 |
235
- | 可复验分 | 0–20 | 其他角色或玩家能否重复检查同一特征;验证步骤必须记录。 |
236
- | 独立佐证分 | 0–20 | 不同来源、不同形成过程且结论相符的佐证;同源材料不重复计分。 |
237
- | 来源链分 | 0–10 | 发现地点、时间、经手人和移交记录完整程度。 |
238
- | 保全分 | 0–10 | 证据从取得至提交未受潮、混合、改写、拆封或遗失。 |
239
- | 污染/胁迫/矛盾扣分 | 0 至 −40 | 污染 −5 至 −20;胁迫口供 −15 至 −30;无法解释的关键矛盾 −10 至 −25。 |
240
-
241
- - 每条证据初次发现时,系统按类型赋基础可靠度:流言 10、未经核实口述 18、二手抄本 22、现场痕迹 25、原始账册或实物 30、可验证的原始记录 35、带完整封存信息的原件最高 40。
242
- - 每完成一种有效复验,按难度加 5–10 分;同一验证动作不得重复加分。可复验分最多 20 分。
243
- - 独立佐证按来源簇计算:同一证人重复讲述、同一账本的抄本、同一经手人转述都属于一个来源簇;每个新的独立来源簇最多加 10 分,独立佐证分最多 20 分。
244
- - 来源链每记录一项加分:发现地点与时刻 +4、发现者 +2、移交记录 +2、封存状态 +2;总计最多 10 分。
245
- - 完整封存 +10;有轻微但可解释的暴露 +5;受潮、混入其他样本或无人记录经手过程不得获得保全分,并按损害程度扣分。
246
- - 胁迫获得的口供不能单独超过 49 分;须有至少一项独立物证或记录佐证,才可突破此上限。
247
- - 证据强度描述“证据可信度与可用于推理的程度”,不直接等同于某人的罪责概率。高分证据仍可支持多个解释。
248
- - 系统不允许靠反复点击、重复询问、重复提交同一物证刷分;每个证据 ID 的每个加分来源只结算一次。
249
-
250
- #### 7.2.2 证据处理结果与分值变化
251
-
252
- | 玩家操作 | 分值变化 | 适用条件 |
253
- |---|---:|---|
254
- | 找到原件而非转抄件 | +8 至 +15 | 原件与副本能明确区分,且原件未被篡改。 |
255
- | 完成笔迹、页码、印泥或纤维对照 | +5 至 +10 | 对照对象有可靠来源,方法在游戏内有说明。 |
256
- | 由独立角色复验实物 | +5 至 +10 | 复验者未参与证据形成或此前的转述。 |
257
- | 找到第二种独立记录相互印证 | +10 | 两份记录非同源抄写,且关键时间或对象一致。 |
258
- | 补齐发现与移交记录 | +2 至 +8 | 仅补记真实发生且可验证的经手过程。 |
259
- | 私自拆封、涂改、混样 | −10 至 −25 | 损伤程度由操作和后续可复验性决定。 |
260
- | 证人口供明显受胁迫 | −15 至 −30 | 口供仍可作线索,但可信度下降。 |
261
- | 出现无法解释的关键矛盾 | −10 至 −25 | 矛盾须涉及时间、地点、身份或物品来源。 |
262
-
263
- #### 7.2.3 算分示例:旧香囊残灰
264
-
265
- - 初始发现:现场实物,基础可靠度 30;尚未验证、来源记录不全,初始强度 30。
266
- - 记录锦瑟院发现地点、当时时刻与发现者:来源链 +6,变为 36。
267
- - 虞蘅独立比对香样,确认灰烬特征:可复验 +8,变为 44。
268
- - 找到花房领用单,记录显示同批香料进入旧库:独立佐证 +10,变为 54。
269
- - 再由顾怀青检查灰烬中混入的药材纤维:可复验 +7,变为 61。
270
- - 将残灰封入独立纸包并登记交接:保全 +10,变为 71(佐证级)。
271
- - 若另找到未经转抄的旧库原始账册并确认批号,再加独立佐证 +10,变为 81(强证据)。
272
- - 若玩家把残灰与另一香样混在一起,先扣 15;后续仍可靠独立账册和未污染残留恢复部分证明力,但不能恢复已失去的样本信息。
273
-
274
- #### 7.2.4 生成提示词必须输出的字段
275
-
276
- 每个证据对象都必须输出:`证据ID���名称、类型、发现位置、发现时间、发现者、基础可靠度、当前强度、来源簇、可复验点、独立佐证、来源链状态、保全状态、污染风险、支持的假设、可能的替代解释、强度变化日志`。
277
- 每次强度变化都必须给出具体原因,例如“+8:完成纸张纤维对照”,不能只显示“证据更可靠”。
278
- 玩家调查界面显示强度等级与变化理由;开发者调试界面显示每项分值。不得向玩家隐藏扣分原因。
279
-
280
- ### 7.3 推理操作
281
-
282
- - 连接两条线索:建立“支持 / 矛盾 / 依赖 / 同源”关系。
283
- - 建立时间线:把时辰、巡更、差事、天气与门禁放到同一刻度。
284
- - 提出人物假设:标明动机、机会、权限和反证。
285
- - 重演路线:在俯视图上验证某人在限定时间内能否到达某地。
286
- - 校验来源:对比笔迹、印泥、页码、纸张批次和经手人。
287
- - 提交推论:选择公开、私下询问、继续调查或暂缓。
288
-
289
- ### 7.4 推理失败处理
290
-
291
- - 错误指认不锁死主线,但会损害关系或令目标转移。
292
- - 证据不足时,系统给出“还缺权限 / 时间 / 独立来源”提示,不透露答案。
293
- - 玩家可撤回假设;已公开的指控不可完全收回,会形成长期名声变量。
294
- - 自动辅助模式会把关键推理分成三问,玩家仍需选择证据。
295
-
296
- ## 8. 礼法、关系与资源 [P04]
297
-
298
- ### 8.1 关系维度
299
-
300
- 每位重要角色包含:信任、好感、戒心、亏欠、秘密知情度五项。信任影响合作,好感影响私人交流,戒心影响盘问,亏欠影响请求,秘密知情度影响结局。
301
- 一次选择可同时提高一项、降低另一项;不得把关系压成单一好感条。
302
-
303
- ### 8.2 宫中名声标签
304
-
305
- - 稳重:公开行动更易被解释为谨慎,秘密调查耗时略增。
306
- - 聪慧:NPC 更愿意谈论事务,敌对者更早注意玩家。
307
- - 温厚:求助门槛降低,强硬盘问难度上升。
308
- - 锋芒:资源调度更快,高位人物戒心增加。
309
- - 守礼:正式场合宽容度增加,临场伪装弹性降低。
310
- - 多疑:线索提取更快,闲谈关系获取变慢。
311
-
312
- ### 8.3 资源经济
313
-
314
- | 资源 | 初始 / 上限 | 用途 | 来源 |
315
- |---|---:|---|---|
316
- | 银钱 | 18 两 / 99 | 购买普通道具、打点非关键杂役 | 月例、支线、赏赐。 |
317
- | 香料 | 3 份 / 12 | 制造气味线索或短暂掩盖药味 | 花房、药房、交易。 |
318
- | 纸墨 | 2 套 / 8 | 拓印、做对照、复写通行单 | 档房、书案、采买。 |
319
- | 人情 | 2 点 / 10 | 请求陪同、递话、拖延盘查 | 帮助角色、兑现承诺。 |
320
- | 药材 | 1 份 / 10 | 辨识配伍、安抚疲劳、验证药包 | 太医院、花房、救助。 |
321
- | 通行凭据 | 0 / 5 | 一次性进入受限门区 | 差事、伪造或借用。 |
322
-
323
- ### 8.4 物资获取规则
324
-
325
- - 普通物资允许探索补足;关键剧情道具不设随机掉落。
326
- - 每章提供至少三种非付费获取途径:探索、关系请求、以物易物。
327
- - 打点不会购买关键人物的沉默,只影响杂务流程或短暂延迟盘问。
328
- - 物资救援:若同伴因玩家选择失去住所或差事,可花 1 人情 + 3 两安排临时落脚。
329
- - 资源不足时仍可继续主线,系统转为耗时、风险或关系代价。
330
- - 商贩每日限量 2–4 件,价格受名声与宫中流言影响,浮动不超过 30%。
331
-
332
- ## 9. 关键角色功能表 [P05]
333
-
334
- 角色数值只服务于玩法判定,不把外形数值变成能力优劣。每位角色有信任、戒心和一个明确的协助专长。
335
-
336
- | 人物 | 属性 / 专长 | 目标与边界 |
337
- |---|---|---|
338
- | 沈知微(玩家) | 洞察9、记忆9、礼仪7、潜行6 | 调查名册;资历浅,强硬行动容易招致盘问。 |
339
- | 许令仪 | 礼仪9、交涉8、洞察7 | 可安排正式拜访、调阅公开账目;不为无证据指控背书。 |
340
- | 虞蘅 | 手作9、听觉8、记忆7 | 可辨香、制作气味标记;重视承诺与安全感。 |
341
- | 顾怀青 | 医术9、观察8、交涉6 | 可验证药材、处理轻伤;不能公开越权行动。 |
342
- | 阿蘩 | 宫规9、记忆8、交涉8 | 核验差事、建议路线;拒绝牵连无辜者。 |
343
- | 裴云舟 | 观察8、骑术8、交涉7 | 提供宫外消息与转递;可用章节有限,暴露关系代价高。 |
344
- | 乔婉宁 | 礼仪10、洞察9、戒备9 | 通过制度与代理人维持秩序;难以亲自核验所有细节。 |
345
- | 闻素绮 | 威势9、洞察7、戒备8 | 高压社交对手;利益可推断,误判会快速升级警戒。 |
346
- | 郑云舒 | 洞察8、记忆8、礼仪8 | 中立信息枢纽;不愿卷入没有退路的公开斗争。 |
347
- | 杜怀礼 | 权限10、记忆9、观察8 | 把关门禁与御前流程;完整手续比情面更有效。 |
348
- | 邵庭章 | 追踪8、威势7、潜行4 | 固定路线巡查者;可被观察与绕行。 |
349
- | 满仓 | 维修8、潜行8、交涉6 | 协助机关与搬运;需要明确撤离方案。 |
350
- | 陆青蘅 | 宫规9、观察8、戒备9 | 盘问与流程审查;无法同时看守多个门区。 |
351
- | 孟兰汀 | 礼仪8、记忆8、交涉7 | 新人引导与低风险信息来源;害怕卷入主位争斗。 |
352
- | 姚素秋 | 记忆6、观察7、胆量3 | 冷宫旧案证人;受惊时会混淆年份与时辰。 |
353
-
354
- 信任低于 -30 时角色拒绝主动协助;高于 40 时提供额外路线;高于 75 时每章可承担一次高风险支援。同伴提供机会,不替玩家完成推理。
355
-
356
- ## 10. 压力来源与应对方式 [P06]
357
-
358
- 本作没有怪兽或外星阵营;对抗对象是由制度、信息、门禁与具体执行者组成的宫廷权力网络。战斗力数值表示对玩家造成的追捕压力,不代表善恶。
359
-
360
- ### 10.1 势力层级
361
-
362
- | 层级 | 单位 / 角色 | 压力值 | 主要能力 | 破解方式 |
363
- |---|---|---:|---|---|
364
- | 顶层 | 皇帝与内廷裁决流程 | 100 | 改变规矩、封锁区域、决定证据是否进入审阅。 | 决定呈递对象与时机,拿出可复验链条。 |
365
- | 高层 | 乔婉宁与闻素绮的代理网络 | 85 | 调度宫人、制造名声压力、提高区域检查。 | 拆分信息来源、转移注意、利用利益差异。 |
366
- | 中高层 | 内务府、尚仪局、太医院主管 | 65 | 控制账册、药材、排班和通行凭据。 | 流程核验、关系交换、找出记录矛盾。 |
367
- | 执行层 | 掌事宫女、值房内侍、巡查太监 | 45 | 盘问身份、封门、追踪脚印与物件。 | 合规差事、伪装、绕路、提交真实单据。 |
368
- | 基层 | 门房、杂役、膳房帮工 | 25 | 提供目击、传话或短时阻拦。 | 礼貌交涉、帮助其完成差事、避免惊动。 |
369
- | 环境 | 宵禁、雨雪、宴席、礼仪流程 | 10–60 | 改变可见度、通行、噪声与时间窗口。 | 观察规律、改路线、借环境掩护。 |
370
-
371
- ### 10.2 应对原则
372
-
373
- 顶层决定规则,中层掌握记录,基层掌握现场。玩家不能靠击败基层解决高层问题;必须逐层取得可验证信息并找到制度上的合法入口。
374
- - 顶层角色不会成为可战斗 Boss。
375
- - 高层威胁通过时间表、门禁和代理命令呈现。
376
- - 中层弱点是手续和交叉记录。
377
- - 基层弱点是重复工作、有限视野与真实的人情需求。
378
- - 环境压力随章节动态变化,不做无提示的随机惩罚。
379
-
380
- ## 11. 轻度动作与非致命冲突 [P07]
381
-
382
- - 玩家可推门、翻越矮墙、攀爬柱廊、滑入桌下、熄灯、转移器物。
383
- - 近身动作仅有挣脱、短暂制服、遮挡视线三类;制服普通看守需 2.5 秒贴近操作。
384
- - 精锐护卫无法被正面制服;玩家应逃跑、藏匿、求援或谈判。
385
- - 武器只出现为环境道具,不提供大规模杀伤能力。
386
- - 冲突后果:受伤降低耐力,制服看守增加 20 警戒并可能损害同伴信任。
387
- - 无伤通过可解锁“无声行者”叙事称号,不给予过强数值优势。
388
-
389
- ## 12. 谜题与机关 [P08]
390
-
391
- ### 12.1 谜题设计原则
392
-
393
- - 每个主要谜题至少有观察、推理、互动三个步骤。
394
- - 谜题线索需在同一区域或前一安全区域可取得;不依赖像素级搜寻。
395
- - 同一目标至少两个解法:身份路线、环境路线、关系路线、证据路线任选其二。
396
- - 谜题答案应可从世界规律推出,不能靠无提示猜数字。
397
- - 失败反馈必须说明错在时间、工具、顺序或前提,不直接给答案。
398
- - 机关不会要求玩家反复做无意义的精确操作。
399
-
400
- ### 12.2 可复用机关组件
401
-
402
- 1. **屏风视线棋盘**:移动屏风改变守卫视线,同时制造可观察的影子缺口。
403
- 2. **漏刻计时机关**:读取水位、滴漏速度和巡更节奏,推算安全窗口。
404
- 3. **纸页拼接**:根据页码、纤维、裁边与墨色恢复被换页顺序。
405
- 4. **香气辨识**:比较燃烧残香、香囊残留与药材气味;需注意通风方向。
406
- 5. **回廊声学**:拍击、脚步与滴水声帮助判断空房、夹墙和门后距离。
407
- 6. **花木位置**:利用花盆移动痕迹、浇水记录和季节差异还原人员路线。
408
- 7. **礼仪顺序**:座次、敬茶顺序、称谓等级构成社交逻辑谜题。
409
- 8. **腰牌校验**:比对铸纹、绳结、磨损和当日发牌登记,识别伪造通行证。
410
- 9. **衣饰替换**:在时间限制内匹配主色、腰牌、鞋履和差事物件。
411
- 10. **账簿对账**:将月例、药材、膳食与值房开支对照,找出不可能的记录。
412
- 11. **锁具结构**:观察锁舌磨损与钥匙齿痕,决定正确开锁工具与方向。
413
- 12. **窗棂投影**:根据月光角度推断某人是否能从庭院看见室内动作。
414
- 13. **烛火风向**:利用火焰偏移和烟迹判断暗门缝隙、通风道或刚有人经过。
415
- 14. **棋局复盘**:把一盘未完棋局对应到人物策略或时间顺序,不考复杂棋力。
416
- 15. **诗笺藏字**:依据纸张折痕和行距发现缺失字句,不使用外部典故知识门槛。
417
- 16. **药包配伍**:核对份量、包纸、封口与领料单,判断药材是否同一批次。
418
- 17. **宴席传菜**:按温度和出菜次序复原谁能在何时接触某道菜。
419
- 18. **园林路线**:以桥、假山、花窗与巡逻路线组成俯视路线推演。
420
-
421
- ## 13. 关卡总览 [P09]
422
-
423
- 各章沿用同一结构:目标、关键谜题、至少两条路线、证据状态与失败补救。每章只保留会改变玩法的独特细节。
424
-
425
- ### 序章:宫门与锦瑟院
426
-
427
- - 区域规模:小型教学场;目标时长:约 12 分钟。
428
- - 人员配置:1 名门房 + 2 名巡逻宫人。
429
- - 谜题:引荐牌、门禁问答、屏风潜行。
430
- - 证据目标:证明玩家已被安排进入锦瑟院。
431
- - 风险与替代:警戒上限 35;被盘问可用引荐牌或拜托孟兰汀。
432
-
433
- ### 第一章:夜香留痕
434
-
435
- - 区域规模:中型寝殿群;目标时长:约 35 分钟。
436
- - 人员配置:3 巡逻 + 1 值夜女官。
437
- - 谜题:香灰比对、窗棂视线、漏刻窗口。
438
- - 证据目标:找到旧库香料批次与被剪断的线头。
439
- - 风险与替代:警戒上限 55;误入寝殿触发盘问,不丢失证据。
440
-
441
- ### 第二章:尚仪局的页码
442
-
443
- - 区域规模:中型办公区;目标时长:约 45 分钟。
444
- - 人员配置:4 巡逻 + 2 书办。
445
- - 谜题:纸页拼接、腰牌校验、值房时间线。
446
- - 证据目标:取得轮值表抄本并识别三名接触者。
447
- - 风险与替代:可借女史身份、夜间翻窗或请求许令仪正式查阅。
448
-
449
- ### 第三章:宫宴无声
450
-
451
- - 区域规模:大型社交区;目标时长:约 50 分钟。
452
- - 人员配置:8 关键 NPC + 6 服务人员。
453
- - 谜题:座次谜题、传膳路线、对话伪装。
454
- - 证据目标:确认宴前药膳记录来源。
455
- - 风险与替代:警戒达到 70 会提前结束宴席但仍有后续补查路线。
456
-
457
- ### 第四章:雨夜太医院
458
-
459
- - 区域规模:垂直探索区;目标时长:约 55 分钟。
460
- - 人员配置:5 巡逻 + 1 护卫。
461
- - 谜题:雨声掩护、药包配伍、屋脊路线。
462
- - 证据目标:取得药材批号与独立证词。
463
- - 风险与替代:湿纸需及时烘干;暴雨减少视线、增加滑落风险。
464
-
465
- ### 第五章:冷宫旧档
466
-
467
- - 区域规模:限时探索区;目标时长:约 60 分钟。
468
- - 人员配置:3 巡查 + 2 证人。
469
- - 谜题:路线重演、火灾旧图、记忆校验。
470
- - 证据目标:恢复原始抄录并安置姚素秋。
471
- - 风险与替代:证人受惊会中断谈话,需先安排安全出口。
472
-
473
- ### 第六章:御前校档
474
-
475
- - 区域规模:多路线终章;目标时长:约 70 分钟。
476
- - 人员配置:高层盘问、巡查网络。
477
- - 谜题:证据排序、呈递策略、双重门禁。
478
- - 证据目标:决定公开范围与承担后果者。
479
- - 风险与替代:警戒并非失败条件;高警戒提高呈递成本与风险。
480
-
481
- ### 尾声:春信
482
-
483
- - 区域规模:叙事收束区;目标时长:约 20 分钟。
484
- - 人员配置:无敌对单位。
485
- - 谜题:信件选择、关系结算。
486
- - 证据目标:展示结局及后续生活。
487
- - 风险与替代:允许玩家查看变量摘要与未解问题。
488
-
489
- ### 13.1 代表性任务与谜题解法
490
-
491
- 每项任务至少包含观察、推理、行动三个环节。表中任一方案均可推进;失败改变代价,不锁死主线。
492
-
493
- | 任务 | 关键线索 | 可行解法 | 失败后的继续方式 |
494
- |---|---|---|---|
495
- | 漏刻不说谎 | 漏刻水位、巡更锣声、未干水痕 | 计时进入;用屏风遮挡;请同伴引开守卫 | 警戒 +10,等待下一轮窗口。 |
496
- | 一页三纸 | 纸纤维、裁边、墨色、装订孔 | 公开查册;找废纸边角;拓印目录对照 | 错拼只形成低可信假设,可重新核对。 |
497
- | 同一枚腰牌 | 铸纹、绳结磨损、登记笔迹 | 正式核验;借差事接近登记册;询问制牌工匠 | 私取被发现会增加怀疑,但证据仍可补查。 |
498
- | 无声的宴席 | 出菜次序、座位图、托盘余温 | 传膳伪装;客人请求;宴后查厨房票据 | 宴席结束后改走厨房补查路线。 |
499
- | 两种香灰 | 灰烬颗粒、香料账单、风向 | 虞蘅比香;玩家做小样;通风处复验 | 误判可由领料单和另一份样本纠正。 |
500
- | 温药三问 | 药包封口、领料簿、煎煮时间、药渣 | 顾怀青协助;医女身份;私下取得配伍表 | 缺表时用药渣复验,耗时增加。 |
501
- | 冷宫里的年份 | 旧历、节令、药方背面、修补痕迹 | 安抚证人;用实物提示;玩家建立时间线提问 | 证人中断谈话,可先安排安全出口再返回。 |
502
- | 最后一盏灯 | 巡更节奏、灯油量、风向、同伴位置 | 调度同伴;合法差事;利用雨声撤离 | 可改走正门呈递,增加关系与时间成本。 |
503
-
504
- 其他机关可复用:窗棂投影推断视线、花盆痕迹复原路线、火漆裂纹判断封档是否开启、礼单座次推断空位��途、回廊回声定位夹墙。每种机关必须给出可观察线索和至少一种替代解法。
505
-
506
- ## 14. 任务与解锁 [P10]
507
-
508
- ### 14.1 主线解锁
509
-
510
- - 序章完成后解锁第一章,无属性门槛。
511
- - 第二章需“香料来源”或“值夜时间”任一线索达到佐证级。
512
- - 第三章需获得一份正式宴席差事,或通过关系建立随行身份。
513
- - 第四章需取得药材批号假设;若未取得,顾怀青可提供补查任务。
514
- - 第五章需两条独立证据指向旧案时间段。
515
- - 第六章不要求满证据;证据越完整,呈递风险越低、结局信息越清晰。
516
-
517
- ### 14.2 支线解锁示例
518
-
519
- - “花房的旧账”:虞蘅信任 ≥ 20 或找到重复香料领用单。
520
- - “许令仪的茶席”:名声 ≥ -10 且完成一次无冲突交涉。
521
- - “冷宫的冬衣”:人情 ≥ 1 或发现旧档房的修缮记录。
522
- - “满仓的暗格”:完成一次环境机关,不需要任何人物关系门槛。
523
- - “医女的夜班”:取得药材批号,或以 2 两雇请短期抄录。
524
- - “失而复得的腰牌”:被盘问一次后开启,以补救制度漏洞。
525
-
526
- ### 14.3 战力 / 能力阶梯
527
-
528
- | 阶段 | 玩家解锁 | 效果 | 对应威胁 |
529
- |---|---|---|---|
530
- | 初入宫门 | 慢走、观察、基础交谈 | 识别视线与差事提示 | 基层盘问 |
531
- | 熟悉规矩 | 稳定伪装、标注巡逻 | 同身份停留时间 +2 分钟 | 执行层巡查 |
532
- | 善察人心 | 证词偏见提示、关系请求 | 辨认回避与矛盾,不自动判真伪 | 中层流程审查 |
533
- | 暗线织网 | 同伴支援、两点同步调度 | 一次任务可安排一名同伴行动 | 高层代理网络 |
534
- | 御前校档 | 证据链呈递、公开范围选择 | 决定由谁先看、哪些可封存 | 顶层裁决流程 |
535
-
536
- ## 15. 选择与结局 [P11]
537
-
538
- ### 15.1 全局变量
539
-
540
- - truth_score:事实链完整度 0–100。
541
- - evidence_integrity:证据保全度 0–100。
542
- - public_exposure:公开程度 0–100。
543
- - ally_safety:同伴安全 0–100。
544
- - court_suspicion:宫廷总体怀疑 0–100。
545
- - compassion_debt:玩家对他人承担的承诺与亏欠 0–10。
546
- - witness_consent:证人是否自愿参与 0–100。
547
-
548
- ### 15.2 结局组
549
-
550
- - 封存的真相:truth ≥ 75、public_exposure < 40、ally_safety ≥ 60;案件被纠正但对外只留下模糊记录。
551
- - 庭前明证:truth ≥ 85、evidence_integrity ≥ 75、public_exposure ≥ 60;真相进入正式审阅,代价是关系与名声震荡。
552
- - 各自求生:truth 45–74 或 ally_safety < 50;玩家保全部分同伴,证据仅能改动局部制度。
553
- - 以名换路:truth ≥ 60、court_suspicion ≥ 75;玩家主动承担流言,换取证人脱身与关键文件保全。
554
- - 宫墙回声:truth < 45;案件未能定论,但玩家留下的线索让后来者仍可继续追查。
555
- - 春信归人(隐藏变体):满足关键人物信任、证人自愿与低伤害路线;尾声额外显示同伴未来去向。
556
-
557
- ### 15.3 结局公平性
558
-
559
- - 结局由累积变量触发,不由最后一个二选一独占决定。
560
- - 结局前显示风险提示:公开范围、可能受影响人物、证据缺口。
561
- - 不设置“全员幸福”单一完美结局;正义、安全、名声和自由之间存在真实取舍。
562
- - 隐藏结局强调过程条件,不要求收集所有物品或完成所有支线。
563
-
564
- ## 16. NPC 与对话规则 [P12]
565
-
566
- - NPC 以日程状态机运行:工作、巡查、休息、社交、应急。
567
- - 重要 NPC 记住三类事件:玩家公开行为、承诺是否兑现、是否牵连无辜者。
568
- - 对话提供观察、询问、转移、求助、试探、结束六类意图。
569
- - 玩家可先观察再对话;观察到的细节可解锁具体问题。
570
- - 角色不会凭空知道玩家在别处做的事,除非有明确传话路径。
571
- - 谣言传播延迟为 10–30 分钟游戏时间;秘密不会瞬间全宫知晓。
572
- - 同一问题重复盘问会使对方戒心 +10,除非出现新证据或新情境。
573
- - NPC 说谎时不使用统一红字提示;玩家需通过矛盾、动机和行为判断。
574
-
575
- ## 17. 美术、声音与 UI [P13]
576
-
577
- ### 17.1 美术方向
578
-
579
- - 色彩:朱红、黛青、暖金、旧纸白;不同区域通过灯温区分。
580
- - 构图:门框、窗格、屏风形成层叠遮挡,强调“看见与被看见”。
581
- - 角色轮廓:身份由袖口、腰牌、发饰与鞋履辨识,不靠头顶阵营色。
582
- - 交互提示:低饱和线框,不使用现代科幻扫描效果。
583
- - 环境变化:雨雪、烛火、雾气要影响玩法但不遮挡关键交互。
584
-
585
- ### 17.2 声音设计
586
-
587
- - 脚步按木廊、石地、地毯、湿地分别设计声纹。
588
- - 巡更锣、滴漏、风铃、膳房铜器形成时间与方向线索。
589
- - 音乐以弦乐、箫、打击乐低频为主;追查时加入节奏但避免持续压迫。
590
- - 证据发现使用短促细节声,不用夸张“正确答案”音效。
591
- - 听觉提示可视化,支持��障玩家;视觉线索也能完成同一谜题。
592
-
593
- ### 17.3 UI 页面
594
-
595
- - 任务页:当前目标、已确认事实、待验证假设分栏。
596
- - 线索板:人物、地点、时间、物证四种节点,可筛选关联。
597
- - 地图:门禁等级、光照、巡逻方向、可用身份和撤离点。
598
- - 伪装页:衣饰匹配度、差事解释、剩余合理停留时间。
599
- - 关系页:信任、戒心、承诺与可请求协助,不展示隐藏剧情变量。
600
- - 资源页:数量、来源、可替代取得方式。
601
- - 无障碍页:字幕、字幕说话人、文本大小、颜色模式、谜题提示强度。
602
-
603
- ## 18. 键盘鼠标操作与无障碍 [P14]
604
-
605
- - 键盘:WASD 移动,E 互动,Q 专注观察,C 蹲伏,Tab 打开线索板,M 打开地图,Shift 快走,Esc 暂停。
606
- - 鼠标:移动观察视角,左键确认 / 使用物件,右键取消 / 退出观察,滚轮切换可交互目标。
607
- - 所有按键可重映射;潜行支持按住或切换模式。
608
- - 提供字幕背景、字号、对比度、色盲方案与高亮交互轮廓。
609
- - 所有听觉线索均有可选视觉方向提示。
610
- - 谜题提示三级:主题提醒、已知条件整理、逐步提示;默认仅主题提醒。
611
- - 限时关卡允许延长 25% 或关闭倒数;不影响结局资格。
612
-
613
- ## 19. 失败恢复与反挫败 [P15]
614
-
615
- - 连续两次潜行失败后,第三次行动会出现可见的替代路线提示。
616
- - 任务关键 NPC 不会因普通警戒永久死亡或消失。
617
- - 同一区域被封锁后,提供等待、求助、合法申请或绕行四类恢复办法。
618
- - 玩家丢失伪装后可在储物间、屏风后或安全屋更换,不强迫重开关卡。
619
- - 证据损坏降低强度但不清除玩家已学到的信息。
620
- - 支线错过可通过后续书信或口述补全背景,避免收藏品式遗漏惩罚。
621
- - 重要状态变化给出具体原因,如“腰牌被登记为失窃”,不只弹出抽象失败提示。
622
-
623
- ## 20. AI 产出顺序 [P16]
624
-
625
- 1. 先生成全局系统:属性、警戒、身份、资源、关系、证据与结局变量。
626
- 2. 生成章节地图草图:标注门、巡逻、遮蔽、入口、撤离点和安全屋。
627
- 3. 为每章生成任务流程图,含至少两条路径与失败后恢复方案。
628
- 4. 生成角色日程表和对话状态,不允许 NPC 全程站桩。
629
- 5. 生成每个谜题的线索投放表、答案逻辑、替代解法与错误反馈。
630
- 6. 生成原创对话,标注说话意图、前置变量、关系变化和信息可信度。
631
- 7. 生成主线与支线的变量影响清单。
632
- 8. 生成难度与无障碍参数。
633
- 9. 最后输出可玩原型优先级:序章 + 第一章核心循环 + 线索板 + 一次结局变量结算。
634
-
635
- ## 21. 原型范围 [P17]
636
-
637
- - 预计游玩时长:25–35 分钟。
638
- - 场景:宫门、锦瑟院前庭、寝殿外廊、花房一角。
639
- - 可玩角色:沈知微;支援角色:孟兰汀、阿蘩。
640
- - 机制:移动、观察、潜行、身份检查、对话、线索板、简单解锁、警戒恢复。
641
- - 主谜题:香囊来源与值夜时间窗口。
642
- - 备选路线:女史差事、夜间屏风潜行、请求孟兰汀带路。
643
- - 最低内容:3 个 NPC 日程、2 套伪装、5 条线索、3 个结论节点、1 个资源兑换。
644
- - 原型完成标准:新玩家能在不看攻略的情况下说明自己为什么选择某条路线。
645
-
646
- ## 22. 范围与正向能力清单(创作边界,v3 正向化)[P18]
647
-
648
- - N01 本作的核心能力:观察、潜行、身份伪装、证据推理、线索板关联、关系经营与礼法交涉;冲突为高风险备选(§11),不构成刷怪或数值碾压玩法。
649
- - N02 谜题难度来自信息推理与空间观察;同一谜题均提供可观察线索与替代解法(§12.1)。
650
- - N03 通关条件仅依赖本作文内信息;不要求任何历史或既有影视作品知识。
651
- - N04 NPC 遵守尊卑、日程与已发生事件的记忆一致性;关系呈现为多维灰度而非善恶二元。
652
- - N05 失败恢复以补救与替代路线为主(§19);关键线索不因失败永久删除,不强迫重玩长段流程;克制路线获得等价内容与奖励。
653
- - N06 能力判定基于证据、身份、环境与关系状态;身高体重不参与潜行、魅力或战斗数值。
654
-
655
- ## 23. 区域数据示例 [P19]
656
-
657
- | 区域 | 可见度 | 噪声基线 | 门禁 | 巡逻密度 | 安全点 |
658
- |---|---:|---:|---:|---:|---|
659
- | 宫门外廊 | 80 | 15 | 低 | 低 | 引荐处 |
660
- | 锦瑟院庭院 | 55 | 10 | 低 | 低 | 内室 |
661
- | 寝殿外廊 | 35 | 8 | 中 | 中 | 屏风后 |
662
- | 花房 | 65 | 20 | 低 | 低 | 工具间 |
663
- | 尚仪局 | 70 | 12 | 高 | 中 | 值房侧室 |
664
- | 膳房 | 45 | 35 | 中 | 中 | 储物间 |
665
- | 宴席偏厅 | 50 | 45 | 高 | 高 | 屏风后廊 |
666
- | 太医院药房 | 60 | 18 | 高 | 中 | 诊候间 |
667
- | 冷宫外院 | 25 | 6 | 低 | 低 | 废置耳房 |
668
- | 御前校档廊 | 75 | 10 | 极高 | 高 | 无永久安全点 |
669
-
670
- ## 24. 证据目录示例 [P20]
671
-
672
- | ID | 证据 | 来源 | 初始强度 | 可验证方式 | 关键关系 |
673
- |---|---|---|---:|---|---|
674
- | E01 | 旧香囊残灰 | 锦瑟院 | 35 | 虞蘅对照香样 | 旧库批次 |
675
- | E02 | 值房轮值签半页 | 锦瑟院 | 45 | 页码与装订孔 | 当夜值守 |
676
- | E03 | 尚仪局排班抄本 | 尚仪局 | 60 | 笔迹与登记章 | 接触权限 |
677
- | E04 | 药材领用单 | 太医院 | 50 | 批号和剂量 | 药包来源 |
678
- | E05 | 宴席传膳单 | 膳房 | 55 | 出菜时辰 | 接触窗口 |
679
- | E06 | 封档火漆拓印 | 档房 | 65 | 裂纹方向与印章 | 档案被开启 |
680
- | E07 | 姚素秋口述片段 | 冷宫 | 25 | 节令与旧历交叉 | 旧案时间 |
681
- | E08 | 墙面修补痕迹 | 冷宫 | 40 | 灰浆层与火烧痕 | 逃生路线 |
682
- | E09 | 门房登记册 | 宫门 | 60 | 腰牌磨损特征 | 身份替换 |
683
- | E10 | 满仓路线草图 | 锦瑟院 | 30 | 现场步行复现 | 空间可达性 |
684
- | E11 | 中宫礼单副本 | 宫宴 | 45 | 版本差异 | 座位安排 |
685
- | E12 | 剪断的香囊线头 | 寝殿外廊 | 30 | 纤维显微对照 | 物品经手人 |
686
-
687
- ## 25. 规则优先级 [P21]
688
-
689
- - 剧情锁定 > 关键证据保护 > NPC 日程 > 玩家临时动作。
690
- - 同一 NPC 同时处于巡逻与对话时,结束对话后从最近合理节点继续日程。
691
- - 玩家在安全屋整理证据不消耗区域巡逻时间,但消耗章节时间推进。
692
- - 区域警戒 ≥75 时门禁升级;警戒下降需要离开视线并等待 60 秒或完成合法差事。
693
- - 怀疑值属于个人记忆,区域警戒属于群体反应;二者不可互相自动清零。
694
- - 借来的制服归还或遗失需记入关系与物资,不允许无限复制。
695
- - 同伴支援不可穿越封锁区瞬间出现;需明确展示其路线与耗时。
696
-
697
- ## 26. 最终生成命令 [P22]
698
-
699
- 请基于本设计先输出一个可玩的垂直切片,并严格遵守以下顺序:
700
- 1. 展示核心操作与一段原创开场剧情。
701
- 2. 提供锦瑟院周边俯视地图及门禁、视线、巡逻、可藏身点。
702
- 3. 实现至少两种潜入路线与一次身份盘问。
703
- 4. 实现香囊来源谜题,呈现可观察线索、线索板推理和错误反馈。
704
- 5. 实现一位同伴的可选支援,但不替玩家自动完成推理。
705
- 6. 结束时记录证据强度、警戒、关系与下一章解锁条件。
706
- 7. 提供开发者调试面板查看数值,但玩家界面不泄露隐藏变量。
707
- 完成垂直切片后,再按章节扩展内容;保持相同数值口径、身份规则、证据规则和人物行为逻辑。
708
- ---
709
- 设计文档版本:1.0|类型:第三人称宫廷潜行解谜互动游戏|语言:简体中文
710
-
711
- ## 27. 跨阶段后果表(L 档专节;早期状态在后续阶段仍可观察)
712
-
713
- - K01 阶段A(序章–第一章把「旧香囊残灰」等证据经营到佐证级,§7.2.3 算分链 30→54 落盘)→ 阶段B(第二章解锁条件「香料来源」或「值夜时间」达佐证级 50 分方可推进,§14.1;后续章节按证据假设解锁)→ 阶段C(第六章呈递风险随证据完整度降低,尾声结局组由 truth_score/evidence_integrity 等全局变量触发,§15;早期证据强度在终局仍可观察);验证方式=跨阶段回溯探针:阶段 A 记录证据强度快照 → 阶段 B 验证章节解锁事件 → 阶段 C 验证结局变量与分支对照。
714
- - K02 阶段A(早期支线与承诺积累关系信任,如虞蘅信任 ≥20 解锁「花房的旧账」,§14.2;关系五维落盘)→ 阶段B(信任 >40 提供额外路线、「暗线织网」阶段解锁同伴支援与两点调度,§9/§14.3)→ 阶段C(尾声 ally_safety 与隐藏结局「春信归人」由关键人物信任、证人自愿与低伤害路线共同决定,§15.2;早期关系投资在终局名单仍可观察);验证方式=阶段 A 关系快照 → 阶段 B 支援解锁事件 → 阶段 C 结局变量对照。
715
-
716
- ## 28. 复杂度档案
717
-
718
- ```yaml
719
- difficulty:
720
- tier: L-Regular
721
- complexity_profile: {coupling: C2, persistence: P3, recovery: R2, runtime_complexity: B1, hidden_state: H2}
722
- dominant_dimensions: [persistence, hidden_state]
723
- rationale: "跨阶段长依赖:证据强度积累→章节解锁→结局分支(K01)与关系信任投资→同伴支援→尾声结局变量(K02)两条 A→B→C 链;章节/全局变量/关系/证据/名声跨章节持久+自动与手动存档(P3);身份伪装-侦测-警戒、证据-推理-交涉多系统耦合(C2);NPC 记忆、谣言延迟与说谎不提示构成多层信息不对称(H2);回合式判定无实时压力(B1);安全屋存档/失败补救/反挫败多级恢复(R2);新增演出、视频与阶段重开均为表现与恢复层,不改变档位"
724
- downgrade_boundary: "删去跨章节持久与全局变量结算(每章独立、结局与早期选择无关)后降为 M"
725
- upgrade_boundary: "加入多人同步或组合状态压力轴并 ≥3 系统相互影响时升 L-Stress"
726
- stress_axes: []
727
- ```
728
-
729
- ## 29. 演出、镜头���反馈补充 [F01–F08](在 §17/§18 基础上冻结事件驱动反馈)
730
-
731
- - F01 视角与镜头:沿用 §5(第三人称)与 P14 键鼠设定;补充:观察/线索板打开时镜头轻微拉近 5%(0.5 秒);被发现追查时镜头轻微拉远 8% 增大视野;镜头边界=区域碰撞体;镜头辅助回正可按无障碍页设置关闭。
732
- - F02 关键操作镜头变化:证据强度提升确认时镜头轻微推近线索 3%+定格 0.3 秒;警戒升到 75(搜查封门)时小幅震动 0.2 秒;抓捕失败黑屏渐入走 §19 恢复流程;无慢动作。
733
- - F03 输入与延迟目标:移动/交互按键到画面反馈 ≤ 80 毫秒;面板操作 ≤ 100 毫秒。
734
- - F04 环境常驻效果(在 §17.1 基础上冻结):灯影摇曳(烛火投射)、屏风透光剪影、纸页风动、宫门灯廓、夜雨瓦檐滴水(雨夜章节)、秘密通道尘埃光柱;均不得遮挡关键交互物(§17.1 口径保留)。
735
- - F05 事件反馈表(事件 → 表现 → 结束条件,全部为冻结需求;音效引用 §17.2 声纹设计):
736
-
737
- | 事件 | 画面表现 | 音效 | 结束条件 |
738
- |---|---|---|---|
739
- | 拾取证据 | 纸页/物证描边高亮+入档飞入线索板 | 短促细节声(§17.2) | 登记完成 |
740
- | 证据强度提升 | 线索节点光色升档+强度数字定格 | 细节确认声 | 0.3s 定格 |
741
- | 伪装换装 | 衣饰逐件着身+匹配度逐项打钩 | 布料摩擦声 | 完成 |
742
- | 盘问对答 | 屏风剪影停顿+对话字幕微缩 | 纸页/衣料声 | 对话结束 |
743
- | 侦测留意 | 目光锥高亮 2 秒(§P02 留意态) | 短促注意音 | 确认或解除 |
744
- | 警戒升级 | 区域灯色升档+更鼓一响 | 更鼓/锣声(对应层级) | 常驻至回落 |
745
- | 挣脱/制服 | 动作残影+看守踉跄 | 布料与闷响 | 动作完成 |
746
- | 关系变化 | 关系页徽记明暗变化 | 玉佩轻碰声 | 常驻 |
747
- | 章节切换 | 夜色转场+章名卷轴展开 | 转场弦乐短句(一次性) | 转场完成 |
748
- | 结局结算 | 判词卷轴展开(按结局组) | 终章箫声短句 | 进入结局面板 |
749
-
750
- - F06 关键音效补充清单(与 §17.2 并存):细节确认声、注意音、更鼓/锣(警戒四级分档)、玉佩轻碰、转场短句、判词短句;循环音开始/结束:环境声按区域载入、追查节奏乐按追查态起止、暂停全部静音。
751
- - F07 VFX 信息性与遮挡约束:目光锥与警戒灯色承担侦测信息(同时附「留意/确认/追查」文字提示);雨雾烛烟不遮挡交互物与线索板(§17.1 口径);追查粒子和烟粒单帧 ≤ 45。
752
- - F08 音画同步:更鼓与警戒事件帧同步;证据确认声与强度定格帧同步;判词短句与结局事件帧同步。
753
-
754
- ## 30. 过场视频 [C01–C05](L 档 ≥4;按章节结构冻结)
755
-
756
- - C01 开场·宫墙夜景(Opening,12 秒,1280×720@24fps)
757
- - 触发时机:首次由标题进入序章之前,每局一次
758
- - 内容与角色场景:宫墙夜色,檐灯次第亮起,沈知微提灯立于锦瑟院外廊,远处更鼓声中一册名册被暗中替换
759
- - 资产对应:宫墙、锦瑟院、沈知微与游戏内模型同源设计
760
- - 可跳过:Esc/Enter 跳过 → 直接进入序章;自然播完 → 同
761
- - 状态一致:播放期间不推进任何游戏计时;进入状态=序章初始
762
- - C02 区域转换·尚仪局的页码(Transition,8 秒,1280×720@24fps)
763
- - 触发时机:第二章解锁事件后的下一帧短暂插入,每局一次
764
- - 内容与角色场景:晨光中的尚仪局书廊,排班册页角被翻起,墨迹未干的抄本旁落着一枚登记章
765
- - 资产对应:尚仪局、抄本、印章与游戏内一致
766
- - 可跳过:Esc 跳过 → 回到探索;自然播完 → 同
767
- - 状态一致:仅在真实解锁后触发,返回时章节状态不变
768
- - C03 区域转换·冷宫旧档(Transition,8 秒,1280×720@24fps)
769
- - 触发时机:第五章(冷宫旧档)进入时短暂插入,每局一次
770
- - 内容与角色场景:冷宫荒院暮色,断墙修补痕迹与新灰浆并置,尘光中一页残档被风掀起
771
- - 资产对应:冷宫、断墙、残档与游戏内一致
772
- - 可跳过:Esc 跳过 → 回到探索;自然播完 → 同
773
- - 状态一致:进入状态与章节解锁条件一致
774
- - C04 真相揭示·御前校档(Event,10 秒,1280×720@24fps)
775
- - 触发时机:第六章御前呈递关键证据(truth_score 达呈递线)事件后短暂插入,每局一次
776
- - 内容与角色场景:御案灯下,篡改页与原档并排铺开,火漆裂纹方向对上封印,殿外雨声骤停
777
- - 资产对应:御前校档廊、证据物、火漆拓印与游戏内一致
778
- - 可跳过:Esc 跳过 → 回到呈递流程;自然播完 → 同
779
- - 状态一致:仅在真实呈递事件后触发,证据状态不变
780
- - C05 结局结算(Ending,12 秒,1280×720@24fps,尾帧按结局组变体)
781
- - 触发时机:任一��局组触发后、结局面板之前,每局一次
782
- - 内容与角色场景:判词卷轴展开;尾帧按 §15.2 六结局组呈现(封存/明证/各自求生/以名换路/回声/春信归人)
783
- - 资产对应:卷轴、宫殿、关键人物与游戏内一致
784
- - 可跳过:Esc/Enter 跳过 → 直接进入结局面板;自然播完 → 同
785
- - 状态一致:尾帧与真实触发的结局组一致
786
-
787
- ## 31. 阶段重开与历史状态保留 [P23]
788
-
789
- - P23.1 章节级重开:任一章节可从章节入口「重开本章节」——本章内证据强度变动、警戒与关系变化回本章开始时快照;已完成章节的证据、关系、名声与全局变量不回退(部分历史状态保留)。
790
- - P23.2 全局重开:「重新开始」经二次确认后建立新局,全量清空(seed=1055 复现抽查口径保留)。
791
- - P23.3 存读档语义:沿用 G03 自动与手动存档;章节入口自动存档点在章节开始帧;读档恢复全量快照(diff=0)。
792
- - P23.4 恢复边界:章节级重开不删除已学到的线索信息(线索板「已确认事实」保留),仅回退数值状态;与 P15 失败补救互不冲突。
793
-
794
- ## 生成任务明确性补充(2026-10-12)
795
-
796
- > 本补充只约束游戏生成输入,不引用 rubric。所有生成器必须以本文件规则和下列 JSONL 任务为准;未列出的装饰细节可以自由发挥,但不得省略已列资产或生命周期。
797
-
798
- ### 故事与流程最低合同
799
- - **玩家身份与当前冲突**:以 §1、§8 的角色/世界/冲突为唯一叙事入口;开局必须让玩家知道“我是谁、现在要解决什么问题”。
800
- - **阶段性目标**:必须有可操作的阶段目标、至少一次状态变化/升级,并在 §7 的胜利、失败、重开入口中闭环;不能只生成一个无结果的场景演示。
801
- - **结果承接**:胜利/失败都要有明确结果文案或演出,并回到可重开、暂停或标题入口;视频、音频和 VFX 不得推进未完成的玩法状态。
802
-
803
- ### 生成任务清单(JSONL 冻结口径)
804
- - scene:1 个;必须生成连贯可游玩的空间,不得出现下半部文字、拼贴面板、物体 lineup 或 UI。
805
- - object:4 个(text-to-3D 2,image-to-3D 2);object 任务资产为:shen_zhiwei, evidence_pouch_ash, rota_manuscript, seal_wax_imprint。每项只允许一个物体;组合物必须拆成独立任务。
806
- - motion:4 个;覆盖动作:idle, walk, crouch_sneak, evidence_examine。
807
- - video/audio/mechanic/UI:分别为 5/10/1/1 个任务;若某项为 0,只能是本案例明确的 N/A,不得默默遗漏。
808
-
809
- ### Video / motion / audio 生命周期
810
- - Video 每项必须写清 `trigger`、`start_condition`、`duration_sec`、`skippable`、`skip_behavior`、`end_condition` 和进入的下一个游戏状态;播放期间不得偷偷消耗计时或改变战斗/任务数值。
811
- - Motion 每项必须写清 `source_asset`、`trigger`、`start_condition`、`duration_sec`、`loop`、`end_condition` 和 `cancel_behavior`;动作要与对应输入/机制事件绑定,不可只写“walk animation”。
812
- - Audio 每项必须写清触发事件、时长、`loop`、`stop_condition`、优先级和 ducking;关键命中/提示音必须与事件帧同步,暂停/结算时按 §9 处理。
813
-
814
- ### Mechanic / UI / 游玩手感
815
- - Mechanic JSON 与 §1–§7 配对:每个输入都要有 precondition、state_change、feedback、next_state;必须说明失败/禁用分支。
816
- - UI JSON 与 §9/F10 配对:HUD、菜单、结算、暂停、错误/禁用态要有 state binding;关键状态不能只用颜色表达。
817
- - 手感冻结:相机跟随/构图、FOV、碰撞/遮挡、镜头调整与镜头震动必须在 §9 说明;震动反馈必须写明触发、强度和持续时间,并提供关闭设置且不改变判定。
818
-
819
- ### VFX 时机合同
820
- 每个特效必须写清 `trigger event`、预备/前摇、active window、持续时间、结束条件、被打断/取消条件,以及和相机震动、音效、UI 的同步关系;命中特效、失败特效、胜利特效不能只写“有粒子”。
821
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/method_selection.md DELETED
@@ -1,152 +0,0 @@
1
- # 优先资产:方法选择与执行边界
2
-
3
- 核对日期:2026-10-15。本文是候选方法建议和开跑前检查,不是实验结果,也不是已验证的
4
- API 配置。评测口径以 `GameFactory-3A/eval/asset_evaluation_plan.md` 为准;
5
- 本次没有改动该方案、原始全集或其他五份现有优先生成 JSONL。
6
-
7
- ## 1. 建议首轮组合
8
-
9
- | 类别 | 首轮建议 | 放在哪组 | 选择理由 / 限制 |
10
- |---|---|---|---|
11
- | 3D Object | Meshy + Tripo + 腾讯云 Hunyuan3D Pro 商用 API | B1 商用生成 API;逐版本核对开放权重状态 | 三家即可;人物/刚体、T2/I2 分组,保持同输入和低模预算 |
12
- | 3D Scene | 新增 World Labs Marble World API;保留 WorldPlay/HY-World 2.1 Scene → WorldMirror → 导出,以及 LLM 布局/组合 | Marble 为 B2 商用 direct 候选;另两路为组合方法对照 | Marble 提供世界生成和 mesh 导出;不拿仅视频输出替代 3D 交付 |
13
- | 自然音 / SFX(看情况吧) | ElevenLabs Sound Effects v2 + Stable Audio 3.0 Large 商用 API | B4 Environment / Foley 子组 | 一个面向音效、一个通用音频基线;先用声音事件和环境音 pilot 验证,不跑成音乐任务 |
14
- | 中文人物语音 | MiniMax Speech(优先验证 `speech-2.8-hd`)+ ElevenLabs TTS 的固定中文支持版本 | B4 Speech 子组 | 当前 38 条都是中文;固定说话人、情绪映射和完整台词 |
15
- | CG | Seedance + MiniMax Hailuo(优先验证 `MiniMax-Hailuo-2.3`) | B3;T2V 与 I2V 分开 | 先跑两家即可,不急着加入第三家;按 API 的时长/分辨率能力冻结共同条件 |
16
- | Motion | MoMask + Puppeteer 对比 LLM 直出;可再加 Meshy Text-to-Motion | 前两路是开放模型链 / 系统对照;Meshy 是 B5 商用 API 候选 | 先满足用户要的两路比较;若 B5 必须有闭源专用生成 API,再加第三路 |
17
- | T-pose | 同一份冻结准备输入 | 不新增主表 | 两个动作方法共享角色输入,不把图像准备差异算成动作模型收益 |
18
-
19
- 商用 API 不自动等于闭源模型。特别是 Hunyuan3D 必须登记云端实际版本,不能把本地
20
- 开放权重版本包装成闭源行;MiniMax 的视频实验也要写清 Hailuo 的具体 ID,不能只写品牌。
21
- MoMask、Puppeteer、WorldPlay、WorldMirror 的源码/模型发布可在官方仓库核对 [R1–R4]。
22
- 上述 API 的能力入口见文末官方来源;账户权限和实际可用版本仍需 pilot 确认。
23
-
24
- **规模建议**:B1 三家、CG 两家、音频按两种任务各两家;场景优先补 Marble,而不是
25
- 为了凑三列加入不能导出场景的模型。B2 暂时只有一个确认能力范围的商用 direct 候选,
26
- 如需至少两家闭源 direct API,仍须找到并实测第二家,不能用组合方法假充。
27
-
28
- ## 2. 为什么按这些方法分组
29
-
30
- 原方案 §1/§4 要求闭源专用生成 API、开放模型与多模型组合分开,首轮主表不混排。
31
- 因此保留用户提出的系统比较,但不悄悄改掉主表含义:
32
-
33
- - B1/B3/B4:按相同任务和输入条件比较商用服务,记录模型的权重开放状态。
34
- - B2:Marble direct 与 WorldPlay + WorldMirror / LLM composition 分 panel 报告。
35
- - B5:MoMask 生成人形动作,Puppeteer 在本项目链路中主要承担绑骨/蒙皮准备;
36
- LLM 直出是另一条系统方法。不能把它们标为两个闭源 motion API。
37
- - 如果研究问题改成“各种资产生成方法谁更适合游戏”,可以增加系统对照表,
38
- 但需要显式修改协议,而不是沿用 `Closed-source method` 表头。
39
-
40
- 为了隔离动作生成能力,建议先冻结一套人形骨架和白模,两路动作采用同一个映射、
41
- retargeter、20 fps 和接触检查;角色网格、绑骨差异另在第三阶段端到端比较。
42
- LLM 直出必须交付轨迹/骨骼动画文件,只有 Blender 脚本、文字或预览视频不算交付成功。
43
-
44
- ## 3. 与当前优先清单对齐的阻塞项
45
-
46
- ### 3D Object:低模与上下文能力
47
-
48
- 现有 200 条保留低模 prompt,且 `decimation_target` 最高 8000。
49
- 建议统一保存原始 API mesh 和固定后处理后的导入 mesh,分别报告;不要只给一家减面,
50
- 也不要把减面后生成的新几何缺陷完全归给原始模型。
51
-
52
- Hunyuan3D Pro 的普通模式中 `Prompt` 与图像输入互斥,不能假设 I2 能同时传游戏梗概。
53
- 这类调用按原方案记为 `reference-only` 辅助组;T2 可以拼接冻结的短版梗概。
54
- 其 3.0/3.1 的 LowPoly 支持不同,先确定使用的版本和后处理政策 [R5]。
55
- 其他 I2 API 也逐一核对文本上下文能力,不能只在 judge 阶段补梗概。
56
-
57
- 不要用 `latest` 当作冻结版本。Meshy 文档还提示其 `lowpoly` 模型类型将在
58
- 2026-10-30 退役;现有清单的 `poly_policy: low_poly_priority` 只是实验政策,
59
- 不等于必须调用这个即将退役的 API 类型 [R6]。
60
-
61
- ### 3D Scene:交付要有真实场景
62
-
63
- 69 条现有场景任���要求 `output_format: glb`。WorldPlay 的视频需经 WorldMirror 重建
64
- 及场景导出后,才与场景任务对齐;只生成视频不能记 Delivery 成功。
65
-
66
- Marble 的世界接口返回 splats 和 collider mesh,并另提供高质量带纹理 GLB 导出。
67
- 建议正式 B2/第三阶段使用后者,保存碰撞网格和世界尺度信息;不要拿无外观的 collider
68
- mesh 冒充完整美术场景,也不要把 splat 文件简单改后缀作为 GLB [R7]。
69
- 生成和导出都计入耗时/成本;若导出不可用,保留该任务的未交付状态,不替换成全景图。
70
-
71
- LLM composition 固定资产池、资产提供者、坐标系统、规划调用数和修补预算;
72
- 若它额外调用对象生成器,费用和调用全部计入系统方法,不能只记 LLM 的 token 成本。
73
-
74
- ### Audio:需要按声音任务选方法,不是一种模型跑全部
75
-
76
- 当前非语音 706 条:623 条 `one_shot`、83 条 `ambient_loop`;语音 38 条,全部中文。
77
- 这是分类字段统计;实际循环要求仍以每条任务的 `loop` / 生命周期字段为准。
78
- 非语音时长为 0.15–45 秒,其中 93 条短于 0.5 秒、281 条短于 1 秒、19 条长于 30 秒。
79
-
80
- ElevenLabs SFX API reference 写 0.5–30 秒,能力总览却写 0.1–30 秒,存在官方文档差异;
81
- 开跑前应实测短音边界并保存实际请求/响应,不自行认定所有短音都支持 [R8, R9]。
82
- Stable Audio 的 API 时长范围也应从具体 endpoint 核对。这里建议验证 3.0 Large 商用服务,
83
- 而不是把已开放权重的 3.0 Medium / Small / Small SFX 当作闭源方法 [R10]。
84
- 若账户仍使用 2.5 API,则固定并单列 2.5 版本,不混写成 3.0 的结果 [R16]。
85
- 共同能力子集用于配对排名;其余任务仍留在第三阶段完整清单中,标注能力缺口。
86
-
87
- 解码、重采样到要求的 48 kHz / 24 kHz 可采用同一确定性工具并保存原件。
88
- 短音裁切、长音拼接/循环、生成式延长不是无代价格式转换:如要使用,须预先冻结规则,
89
- 单列 pipeline 结果及成本,不能悄悄改任务时长来让 API 通过。
90
-
91
- 语音不得把“游戏梗概 + 角色介绍”拼进需要念出的台词。检查 API 是否有独立的
92
- instruction/style 通道;没有时仅可使用冻结的 voice/emotion 参数映射,单列标准 TTS
93
- 条件,不宣称模型直接读取了梗概。MiniMax 的 TTS 参数和 ElevenLabs 版本入口见 [R11, R12]。
94
-
95
- ### CG:现有任务与 reference-image 协议不一致
96
-
97
- 194 条中 190 条 `text_to_video`、4 条 `image_to_video`,但当前全集和优先清单都没有
98
- 参考图路径字段。不能用这批数据直接声称完成“194 条 reference-image-to-video”评测。
99
-
100
- 建议明确选择:保留 T2V 子表、Q1 预声明 N/A;同时为 I2V 冻结真实的共同参考帧,
101
- 补齐路径和来源后再运行 Q1。若要全部改成 I2V,应生成并冻结共享参考图,然后形成新的
102
- 任务版本;不能各视频方法自行画图作为自己的参考。
103
-
104
- 现有时长 4–12 秒、输出 1280×720/24 fps;Hailuo 2.3 文生接口的原生规格是
105
- 768P 的 6/10 秒或 1080P 的 6 秒,不能把原 JSON 原样发出去 [R13]。
106
- 先冻结共同能力 pilot 子集;非原生任务保留并标注,不隐式拉伸视频来满足事件时长。
107
- Seedance 使用已接入的实际 model ID,先验证能力矩阵,不将品牌名当成版本。
108
-
109
- ### Motion:保留非人形与短动作
110
-
111
- 新增 78 条中,76 条人形、G040 一条 vehicle、G030 一条 mechanical。
112
- 两个非人形任务必须保留用于第三阶段,不能因为 MoMask 主要面向人形就删去。
113
- 按 rig group 单列,不支持时明确记录;需要 procedural 交付时另列方法,不冒充 MoMask 结果。
114
-
115
- Meshy Text-to-Motion 的官方接口提供自然语言动作和 FBX/BVH,当前时长要求为
116
- 2–10 秒且按 0.5 秒步长 [R14]。它适合补商用动作 API 对照,但不是现有所有短动作的
117
- 无缝替换。LLM 直出也固定同一骨架、帧数和时长,不只比较好看的视频。
118
-
119
- ## 4. 最小执行顺序
120
-
121
- 1. 先为 13 个 L 级游戏准备第三阶段的全部既有资产任务;完整性与模型共同能力子集是两件事。
122
- 2. 选 G003/G025/G040 等 3–5 个游戏做 pilot,覆盖中文语音、RTS 压力、机械/载具边界。
123
- 每类先挑少量代表任务;这不缩减 L 级正式导入清单。
124
- 3. 冻结游戏梗概、任务/参考输入、适用指标、API 版本、调用次数、超时重试和后处理。
125
- 4. 报 Delivery、C、Q1/Q2/Q3、Coverage 和成本,按游戏宏平均;不挑 best-of,也不将
126
- 未测能力缺口伪装成 API 生成失败。正式分母政策在开跑前固定。
127
- 5. 第三阶段各 repair 方法共享同一冻结资产包。提前固定每类资产的交付来源,不能按
128
- 单条 judge 最高分挑选,也不能悄悄用 LLM 补产物来遮盖某个生成方法的失败。
129
- 若比较资产来源,另设因子并保持 repair 方法与预算一致。
130
-
131
- ## 官方来源
132
-
133
- - [R1: MoMask](https://github.com/EricGuo5513/momask-codes)
134
- - [R2: Puppeteer](https://github.com/Seed3D/Puppeteer)
135
- - [R3: HY-WorldPlay](https://github.com/Tencent-Hunyuan/HY-WorldPlay)
136
- - [R4: HunyuanWorld-Mirror](https://github.com/Tencent-Hunyuan/HunyuanWorld-Mirror)
137
- - [R5: Hunyuan3D Pro API](https://intl.cloud.tencent.com/ind/document/product/1284/75540)
138
- - [R6: Meshy Text-to-3D](https://docs.meshy.ai/en/api/text-to-3d)
139
- - [R7: World Labs World API](https://docs.worldlabs.ai/api)
140
- - [R8: ElevenLabs Sound Effects API](https://elevenlabs.io/docs/api-reference/text-to-sound-effects/convert)
141
- - [R9: ElevenLabs Sound Effects 能力总览](https://elevenlabs.io/docs/overview/capabilities/sound-effects)
142
- - [R10: Stable Audio 模型及商用 API 入口](https://stability.ai/stable-audio)
143
- - [R11: MiniMax TTS](https://platform.minimax.io/docs/api-reference/speech-t2a-http)
144
- - [R12: ElevenLabs 模型](https://elevenlabs.io/docs/overview/models)
145
- - [R13: MiniMax Hailuo T2V](https://platform.minimax.io/docs/api-reference/video-generation-t2v)
146
- - [R14: Meshy Text-to-Motion](https://docs.meshy.ai/en/api/text-to-motion)
147
- - [R15: Tripo 开发者入口](https://developers.tripo3d.ai/)
148
- - [R16: Stable Audio 2.5 官方发布与 API 入口](https://stability.ai/news-updates/stability-ai-introduces-stable-audio-25-the-first-audio-model-built-for-enterprise-sound-production-at-scale)
149
-
150
- Seedance 的本地封装 `GameFactory-3A/models/gen_cg_video/seedance_model.py` 已登记
151
- model ID 和官方文档入口;本次网页抓取未能直接读取该火山引擎文档,因此不在本文
152
- 宣称已核对其全部时长/分辨率规格,开跑前仍需 API pilot。
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/object_requirement.txt DELETED
@@ -1,9 +0,0 @@
1
- # G055_inner_court_investigation / 3D_object requirement
2
-
3
- Source of truth: ../general_requirement.txt
4
-
5
- Frozen task inputs:
6
- - shen_zhiwei: Original young court investigator Shen Zhiwei: understated dark-teal commoner robe with faint embroidery, hair in a plain bun with wooden pin, cloth shoes, sharp calm gaze, stylized semi-realistic ancient court style, vermilion-and-indigo palette, T-pose, game-ready, PBR textures
7
- - evidence_pouch_ash: Small scorched silk incense pouch with charred rim and faint residual ash, tied with a cut thread, ancient court stylized, game-ready, PBR textures
8
- - rota_manuscript: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures
9
- - seal_wax_imprint: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/localization_integrity.json DELETED
@@ -1,211 +0,0 @@
1
- {
2
- "created_at": "2026-10-16T11:11:05+00:00",
3
- "parent_plan_sha256": "5c287b8399f00fbb0dddf94c9e3f5ffcbcc2e679ff76882bc8fc369d6eae840d",
4
- "original_records_archive": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/records_before_localization_35f1ef02271bc9d9.zip",
5
- "archive_sha256": "3a4589dae5b7ff0df7e9065c9ca973a213a8a195d8bce0c8a6fa3fa74695ed80",
6
- "records": [
7
- {
8
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/meshy/obj200_meshy_tripo_20261015_v2/asset_task_index.json",
9
- "before_sha256": "d149b0e5f4f50fb1c2c385fb58dbf86013a76e4dc2d83ab4265b0b33105f365f",
10
- "after_sha256": "d149b0e5f4f50fb1c2c385fb58dbf86013a76e4dc2d83ab4265b0b33105f365f"
11
- },
12
- {
13
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/tripo/obj200_meshy_tripo_20261015_v2/asset_task_index.json",
14
- "before_sha256": "b0d21e024c4d8295e97ec0b30e0e048ee0e56a494c8e02882a38434d2dca932d",
15
- "after_sha256": "b0d21e024c4d8295e97ec0b30e0e048ee0e56a494c8e02882a38434d2dca932d"
16
- },
17
- {
18
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_001/meshy/meta.json",
19
- "before_sha256": "2e9521eefcf93af399d7ed6211d7fb10e61e5340aa9c4ac26e82cc76da4cb488",
20
- "after_sha256": "2e9521eefcf93af399d7ed6211d7fb10e61e5340aa9c4ac26e82cc76da4cb488"
21
- },
22
- {
23
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_001/tripo/meta.json",
24
- "before_sha256": "0a1a78c78deca90cd96baea99b6517f284da9572b4d633fd3590188479aafdbb",
25
- "after_sha256": "430c3d67b91e4259672399707c886659d6a3a94456d303cbcbb5c569cfdd6613"
26
- },
27
- {
28
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_002/meshy/meta.json",
29
- "before_sha256": "06b65bfe9aded33ff643118a6a4383eb7811a1d1d2f5a500e42478052cf690d4",
30
- "after_sha256": "06b65bfe9aded33ff643118a6a4383eb7811a1d1d2f5a500e42478052cf690d4"
31
- },
32
- {
33
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_002/tripo/meta.json",
34
- "before_sha256": "94f5b15a354d1f392d09120e822b4b33bd50647924ef697ef82fcc36021611f2",
35
- "after_sha256": "9a9d569bf378dc4d5adf10e1e8b8d0150a79df870f9fa97e1b79de61b6372d19"
36
- },
37
- {
38
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_003/meshy/meta.json",
39
- "before_sha256": "417218fe9c0192bef8b4118407f8d50ce54250a2fd56c23b96776153614d74d1",
40
- "after_sha256": "417218fe9c0192bef8b4118407f8d50ce54250a2fd56c23b96776153614d74d1"
41
- },
42
- {
43
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_003/tripo/meta.json",
44
- "before_sha256": "5bcdd910b7a7ad1dee58b66ae31397b126e9368e6d4ce7f98e27cfc837778f9d",
45
- "after_sha256": "d704917d50c76d11294c20485800ce810d343c43969aaab7478694478e08a7a3"
46
- },
47
- {
48
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_004/meshy/meta.json",
49
- "before_sha256": "94057a934a42351449349015d7dbc7de03eee4a0bc7b05424f1d95b5b78b2a11",
50
- "after_sha256": "94057a934a42351449349015d7dbc7de03eee4a0bc7b05424f1d95b5b78b2a11"
51
- },
52
- {
53
- "path": "bench/test_data/outputs/G055_inner_court_investigation/assets/3d_object/G055_inner_court_investigation_obj_004/tripo/meta.json",
54
- "before_sha256": "9e880e77fad00b8adedb997f9f21149cd0c667787e5f4ef23ac1271305c2ee4d",
55
- "after_sha256": "2bacfcb8b5d9769920e36d3eff8ee622b850ab2efcdb1cbaea1a8ce58fa0bca8"
56
- },
57
- {
58
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/render_meta.json",
59
- "before_sha256": "bae92d75d635e458d01bf2cc3faf7c612598204a3bd004e854bd2071ea635515",
60
- "after_sha256": "bae92d75d635e458d01bf2cc3faf7c612598204a3bd004e854bd2071ea635515"
61
- },
62
- {
63
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/native/render_meta.json",
64
- "before_sha256": "6d9d09c8fedefd8565f7a36d8290e2fd59ac9d7d2e7284c9a3034c0c65bef9b5",
65
- "after_sha256": "6d9d09c8fedefd8565f7a36d8290e2fd59ac9d7d2e7284c9a3034c0c65bef9b5"
66
- },
67
- {
68
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/qa.json",
69
- "before_sha256": "86ed4bbec1e754c198b512db227ead3ecba746afc8a892adccde44b3c1834b6c",
70
- "after_sha256": "86ed4bbec1e754c198b512db227ead3ecba746afc8a892adccde44b3c1834b6c"
71
- },
72
- {
73
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/tripo/previews/import/render_meta.json",
74
- "before_sha256": "7295be4167d266e3a6ce7431eb5a4dc9b1dd178513c7e3a52dd2fd39960966d5",
75
- "after_sha256": "7295be4167d266e3a6ce7431eb5a4dc9b1dd178513c7e3a52dd2fd39960966d5"
76
- },
77
- {
78
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/tripo/previews/native/render_meta.json",
79
- "before_sha256": "ccd82d6c47a10d9316346ce16838e66f8df8cd69cba2afb048059372ec34bba3",
80
- "after_sha256": "ccd82d6c47a10d9316346ce16838e66f8df8cd69cba2afb048059372ec34bba3"
81
- },
82
- {
83
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/tripo/qa.json",
84
- "before_sha256": "dbc2a797f167dc5a3bd705833f12d8abe9b862ead78af7cefaf804fc0430f606",
85
- "after_sha256": "dbc2a797f167dc5a3bd705833f12d8abe9b862ead78af7cefaf804fc0430f606"
86
- },
87
- {
88
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_002/meshy/previews/import/render_meta.json",
89
- "before_sha256": "8b05c7048d95fa0ab98fe8e169c5ab0c8753321427c8511b9587530a03994f36",
90
- "after_sha256": "8b05c7048d95fa0ab98fe8e169c5ab0c8753321427c8511b9587530a03994f36"
91
- },
92
- {
93
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_002/meshy/previews/native/render_meta.json",
94
- "before_sha256": "1c416380f959ce6e2f061493478fcdebdbda3360478c79f7dd3304f254b636a8",
95
- "after_sha256": "1c416380f959ce6e2f061493478fcdebdbda3360478c79f7dd3304f254b636a8"
96
- },
97
- {
98
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_002/meshy/qa.json",
99
- "before_sha256": "19495bb510ec72bc99ec8cbcbcbdac5bad8715882ea6e1873b70b9673d5debe3",
100
- "after_sha256": "19495bb510ec72bc99ec8cbcbcbdac5bad8715882ea6e1873b70b9673d5debe3"
101
- },
102
- {
103
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_002/tripo/previews/import/render_meta.json",
104
- "before_sha256": "0fce4caa1517ddff8a159da66367db25beee202cd3a80cee32e7de44716c9e11",
105
- "after_sha256": "0fce4caa1517ddff8a159da66367db25beee202cd3a80cee32e7de44716c9e11"
106
- },
107
- {
108
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_002/tripo/previews/native/render_meta.json",
109
- "before_sha256": "93aac75e5b3830594b4279be9176c868c0f9760ceda6b5328fa5120cdab0bd02",
110
- "after_sha256": "93aac75e5b3830594b4279be9176c868c0f9760ceda6b5328fa5120cdab0bd02"
111
- },
112
- {
113
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_002/tripo/qa.json",
114
- "before_sha256": "d72ca5fbdcb8c286a7795d7c924bbfc2f5b9798ff4c17a0cf24552d3cc2c4b74",
115
- "after_sha256": "d72ca5fbdcb8c286a7795d7c924bbfc2f5b9798ff4c17a0cf24552d3cc2c4b74"
116
- },
117
- {
118
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_003/meshy/previews/import/render_meta.json",
119
- "before_sha256": "dd8922198b86b694f46437dbf8ca3079d1b849fa0695d16817be6b9f1c9dac61",
120
- "after_sha256": "dd8922198b86b694f46437dbf8ca3079d1b849fa0695d16817be6b9f1c9dac61"
121
- },
122
- {
123
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_003/meshy/previews/native/render_meta.json",
124
- "before_sha256": "5b6eeda9b692f3d099c7eddc8facc54540f259eb71f83482b26d9af6cba0cf6f",
125
- "after_sha256": "5b6eeda9b692f3d099c7eddc8facc54540f259eb71f83482b26d9af6cba0cf6f"
126
- },
127
- {
128
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_003/meshy/qa.json",
129
- "before_sha256": "94e9e092383298f309f53b39fa2446dd5b49ac47ce1a2565b03ad228215272b4",
130
- "after_sha256": "94e9e092383298f309f53b39fa2446dd5b49ac47ce1a2565b03ad228215272b4"
131
- },
132
- {
133
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_003/tripo/previews/import/render_meta.json",
134
- "before_sha256": "60ce7d88f68108dba0dd451875422cfa0754f3ab25ea06a167c92049800d74bd",
135
- "after_sha256": "60ce7d88f68108dba0dd451875422cfa0754f3ab25ea06a167c92049800d74bd"
136
- },
137
- {
138
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_003/tripo/previews/native/render_meta.json",
139
- "before_sha256": "9362abafab97078703cf606c2317ec8c0c42976760ee2ad7e6a65a155e9dfc92",
140
- "after_sha256": "9362abafab97078703cf606c2317ec8c0c42976760ee2ad7e6a65a155e9dfc92"
141
- },
142
- {
143
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_003/tripo/qa.json",
144
- "before_sha256": "e02387a10048f5a249e78fc1b86faa532d2b97c18996539a8dc4d8f794991f30",
145
- "after_sha256": "e02387a10048f5a249e78fc1b86faa532d2b97c18996539a8dc4d8f794991f30"
146
- },
147
- {
148
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_004/meshy/previews/import/render_meta.json",
149
- "before_sha256": "b58da74630c5bf0ba130856d2d44d5931410e195641390ba87c14ce3739fadd9",
150
- "after_sha256": "b58da74630c5bf0ba130856d2d44d5931410e195641390ba87c14ce3739fadd9"
151
- },
152
- {
153
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_004/meshy/previews/native/render_meta.json",
154
- "before_sha256": "0bdf4845b4c02a1713186662ca0a6714930a3a23e6876a010f4304bcab8d909e",
155
- "after_sha256": "0bdf4845b4c02a1713186662ca0a6714930a3a23e6876a010f4304bcab8d909e"
156
- },
157
- {
158
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_004/meshy/qa.json",
159
- "before_sha256": "8eb5f7aeb0139ca4bf8817f7166e4cf996301827d9244c04c5f6a84fb04d9aff",
160
- "after_sha256": "8eb5f7aeb0139ca4bf8817f7166e4cf996301827d9244c04c5f6a84fb04d9aff"
161
- },
162
- {
163
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_004/tripo/previews/import/render_meta.json",
164
- "before_sha256": "417683b006498a2f4fee486a6a967746640abfc4cf5b3bfb8537cfbe5444ace0",
165
- "after_sha256": "417683b006498a2f4fee486a6a967746640abfc4cf5b3bfb8537cfbe5444ace0"
166
- },
167
- {
168
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_004/tripo/previews/native/render_meta.json",
169
- "before_sha256": "6d4ac7f84931da33ce5e31222fd994e5ff85a63cc8e14adf61d906dcea51f307",
170
- "after_sha256": "6d4ac7f84931da33ce5e31222fd994e5ff85a63cc8e14adf61d906dcea51f307"
171
- },
172
- {
173
- "path": "bench/test_data/outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_004/tripo/qa.json",
174
- "before_sha256": "82efab736ffa6388700219103729837b18760545f4a16dff23941f07894518a5",
175
- "after_sha256": "82efab736ffa6388700219103729837b18760545f4a16dff23941f07894518a5"
176
- },
177
- {
178
- "path": "bench/test_data/outputs/G055_inner_court_investigation/pipeline/G055_inner_court_investigation_pipeline/obj200_meshy_tripo_20261015_v2/meshy/asset_generation_report.json",
179
- "before_sha256": "8fedad56b1050c480d6f33331fe3399028598be7070ae2e562e8d567ff8cbaf1",
180
- "after_sha256": "8fedad56b1050c480d6f33331fe3399028598be7070ae2e562e8d567ff8cbaf1"
181
- },
182
- {
183
- "path": "bench/test_data/outputs/G055_inner_court_investigation/pipeline/G055_inner_court_investigation_pipeline/obj200_meshy_tripo_20261015_v2/meshy/asset_manifest.json",
184
- "before_sha256": "0a08f6d0ce09333ad9a63969597819079e2b60a047bca3d28dcc4f0a0b3cef4a",
185
- "after_sha256": "d6f0b8013f628c5d0d8c756db473063b2af29db5b2c162b958f1f34c321e9657"
186
- },
187
- {
188
- "path": "bench/test_data/outputs/G055_inner_court_investigation/pipeline/G055_inner_court_investigation_pipeline/obj200_meshy_tripo_20261015_v2/meshy/meta.json",
189
- "before_sha256": "1444bc1509f5867fc0129bd39aa84bd0dca3640910641b193fc10a12019f7f88",
190
- "after_sha256": "1444bc1509f5867fc0129bd39aa84bd0dca3640910641b193fc10a12019f7f88"
191
- },
192
- {
193
- "path": "bench/test_data/outputs/G055_inner_court_investigation/pipeline/G055_inner_court_investigation_pipeline/obj200_meshy_tripo_20261015_v2/tripo/asset_generation_report.json",
194
- "before_sha256": "8fedad56b1050c480d6f33331fe3399028598be7070ae2e562e8d567ff8cbaf1",
195
- "after_sha256": "8fedad56b1050c480d6f33331fe3399028598be7070ae2e562e8d567ff8cbaf1"
196
- },
197
- {
198
- "path": "bench/test_data/outputs/G055_inner_court_investigation/pipeline/G055_inner_court_investigation_pipeline/obj200_meshy_tripo_20261015_v2/tripo/asset_manifest.json",
199
- "before_sha256": "1089634e22ddb45ab3eb259eadda7d0405be4c3af3d10cb19d3d9df580da3428",
200
- "after_sha256": "5f346d63c9d476e98b3a1e65b35fb4178efff5b00f544be5d7e60d216625aa38"
201
- },
202
- {
203
- "path": "bench/test_data/outputs/G055_inner_court_investigation/pipeline/G055_inner_court_investigation_pipeline/obj200_meshy_tripo_20261015_v2/tripo/meta.json",
204
- "before_sha256": "a9f5d3e9db3e4775452200749d3efb3f5ae4e05cbd35743e778e061feeca470e",
205
- "after_sha256": "a9f5d3e9db3e4775452200749d3efb3f5ae4e05cbd35743e778e061feeca470e"
206
- }
207
- ],
208
- "api_calls": 0,
209
- "meshes_changed": false,
210
- "absent_artifact_links_cleared_for_undelivered_inventory_rows": true
211
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/frozen_parent_plan_original.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:a72e49a29033d161d0ec0abd27dc003f34c8b7a190201c7e8b514e0afc4bce45
3
- size 60760
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/records_before_localization_35f1ef02271bc9d9.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:3a4589dae5b7ff0df7e9065c9ca973a213a8a195d8bce0c8a6fa3fa74695ed80
3
- size 57079
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/records_before_localization_967e45174af30b60.zip DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:75b3712273ecfbb897ff5e0470756ca6e4a3a876768c291bf8d0c8ecc3010a4c
3
- size 34153
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_001.json DELETED
@@ -1,56 +0,0 @@
1
- {
2
- "submissions": {
3
- "a14aeba402818b8298faaaa22ba37f7edb17e17812022190f3572be08421d57c": {
4
- "endpoint": "/openapi/v1/image-to-3d",
5
- "reserved_credits": 30,
6
- "status": "created",
7
- "remote_id": "01a14308-6fee-75e4-bd70-e2f54b0bc11d",
8
- "response": {
9
- "result": "01a14308-6fee-75e4-bd70-e2f54b0bc11d"
10
- }
11
- }
12
- },
13
- "results": {
14
- "/openapi/v1/image-to-3d/01a14308-6fee-75e4-bd70-e2f54b0bc11d": {
15
- "id": "01a14308-6fee-75e4-bd70-e2f54b0bc11d",
16
- "type": "image-to-3d",
17
- "name": "",
18
- "art_style": "",
19
- "object_prompt": "",
20
- "style_prompt": "",
21
- "negative_prompt": "",
22
- "texture_prompt": "",
23
- "texture_image_url": "",
24
- "text_style_prompt": "",
25
- "image_style_url": "",
26
- "status": "SUCCEEDED",
27
- "created_at": 1792125990203,
28
- "progress": 100,
29
- "started_at": 1792125990207,
30
- "finished_at": 1792126166707,
31
- "expires_at": 1792385366707,
32
- "task_error": null,
33
- "model_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.glb?Expires=1792385366&Signature=Z2zAoyNJYhNNYDWkUcmNYfp2X3zG4hz4zlM~yPJoi5OKmE1KrIJ3RQWSpCO7G2bFxgU0XPVj5YqiWAB4WvKTY5OH2gwau3v7jBCMkTlJ0WzhioqpRoVGpo7po94vYNCnk0EDLYiyTurhPMKA2XtkeGUBns74HqDCGALBtCxiATsFlIAYgwQeNxM~c7zfhOtapU8yushXFgPtsfNKeSCp90kAyQR0H8LdD7H-nhjgtXqM70zjbjGr2hhqnTnSDVSis6rGajtRzHI6S9HyML~4GyAGYB4Lsl0iPSqIFrCw9JPjzfZilnXfOxIGKOQ7RbT2~ztMZJzZ2lm16T4MosJMZQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
34
- "model_urls": {
35
- "glb": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.glb?Expires=1792385366&Signature=Z2zAoyNJYhNNYDWkUcmNYfp2X3zG4hz4zlM~yPJoi5OKmE1KrIJ3RQWSpCO7G2bFxgU0XPVj5YqiWAB4WvKTY5OH2gwau3v7jBCMkTlJ0WzhioqpRoVGpo7po94vYNCnk0EDLYiyTurhPMKA2XtkeGUBns74HqDCGALBtCxiATsFlIAYgwQeNxM~c7zfhOtapU8yushXFgPtsfNKeSCp90kAyQR0H8LdD7H-nhjgtXqM70zjbjGr2hhqnTnSDVSis6rGajtRzHI6S9HyML~4GyAGYB4Lsl0iPSqIFrCw9JPjzfZilnXfOxIGKOQ7RbT2~ztMZJzZ2lm16T4MosJMZQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
36
- "fbx": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.fbx?Expires=1792385366&Signature=d2GmNfNzvAFCV1w-QI~5R1OrBOP-2bhQpIMikvgqXAsZLtdiw7WLhVOmWdzThecRX6PcU71Ctoyxj6KlajIZUvvBQKzMpy0sW6kqaSBmzVLx9FxNhJZPUQjEaPrhx~lOP-rKQtBLiRx7AxdiZHU0EQ8X05n44NdwlL6GUoRkTuag32pstNFwUsWGdqZmm07H~4Zce~H8smv8xrdmZW8s8rrGQInu6i1Xg64UcB9bNTko~XZBg-NQTZxe1-s7-pz2hEZRjJRND7xqz36xTpla2N9ussOa58yoPF~f10JfUmz08l2yud-MIlmoHSn6lMTQ6oxMcGNn8B6QiVxPv2Yzqg__&Key-Pair-Id=K1VGYTHIYLM9UM",
37
- "usdz": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.usdz?Expires=1792385366&Signature=j0TCZxAJ995PlY7~epVBnOp8p10-uvANq3ehv2fPmIRzCYvlO5vKOF0Q5kkhElsUq8VEfT5CspSgNUZjKJCqJr7QgUn4oFtCpBXNJhbzkPWAYxne5smDLZCyk2Wu8-otY18KsuNRw0gUoMXLsPk~B-6BLVmMmf3erXOeqNrnQKBsyjUZ9Q~CJB~CmIMazjUCzuqJTxoM0ZMe25FNoXPdP~NFi66osTr9IePKlduIYZQvbJFBqZr3kdSAbjnOLkwEpVMRJM6RUDku9bWYSRqy11BuQHywq~XPsVR991DyVdrWMnlzPaTpB1F-FxD44b~NMEcz37NIEmTeNrRA1XMLRw__&Key-Pair-Id=K1VGYTHIYLM9UM",
38
- "obj": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.obj?Expires=1792385366&Signature=fGOevi-EUQZLKw1RSQACVHNTvt7G28vyhQGyLU25UpP50kI8Ops798Z7-cLDh2P2fcVbEVqPtSL69G6suFYm2p-jMyAhGLtWJqqgg8mg4d-O-cPNiTbxEytlFjvMBh8Q3gDnAOkLNzgZZhISyAY0LW-mdJ3uHnPDNt18GD5vgPIi1m2CbplXWoR-EiujdKwWgVyCQxzvtTXs7BW9F1Hk~BgC1Y0ih7xWIrZzMbWsc5IxnT-mVT8pS1MXeaMU3SHfpCP~LZgg2hIcMbWf~sCkh87fou4NliXbwtPl92yIeZLu5T1d25kaDAMTlpd5kxg2-hnmOee8YgQ4cX425sZrJA__&Key-Pair-Id=K1VGYTHIYLM9UM",
39
- "mtl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.mtl?Expires=1792385366&Signature=UCmMMRfe2LuByNuzETdANmsRoFNY6dGIT3l7unippSxSPTonB8RWbNHry8IakDPq4EUJSyAmZtUxiaVm0ubYtBE17cpJlgC2x0YjXFshdJDNd9IrN6Z-aDrzgwN2AekXXhQzjtCJC~lmnkiW5riiABNln00avHEYvgnG4gxva1gnqv-n82OtlQ11QUoYEPxvK8FAWkX63SLpdk4-JA0G0To2Kh4DESOjmEyhYLUejh-4MapghJGIfn5sjWoUQoFHQFpMjMJZXH24GObQKAEx3bUhZ-92dMeXgYKhTjE0zx43X0F3z1h1mOsRUlEkkRbQ1UU3wNLwLAnQwzwwoTBdDA__&Key-Pair-Id=K1VGYTHIYLM9UM",
40
- "stl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/model.stl?Expires=1792385366&Signature=VOtx1wf9rIFJNmLxiC0m2h5Jgjq7Jpwy9vIItURJlSzKFrkLP9A0t1TAVtoWlawYG0a097KNp30MMWGqzyLSIGIE1fYUBGpj~xRVNrJGYVzF3UsM-8MpUOlEQ3aa2IzYgDiNXzUUkkD82YPiL93SFi-4cAf1TaRHnyZomsjzX8beGGTTI795kt~qv6VIJLL0nHq4KeEHcl69iQhjE~5cuL0EhqSm6-hsadCH3472LHF3433evveo7Zw5z26bOpl5ugbBwQ6PQHJ8mRMXrtV1jt~FHL2NNtuVUOyrLSbprzfoa4FSiQOwfV4o69r~8OKd6GhZlnrcMxm07q4YU6-~Ww__&Key-Pair-Id=K1VGYTHIYLM9UM"
41
- },
42
- "thumbnail_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/preview.png?Expires=1792385366&Signature=JCBow1nbHcXUILX0Ze5-sweTyv5aCahW8QWRW-~2KRAQQ8GDbtTA8b2d1WL5umsbmuBAuwtslyEQqf7rUY0GJudsYHexfjpZgMqlmn2LfnXrF1o4AcLREOy4wsBrDrYhbQbstd0jHpTV7bppRFdHE3wJrQk1eiinMPGjYS14jbGTjReff0-LAYhQdiLRqwLpq-8wOJ1SMoAHNuyeefbZYzqLo3HjkxITbPB5EmIocqSU16nihzPBBJsHFL1TqOdmZR-IsDPIqbqDWtAE6wlGTil763C2V5qMrGWfHDKXEbdLjSXi4hV1AJaS~4Dp6~nGze0fAOvYEd1Pz-jXg173cw__&Key-Pair-Id=K1VGYTHIYLM9UM",
43
- "texture_urls": [
44
- {
45
- "base_color": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/texture_0.png?Expires=1792385366&Signature=MhHug7YVsk1GTvXUmmPqeOOg7hjT55DtarE1PwVs9O9e04fihwS2Dw5IKixY0xNmzyCLpFF5GTk~AY42gOQt2-tNdufCP8nqisRNRj-hcp0w33qe1zrZgeoT~UZfB2gxHrpS0rhqkLWq1sLfusSafbnT94mFkL9HrrIU5HsdxZ34g6AvvR6P1iZj7UgsSZ3HF5BfINsnHJ6u8wVHn5Yi4SWD64Mm6lHpHkSR2NOII4WrSjrM2ANet4wGJmCg43gWl2fhhU1feljICj5Gew3ric6OnmXrhI-9L-5mechtbzBPF~3SS07HdHxRJN~Xc2QmMzVrjGmBD49EEaRzqGHcCA__&Key-Pair-Id=K1VGYTHIYLM9UM",
46
- "metallic": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/texture_0_metallic.png?Expires=1792385366&Signature=pzImMdd-1Y68no5yiKSSBEpb4SnWXTrYWhD-jkVzQE3fg00Q-4wIXOg3hSXvHQqIfT12qPxEVz4ysKLGXIJZSKuaKs6Gp3X41MpW5nUk15V0ViVoCgIfRzKG0Z9rqxXAcgfAYy8L~nXL8ds~tc8KwPUh39ER2B0D81pFCYGvRGRh--OtH2R1ooPHcnoRZaRcsbUc9IEAYfPEPGMYkuCL7z2vDHyP6~TmBh87N2igiX0lm3QvqqH1vdfyxj4M4JfBUA-GynoQXSq4Xxvko-sdZ8o4fjEx1XjDZ7siVwrU20iR-XoW0slakkEB-sLxCdL7AAFsZxmERdEAjsHLQLLP4g__&Key-Pair-Id=K1VGYTHIYLM9UM",
47
- "roughness": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/texture_0_roughness.png?Expires=1792385366&Signature=mnoriLymHP5-wCEKnXELKwK1czCoJ~LmBQwTeOQOPL4KL7zYMMAU6eQLrt0DOWf89PIsq~6I-y91hze2EpgCDsLku3fm~GEbKTLjQQfVE97Cootsbwii3VM6W5S5VgLyEEt6lehYkqrCZn6hEsNU6xR7taZqbRx-mXdGuebG4--jHqspXRn5ntAUJ3lOJMqYQYRhWFaakSudQOGBsbrBk9bS3lpdcNlXtnLWUlr7zcxI0ldhdxpaPeRICO54MqXRcMR-lEaqO4GuB2t~Gjg~dhAT5r-prAOnACrlDoI-wgsNVa1jOp3NLw39p9fTeAf~U7BlpUnWb76TpysYvxKJqg__&Key-Pair-Id=K1VGYTHIYLM9UM",
48
- "normal": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/texture_0_normal.png?Expires=1792385366&Signature=nhOY8w9J~afpMOp~nImzjOxkknfFtgjpjvk6ijdjzb~2-KyDjeEeP9qUNxAGVUNepJPDM5BDyQXhz9JnnBu3Z3SuTmZyIvle23VtxPK6h2Dcb0BEwlKaI6hr3eV8IMxoMmwI~Ps9IRcHzAcRV6T~i2LckV4rDc2G5r55k26DpfM56651l8kmwjfTX089EVhYCcQWHnlw3kn59TtdDvZsryQJSEDxLVONXFjgDsTWiWP51nqvCAWPROGyVOWgfNPRxB6FGDDu5Vtr~4toChd472rFHjqG8Ca8TeD7TqlVb5uFYLGtPA82EqlyLAZE1vbTPs7ZdEQl1Vq4ePWS9azamA__&Key-Pair-Id=K1VGYTHIYLM9UM",
49
- "emission": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14308-6fee-75e4-bd70-e2f54b0bc11d/output/texture_0_emission.png?Expires=1792385366&Signature=FGlfNoDp8kX7bQXpRrs6FSWt3Zyg7a2mPnOOMVkPZzeS44OIgaEKVrI4ktkrp926Viw-iXd4OpRaEnxt6VSbz16pJM~Yc93XXt1Ne9n-U4do6QnfgVi75BnlRunKGYgTrY3~ZhLJmmlYEPck6ydxbIPoYg-YsyVIqL4EreMAn5PUVBiXMbL7rc1oYsf-1Q69s9mbuFfceZpX868VNc1nnjIVoedgLJ~P7MrhqgjvR7wfXDnJuFj-GXjhcw1HU4m5w2qx3esnDPIAhGjtoJu9oDC2TLjLwL6mhob3uVvsaiUNzPWBNwbzVISlllnDq8GwqC4nxVcITw2qVdsBNbZzhg__&Key-Pair-Id=K1VGYTHIYLM9UM"
50
- }
51
- ],
52
- "remove_lighting": true,
53
- "consumed_credits": 30
54
- }
55
- }
56
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_002.json DELETED
@@ -1,56 +0,0 @@
1
- {
2
- "submissions": {
3
- "13326667a29b2cb3b541bb6794c983cf8b86e52adffe63e3436e37d38a03324e": {
4
- "endpoint": "/openapi/v1/image-to-3d",
5
- "reserved_credits": 30,
6
- "status": "created",
7
- "remote_id": "01a14309-545d-72a9-bc79-327f44ab8d13",
8
- "response": {
9
- "result": "01a14309-545d-72a9-bc79-327f44ab8d13"
10
- }
11
- }
12
- },
13
- "results": {
14
- "/openapi/v1/image-to-3d/01a14309-545d-72a9-bc79-327f44ab8d13": {
15
- "id": "01a14309-545d-72a9-bc79-327f44ab8d13",
16
- "type": "image-to-3d",
17
- "name": "",
18
- "art_style": "",
19
- "object_prompt": "",
20
- "style_prompt": "",
21
- "negative_prompt": "",
22
- "texture_prompt": "",
23
- "texture_image_url": "",
24
- "text_style_prompt": "",
25
- "image_style_url": "",
26
- "status": "SUCCEEDED",
27
- "created_at": 1792126047522,
28
- "progress": 100,
29
- "started_at": 1792126047528,
30
- "finished_at": 1792126213676,
31
- "expires_at": 1792385413676,
32
- "task_error": null,
33
- "model_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.glb?Expires=1792385413&Signature=cB4rn0nW0PeJiYvCgJk853m8KYdmHTpszX13VPiuyVHgguzl5hHLW60Ooxd1GdxOCHTMvEvfGLgHLemqEjdVrm5FrnMjMKUcEDYYOaixqXS7qniySBnLH7IhM8z7XoBjzmv148aUPHq19Fj6b0YFn2kCzarq0H6p6Yxcog0JUREL3k8Qh1MMeiqCoGg7zCjrt0xl9QpmQfCoovnmjD19FPezMU1A9k4xBhx9KK7-vOOg8UdXqOFYcZiWRY6kXP4YBQ9D0Hmaxf0p~efGqKjjr3D7DZzccTxwt7M~UTEAYERXBP~b-I0CnURkPJchs3WQbnmUv3574tOMYPnPLAS6Gg__&Key-Pair-Id=K1VGYTHIYLM9UM",
34
- "model_urls": {
35
- "glb": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.glb?Expires=1792385413&Signature=cB4rn0nW0PeJiYvCgJk853m8KYdmHTpszX13VPiuyVHgguzl5hHLW60Ooxd1GdxOCHTMvEvfGLgHLemqEjdVrm5FrnMjMKUcEDYYOaixqXS7qniySBnLH7IhM8z7XoBjzmv148aUPHq19Fj6b0YFn2kCzarq0H6p6Yxcog0JUREL3k8Qh1MMeiqCoGg7zCjrt0xl9QpmQfCoovnmjD19FPezMU1A9k4xBhx9KK7-vOOg8UdXqOFYcZiWRY6kXP4YBQ9D0Hmaxf0p~efGqKjjr3D7DZzccTxwt7M~UTEAYERXBP~b-I0CnURkPJchs3WQbnmUv3574tOMYPnPLAS6Gg__&Key-Pair-Id=K1VGYTHIYLM9UM",
36
- "fbx": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.fbx?Expires=1792385413&Signature=gBnSZvfDOF9XnOFut4VS~RqYuqeq0m7kIl6yyA38UnNyZylJ6tzFP6GZwhImww-EPRTLtfdGdgr1Uz47GfvTEdoZPOGkxc7nPxDdvNo6geiKppjqLggo7myD~QWw1itQBYHs0e5oRYP3C13Jff9msaiM4Ue2lvbTFLqxEQbaTdHaf~VxKr3KGwvOtx5evT~Rckfa-dYZwoQVLZIyVMmRdZNSKWaqZi-bz33k~Ijwx1eCsvfshHAO2tEtdGol-Vmsj-XwaVIjwjuTnpTp~y4G38FS5byxhKH6looN3bNSLYelfU2CLkR-7-Ps~n8LsxodF4x9jOKugF6VPgmNznvE4A__&Key-Pair-Id=K1VGYTHIYLM9UM",
37
- "usdz": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.usdz?Expires=1792385413&Signature=I9ncFMNpnP~LVsh~fS6FsrzseuqKzbk483ZRaxR0DfiblPMqSaJjXd~uWTfPlP7comAmDEUcU8uA8JxacXlsRtlKJEc24ouwfSYY5RMhsj~wo7aiyIgFmI1ovoF~J2DEpApWaM2CMIw11ah8prXdwPr4AzEz3x58j83wfnyXArQYZRPdgYgwofcvP4oaNRjoCkd2Onqv5S2x3s0JapGkidm8LwrfB1o9ZwPPbU8TmFFspuGf10SMHDHgrT7hDHGSYZAu2b5secYporhbrsYdX9AGYfgpi7LpYf4-cN0aObnHSOmM8Ue-2VWkh8poO77GYosYSLiqrIfbbOs4LbvXJg__&Key-Pair-Id=K1VGYTHIYLM9UM",
38
- "obj": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.obj?Expires=1792385413&Signature=BzQSR2U~etuhBA1zGYtvR~drExMoQ-fiCZwXnzlWx3dS6Q9guKWCKP~sjbVrzPVhbk9kNqSjrLN2ASorwjB1CgqjLjn~qKHxCCLnO1hBq6C7Ia2vHKcImSC6msO6EcINdzWGa~euvVCIr5McVWiYJgIroHjuEDZrCkMzCEYi0~8rJ7KqUkylccChKtaomi3JTRHD4mmqE2gyXTWP3VGUs-OqRZw4PF~TTVSmWfkTpZYIXDdY4W~PMslHWxuuPetHERMvRb0jZd6Qz5VA0f-NLtT7pI2lxRJsfYAHcBC~x40HgVUzPKyfLzSn7a8jzsaL2Xcq-IyFdBG0prv6xLeaNw__&Key-Pair-Id=K1VGYTHIYLM9UM",
39
- "mtl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.mtl?Expires=1792385413&Signature=AbOAbNYR10tm66ITsuaEGXAhSMOkdj6rpHViIxUqPidny82TbFI9SpQPhjfqlFulUK3boSaBF8F0xPFNlQuqRSf4U4nB-BhXMcJ-F2pp09Tb7NRBW-QG8q4a-~ilqReW~EhYuaGRtlj6bMUOEr6TD5vAxK3kE-mAopSlkce2rmpFRpLwhGXetXgNyoqvMy4YFqFIV6qD2NRDVh4~e5Yl6z5F0NeK2fyiX6LMUhmU6CVm1zBH6Jpt60RlFUkI4uyFRep4TEezzZ4E2AJ5x2QGFvg2tgBCz2x80SeG~iUhRUSRNkJP7xW8xf4U2Gai~Xhbx09rW~pwHTqm7pjGKOMGzQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
40
- "stl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/model.stl?Expires=1792385413&Signature=hPSffBdbF2ewSb0RD4~rbp63aSn2KS2lIuw8-5hg9ypCB8InafXSKwClLQIaWQg0F4orQ8dDemakdb00DnzMOEa9UFppQwbI8tRU2JO-yWp9xlBpDnOtqHFeXGbjR1rbuEohmAkLk3Fx4cqw7jOkQZ4clFbLQGs39w39TZ26yeXJZifwhfkPpzCKGpNrZrIm7eZzl6gqDIrv9ktlxcyEUkz1VTpEOZw-S5UKWvrblNFsCWQtPOb3SdM76lIRCjTxekC3NQXK3zAW7PcGDuQoMjUhwId-~HrrE-AUo03soKrE~gwjoeIQWY6oP6Hqb7~g85OUFItoiRaxBtfb0egL~A__&Key-Pair-Id=K1VGYTHIYLM9UM"
41
- },
42
- "thumbnail_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/preview.png?Expires=1792385413&Signature=jMWVrIkhbtIEjx~kBiuMo2YxpqdR91XA1glI~3O95za1sYO7VUOVfUbULMHfJlcU18Krtykton8wYOFbKUN7TcH-32RkA7VWyF2Sg-mv7bm2iUYPNkRtw6rXu0TVxrhZ8Y3hnrsLS7h2tUDH8gIwQGQam3iv9LBW1tsYRGTKnbbW6CWf13rXJPicYJqI9n7NeYCoIreCrXcXzHec1WNNSOXmmeMPhRL1oQukn7mGQtJ8GohFLvinIFWxIuPphNWVxZOv-YXGuzsjThRATPVyCICTnPiolTNCZF5Xra245ssmtblQGQKslpQg7LkACoAgqdotacCtgJ5gyx6YybJitA__&Key-Pair-Id=K1VGYTHIYLM9UM",
43
- "texture_urls": [
44
- {
45
- "base_color": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/texture_0.png?Expires=1792385413&Signature=gNKasEdSYfDMtbUoNJXTh8M5MlD7PoHetyV5ElJkhUhXRgYCugxBLk2SgG6RlXNoDMT0GI7t0c8QYv-sS338fHSHFVbirvZioBqXfClRa3S8cWntvORZd8-NW2hEiljOCvxO52Y2-cOly4wmYz22-VgtE41w07SQ7E3wnhhT2GicK-CfvOIOsqwnQ2BEtR27UvO0gcWbVuAByT2waCPY3MoeojhQ~mNPEiV1CwMLD~38TC0Itj7xmO6mSs8Wbeqjnh3mGzBUG8W2wz2Wt20U7ttk166f8aiIspdeJcrxEQiUkaPC3hHL13yS-uvHcOri1Hc2~N2UERmUKJ7eZBSSCg__&Key-Pair-Id=K1VGYTHIYLM9UM",
46
- "metallic": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/texture_0_metallic.png?Expires=1792385413&Signature=EBmqTgz7NTQNVgYTgd2gqa0JO1QVdlGxNIU21ZpUnUKamhpfWkjmH4CdcebCMbmysuDXdNbd3Iki8gCvCW7Lbz~cZNp~ttA25cGyEa6Xp4O4y~LZs6pM3SoxSnIXjZFEbjuvNbOnT6BunEYfNa3PJSKbOunQG8qRIoNhH7hKJdkgJkfT3zeFgeR1Of9sgJ9QCcasORUvjtYHGN5dnW08Fq8Rnl2GsrRqiCULN6Aq5KfCfjbl~wg-MPAAghb-R0cxReirfe8sTj6sZ16pp3exIYHDGU8CeCQW4fYrrYPQtXlboGyfDUv5laF0spcIjDBgiHd0hSe55WfLcpF-MoFdhA__&Key-Pair-Id=K1VGYTHIYLM9UM",
47
- "roughness": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/texture_0_roughness.png?Expires=1792385413&Signature=ItUS-O8~Crj2vyHMRSKVAoTKrNZfvXOpcFuzLRZAZKoyAMPI~G3bHsl-XP2bati81L~dOTc656zqZhGHe3QqsP0YF6~pCMoibeHqOo0B9lg~35hv79oYOXi7qY6Gx29~2wwnmkiGGPDV9cI6mtCA7KZ7107g3s8pKDtKqWS0fA5-vpz8pZvkKcHXKUAXRj0GRzp2y4ZEkmHAFtFStP0njJD8tMNxe3cUBtcWgYE-WAZfzzcgREax67gRInj~7FwWZsyCUEEwIuYW2Sb2Hg0zUS9TLYCXe39ukdC0HFPPvZicu3H8iSVrbbxnHzvTLzBU0VCZXJxO5eZgwzUORvjQyw__&Key-Pair-Id=K1VGYTHIYLM9UM",
48
- "normal": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/texture_0_normal.png?Expires=1792385413&Signature=cDso5H09hZ5Z4LyOu8gUfkP1AiRLKi3Kr-vhJTTI~2sjr2pYuWQLcuHKbV5Tj4RfpICYAZmOsdNJPMUFsts~Y-sSySCbwWk8w4aUJKTiUaLJbeZPZOgDcKR0LO0B5uUxLa28GNidJLb-k9kLn73ufTIj4E8bOayZieg2WZPbGFkDd58cd6WFzDKWbgmRZZq710K2SaWO6zxoT1UzgwpMcIIrR1mu~RLZuMvx9Mh-aFA5WyY2~FWsy~Op3TMwLVRmYBUSUPMOBmvgtRi03QlaYjg8JOhoZAeEtSAUYp5rXBBeBMkcmaoTklDIyTPM6wsnHOeJ-d7Ya7qEWmMN-9kuqQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
49
- "emission": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-545d-72a9-bc79-327f44ab8d13/output/texture_0_emission.png?Expires=1792385413&Signature=AvrXmjx9AOiQ-0pZsfGwAwtJt6hSOm1fntBkR1F~C~V8OBT2tmJTIttYY015JerAMKHMX5bll28CzULiSSgugkbciAwSDqYtl-N2k8dv0RNt126rP1NYzyJfTmJQfSZWScyeF4DbL5wTujOgu1PCEeSY2NwTCeIfTQHrIjSfPpPasba-me0kVVcCsdO6NOsWUpq5CDQvQM6m-Iz8WSKK0SkOc~Zb5MxIAQH8UUtsPpgkwjrlD2IPljqWOY4P4kNGy4mvedrUpNlXBw-UAeSwqhmgksT5T5b33dFZuzFuxMgk-L0kB8CoTAX19PKYny~9BorgDUmMwGwS01Hlq6HZfQ__&Key-Pair-Id=K1VGYTHIYLM9UM"
50
- }
51
- ],
52
- "remove_lighting": true,
53
- "consumed_credits": 30
54
- }
55
- }
56
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_003.json DELETED
@@ -1,88 +0,0 @@
1
- {
2
- "submissions": {
3
- "7acaf580cb488901c29de2f7d09933255a3ecc868e75db9a8f1893a43b32a851": {
4
- "endpoint": "/openapi/v2/text-to-3d",
5
- "reserved_credits": 20,
6
- "status": "created",
7
- "remote_id": "01a14309-c130-7684-ae32-797c57e1dd4b",
8
- "response": {
9
- "result": "01a14309-c130-7684-ae32-797c57e1dd4b"
10
- }
11
- },
12
- "ceae924f4c99b753b966dce01aa394fe28c4d3102d75aa934d6f61c789b9b05f": {
13
- "endpoint": "/openapi/v2/text-to-3d",
14
- "reserved_credits": 10,
15
- "status": "created",
16
- "remote_id": "01a1430a-f6b4-747f-a843-36a9fb07347c",
17
- "response": {
18
- "result": "01a1430a-f6b4-747f-a843-36a9fb07347c"
19
- }
20
- }
21
- },
22
- "results": {
23
- "/openapi/v2/text-to-3d/01a14309-c130-7684-ae32-797c57e1dd4b": {
24
- "id": "01a14309-c130-7684-ae32-797c57e1dd4b",
25
- "type": "text-to-3d-preview",
26
- "mode": "preview",
27
- "name": "",
28
- "model_type": "standard",
29
- "seed": 1553,
30
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
31
- "texture_prompt": "",
32
- "texture_image_url": "",
33
- "status": "SUCCEEDED",
34
- "created_at": 1792126077378,
35
- "progress": 100,
36
- "started_at": 1792126077383,
37
- "finished_at": 1792126152168,
38
- "task_error": null,
39
- "model_urls": {
40
- "glb": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-c130-7684-ae32-797c57e1dd4b/output/model.glb?Expires=4945708800&Signature=qfay~xTGn5NkS3mnFN91vBnxGUwfkLzhqiqyw3hIuJPi12LfuDWDz91LqBA~Qgmbwe7EyQUUpDywpdjjh5UfXKTl1irolA-WVzIpBVUecq0jXbSE2xeDCCkHoi2V4vuFGb8u8y70uFLgJ2S-NumADKEHMEEp4YziYmc-f8aFc0QybVe0g3HiwkSHJITt3csm5JW3--U887V9fyi6d2AwbCag91pCi2o7kM93M3JOaL~Ws7UHlKyF2qXC6c8xkOcpM~cCimjD67GJ3U~vo8pYuGD3ZXmjbWD2JsI3uCdD5A~fqrHPTWuX~7WiDZOCyj7tOQfNykqfCo8k84UJoTtbvQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
41
- "fbx": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-c130-7684-ae32-797c57e1dd4b/output/model.fbx?Expires=4945708800&Signature=Da3Vyl8UXhV-Zm8k2vFjjlEnhSs8KAsgIYNwDDwccjEaHKCfkvm39d6RBiClBNfjz8ZwuzU5NBEQ~awP9G5--e0h61JuL476d5izkCkC3NcWJyh7x2zJ8-wYFNUd1z6raQnhlndlSJFHmEBugrpF37ZY2ZH9GHi5~erChu7Znm893a8F-9Xkkr~7I1PFk3TMecKuJLwle1jbVlHvY8EiDvyiWqgfMRch-yQH~FFrrVc5RJyQnd0Z0IbHBQGQuyqOH6b2BVC5IRslxDgL8fUVWny~2huAl7QQbSbAT7NowIXbuFouIxlVRAhJMycpGE5lul2JnSoJvRrdXXWrmJDkUQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
42
- "usdz": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-c130-7684-ae32-797c57e1dd4b/output/model.usdz?Expires=4945708800&Signature=SxSMmTrETOVqiZidhFHg-vYJMQNbsaKnZ6WS4fafoNvprj1X04L6XLN4EYcYM2B5h35KicR1fZnl79BJF2sbd9pLTq9OUxATHE-Q3DNVk1bRq4wThR7qXd5W9riCrHiTWtRH0K6iiyGYKEngt4vxmbgSxsu7lWUV2QIu83h2WwahkvOh54dToraSbMKH0pJdUVjRUJfez2DYaqfeNP2pePZEvQHYsJBiQiEVedlWqr4KIv4eoyOFhS6zo8w7wgFsbIaVdGQ4KvCS0o87lxYnL-yU87MG4MoY~XvoP3IZbaNNJMKMmmgwDFiH9YFcyodCirsgxMS8DSMNe4V1axJt8Q__&Key-Pair-Id=K1VGYTHIYLM9UM",
43
- "obj": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-c130-7684-ae32-797c57e1dd4b/output/model.obj?Expires=4945708800&Signature=G-DxVDqXS~5xtcGOUDCuVQS-TlASMJt048w60LuaZkXQEtc-F2GA~lmldkFnGtCzzIyrMTCe7xl4mVtDmstJV0WoZmJ01LTD1Yie~S0X7y0cxVrnp300BYiQbvuhMK~UXWN4oZury1oVF-AzK-VIpJ-3jBFWWPB9GJUCB7ggPIkxofLTQStAcfuDAeiWmkpVALgPIsGi5bkwWB-4J9MNpDCY-Hp8Jt1UA-QnIj~C9bepHsIlQpUBVyk0dYJLWIlSt3xJXqI1nCexiQYKI2CR2W7OIrNkDtejjl9-UOEi8IdEadNo1OrpEEv-E9TcHas0C9JxMQ8o3ylQdejUAnSiMg__&Key-Pair-Id=K1VGYTHIYLM9UM",
44
- "stl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-c130-7684-ae32-797c57e1dd4b/output/model.stl?Expires=4945708800&Signature=p7jaW6z6RiBCyxsEv7QkpoeqfkOczx6GGNXtNlKZpMYVYa9vsSX0fpjOy267KQ7upn5VzlefguOUNM2wajBpXOMarsm~TvujYmb1H0olcLD2xh1Dsuhuh5XXANlfI8QYZXagFQ1Q1pYp0IlnRg-U1gFhSZteYcd07XmqJKH~fHWEe8M6SMttli12X4rSoz7OfJSbjJ21ORIa2OGS-cBI709dCr9Q4V~rf3HtTRGvTPSxyBdXbVsRiQ4Dq0Hs92dVhYOKQc6cwu13SJTeD9AtL5i0OSnh5w08re0vvIAVqGhw8isIuK8SCqsOH5XP227Hl7F3mNPVMWPo1uez3AJlDw__&Key-Pair-Id=K1VGYTHIYLM9UM"
45
- },
46
- "thumbnail_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a14309-c130-7684-ae32-797c57e1dd4b/output/preview.png?Expires=4945708800&Signature=aT8DWVDOZocCu702bRS14uy6aHa3jDyMCrA8ZNGjgdguIJq~IsCPggdfFOCMNob5jKlRn1-Sgmeci8aL2Ejw70ThUGou7SaM7TAvhB1DcHhtEB-IofNUl6~lzDHlFUdTUZ0LfVatVSVzZ9w7wIQ3TwOA9fWvnFXSiHVwpNb8uCOdVBRsSjPfDDhneYOw-yfangTwJMeViE5IaOlkSKolLq4-nI~qmdY7TkHApg7ejNrHLkdbqoqa5yffIWyXxU8eZQJbBkKW1jYv6LwVf~1A-wqnqzOmLQJAbSnK~Husrancf46IQ07RKnGWywYuHO-NRqqnC5tiHEcisBrmY1Wgsw__&Key-Pair-Id=K1VGYTHIYLM9UM",
47
- "texture_urls": [],
48
- "consumed_credits": 20
49
- },
50
- "/openapi/v2/text-to-3d/01a1430a-f6b4-747f-a843-36a9fb07347c": {
51
- "id": "01a1430a-f6b4-747f-a843-36a9fb07347c",
52
- "type": "text-to-3d-refine",
53
- "mode": "refine",
54
- "name": "",
55
- "model_type": "standard",
56
- "seed": 1553,
57
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
58
- "texture_prompt": "",
59
- "texture_image_url": "",
60
- "status": "SUCCEEDED",
61
- "created_at": 1792126154430,
62
- "progress": 100,
63
- "started_at": 1792126154437,
64
- "finished_at": 1792126252009,
65
- "task_error": null,
66
- "model_urls": {
67
- "glb": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/model.glb?Expires=4945708800&Signature=exGo163q167rC8EJqfA5y-zVoUJsR1VrYG7oNcSiF1ZDyhGgJzxYePuELz9-ZrPBuuTlXjLeP0CdM4et5HTf-YvYrHs8WMlrdeoW~Ato5bnTfKKkzaaNrpWa~aa3nEx472QDsb5lMCC44vW5B1unhUVoEInTf9Q8xt--99~3p~UQT2m4rCBzBCtyeGR2tJFfOb0BnsLSKl6IdkcVXJRJfW2k8kRJ2zpdZqYTmd-92u985gCf1dustzw4BmSNseCn6zPeFhudRcj8UrC9vsKd0tWyQSROysSzS-QERacWm1-jywvfHXZpLrKqdAIZ6BVkfEbW59gbGe2BjP~HN8sX1A__&Key-Pair-Id=K1VGYTHIYLM9UM",
68
- "fbx": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/model.fbx?Expires=4945708800&Signature=Y4x9rGwzZu8uYwB~OuGAvE5QIAyRHRJQD~OUE6EXua01P9TrXLlDnRI-D1tZ-YI6L~b4ur~38i8sR2r-1ul-75Uc1-M6ukg-jUfhzWc8vnIVUScJEbExQ06uN5Dv96pfQyeDJEx7IA-3s4HVk2NiJ~95oTpFTtxX12s6o681C8xrgcJBT0W7Rt5WLQPaJH5bZxxUKHH~DXuf9UjjTRPsaZgU-SR4hDJOn4lkg4UPDiLOrv8zupoG6NUc14s-ahlqj9zJQCpTW17Xe7NZr~UVU1rUchH0L3YQ3lYRfsPcFNR6qyE5AkhxilyzU3MpNu1fy4-srH80KiRGQytPETtYYA__&Key-Pair-Id=K1VGYTHIYLM9UM",
69
- "usdz": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/model.usdz?Expires=4945708800&Signature=ecYxp7o84jT7c3LbmywTOeOXtxwZumeNIjIqr4HIcfgQlohltGVlFBZ3G6Azgm4Fnk3kLFRZLD6QEVLRErGDIU9Lm4oRGbyMxn8bmfqxoxf-NCqf2dkshrHqAqoLmZiKETTBPSwQpVN2SZqz1CR7ifrE5uoaILyojieygUk9z1~NYmam9emc5AbivtkrUHiVZFZ3xA3rqtMd0UZwelhnH7C2YzgBIBW0SwxxpyRa9ZCUDfQocSSaAZKWZEIpwIp0Fx2SqXatjKJ4-Lpqu~z9lKvVAfrxh62RrDIN4O~WHYA1BDLemp6y7V7IqNkLD91twcAaLI3py8Nwpw~S~7CopQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
70
- "obj": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/model.obj?Expires=4945708800&Signature=AenT7LwYvnXj99~kbVZ-v0efmfSb8GT1iE5gHG0Orp9l1FFOgaCEhQmPBYDThASLC~kq0WPZovwdmdyjZxqunOHGYotqOLZJ~efpHuNOW6eaUhGtwHuIBASEz1xme1l2kRpFxUUiTpEIjEFsgOw0l-HRHZhzg3nGmQFFROAY0FPE3rTXTeSfcgvT6d2U5CRtv6pLM0WuunFHIyRz5V2sk4DcXrTXoAPCL9MTZ~OERhfQ7ifyTAXEm8qPjsK-J46qkYcwgp98P5VNGUvhnR0oU-EHvdeLkOyJE80ywgGWvt9iySXi-MgriAGEhbkP0FHoMzZmxPg0FHXEdXgzEmHhmw__&Key-Pair-Id=K1VGYTHIYLM9UM",
71
- "mtl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/model.mtl?Expires=4945708800&Signature=b2yHyqWyo6wE2723rNFoV0sYBsFxp27UOK1wL0sOtd1uIeUtJIvymqULH5clcGcQNehnhrha~Y2fvth2PY4gFmP5HpYAyIyPJEylgMY3YUom7CTnuxc2wrl5~5EQYGeYTaJ5ltn76jO0RKuLihrVZPbWFE8To81Dr10sUKlmCTvAfnyJcFUmaSStVPJju12V7mK2K7qwwvxv9-cAiXA1qIo5QZ0cbiEhy7rOIb9SC3dw-mL2SEKejh-tOetbPZtTrTsjyaB4Az5OKeZRQCnljMREA6dQPUcH0ojHEaatpjphA5Cv1wDDa-RjMNxdQlzYixP7qxkN3SfDSFxyE1se4A__&Key-Pair-Id=K1VGYTHIYLM9UM",
72
- "stl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/model.stl?Expires=4945708800&Signature=CtF18J7L9YMg1EGeE6mi~jgbJwrq1~3D6bo6vfrMBbPGVb5xNUFXyMk1U9bYGPaFhFWejg~pC8i8adE9Q~lRqYb3k8llCGki6oTSk91wyiu1xK3Egv9Lpn7omiBdObKAyOzINBRGQLUzxRRMjYS3mEX14sn77wsCgLaTX1Czf-o8Aalne3VQ1QZ9WV6M4MIjr~ZUk5TDRXW-r8dVJ6Rsjb9jMiyjqw9VsLdcoEn59X0ghya6bOdOC6t7HJkk59vivZVqpEL70zfKPIX6js0Lr033Dl59mLl-0lvHjkEjNwy0vbojZma~~0M648SiyB0LHt6RRjHBPFXl4q~BJDKkeA__&Key-Pair-Id=K1VGYTHIYLM9UM"
73
- },
74
- "thumbnail_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/preview.png?Expires=4945708800&Signature=Q5TlMTiuz4nC9lIV5eo3Iw7Ynk3ep0au8qg-GPkaYvxx378pwu5I8z1ofbJoGjnBycSeBytb-zPselBJcl9A4~dRDDP9wprNW96EIp~xd-dwaoBJiVvv4slTXz2DkiCv-T0NY1CefkTsor3Exv5uJ59WOLHQ8AJQs4pImAc5nRlTUjsPQh1ZcaUET4a4xGX4EnSgGRBafvzIVrVGmvXJAZInML4Y4BnwQlYsHrUAifFE2f6goNfNajqLDuHkOWGVCFJf8snYTVXJ-oARhDJlrUuO-xnfELJKSYPOrbw0MJLf3r9k6z2rPupzJpqeFsKxrn4v~l5YZPkJwnowuAZEoQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
75
- "texture_urls": [
76
- {
77
- "base_color": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/texture_0.png?Expires=4945708800&Signature=kcdQUmIJAcxGqrWHFzhZySP3jEW-5NHfefHfL0Y2z2TJ4cBHaekkk8E6n7ADQPp-Zpy4R~K-ZQ1a~oQHkSVPlAhvksRaqhsDhUWnLYjv2KmltrM-K5NAnfWfhWw2X3gh0nWXNJNhoYLwWdK6AylSXfSQMHOkNxWeCwP4pK535lr7MKWzz56bwcYXTkVSAMSajxFIg22YuaUfoqybZbCjw~DQhgBtJ6C7uVd2E6rvvEC7C~MZc5RndodMYtPGY6TuSBIJ7S7dICWZqUlcluYZPZ~z6bl3SRws81PoABSSghz9N4H51EXMEwpSVW8lHukMKnzz2gdL1MInih6gUh0Ibg__&Key-Pair-Id=K1VGYTHIYLM9UM",
78
- "metallic": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/texture_0_metallic.png?Expires=4945708800&Signature=XNFWaqjneoPxs9R0aDJBU5H0DiaEP0aS9kV1pQpp4z7TZvA5ePG1~R307kf6gP3jYNeIwWdA4TNt3-uD13Bv4NggurAG6zapQXStrDlrW6Gzmu9s4KvLBY13lA-wQdaIb14kDJtPUrWm6F2kOGnKMNZn614QIw5mhHRvekNSfdTJTI-fFiDs4~8R5A3m54T-I-vU5DnTv5Tg7X7CKoyyWWOo60fL~mHgseuGQ15QhHwv7ekpOnSsrVycozHTnLYTmf5uNZg6izwZwG5TPSGPZfI1EPeIfFh~gYJEFFF3jWuhqf7H4Vt0Zb~4R0PBrh-MLDu9KkZZXL7tQVA-zXD5QA__&Key-Pair-Id=K1VGYTHIYLM9UM",
79
- "roughness": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/texture_0_roughness.png?Expires=4945708800&Signature=BLrhcT2qIb6Y68CeprpgAjHC3MBuI0mAueVk-K3qplfu58EPAihXiHKOEs9xV9FzwrkJv5jMQ-zoqeazdHoMQLeML8d47rg14YgF2M5oj8CmFIswwyyvS02~-sWbRHVbnu8v~U0-B-BrsPGwAnD9UCeVe6j7hE86AeXAbH6vicSuC5mbIvmMZ1BE0ADfqpC6wID3-oTrvoZ34YPHjycdoHoDks2WBJe~bYot84ybp0liWnmqoeDKyErmqnH~TrgSdaHei6BtdaWtbySCmMjh7MK2Om720nVXik~GTuM6tAqXESokeQktUTCq7cUTW-19KvPAygwM3z00wKqc3gmlJQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
80
- "normal": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/texture_0_normal.png?Expires=4945708800&Signature=f3o7XpquYFUp8zlKC1hgPeF~7ywaXMcfna2SV3S78-UKJQvAYNHLhxuvNLU7zxhUHA5hNLKWzGxczEUXL~I29mhYmv95ooZo7ROijohvoJml1kOWQQCLAHIzaTlEwZBa4LBKm3M-3yjnw87o4RZbJp5nNSPFlci~39EtOq~Qxb2eMfZ7nJYU6jIdsoIBywqGldkB~neLzX55e~VDPt5fhYH3QJJr~~FhjCsTHfJ6hAKS3YyAfHJrlwEnMDB4j5c~Rp5TsBjfxrNA-AHn2NUEuOSGZkbP4CTD6y3TnXGByL5HuNB1IP2SN7T5P5BtjvrFtroiyWCsu8Iq8Lh39w4QEg__&Key-Pair-Id=K1VGYTHIYLM9UM",
81
- "emission": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-f6b4-747f-a843-36a9fb07347c/output/texture_0_emission.png?Expires=4945708800&Signature=bdjaPfQHiBJQQh02LCMQ2-ar3xL3wqNakpKozhJhkNUGTKvJoNwDkZ5yvBKlLkt8Sz6IG4Erq51QKlvJ3DUqjYsiqrc6adEeLjwVXxMatVjhp37fBTEFhbu44c3od5TN6LvBXx2vXW4kQpnQWromnTONrbrCG5Hs9fIUnNErap2wFZ9ty1QWaDFkWMTFAtlm~S0rJqpkunZR3a1PztrwTl7eePrpnUgw3gVJDZ8HIB1OoHaYuAjsGXFoDDC7lYw4MKKUf9YMpRiC26rYgH7tzl7kL8YhdARJNthcmiv~A5jvQnqzjTuu1Z3kD2FqU~CpDKzWWZtGylKnRevjHUldeg__&Key-Pair-Id=K1VGYTHIYLM9UM"
82
- }
83
- ],
84
- "remove_lighting": true,
85
- "consumed_credits": 10
86
- }
87
- }
88
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_004.json DELETED
@@ -1,88 +0,0 @@
1
- {
2
- "submissions": {
3
- "796f6e881afe734b13e23853be0b85d154ce6ca430d895abc7268866e9a18b59": {
4
- "endpoint": "/openapi/v2/text-to-3d",
5
- "reserved_credits": 20,
6
- "status": "created",
7
- "remote_id": "01a1430a-442d-768f-90d5-a63534dfd811",
8
- "response": {
9
- "result": "01a1430a-442d-768f-90d5-a63534dfd811"
10
- }
11
- },
12
- "81fa2f6d4de920633838b3dc3c126325dd8f3bea02cc95a7cdd1bce2232f93fc": {
13
- "endpoint": "/openapi/v2/text-to-3d",
14
- "reserved_credits": 10,
15
- "status": "created",
16
- "remote_id": "01a1430b-df46-71b4-9dbf-41a42f98f5be",
17
- "response": {
18
- "result": "01a1430b-df46-71b4-9dbf-41a42f98f5be"
19
- }
20
- }
21
- },
22
- "results": {
23
- "/openapi/v2/text-to-3d/01a1430a-442d-768f-90d5-a63534dfd811": {
24
- "id": "01a1430a-442d-768f-90d5-a63534dfd811",
25
- "type": "text-to-3d-preview",
26
- "mode": "preview",
27
- "name": "",
28
- "model_type": "standard",
29
- "seed": 1554,
30
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
31
- "texture_prompt": "",
32
- "texture_image_url": "",
33
- "status": "SUCCEEDED",
34
- "created_at": 1792126111044,
35
- "progress": 100,
36
- "started_at": 1792126111047,
37
- "finished_at": 1792126209979,
38
- "task_error": null,
39
- "model_urls": {
40
- "glb": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-442d-768f-90d5-a63534dfd811/output/model.glb?Expires=4945708800&Signature=Xqgj84uzrjUF1GZGEaEtY-EotUP6LLxu3jXzY8SQ2Y3AJfrtwR8dvGCXomkir1UqXzjhJcQrubyeVpY45rmBf7qYAjnV97kb2QZKleOBLLaug1DowcyER5lFT~ll1VoAL6GmsPKM9xAjmxPWWqRTm-QSniFZ2Di1wK4TjbzYhnQdGJLwXZmVI1VUmmATLW97MWuwVG9AG7z2kVnWu3vgijV6vnnkK2MA~1wMxALw~na9OUdckvsm33Ii4PeJAkZk98~BfVMZ1wXt2Ms9UNlmiWRHZ8Ils9nAGcE7gZO2P2ZvIio0E~uXKfTF6rnT9Ctm4E4TKVCgLjaAbZcCXiGyhw__&Key-Pair-Id=K1VGYTHIYLM9UM",
41
- "fbx": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-442d-768f-90d5-a63534dfd811/output/model.fbx?Expires=4945708800&Signature=JpoVpuXukzpNNOWBsw2yvlkBzMuvtPe4~GhZu8ctoa1PhOiRCAHu1UUimCm5KGLtlcBGnNHspJeNZUZN7P1lH9ATb6B4uvAn~WzxtleUgxG-l6qt4vgVs9MRn7HGwgXcBXBQiCCo5VWyAQk5LITKZ7lhRbz91r90fUbNK60So2csTHT1-fnn8KL4pVN2sZN9W3AfoJaK6mApQKqCtxepp9IEOEPxnWayeEC3eR1OZoEx4J5vXz4VRa3bCHEQHDqPSu90jLQk-7urphrshPfYqKUToCn2SSNEpV~jFCJM1CIfx~HexeObHGyIK8qZiLsTjhbnjW1Osc~nLG6gX5LSzQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
42
- "usdz": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-442d-768f-90d5-a63534dfd811/output/model.usdz?Expires=4945708800&Signature=hntgr9JNZu5KX1jykka3chVMSwUk~ABdkYfrOhftnIaozZIbSXcs4x3yDs-2mKzAZgcBxLwadYO3NI3NlFM~AI~9gufd8n8qdmrEZv1Nv1qsNZ1Sz9eqJW7YSbLY7SwuOrHushAJ-UHBHUYvCt7jTdzrYvuuQsut2gERYH37vuvXPAnquuh0OgaJjEdnDMtelguPnp1T-KKIaeGAcJ5T7EzDd2SHmm4Sg45JgkWQPW6dK5OphNAsCIDvXCADbDaOucQv9rOu~AFhgj3zCC7~1dizfdtq2JgSzKMNklyYPyi~vuvF3nBO0smS0ofpAsY63vyKLxNq7o98oE0m9YVbag__&Key-Pair-Id=K1VGYTHIYLM9UM",
43
- "obj": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-442d-768f-90d5-a63534dfd811/output/model.obj?Expires=4945708800&Signature=KjtCoAr1eSN2Z70LG03hK78UhINUk9dsD83sT4GAI~UVlrXI0xsuoritx9wl04a2IRw2klQi8bORMTKr4jhx-oXVar2TLycorNV0XVemNZ~3~RNzvGQ7TiMv8T-2TUvHtMNCxcn87rqie1u11H19ue0vj1STzwkVQ2X5282MfQtwf3QpQZUhCjau7zCKejkrHboSMaLm0yZgFCrIHGU4mId9ns4E-slMAyCg6x8ISOC3IWN8ao6GCQIfCfsrjk9sUlyXiUHdf5w7OedjGz7IQ9CIjB-JyP1In11GShh3X90pBqjb2oGF-KgeOiCK8nhvo6Vqhmqdi8AfNJIXY0uzVA__&Key-Pair-Id=K1VGYTHIYLM9UM",
44
- "stl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-442d-768f-90d5-a63534dfd811/output/model.stl?Expires=4945708800&Signature=MOk9RJua-xfCiCuuq6rBYDnmsd9IEDdcgml4x5hW8cNWfjVXMbJF-23ZYulwzX-24PBeh~qBJS-5Hf94pPiMb5qFZlc1ZKvTXPPK6KEJA3AY2MK2R06rJC6cuaO1-20JW-WwrAJwo3IJaODiL-6pX-JEORq9Tke1tEEaXO444CIpoeoYxmzkI5Q6auU77CKwHZNQrSnobF7h6tXAxBuUDBNJ3zVCuayxoHn4iqhkO-7HuzfNertUtYgLmkQ5bN8QmsrWTZZRhcZ7R6iTg3u~n59-Oy70hFEetHizFujYNOl6hKIkgrq0bKQgB4Vl~Gc3TiLlRXCZTM4OPlAtd3nunw__&Key-Pair-Id=K1VGYTHIYLM9UM"
45
- },
46
- "thumbnail_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430a-442d-768f-90d5-a63534dfd811/output/preview.png?Expires=4945708800&Signature=ak5dXiVwX0-U-iVAiK6ADEW5WMGdwLNoSRweOntNKPjye4Z~mOLiD6z00yNtYSP9yIXgEaR~xlr~C4mls5Pj4Kb1eIOBJunh71HwLWCRbtPZy33bzGOouV2bSq8fR5PkVHRI5FRmlC8bHaljQkDWDpKTCBsrtOx4ncJWRFBnFbVmo3llDY1K3225kunkDK2nW51~Sc0V5~7GhzM8O7uQ~zH3E0YYRo8bPsj-0-SC~6M-0XYAdPbveBmojErwy7jRwKoRlhyI7f4fFimEnaYpqm9ZZ0tNq82Q3lxWujezANImY1U3H-crwMW2gIrmTNudmiqJ~MqaCGDOLYvmLXkR0A__&Key-Pair-Id=K1VGYTHIYLM9UM",
47
- "texture_urls": [],
48
- "consumed_credits": 20
49
- },
50
- "/openapi/v2/text-to-3d/01a1430b-df46-71b4-9dbf-41a42f98f5be": {
51
- "id": "01a1430b-df46-71b4-9dbf-41a42f98f5be",
52
- "type": "text-to-3d-refine",
53
- "mode": "refine",
54
- "name": "",
55
- "model_type": "standard",
56
- "seed": 1554,
57
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
58
- "texture_prompt": "",
59
- "texture_image_url": "",
60
- "status": "SUCCEEDED",
61
- "created_at": 1792126213966,
62
- "progress": 100,
63
- "started_at": 1792126213973,
64
- "finished_at": 1792126327872,
65
- "task_error": null,
66
- "model_urls": {
67
- "glb": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/model.glb?Expires=4945708800&Signature=lEVLPxabJbfC3tYsMSO5P7JVm03FUYuj2CD8kUezk850MRDsqCaszpTP2aUlz6w2FviYOw0jqxSebnOesXo4MSjjnNW4LyOvRUXbbEjqWx7lWpU0926XLfRy3ToGyDEAXkteSuNMDVz7q4K5BmCQTyi2NpxhZQIWVAdRHSYY4xngZ2uFGRnU0JYCiLQapKi7rNnq-EUQvi1ENLQuhLbuAoihOmjol0my7pGF18WK4HJHWGKslFKlH7b0zks8zAP1QwaEChXM3Qxf-MMo0cZp2xOnCh9JYqUpHx6ueaojW8G1WGsT16RzNa5LI-cr9bjWfW~rNBN6zdZzgPo54wCjPw__&Key-Pair-Id=K1VGYTHIYLM9UM",
68
- "fbx": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/model.fbx?Expires=4945708800&Signature=c5lyQkF-HFcJO~aWJ2yuIGCSeoqfdjlz3bSNPYuIvWjTTEN8VXVtUDrq45MLPT-TjoFVyXzrKg8-jvrE5FqLVOsjyi~TNirZyJUrPD4e0GUzloLDF27LKLKJutlRw2OmUtfAKAb3lfby0YLerIPHsFaq4x-eeFhoEn~8ZmvAdRA1ZLIkABDDhr716zQAPeX4zMNdGMIwKTilpSpZ6Oh3hvV5t3pGz53J3PLJlpaGY7beIC0ujX6gncEmtjwkI9-KK460LjVivzhWydZ0v35kVtpSiqcQXChwKu8WAp9eiGAGAQ3HYaoEk3agoVLmhlHv3or3V1BnNMenjSW5FRpu8w__&Key-Pair-Id=K1VGYTHIYLM9UM",
69
- "usdz": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/model.usdz?Expires=4945708800&Signature=BDEQDd9WTjpAfZwNgBnCKOtZTsW3JombVs9JDsJxE1mzjXw9eqJMTNDpvfbUhKNNu-tKCHkXwiOCXtRg83YH-21hzaFepZ8YIw4sRTAOSchfsLaLJRwt0bXUHwTw8mjOeVGs-vDMnatWlBA0ZvdohaPgGXDcHo4xoBI37H3xOmwWc0FuFNKX6N1PDtwvtTenBZMk~N2Z5OgJ9oMWS6n7fFHTR1s3aZSVzDE3OC1JKccZXuVEYSfH9~MT0AHvkfTc67bH9d-Xz0m3fjkJuV4xiJPZnq67ePLx40SixHnPEBVwxFub2fEKUBObAo~1fHVcfI1LZQgg7mylli~czF3lIw__&Key-Pair-Id=K1VGYTHIYLM9UM",
70
- "obj": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/model.obj?Expires=4945708800&Signature=e4rOm2hjviF~cpCb9SPIxP12Jz-AgpV78jyTfsSMZ5~BxHfVaYKeuTZs~IFunhFeg1asKS-m1ZMWEFn2P8TQgkjcyMPR2RPSYoOjanml0CQYsc9sro-bY8wHiXBLO61r0rvlkUaMSwbHCd6You-E2rjYkzqanChhTI-JPizCRvpyF-XmMQ14nv5GTrqBlV7MUurNqszMRIWh4~EH041h~bOnCvxBusniM~l1X0LZVcL19fqW66drpSWA0yOarnbUvS-yhx1IaVrxMo7CTYH-dDAOqoTLdqd52XjlY4ulpNh4zOMOAjJlwm0Ex7gb3tKIsCc97JG2KkA9Vnmb7puObg__&Key-Pair-Id=K1VGYTHIYLM9UM",
71
- "mtl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/model.mtl?Expires=4945708800&Signature=piSGkOhYzpsJUFBk87iZX2DYR3XOsaPzGfVUXT-drDVJgmJHGzCZJJOd67Jpdkj-7NMRiSYWZqO2k2OxzyhVkbfRDjYs3zOvtIx8TkCCui4B2Fjv5Y5rWiSCktHx1kkiQTmy7OfoO728zhOqp2x85abJcZIEzDrOJs430PTV6RuQOJcr~WeE7jT7nreN1E1EF1IrRjpB0XEeurXLdqbTbUUGpL3Day6D4vYuLxArpl1rgiEC4aSyV05Cgf7fmXAVXRzAu5OkIh1rY1oC~jLU9f8ceW5iXCtS4ykhAehbVLfD82sYIHeVq~ynTUiLnwqT3uGqWwWNp3C7huO5kitLFQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
72
- "stl": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/model.stl?Expires=4945708800&Signature=GtTU-esE7rV6LQmkHSF16YKdSb6mpqrUNwwQKLn30g0uwXHn~X-iLo-jTmn7iDapDSxMNgqUhq0EqtpKPLCYxObjTAYh4tWxT4tBOv6Zpo56O1kpfj-dbAOooc7oGoTnxeV3NC6S3XEDoVwjLVyvr4kVxukHVO7uZnnXmNbJ7mdUHj5scyKThH-3ZlGGmYqKw3IqfXicEG0A5heZwtxvN5AgYW~w7-~ujz4dARDS5h1~XtRRQqpv0MypocrPMWsJ863k2zm-xYVjnpTobp2WRQNbioIxPw8hoWkdUWPiYOiB0yYa2~E0M0oinHw1hS8LkglimOZ8MDK0kg2Hm12Hbg__&Key-Pair-Id=K1VGYTHIYLM9UM"
73
- },
74
- "thumbnail_url": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/preview.png?Expires=4945708800&Signature=V7dBr~D1mpC9OpiQk-3bdFqBSVPrBNRo7AIzc11fHefYdYEz57nOpRmSVZL-04znHaNaQmyGxDBeNwdCWHHcneuB0iorcOSQZeX1BAAcLFIlM7~AZ7ixbCzUbIVaGeRwHAz~AWl0JzjnPUqvY-1UN1r6wTGzQSOcfLromVilsfRPiYBvKha3RVAL7yYwNgLGbhgoJe54UC4gAJDBhTQcZWYdS8j3Wxg~d-OOin-AwxBQ386aKGLqqlDNbnakUGVR6gMNlMVQ1JVRXTsIdfCL-wqy-SyZAGPXIAZrEUrFhg39jnLPppzBagkzit-Qwo2NHAcM7sXIMRGTYzGEI9c-aQ__&Key-Pair-Id=K1VGYTHIYLM9UM",
75
- "texture_urls": [
76
- {
77
- "base_color": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/texture_0.png?Expires=4945708800&Signature=TbqQu5YJhub~Yzrx0J1BzrS1W5jqdeCSd4IRVis89uU49RSsBWGtxD9JGDmhGvvOjVXr3JDtlT8qff2z0Y36gQfNhD5gleBC2MK7Xs05SigdWXxcGtHscZYLBuke5voRRqAuNOMLfxBJAe~DtfGp7P8xZoMR5EMz2dq3P4U7eFGRnwTlwzxeHKQFMDL4hlFEmHDcX~5LzgvBpkDOjjI2IGW~tyB8P4EUGIJ0dkBRuhEIK0~U8ieMqJiO7TvzSJMKXX73BLeL0V-Nez9Fs7Lp0onLwGwyMGwVW2rd3kvlsKDJrOlaAuSG2QjjHQVLpW5ynlrB-2Hp9swxnksCtSTTvw__&Key-Pair-Id=K1VGYTHIYLM9UM",
78
- "metallic": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/texture_0_metallic.png?Expires=4945708800&Signature=MMRDq64VlX-rY~qA7L61U5HEDG~AN3l3MUPCaHs-ilJSY1oHmxiWg9moEOu624FRdLpcSyROecL56M2PWpiAsCtbv4N0dp3KcTUfSvz7J8CVc7upXcOE9D2AiSu7ei1V9hbE19nyA24FOuBz1aILfuX~KJA37~~FsFIlHvJcpUrcANt8XG0~kWlDjWOtzOhZOUyKbP3PCgJYJDIw7TVri7IIrUaoJisgahDPEXd0kkQqkk~tWFXfNdbehyp76YC1w2RqaTiacUOggUef-jmDiLZIGt2pM7IM8KoGzZXW3Y2S1ejAMD-zFuac5coZFToClW-Kyqy12tLOtEBRoLLG6Q__&Key-Pair-Id=K1VGYTHIYLM9UM",
79
- "roughness": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/texture_0_roughness.png?Expires=4945708800&Signature=AkUlQSK0Z-l0pI8NpZF8E4PLQ1p~-CYs4MZSRM2hr8~ohxIDfLQRJirNjPgbfdjVJdRLCJhfLaO2GVbXrhiO1Lzvx4KrpK-aPVzM6mfYz~t~HcrgXp5mrkdSmnnZQKqvaLq4bap7kiPZxRpMS4F-9ML60YFG0RR1t8SHpwELV6hg18sjBKKlQUGjU5bnm1FBfAixRumkj~kbPQQMdnrVwzUSxk006RRKtgg6-DgLnuutuZRt6VJdC14YVo45WWb-0weIXnFFTldtIPf-LQMnIGSi2zyEntNgX71hSjcGDSVx-tAznoxmXUF5dVpO25Zll92TjpydqZGRcwnpZgOa4A__&Key-Pair-Id=K1VGYTHIYLM9UM",
80
- "normal": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/texture_0_normal.png?Expires=4945708800&Signature=CEb7d5FZBZzDQPifiJhKpvEFDpAGV~vkerE6UQUvFgAjeH3D00g4f4MYMewjcqc5pnOoZema6EErlcEaJufwbMuUA~TzOWDcMkVWyVgsGePNhls~EPM1rRI9Gplnj-BSKsmqHt3YR9dTGC8ZQiowMm-8zFfKgr72HnuM4Zm8gI~8Hi44-iOIq7IqHQnluVQOPtmZr84dMTMVWIf4MLsoWwxYCdBb~0MzMYGeXCZpHsvBit8KqfKGmbXfujDyLIgNi0CfbLIAODK-Kpss9eKMMUjcELY3BX9ROcbfmLPUiLe3bzBZhuCOwKICecDfYE4IDDULXWixCuJTyDhNfe-Wmg__&Key-Pair-Id=K1VGYTHIYLM9UM",
81
- "emission": "https://assets.meshy.ai/8f6e5ce2-d8d8-44a6-a7ed-c92b05e979af/tasks/01a1430b-df46-71b4-9dbf-41a42f98f5be/output/texture_0_emission.png?Expires=4945708800&Signature=UqCOytIrxKiEL06KXACNbvSYAzaBz0PtTuvQiHvzZ61qGyvsy~1rwKDyFmhyCibCG8FHCR2rf1Si~cuqdxPuZOen~enIoxg~vFIsTCH9sI3zgGhhohO24ydCV5Q45Oqw7fN7J~mSmfq8iWVYvcocM2v3LvLRyuzuXL1pWNTNuD9OeGpR8aIv7SAiJkG7bEs7c3itfubemyAuut9ODL8qbQpWoM-RLZ-vsqPBLrPI40s4-blnJK1raCf3xO2zLK6mJE9qgnOCe8vt2oyoaeBIFHrZy8cTR6w7Wev1hzr-kA7cUuzwrx7ZC5buOkgMKC3HSj-anyYRAx4WnlXho0S~jg__&Key-Pair-Id=K1VGYTHIYLM9UM"
82
- }
83
- ],
84
- "remove_lighting": true,
85
- "consumed_credits": 10
86
- }
87
- }
88
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_001.json DELETED
@@ -1,72 +0,0 @@
1
- {
2
- "submissions": {
3
- "9851529c7194760542dccf8d290374fbb054cd406ce15aa6e032faecff369081": {
4
- "endpoint": "/files",
5
- "reserved_credits": 0,
6
- "status": "created",
7
- "remote_id": "file_00cd8007-e2d1-4126-9443-5eb00661da5a",
8
- "response": {
9
- "code": 0,
10
- "status": "success",
11
- "data": {
12
- "file_token": "file_00cd8007-e2d1-4126-9443-5eb00661da5a"
13
- }
14
- }
15
- },
16
- "63b27c86dfa61ce6d967b89ee48cb0e776f899885909b9deb0020f19d9e3e8d0": {
17
- "endpoint": "/generation/image-to-model",
18
- "reserved_credits": 40,
19
- "status": "created",
20
- "remote_id": "735f2f70-804b-4563-983e-7ce17bfa0be8",
21
- "response": {
22
- "code": 0,
23
- "status": "success",
24
- "data": {
25
- "type": "image_to_model",
26
- "status": "queued",
27
- "task_id": "735f2f70-804b-4563-983e-7ce17bfa0be8"
28
- }
29
- },
30
- "reported_credits": 40.0
31
- }
32
- },
33
- "results": {
34
- "/tasks/735f2f70-804b-4563-983e-7ce17bfa0be8": {
35
- "code": 0,
36
- "status": "success",
37
- "data": {
38
- "type": "image_to_model",
39
- "status": "success",
40
- "progress": 100,
41
- "input": {
42
- "pbr": true,
43
- "file": {
44
- "type": "jpeg",
45
- "object": {
46
- "key": "tcli_8a234d301b1e472ebe10d73adcd43336/20261016/735f2f70-804b-4563-983e-7ce17bfa0be8/image.jpeg",
47
- "bucket": "cn-openapi"
48
- }
49
- },
50
- "texture": true,
51
- "export_uv": true,
52
- "face_limit": 8000,
53
- "model_seed": 5501,
54
- "texture_seed": 5501,
55
- "model_version": "v3.1-20260211",
56
- "smart_low_poly": true,
57
- "texture_quality": "standard",
58
- "geometry_quality": "standard",
59
- "enable_image_autofix": false
60
- },
61
- "output": {
62
- "model_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/735f2f70-804b-4563-983e-7ce17bfa0be8/tripo_pbr_model_735f2f70-804b-4563-983e-7ce17bfa0be8.glb?auth_key=1792227387-bwqqthph-0-e780590f90033e346239dcf730c6f4f7",
63
- "rendered_image_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/735f2f70-804b-4563-983e-7ce17bfa0be8/legacy_mesh.webp?auth_key=1792227387-tddtsndb-0-a3b81fd0aaa003c2df1532eedb83315e"
64
- },
65
- "task_id": "735f2f70-804b-4563-983e-7ce17bfa0be8",
66
- "created_at": "2026-10-16T08:52:31.052577Z",
67
- "completed_at": "2026-10-16T08:56:24.535979Z",
68
- "credits_consumed": 40.0
69
- }
70
- }
71
- }
72
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_002.json DELETED
@@ -1,72 +0,0 @@
1
- {
2
- "submissions": {
3
- "96e532650443ed9b6068e8fbe385143f196d21feb0dbf40b0b6c429a81168623": {
4
- "endpoint": "/files",
5
- "reserved_credits": 0,
6
- "status": "created",
7
- "remote_id": "file_c4378b18-bced-45a0-a2b6-f186bc0578e7",
8
- "response": {
9
- "code": 0,
10
- "status": "success",
11
- "data": {
12
- "file_token": "file_c4378b18-bced-45a0-a2b6-f186bc0578e7"
13
- }
14
- }
15
- },
16
- "ae90ec7936cfeebc1c984cd55d4fb36654b15eedb452ebbe27c733ffec13d29f": {
17
- "endpoint": "/generation/image-to-model",
18
- "reserved_credits": 40,
19
- "status": "created",
20
- "remote_id": "51fc3810-e559-4357-8674-9f24c13383dc",
21
- "response": {
22
- "code": 0,
23
- "status": "success",
24
- "data": {
25
- "type": "image_to_model",
26
- "status": "queued",
27
- "task_id": "51fc3810-e559-4357-8674-9f24c13383dc"
28
- }
29
- },
30
- "reported_credits": 40.0
31
- }
32
- },
33
- "results": {
34
- "/tasks/51fc3810-e559-4357-8674-9f24c13383dc": {
35
- "code": 0,
36
- "status": "success",
37
- "data": {
38
- "type": "image_to_model",
39
- "status": "success",
40
- "progress": 100,
41
- "input": {
42
- "pbr": true,
43
- "file": {
44
- "type": "jpeg",
45
- "object": {
46
- "key": "tcli_8a234d301b1e472ebe10d73adcd43336/20261016/51fc3810-e559-4357-8674-9f24c13383dc/image.jpeg",
47
- "bucket": "cn-openapi"
48
- }
49
- },
50
- "texture": true,
51
- "export_uv": true,
52
- "face_limit": 4000,
53
- "model_seed": 1552,
54
- "texture_seed": 1552,
55
- "model_version": "v3.1-20260211",
56
- "smart_low_poly": true,
57
- "texture_quality": "standard",
58
- "geometry_quality": "standard",
59
- "enable_image_autofix": false
60
- },
61
- "output": {
62
- "model_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/51fc3810-e559-4357-8674-9f24c13383dc/tripo_pbr_model_51fc3810-e559-4357-8674-9f24c13383dc.glb?auth_key=1792227420-sbdmbrqw-0-a6be8ef5dac2677bd4f4976def766c63",
63
- "rendered_image_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/51fc3810-e559-4357-8674-9f24c13383dc/legacy_mesh.webp?auth_key=1792227420-bdngxogf-0-f36c118791ad76fe1084b93cc391f5e3"
64
- },
65
- "task_id": "51fc3810-e559-4357-8674-9f24c13383dc",
66
- "created_at": "2026-10-16T08:53:23.176572Z",
67
- "completed_at": "2026-10-16T08:56:56.708797Z",
68
- "credits_consumed": 40.0
69
- }
70
- }
71
- }
72
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_003.json DELETED
@@ -1,54 +0,0 @@
1
- {
2
- "submissions": {
3
- "5acb82382f3101bc77c2466640c0b4ae188fb8ad3d2ac82f879be07fcaf16daf": {
4
- "endpoint": "/generation/text-to-model",
5
- "reserved_credits": 30,
6
- "status": "created",
7
- "remote_id": "e4c923ad-3363-4403-9067-bcb9ebb5b3f0",
8
- "response": {
9
- "code": 0,
10
- "status": "success",
11
- "data": {
12
- "type": "text_to_model",
13
- "status": "queued",
14
- "task_id": "e4c923ad-3363-4403-9067-bcb9ebb5b3f0"
15
- }
16
- },
17
- "reported_credits": 30.0
18
- }
19
- },
20
- "results": {
21
- "/tasks/e4c923ad-3363-4403-9067-bcb9ebb5b3f0": {
22
- "code": 0,
23
- "status": "success",
24
- "data": {
25
- "type": "text_to_model",
26
- "status": "success",
27
- "progress": 100,
28
- "input": {
29
- "pbr": true,
30
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
31
- "texture": true,
32
- "export_uv": true,
33
- "face_limit": 5000,
34
- "model_seed": 1553,
35
- "texture_seed": 1553,
36
- "model_version": "v3.1-20260211",
37
- "smart_low_poly": true,
38
- "texture_quality": "standard",
39
- "geometry_quality": "standard",
40
- "texture_alignment": "geometry"
41
- },
42
- "output": {
43
- "model_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/e4c923ad-3363-4403-9067-bcb9ebb5b3f0/tripo_pbr_model_e4c923ad-3363-4403-9067-bcb9ebb5b3f0.glb?auth_key=1792227597-cfacvupw-0-7d9d5e51ae26e40ef070f5cc5ca8a8ef",
44
- "rendered_image_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/e4c923ad-3363-4403-9067-bcb9ebb5b3f0/legacy_mesh.webp?auth_key=1792227597-ugaktvjq-0-c50629aab104ee767233ce8dd04c1651",
45
- "generated_image_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/e4c923ad-3363-4403-9067-bcb9ebb5b3f0/text2image_e4c923ad-3363-4403-9067-bcb9ebb5b3f0.jpeg?auth_key=1792227597-yzvnkutm-0-43e7b20d320c8747be3ae6e13a061a74"
46
- },
47
- "task_id": "e4c923ad-3363-4403-9067-bcb9ebb5b3f0",
48
- "created_at": "2026-10-16T08:56:31.308384Z",
49
- "completed_at": "2026-10-16T08:59:57.033628Z",
50
- "credits_consumed": 30.0
51
- }
52
- }
53
- }
54
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_004.json DELETED
@@ -1,54 +0,0 @@
1
- {
2
- "submissions": {
3
- "a00933af721ccd0afc80581bfde7972a08960512eb87a3520bfb8d8e3474bb35": {
4
- "endpoint": "/generation/text-to-model",
5
- "reserved_credits": 30,
6
- "status": "created",
7
- "remote_id": "f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281",
8
- "response": {
9
- "code": 0,
10
- "status": "success",
11
- "data": {
12
- "type": "text_to_model",
13
- "status": "queued",
14
- "task_id": "f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281"
15
- }
16
- },
17
- "reported_credits": 30.0
18
- }
19
- },
20
- "results": {
21
- "/tasks/f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281": {
22
- "code": 0,
23
- "status": "success",
24
- "data": {
25
- "type": "text_to_model",
26
- "status": "success",
27
- "progress": 100,
28
- "input": {
29
- "pbr": true,
30
- "prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
31
- "texture": true,
32
- "export_uv": true,
33
- "face_limit": 4000,
34
- "model_seed": 1554,
35
- "texture_seed": 1554,
36
- "model_version": "v3.1-20260211",
37
- "smart_low_poly": true,
38
- "texture_quality": "standard",
39
- "geometry_quality": "standard",
40
- "texture_alignment": "geometry"
41
- },
42
- "output": {
43
- "model_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281/tripo_pbr_model_f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281.glb?auth_key=1792227624-xrpjkkdl-0-40ede17ac958fa4119066802c88541ac",
44
- "rendered_image_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281/legacy_mesh.webp?auth_key=1792227624-toqzmsks-0-e5c3ed58fcb6bc157d98452f5c922142",
45
- "generated_image_url": "https://openapi.cdn.tripo3d.com/tcli_8a234d301b1e472ebe10d73adcd43336/20261016/f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281/text2image_f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281.jpeg?auth_key=1792227624-qrgaiuhi-0-22348e6382c3ca6947ca1b08251deb48"
46
- },
47
- "task_id": "f8eb5a08-fd79-47f2-a8da-2a6ae7ad8281",
48
- "created_at": "2026-10-16T08:57:02.403517Z",
49
- "completed_at": "2026-10-16T09:00:19.912948Z",
50
- "credits_consumed": 30.0
51
- }
52
- }
53
- }
54
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png DELETED

Git LFS Details

  • SHA256: b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d
  • Pointer size: 129 Bytes
  • Size of remote file: 4.87 kB
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png DELETED

Git LFS Details

  • SHA256: ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55
  • Pointer size: 131 Bytes
  • Size of remote file: 687 kB
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/cloud_api.py DELETED
@@ -1,557 +0,0 @@
1
- """
2
- models/common/cloud_api.py
3
-
4
- Shared plumbing for closed-source ("cloud API") model wrappers — the part of
5
- Tripo / Meshy / Rodin / ElevenLabs / Kling wrappers that is identical no matter
6
- what the provider generates.
7
-
8
- Contract: `agent_skills/develop_harness/api_model_require.md` R9.
9
- This module owns R9.6 (submit → poll → download behind a blocking call),
10
- R9.7 (API key from the environment, at call time), R9.8 (response cache, so a
11
- re-run is not re-billed) and R9.9 (retry with backoff, retryable vs. terminal).
12
-
13
- It knows nothing about 3D, audio or video — only HTTP, money and failure modes.
14
-
15
- Dependencies: stdlib at import time; `requests` is imported lazily on first use
16
- so `test/harness/smoke.py` keeps working on a machine that does not have it.
17
-
18
- Usage:
19
- from models.common import cloud_api
20
-
21
- key = cloud_api.require_api_key(None, "TRIPO_API_KEY",
22
- "https://platform.tripo3d.ai/api-keys")
23
- client = cloud_api.CloudAPIClient("https://openapi.tripo3d.ai/v3", key)
24
- data = client.request("POST", "/generation/text-to-model", json={...})
25
- """
26
- from __future__ import annotations
27
-
28
- import hashlib
29
- import json
30
- import logging
31
- import os
32
- import time
33
- from pathlib import Path
34
- from typing import Any, Callable, Iterable, Optional
35
-
36
- logger = logging.getLogger(__name__)
37
- logger.addHandler(logging.NullHandler())
38
-
39
- #: HTTP statuses worth retrying — transient by nature.
40
- RETRYABLE_STATUS: frozenset[int] = frozenset({408, 409, 425, 429, 500, 502, 503, 504})
41
-
42
- #: HTTP statuses that will never succeed on retry: bad request, auth, quota.
43
- TERMINAL_STATUS: frozenset[int] = frozenset({400, 401, 402, 403, 404, 405, 413, 415, 422})
44
-
45
-
46
- # ── Errors ────────────────────────────────────────────────────────────────────
47
-
48
-
49
- class CloudAPIError(RuntimeError):
50
- """
51
- Base class for every cloud-API failure.
52
-
53
- Attributes:
54
- status: HTTP status code, when the failure came from a response.
55
- task_id: Provider task id, when one had already been created — keep it,
56
- a timed-out task usually finishes anyway and can be fetched
57
- later (`api_model_require.md` R9.6).
58
- payload: Decoded response body, truncated for readability.
59
- """
60
-
61
- retryable = False
62
-
63
- def __init__(self, message: str, *, status: Optional[int] = None,
64
- task_id: Optional[str] = None, payload: Any = None):
65
- super().__init__(message)
66
- self.status = status
67
- self.task_id = task_id
68
- self.payload = payload
69
-
70
-
71
- class CloudAPIAuthError(CloudAPIError):
72
- """401 / 403 — the key is missing, wrong, or lacks the entitlement."""
73
-
74
-
75
- class CloudAPIQuotaError(CloudAPIError):
76
- """402 or a provider-specific "insufficient credits" body. Terminal."""
77
-
78
-
79
- class CloudAPIRequestError(CloudAPIError):
80
- """400 / 404 / 422 — the request itself is wrong. Terminal."""
81
-
82
-
83
- class CloudAPIServerError(CloudAPIError):
84
- """5xx and connection-level failures. Retryable."""
85
-
86
- retryable = True
87
-
88
-
89
- class CloudAPIRateLimitError(CloudAPIServerError):
90
- """429 — retryable, honour `Retry-After` when present."""
91
-
92
-
93
- class CloudTaskFailed(CloudAPIError):
94
- """The provider accepted the task and then reported failure."""
95
-
96
-
97
- class CloudTaskTimeout(CloudAPIError):
98
- """The task did not reach a terminal state within `timeout` seconds."""
99
-
100
-
101
- # ── Credentials ───────────────────────────────────────────────────────────────
102
-
103
-
104
- def require_api_key(value: Optional[str], env_var: str, signup_url: str,
105
- *, who: str = "This model") -> str:
106
- """
107
- Resolve an API key, failing fast with an actionable message (R1.6 / R9.7).
108
-
109
- Args:
110
- value: Explicit key passed to the constructor, or None.
111
- env_var: Environment variable to fall back to.
112
- signup_url: Where a human gets a key.
113
- who: Name used in the error message.
114
-
115
- Returns:
116
- The key.
117
-
118
- Raises:
119
- CloudAPIAuthError: no key available.
120
- """
121
- key = value or os.environ.get(env_var)
122
- if key:
123
- return key.strip()
124
- raise CloudAPIAuthError(
125
- f"{who} needs an API key and none was found.\n"
126
- f" 1. get one at {signup_url}\n"
127
- f" 2. export {env_var}=<your key> # never commit it\n"
128
- f" (or pass api_key=... to the constructor)"
129
- )
130
-
131
-
132
- def redact(key: Optional[str]) -> str:
133
- """Key fingerprint safe to put in a log line or `meta.json`."""
134
- if not key:
135
- return "<unset>"
136
- return f"{key[:4]}…{key[-2:]} (len {len(key)})"
137
-
138
-
139
- # ── HTTP ───────���──────────────────────────────────────────────────────────────
140
-
141
-
142
- def _requests():
143
- """Import `requests` lazily with an actionable message (R1.5 / R1.6)."""
144
- try:
145
- import requests # noqa: PLC0415
146
- except ImportError as e: # pragma: no cover - environment dependent
147
- raise ImportError(
148
- "Cloud-API model wrappers need the `requests` package:\n"
149
- " bash scripts/asset_env_setup/cloud_api_install.sh\n"
150
- " # or simply: python -m pip install requests\n"
151
- "(only needed for the API backends; the local-weight models and "
152
- "test/harness/smoke.py do not import it.)"
153
- ) from e
154
- return requests
155
-
156
-
157
- def _truncate(payload: Any, limit: int = 600) -> Any:
158
- text = payload if isinstance(payload, str) else json.dumps(payload, default=str)
159
- return text if len(text) <= limit else text[:limit] + f"… (+{len(text) - limit} chars)"
160
-
161
-
162
- def classify_http_error(status: int, body: Any, *,
163
- task_id: Optional[str] = None) -> CloudAPIError:
164
- """
165
- Map an HTTP status onto the right exception class (R9.9).
166
-
167
- The body is inspected too: providers happily answer 200/400 with a
168
- "not enough credits" payload, which must not be retried.
169
- """
170
- text = _truncate(body)
171
- low = str(text).lower()
172
- # Providers phrase "you are out of money" in prose, and the status code alone
173
- # does not distinguish it from a bad key: Tripo answers **403** with
174
- # "You don't have enough credit to create this task". Matching the body first
175
- # is what keeps "wrong key" and "no money" apart for the caller.
176
- if any(s in low for s in ("insufficient", "enough credit", "enough credits",
177
- "out of credit", "purchase more credit",
178
- "quota exceeded", "no credits", "balance")):
179
- return CloudAPIQuotaError(
180
- f"HTTP {status}: out of credits — {text}", status=status,
181
- task_id=task_id, payload=body)
182
- if status in (401, 403):
183
- return CloudAPIAuthError(
184
- f"HTTP {status}: authentication rejected — {text}", status=status,
185
- task_id=task_id, payload=body)
186
- if status == 402:
187
- return CloudAPIQuotaError(
188
- f"HTTP {status}: payment required — {text}", status=status,
189
- task_id=task_id, payload=body)
190
- if status == 429:
191
- return CloudAPIRateLimitError(
192
- f"HTTP {status}: rate limited — {text}", status=status,
193
- task_id=task_id, payload=body)
194
- if status in TERMINAL_STATUS:
195
- return CloudAPIRequestError(
196
- f"HTTP {status}: request rejected — {text}", status=status,
197
- task_id=task_id, payload=body)
198
- if status in RETRYABLE_STATUS or status >= 500:
199
- return CloudAPIServerError(
200
- f"HTTP {status}: server error — {text}", status=status,
201
- task_id=task_id, payload=body)
202
- return CloudAPIError(
203
- f"HTTP {status}: unexpected response — {text}", status=status,
204
- task_id=task_id, payload=body)
205
-
206
-
207
- class CloudAPIClient:
208
- """
209
- Thin retrying HTTP client for one provider.
210
-
211
- Args:
212
- base_url: Root of the API, without a trailing slash.
213
- api_key: Bearer token.
214
- timeout: Per-request socket timeout in seconds (not the task timeout).
215
- max_retries: Extra attempts after the first one, for retryable failures.
216
- backoff: Base of the exponential backoff, seconds.
217
- auth_header: Header name / value template, for providers that do not
218
- use `Authorization: Bearer <key>`.
219
- user_agent: Sent as-is; helps providers attribute traffic.
220
- """
221
-
222
- def __init__(
223
- self,
224
- base_url: str,
225
- api_key: str,
226
- *,
227
- timeout: int = 60,
228
- max_retries: int = 3,
229
- backoff: float = 1.5,
230
- auth_header: str = "Authorization",
231
- auth_template: str = "Bearer {key}",
232
- user_agent: str = "3AGameFactory/0.1 (+https://github.com/OpenDCAI/GameFactory-3A)",
233
- ):
234
- self.base_url = base_url.rstrip("/")
235
- self._api_key = api_key
236
- self.timeout = timeout
237
- self.max_retries = max_retries
238
- self.backoff = backoff
239
- self.auth_header = auth_header
240
- self.auth_template = auth_template
241
- self.user_agent = user_agent
242
- self._session = None
243
-
244
- # -- session ---------------------------------------------------------------
245
-
246
- @property
247
- def session(self):
248
- """Lazily created `requests.Session` (kept alive across polls)."""
249
- if self._session is None:
250
- requests = _requests()
251
- self._session = requests.Session()
252
- self._session.headers.update({
253
- self.auth_header: self.auth_template.format(key=self._api_key),
254
- "User-Agent": self.user_agent,
255
- })
256
- return self._session
257
-
258
- def close(self) -> None:
259
- """Idempotent — safe after a failed construction and safe to call twice."""
260
- if self._session is not None:
261
- try:
262
- self._session.close()
263
- finally:
264
- self._session = None
265
-
266
- # -- requests --------------------------------------------------------------
267
-
268
- def request(
269
- self,
270
- method: str,
271
- path: str,
272
- *,
273
- json_body: Optional[dict] = None,
274
- files: Optional[dict] = None,
275
- data: Optional[dict] = None,
276
- params: Optional[dict] = None,
277
- expect_json: bool = True,
278
- ) -> Any:
279
- """
280
- Perform one API call, retrying retryable failures (R9.9).
281
-
282
- Args:
283
- method: "GET" / "POST" / …
284
- path: Path appended to `base_url`, or a full URL.
285
- json_body: JSON payload.
286
- files/data: multipart payload (`requests` semantics).
287
- params: Query string.
288
- expect_json: Parse the response as JSON; otherwise return `bytes`.
289
-
290
- Returns:
291
- Parsed JSON (dict / list) or raw `bytes`.
292
-
293
- Raises:
294
- CloudAPIError subclass — terminal errors immediately, retryable ones
295
- after `max_retries` attempts.
296
- """
297
- requests = _requests()
298
- url = path if path.startswith("http") else f"{self.base_url}/{path.lstrip('/')}"
299
-
300
- last: Optional[CloudAPIError] = None
301
- for attempt in range(self.max_retries + 1):
302
- try:
303
- resp = self.session.request(
304
- method, url, json=json_body, files=files, data=data,
305
- params=params, timeout=self.timeout,
306
- )
307
- except (requests.ConnectionError, requests.Timeout) as e:
308
- last = CloudAPIServerError(f"{method} {url} failed at transport level: {e}")
309
- else:
310
- if resp.ok:
311
- if not expect_json:
312
- return resp.content
313
- try:
314
- return resp.json()
315
- except ValueError as e:
316
- raise CloudAPIError(
317
- f"{method} {url} returned {resp.status_code} with a "
318
- f"non-JSON body: {_truncate(resp.text)}"
319
- ) from e
320
- try:
321
- body: Any = resp.json()
322
- except ValueError:
323
- body = resp.text
324
- err = classify_http_error(resp.status_code, body)
325
- if not err.retryable:
326
- raise err
327
- last = err
328
- retry_after = resp.headers.get("Retry-After")
329
- if retry_after:
330
- try:
331
- time.sleep(min(float(retry_after), 60.0))
332
- except ValueError:
333
- pass
334
-
335
- if attempt < self.max_retries:
336
- delay = self.backoff * (2 ** attempt)
337
- logger.warning("[cloud_api] %s %s attempt %d/%d failed (%s); retrying in %.1fs",
338
- method, url, attempt + 1, self.max_retries + 1, last, delay)
339
- time.sleep(delay)
340
-
341
- assert last is not None
342
- raise last
343
-
344
- def download(self, url: str, *, max_bytes: Optional[int] = None) -> bytes:
345
- """
346
- Fetch a generated artifact. Signed asset URLs are short-lived — call this
347
- immediately after the task reports success.
348
-
349
- Args:
350
- url: Absolute URL from the task output.
351
- max_bytes: Reject anything larger, so a wrong URL cannot fill a disk.
352
-
353
- Returns:
354
- The file content.
355
- """
356
- content = self.request("GET", url, expect_json=False)
357
- if max_bytes is not None and len(content) > max_bytes:
358
- raise CloudAPIError(
359
- f"downloaded artifact is {len(content)} bytes, over the "
360
- f"{max_bytes} byte limit: {url}")
361
- if not content:
362
- raise CloudAPIError(f"downloaded artifact is empty: {url}")
363
- return content
364
-
365
-
366
- # ── Polling ───────────────────────────────────────────────────────────────────
367
-
368
-
369
- def poll_task(
370
- fetch: Callable[[], dict],
371
- *,
372
- task_id: str,
373
- timeout: float = 300.0,
374
- poll_interval: float = 3.0,
375
- done_states: Iterable[str],
376
- failed_states: Iterable[str],
377
- state_key: str = "status",
378
- progress_key: str = "progress",
379
- verbose: bool = False,
380
- ) -> dict:
381
- """
382
- Block until a provider task reaches a terminal state (R9.6).
383
-
384
- Args:
385
- fetch: Callable returning the decoded task object.
386
- task_id: Only used for messages — every error carries it.
387
- timeout: Wall-clock budget in seconds.
388
- poll_interval: Seconds between polls.
389
- done_states: States meaning success (compared case-insensitively).
390
- failed_states: States meaning failure.
391
- state_key: Field holding the state.
392
- progress_key: Field holding 0-100 progress, when the provider sends one.
393
- verbose: Print progress; otherwise only DEBUG logs (R3.7).
394
-
395
- Returns:
396
- The task object in its successful terminal state.
397
-
398
- Raises:
399
- CloudTaskFailed: provider reported failure.
400
- CloudTaskTimeout: budget exhausted; the task may still finish server-side,
401
- the error carries `task_id` so it can be fetched later.
402
- """
403
- done = {s.lower() for s in done_states}
404
- failed = {s.lower() for s in failed_states}
405
- deadline = time.time() + timeout
406
- last_progress = -1
407
-
408
- while True:
409
- task = fetch()
410
- state = str(task.get(state_key, "")).lower()
411
- progress = task.get(progress_key)
412
-
413
- if progress != last_progress:
414
- last_progress = progress
415
- logger.debug("[cloud_api] task %s: %s (%s%%)", task_id, state, progress)
416
- if verbose:
417
- print(f"[cloud_api] task {task_id}: {state} ({progress}%)")
418
-
419
- if state in done:
420
- return task
421
- if state in failed:
422
- raise CloudTaskFailed(
423
- f"task {task_id} ended as {state!r}: {_truncate(task)}",
424
- task_id=task_id, payload=task)
425
-
426
- remaining = deadline - time.time()
427
- if remaining <= 0:
428
- raise CloudTaskTimeout(
429
- f"task {task_id} still {state!r} after {timeout:.0f}s. It may still "
430
- f"finish server-side — query it with the task id above, or raise "
431
- f"`timeout=` on the model.",
432
- task_id=task_id, payload=task)
433
- time.sleep(min(poll_interval, remaining))
434
-
435
-
436
- # ── Response cache ────────────────────────────────────────────────────────────
437
-
438
-
439
- def request_hash(payload: dict) -> str:
440
- """
441
- Stable sha256 over a request description (R9.8).
442
-
443
- The payload must contain everything that changes the output — model version,
444
- prompt, image bytes hash, every generation parameter, the output format.
445
- Never put the API key in it.
446
- """
447
- canonical = json.dumps(payload, sort_keys=True, ensure_ascii=False, default=str)
448
- return hashlib.sha256(canonical.encode("utf-8")).hexdigest()
449
-
450
-
451
- def bytes_hash(data: bytes) -> str:
452
- """sha256 of a binary input, e.g. the reference image."""
453
- return hashlib.sha256(data).hexdigest()
454
-
455
-
456
- class ResponseCache:
457
- """
458
- Content-addressed cache of generated artifacts.
459
-
460
- A cloud call costs credits, so an identical request must never be sent twice
461
- (R9.8). Layout::
462
-
463
- <cache_dir>/<provider>/<hash>.<ext> the artifact
464
- <cache_dir>/<provider>/<hash>.meta.json request description + task id
465
-
466
- Args:
467
- cache_dir: Root directory. `None` disables caching entirely.
468
- provider: Sub-directory name, so two providers never collide.
469
- """
470
-
471
- def __init__(self, cache_dir: Optional[str | Path], provider: str):
472
- self.root = Path(cache_dir).expanduser() / provider if cache_dir else None
473
- self.provider = provider
474
-
475
- @property
476
- def enabled(self) -> bool:
477
- return self.root is not None
478
-
479
- def _paths(self, key: str, ext: str) -> tuple[Path, Path]:
480
- assert self.root is not None
481
- return self.root / f"{key}.{ext}", self.root / f"{key}.meta.json"
482
-
483
- def get(self, key: str, ext: str = "glb") -> Optional[bytes]:
484
- """Cached artifact bytes, or None. Performs no network I/O, ever."""
485
- if not self.enabled:
486
- return None
487
- blob, _ = self._paths(key, ext)
488
- if blob.is_file() and blob.stat().st_size > 0:
489
- logger.info("[cloud_api] cache hit (%s): %s", self.provider, blob.name)
490
- return blob.read_bytes()
491
- return None
492
-
493
- def put(self, key: str, data: bytes, meta: dict, ext: str = "glb") -> Path:
494
- """Store an artifact plus the request that produced it. Returns its path."""
495
- if not self.enabled:
496
- return Path()
497
- blob, meta_path = self._paths(key, ext)
498
- blob.parent.mkdir(parents=True, exist_ok=True)
499
- blob.write_bytes(data)
500
- meta_path.write_text(
501
- json.dumps({**meta, "cache_key": key, "bytes": len(data)},
502
- indent=2, ensure_ascii=False, default=str),
503
- encoding="utf-8",
504
- )
505
- return blob
506
-
507
-
508
- # ── Input encoding ────────────────────────────────────────────────────────────
509
-
510
-
511
- def image_to_png_bytes(image: Any) -> bytes:
512
- """
513
- Normalize an image input to PNG bytes.
514
-
515
- Accepts a `PIL.Image.Image`, raw `bytes` (returned unchanged), or a
516
- `np.ndarray`. Paths are rejected on purpose — `model_require.md` R3.2 says a
517
- model takes the object, not a path.
518
-
519
- Returns:
520
- PNG-encoded bytes (RGBA preserved when present).
521
- """
522
- if isinstance(image, (bytes, bytearray)):
523
- return bytes(image)
524
- if isinstance(image, (str, Path)):
525
- raise TypeError(
526
- "models take an image object, not a path (model_require.md R3.2). "
527
- "Open it in the operator: Image.open(path).convert('RGBA')"
528
- )
529
-
530
- from io import BytesIO # noqa: PLC0415
531
-
532
- from PIL import Image as PILImage # noqa: PLC0415
533
-
534
- if not isinstance(image, PILImage.Image):
535
- try:
536
- import numpy as np # noqa: PLC0415
537
- except ImportError:
538
- np = None # type: ignore[assignment]
539
- if np is not None and isinstance(image, np.ndarray):
540
- image = PILImage.fromarray(image)
541
- else:
542
- raise TypeError(
543
- f"unsupported image type {type(image).__name__}; expected "
544
- f"PIL.Image.Image, numpy.ndarray or bytes"
545
- )
546
-
547
- buf = BytesIO()
548
- image.save(buf, format="PNG")
549
- return buf.getvalue()
550
-
551
-
552
- def image_to_data_uri(image: Any, mime: str = "image/png") -> str:
553
- """PNG data URI (`data:image/png;base64,…`) for providers that take one inline."""
554
- import base64 # noqa: PLC0415
555
-
556
- payload = base64.b64encode(image_to_png_bytes(image)).decode("ascii")
557
- return f"data:{mime};base64,{payload}"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/glb_utils.py DELETED
@@ -1,578 +0,0 @@
1
- """
2
- models/common/glb_utils.py
3
-
4
- Dependency-free inspection of binary glTF (`.glb`) buffers.
5
-
6
- Why this exists instead of `trimesh.load(...)`: the API wrappers must report the
7
- triangle count of what they just paid for (`api_model_require.md` R9.8), and
8
- that number has to be available on a CPU-only box with nothing but the stdlib
9
- installed. Parsing the JSON chunk of a GLB is enough for counting.
10
-
11
- `glb_summary` and `glb_triangle_count` read the JSON chunk only. `read_glb_mesh`
12
- decodes vertex data as well, because a generated part has to be *composed* with
13
- primitives rather than merely counted, and that needs its triangles in the same
14
- coordinate frame as everything else.
15
-
16
- Usage:
17
- from models.common.glb_utils import glb_summary, read_glb_mesh
18
- info = glb_summary(open("model.glb", "rb").read())
19
- print(info["triangles"], info["materials"])
20
-
21
- mesh = read_glb_mesh(open("part.glb", "rb").read())
22
- print(len(mesh["positions"]), mesh["low"], mesh["high"])
23
- """
24
- from __future__ import annotations
25
-
26
- import json
27
- import math
28
- import struct
29
- from typing import Any, Optional, Sequence
30
-
31
- GLB_MAGIC = b"glTF"
32
-
33
- #: glTF primitive mode 4 == TRIANGLES. Only triangle primitives are counted.
34
- _MODE_TRIANGLES = 4
35
-
36
- #: glTF componentType -> (struct format, byte size). Indices are allowed to be
37
- #: any of the unsigned types; positions are float in every file worth reading.
38
- _COMPONENT_TYPES: dict[int, tuple[str, int]] = {
39
- 5120: ("b", 1), # BYTE
40
- 5121: ("B", 1), # UNSIGNED_BYTE
41
- 5122: ("h", 2), # SHORT
42
- 5123: ("H", 2), # UNSIGNED_SHORT
43
- 5125: ("I", 4), # UNSIGNED_INT
44
- 5126: ("f", 4), # FLOAT
45
- }
46
-
47
- #: glTF accessor type -> component count.
48
- _TYPE_COUNTS: dict[str, int] = {
49
- "SCALAR": 1, "VEC2": 2, "VEC3": 3, "VEC4": 4,
50
- "MAT2": 4, "MAT3": 9, "MAT4": 16,
51
- }
52
-
53
-
54
- class GLBParseError(ValueError):
55
- """The buffer is not a GLB we can read."""
56
-
57
-
58
- def is_glb(data: bytes) -> bool:
59
- """True when `data` starts with the 12-byte GLB header."""
60
- return len(data) >= 12 and data[:4] == GLB_MAGIC
61
-
62
-
63
- def glb_json_chunk(data: bytes) -> dict[str, Any]:
64
- """
65
- Return the parsed JSON chunk of a GLB.
66
-
67
- Args:
68
- data: Whole `.glb` file content.
69
-
70
- Returns:
71
- The glTF document as a dict.
72
-
73
- Raises:
74
- GLBParseError: bad magic, truncated file, or a missing JSON chunk.
75
- """
76
- if not is_glb(data):
77
- raise GLBParseError(
78
- f"not a GLB: expected magic {GLB_MAGIC!r}, got {data[:4]!r}"
79
- )
80
- if len(data) < 20:
81
- raise GLBParseError(f"GLB truncated: {len(data)} bytes")
82
-
83
- _magic, version, total_len = struct.unpack_from("<4sII", data, 0)
84
- if total_len > len(data):
85
- raise GLBParseError(
86
- f"GLB header claims {total_len} bytes, buffer holds {len(data)}"
87
- )
88
-
89
- offset = 12
90
- while offset + 8 <= len(data):
91
- chunk_len, chunk_type = struct.unpack_from("<I4s", data, offset)
92
- offset += 8
93
- chunk = data[offset:offset + chunk_len]
94
- offset += chunk_len
95
- if chunk_type == b"JSON":
96
- try:
97
- return json.loads(chunk.decode("utf-8"))
98
- except (UnicodeDecodeError, json.JSONDecodeError) as e:
99
- raise GLBParseError(f"GLB JSON chunk is not valid JSON: {e}") from e
100
- raise GLBParseError(f"GLB (version {version}) has no JSON chunk")
101
-
102
-
103
- def glb_triangle_count(data: bytes) -> Optional[int]:
104
- """
105
- Triangle count of a GLB, or None when it cannot be determined.
106
-
107
- Counts every triangle-mode primitive of every mesh once, i.e. the *asset's*
108
- triangle budget, not the per-instance count after node duplication.
109
-
110
- Returns:
111
- int triangles, or None if the JSON chunk is unreadable.
112
- """
113
- try:
114
- doc = glb_json_chunk(data)
115
- except GLBParseError:
116
- return None
117
-
118
- accessors = doc.get("accessors", [])
119
- total = 0
120
- for mesh in doc.get("meshes", []):
121
- for prim in mesh.get("primitives", []):
122
- if prim.get("mode", _MODE_TRIANGLES) != _MODE_TRIANGLES:
123
- continue
124
- idx = prim.get("indices")
125
- if idx is not None and idx < len(accessors):
126
- count = accessors[idx].get("count", 0)
127
- else:
128
- pos = prim.get("attributes", {}).get("POSITION")
129
- count = accessors[pos].get("count", 0) if pos is not None and pos < len(accessors) else 0
130
- total += count // 3
131
- return total
132
-
133
-
134
- def glb_summary(data: bytes) -> dict[str, Any]:
135
- """
136
- Cheap description of a GLB, for logs and `meta.json`.
137
-
138
- Returns:
139
- dict with keys:
140
- bytes (int) — buffer size
141
- triangles (int | None) — see `glb_triangle_count`
142
- meshes, materials, textures, images, nodes (int)
143
- generator (str | None) — glTF `asset.generator`
144
- error (str) — present only when the JSON chunk is unreadable
145
- """
146
- out: dict[str, Any] = {"bytes": len(data)}
147
- try:
148
- doc = glb_json_chunk(data)
149
- except GLBParseError as e:
150
- out.update({"triangles": None, "error": str(e)})
151
- return out
152
-
153
- out.update({
154
- "triangles": glb_triangle_count(data),
155
- "meshes": len(doc.get("meshes", [])),
156
- "materials": len(doc.get("materials", [])),
157
- "textures": len(doc.get("textures", [])),
158
- "images": len(doc.get("images", [])),
159
- "nodes": len(doc.get("nodes", [])),
160
- "generator": (doc.get("asset") or {}).get("generator"),
161
- })
162
- return out
163
-
164
-
165
- # --------------------------------------------------------------------------
166
- # Vertex decoding
167
- #
168
- # Needed to *compose* a generated part with primitives, as opposed to merely
169
- # reporting on it: composition means putting its triangles into the same
170
- # coordinate frame as the rest, and that requires reading them.
171
- # --------------------------------------------------------------------------
172
-
173
-
174
- def glb_binary_chunk(data: bytes) -> bytes:
175
- """Return the BIN chunk of a GLB, or b"" when there is none."""
176
- if not is_glb(data):
177
- raise GLBParseError(f"not a GLB: got magic {data[:4]!r}")
178
- offset = 12
179
- while offset + 8 <= len(data):
180
- chunk_len, chunk_type = struct.unpack_from("<I4s", data, offset)
181
- offset += 8
182
- if chunk_type == b"BIN\x00":
183
- return data[offset:offset + chunk_len]
184
- offset += chunk_len
185
- return b""
186
-
187
-
188
- def _read_accessor(doc: dict[str, Any], binary: bytes, index: int) -> list[tuple]:
189
- """Decode one accessor into a list of component tuples.
190
-
191
- Honours ``byteStride``. Interleaved buffers are normal out of an
192
- exporter, and reading one as tightly packed yields positions that are a
193
- silent mixture of coordinates and normals — it parses, it renders, it is
194
- garbage.
195
- """
196
-
197
- accessor = doc["accessors"][index]
198
- component_type = accessor["componentType"]
199
- if component_type not in _COMPONENT_TYPES:
200
- raise GLBParseError(f"unsupported componentType {component_type}")
201
- fmt, component_size = _COMPONENT_TYPES[component_type]
202
- per_element = _TYPE_COUNTS[accessor["type"]]
203
- count = accessor["count"]
204
-
205
- view_index = accessor.get("bufferView")
206
- if view_index is None:
207
- # A sparse or zero-filled accessor. Legal, and not something a mesh
208
- # we are about to weld into another asset should be relying on.
209
- return [tuple([0] * per_element)] * count
210
-
211
- view = doc["bufferViews"][view_index]
212
- if view.get("buffer", 0) != 0:
213
- raise GLBParseError("only single-buffer GLBs are supported")
214
- base = view.get("byteOffset", 0) + accessor.get("byteOffset", 0)
215
- packed = component_size * per_element
216
- stride = view.get("byteStride") or packed
217
-
218
- out: list[tuple] = []
219
- element_format = "<" + fmt * per_element
220
- for element in range(count):
221
- start = base + element * stride
222
- if start + packed > len(binary):
223
- raise GLBParseError(
224
- f"accessor {index} reads past the {len(binary)}-byte buffer"
225
- )
226
- out.append(struct.unpack_from(element_format, binary, start))
227
- return out
228
-
229
-
230
- def _node_matrix(node: dict[str, Any]) -> list[float]:
231
- """A node's local transform as a 16-float column-major matrix.
232
-
233
- Handles both an explicit ``matrix`` and a TRS triple, since a file may use
234
- either. Ignoring the transform looks like a modelling error: a mesh
235
- authored Z-up whose node carries the -90 that stands it upright arrives on
236
- its side, and the natural reaction is to rotate it back in the spec —
237
- hard-coding a correction for something the file already stated.
238
- """
239
-
240
- if "matrix" in node:
241
- return [float(value) for value in node["matrix"]]
242
-
243
- tx, ty, tz = [float(value) for value in node.get("translation", (0, 0, 0))]
244
- qx, qy, qz, qw = [float(value) for value in node.get("rotation", (0, 0, 0, 1))]
245
- sx, sy, sz = [float(value) for value in node.get("scale", (1, 1, 1))]
246
-
247
- # Quaternion to rotation matrix.
248
- xx, yy, zz = qx * qx, qy * qy, qz * qz
249
- xy, xz, yz = qx * qy, qx * qz, qy * qz
250
- wx, wy, wz = qw * qx, qw * qy, qw * qz
251
- rotation = (
252
- (1 - 2 * (yy + zz), 2 * (xy + wz), 2 * (xz - wy)),
253
- (2 * (xy - wz), 1 - 2 * (xx + zz), 2 * (yz + wx)),
254
- (2 * (xz + wy), 2 * (yz - wx), 1 - 2 * (xx + yy)),
255
- )
256
- scale = (sx, sy, sz)
257
- return [
258
- rotation[0][0] * scale[0], rotation[0][1] * scale[0], rotation[0][2] * scale[0], 0.0,
259
- rotation[1][0] * scale[1], rotation[1][1] * scale[1], rotation[1][2] * scale[1], 0.0,
260
- rotation[2][0] * scale[2], rotation[2][1] * scale[2], rotation[2][2] * scale[2], 0.0,
261
- tx, ty, tz, 1.0,
262
- ]
263
-
264
-
265
- def _multiply(a: Sequence[float], b: Sequence[float]) -> list[float]:
266
- """Column-major 4x4 product, applying `b` then `a`."""
267
- out = [0.0] * 16
268
- for column in range(4):
269
- for row in range(4):
270
- out[column * 4 + row] = sum(
271
- a[k * 4 + row] * b[column * 4 + k] for k in range(4)
272
- )
273
- return out
274
-
275
-
276
- _IDENTITY = [1.0, 0, 0, 0, 0, 1.0, 0, 0, 0, 0, 1.0, 0, 0, 0, 0, 1.0]
277
-
278
-
279
- def read_glb_mesh(data: bytes) -> dict[str, Any]:
280
- """Decode every triangle of a GLB into one flat, world-space mesh.
281
-
282
- Walks the scene graph and applies each node's accumulated transform, so
283
- the result is in the file's own world coordinates rather than in the
284
- local frame of whichever mesh happened to be first.
285
-
286
- The node hierarchy is deliberately *not* preserved. A part fetched from
287
- a generator is one object as far as the composition is concerned — the
288
- thing worth keeping addressable is the spec's part id, which the writer
289
- attaches, and a generator's internal node names ("mesh_0", "Object_2")
290
- are not names anything downstream can use.
291
-
292
- Returns:
293
- dict with ``positions`` and ``normals`` as lists of 3-tuples,
294
- ``uvs`` as 2-tuples, ``indices`` as a flat list, ``low``/``high`` as
295
- the axis-aligned bounds, and ``images`` / ``material_runs`` carrying
296
- whatever textures the file brought with it. Normals are generated
297
- per-face when the file omits them.
298
-
299
- Textures come across because a generated part's whole contribution is
300
- surface. Dropping them and letting the part inherit the spec's flat PBR
301
- factors is consistent but throws away the thing that was paid for: the
302
- stippling on a grip, the panel creases and decals on a car shell. That
303
- needs three things travelling together — UVs per vertex, the image
304
- bytes, and which triangles use which image — so ``material_runs``
305
- records index ranges rather than a single material for the part.
306
- """
307
-
308
- doc = glb_json_chunk(data)
309
- binary = glb_binary_chunk(data)
310
- if not binary:
311
- raise GLBParseError(
312
- "the GLB has no BIN chunk, so its vertex data is external; "
313
- "only self-contained GLBs can be composed"
314
- )
315
-
316
- positions: list[tuple[float, float, float]] = []
317
- normals: list[tuple[float, float, float]] = []
318
- uvs: list[tuple[float, float]] = []
319
- indices: list[int] = []
320
- material_runs: list[dict[str, Any]] = []
321
-
322
- def add_mesh(mesh_index: int, matrix: Sequence[float]) -> None:
323
- # The inverse transpose of the upper 3x3, for the normals. A non-uniform
324
- # node scale makes the forward matrix wrong for them, and the symptom is
325
- # lighting that reads as a modelling error rather than as a bad import.
326
- normal_matrix = _inverse_transpose_3x3(matrix)
327
-
328
- for primitive in doc["meshes"][mesh_index].get("primitives", []):
329
- if primitive.get("mode", _MODE_TRIANGLES) != _MODE_TRIANGLES:
330
- continue
331
- attributes = primitive.get("attributes", {})
332
- if "POSITION" not in attributes:
333
- continue
334
-
335
- local = _read_accessor(doc, binary, attributes["POSITION"])
336
- base = len(positions)
337
- for x, y, z in local:
338
- positions.append((
339
- matrix[0] * x + matrix[4] * y + matrix[8] * z + matrix[12],
340
- matrix[1] * x + matrix[5] * y + matrix[9] * z + matrix[13],
341
- matrix[2] * x + matrix[6] * y + matrix[10] * z + matrix[14],
342
- ))
343
-
344
- if "NORMAL" in attributes:
345
- for nx, ny, nz in _read_accessor(doc, binary, attributes["NORMAL"]):
346
- tx = normal_matrix[0] * nx + normal_matrix[3] * ny + normal_matrix[6] * nz
347
- ty = normal_matrix[1] * nx + normal_matrix[4] * ny + normal_matrix[7] * nz
348
- tz = normal_matrix[2] * nx + normal_matrix[5] * ny + normal_matrix[8] * nz
349
- length = math.sqrt(tx * tx + ty * ty + tz * tz) or 1.0
350
- normals.append((tx / length, ty / length, tz / length))
351
- else:
352
- normals.extend([(0.0, 0.0, 0.0)] * len(local))
353
-
354
- # UVs are per-vertex and must stay aligned with positions, so a
355
- # primitive without them still contributes placeholders — a short
356
- # UV list would silently shift every following part's texturing.
357
- if "TEXCOORD_0" in attributes:
358
- uvs.extend(
359
- (float(u), float(v))
360
- for u, v in _read_accessor(doc, binary, attributes["TEXCOORD_0"])
361
- )
362
- else:
363
- uvs.extend([(0.0, 0.0)] * len(local))
364
-
365
- if "indices" in primitive and primitive["indices"] is not None:
366
- local_indices = [
367
- value[0] for value in
368
- _read_accessor(doc, binary, primitive["indices"])
369
- ]
370
- else:
371
- local_indices = list(range(len(local)))
372
-
373
- run_start = len(indices)
374
- indices.extend(base + value for value in local_indices)
375
- material_runs.append({
376
- "start": run_start,
377
- "count": len(local_indices),
378
- "material": primitive.get("material"),
379
- })
380
-
381
- def walk(node_index: int, parent: Sequence[float]) -> None:
382
- node = doc["nodes"][node_index]
383
- matrix = _multiply(parent, _node_matrix(node))
384
- if node.get("mesh") is not None:
385
- add_mesh(node["mesh"], matrix)
386
- for child in node.get("children", ()):
387
- walk(child, matrix)
388
-
389
- scenes = doc.get("scenes") or []
390
- scene = scenes[doc.get("scene", 0)] if scenes else None
391
- roots = scene.get("nodes", ()) if scene else range(len(doc.get("nodes", [])))
392
- for root in roots:
393
- walk(root, _IDENTITY)
394
-
395
- if not positions:
396
- raise GLBParseError("the GLB contains no triangle geometry")
397
-
398
- # Fill in any face normals the file left out, so a consumer never has to
399
- # ask whether normals are present.
400
- _fill_missing_normals(positions, normals, indices)
401
-
402
- low = tuple(min(point[axis] for point in positions) for axis in range(3))
403
- high = tuple(max(point[axis] for point in positions) for axis in range(3))
404
- return {
405
- "positions": positions,
406
- "normals": normals,
407
- "uvs": uvs,
408
- "indices": indices,
409
- "triangles": len(indices) // 3,
410
- "low": low,
411
- "high": high,
412
- "material_runs": material_runs,
413
- "materials": _read_materials(doc, binary),
414
- "has_uvs": any(u != 0.0 or v != 0.0 for u, v in uvs),
415
- }
416
-
417
-
418
- def _read_materials(doc: dict[str, Any], binary: bytes) -> list[dict[str, Any]]:
419
- """Each material's PBR factors, its base-colour image and its MR image.
420
-
421
- Base colour and metallic-roughness. Not the normal or occlusion maps:
422
- those would give the composed asset a texture set the primitives beside it
423
- lack, and a car body with a normal map next to a flat-shaded spoiler reads
424
- as two assets. Metallic-roughness is different in kind — it is not detail
425
- on top of a surface, it is *which surface it is*.
426
-
427
- Taking the map is what makes the factors trustworthy. An exporter that puts
428
- those channels in a texture leaves the factors at glTF's default ``1.0`` to
429
- multiply it, so the two only mean anything together. This function used to
430
- take the colour alone and substitute dielectric defaults (``0.0``/``0.55``)
431
- on the grounds that inherited factors would render a polymer grip as
432
- sandblasted steel — true, but it threw away the answer to keep from
433
- guessing wrong. Plate armour came out as grey plastic: every generated
434
- piece of the knight had ``metallicFactor 1.0`` and an MR map in the file,
435
- and both were discarded.
436
-
437
- ``factors_from`` is still recorded, because a material with no map at all
438
- still has to fall back and a caller needs to know which happened.
439
- """
440
-
441
- out: list[dict[str, Any]] = []
442
- images = doc.get("images", [])
443
- textures = doc.get("textures", [])
444
- views = doc.get("bufferViews", [])
445
-
446
- def image_bytes(reference: dict[str, Any]) -> tuple[bytes | None, str | None]:
447
- index = reference.get("index")
448
- if index is None or index >= len(textures):
449
- return None, None
450
- image_index = textures[index].get("source")
451
- if image_index is None or image_index >= len(images):
452
- return None, None
453
- image = images[image_index]
454
- view_index = image.get("bufferView")
455
- if view_index is not None and view_index < len(views):
456
- view = views[view_index]
457
- start = view.get("byteOffset", 0)
458
- return (
459
- bytes(binary[start:start + view["byteLength"]]),
460
- image.get("mimeType") or "image/png",
461
- )
462
- if str(image.get("uri", "")).startswith("data:"):
463
- # A data URI is legal and some exporters emit one even in a GLB.
464
- # Decoded here so the caller never has to care which form was used.
465
- import base64
466
-
467
- header, _, payload = image["uri"].partition(",")
468
- return (
469
- base64.b64decode(payload),
470
- header.split(":")[1].split(";")[0] if ":" in header
471
- else "image/png",
472
- )
473
- return None, None
474
-
475
- for material in doc.get("materials", []):
476
- pbr = material.get("pbrMetallicRoughness") or {}
477
- colour_image, colour_mime = image_bytes(pbr.get("baseColorTexture") or {})
478
- mr_image, mr_mime = image_bytes(
479
- pbr.get("metallicRoughnessTexture") or {})
480
-
481
- if mr_image is not None:
482
- # The map is coming with us, so the factors mean what they say:
483
- # they scale it. glTF's default is 1.0 for exactly this case.
484
- metallic = float(pbr.get("metallicFactor", 1.0))
485
- roughness = float(pbr.get("roughnessFactor", 1.0))
486
- provenance = "file"
487
- elif (pbr.get("metallicRoughnessTexture") or {}).get("index") is not None:
488
- # A map is declared but its bytes could not be read. Falling back
489
- # to dielectric defaults, because inheriting factors written to
490
- # scale a map we do not have renders a polymer grip as steel.
491
- metallic, roughness, provenance = 0.0, 0.55, "assumed"
492
- else:
493
- metallic = float(pbr.get("metallicFactor", 1.0))
494
- roughness = float(pbr.get("roughnessFactor", 1.0))
495
- provenance = "file"
496
-
497
- out.append({
498
- "name": material.get("name") or "",
499
- "baseColor": [
500
- float(value)
501
- for value in pbr.get("baseColorFactor", [1.0, 1.0, 1.0, 1.0])
502
- ],
503
- "metallic": metallic,
504
- "roughness": roughness,
505
- "factors_from": provenance,
506
- "image": colour_image,
507
- "mimeType": colour_mime,
508
- "mrImage": mr_image,
509
- "mrMimeType": mr_mime,
510
- })
511
- return out
512
-
513
-
514
- def _inverse_transpose_3x3(matrix: Sequence[float]) -> list[float]:
515
- """Inverse transpose of a column-major 4x4's rotation/scale block."""
516
- a = (matrix[0], matrix[1], matrix[2])
517
- b = (matrix[4], matrix[5], matrix[6])
518
- c = (matrix[8], matrix[9], matrix[10])
519
- determinant = (
520
- a[0] * (b[1] * c[2] - b[2] * c[1])
521
- - a[1] * (b[0] * c[2] - b[2] * c[0])
522
- + a[2] * (b[0] * c[1] - b[1] * c[0])
523
- )
524
- if abs(determinant) < 1e-20:
525
- return [1.0, 0, 0, 0, 1.0, 0, 0, 0, 1.0]
526
- # cofactor / determinant, already the inverse transpose.
527
- return [
528
- (b[1] * c[2] - b[2] * c[1]) / determinant,
529
- (b[2] * c[0] - b[0] * c[2]) / determinant,
530
- (b[0] * c[1] - b[1] * c[0]) / determinant,
531
- (a[2] * c[1] - a[1] * c[2]) / determinant,
532
- (a[0] * c[2] - a[2] * c[0]) / determinant,
533
- (a[1] * c[0] - a[0] * c[1]) / determinant,
534
- (a[1] * b[2] - a[2] * b[1]) / determinant,
535
- (a[2] * b[0] - a[0] * b[2]) / determinant,
536
- (a[0] * b[1] - a[1] * b[0]) / determinant,
537
- ]
538
-
539
-
540
- def _fill_missing_normals(
541
- positions: list[tuple[float, float, float]],
542
- normals: list[tuple[float, float, float]],
543
- indices: list[int],
544
- ) -> None:
545
- """Replace zero-length normals with the area-weighted face average."""
546
-
547
- missing = [
548
- index for index, normal in enumerate(normals)
549
- if normal[0] == 0.0 and normal[1] == 0.0 and normal[2] == 0.0
550
- ]
551
- if not missing:
552
- return
553
-
554
- accumulated = {index: [0.0, 0.0, 0.0] for index in missing}
555
- needed = set(missing)
556
- for triangle in range(len(indices) // 3):
557
- ia, ib, ic = indices[triangle * 3:triangle * 3 + 3]
558
- if not (ia in needed or ib in needed or ic in needed):
559
- continue
560
- pa, pb, pc = positions[ia], positions[ib], positions[ic]
561
- e1 = (pb[0] - pa[0], pb[1] - pa[1], pb[2] - pa[2])
562
- e2 = (pc[0] - pa[0], pc[1] - pa[1], pc[2] - pa[2])
563
- cross = (
564
- e1[1] * e2[2] - e1[2] * e2[1],
565
- e1[2] * e2[0] - e1[0] * e2[2],
566
- e1[0] * e2[1] - e1[1] * e2[0],
567
- )
568
- for index in (ia, ib, ic):
569
- if index in accumulated:
570
- for axis in range(3):
571
- accumulated[index][axis] += cross[axis]
572
-
573
- for index, vector in accumulated.items():
574
- length = math.sqrt(sum(value * value for value in vector))
575
- normals[index] = (
576
- (vector[0] / length, vector[1] / length, vector[2] / length)
577
- if length > 1e-20 else (0.0, 1.0, 0.0)
578
- )
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/meshy_model.py DELETED
@@ -1,463 +0,0 @@
1
- """
2
- models/gen_3d_object/meshy_model.py
3
-
4
- MeshyModel — wrapper around Meshy's cloud 3D-generation API (image-to-3D and
5
- text-to-3D), producing GLB / FBX / OBJ / USDZ / STL.
6
-
7
- Reference: https://docs.meshy.ai/ (base https://api.meshy.ai)
8
-
9
- Second backend for the `gen_3d_object` slot. Its `infer` / `infer_and_save`
10
- signatures are identical to `TripoModel`'s and positionally identical to
11
- `Trellis2Model`'s, so `Gen3DObjectOperator` dispatches on class name alone
12
- (model_require.md R6).
13
-
14
- CONTRACT DEVIATIONS (model_require.md targets local-weight models; the rules that
15
- replace it live in agent_skills/develop_harness/api_model_require.md):
16
- C1 R2.1/R2.2 → R9.1 `model_path` carries a *model version id* ("meshy-6").
17
- C2 R2.3 → R9.2 `device` accepted for interface parity and ignored.
18
- C3 R3.6/R5.5 → R9.3 no @torch.no_grad(); torch is not involved.
19
- C4 R4.1-R4.3 → R9.4 `unload()` is a no-op that closes the HTTP session.
20
- C5 R3.3/R3.4 → R9.5 `seed` is forwarded for text-to-3D only, where the API
21
- accepts it; image-to-3D has no seed. Reproducibility is
22
- NOT guaranteed server-side and is never faked locally.
23
- C6 R3.1 → R9.6 task-based API (submit → poll → download); `infer()`
24
- blocks until completion or `timeout` (default 300 s).
25
- Text-to-3D runs two tasks (preview → refine) when a
26
- texture is requested; both ids are logged.
27
- C7 R1.2 uses the documented `infer_and_save(..., output_path)`
28
- exception; the bytes-returning `infer()` exists alongside.
29
- C8 (new) → R9.8 `cache_dir` prevents re-billing an identical request.
30
- C9 (new) → R9.9 `max_retries` + exponential backoff; retryable and
31
- terminal failures are split.
32
-
33
- Environment:
34
- MESHY_API_KEY required at the first infer() call, not at construction.
35
- Get one at https://www.meshy.ai/api (a `msy_dummy_...`
36
- test-mode key returns canned assets without spending credits).
37
- MESHY_API_BASE optional override of the API root.
38
- `pip install requests` is needed for the API backends only.
39
-
40
- Usage:
41
- from models.gen_3d_object.meshy_model import MeshyModel
42
- model = MeshyModel(model_path="meshy-6", output_format="fbx")
43
- path = model.infer_and_save(image, output_path="out/model.fbx", seed=42)
44
- """
45
- from __future__ import annotations
46
-
47
- import logging
48
- import os
49
- import time
50
- from pathlib import Path
51
- from typing import Any, Optional
52
-
53
- from models.common import cloud_api
54
- from models.common.glb_utils import glb_summary
55
-
56
- logger = logging.getLogger(__name__)
57
-
58
- SIGNUP_URL = "https://www.meshy.ai/api"
59
-
60
- DEFAULT_API_BASE = "https://api.meshy.ai"
61
-
62
- #: Short names accepted for `model_path`. Unknown values pass through untouched.
63
- MODEL_ALIASES: dict[str, str] = {
64
- "meshy": "meshy-6",
65
- "meshy-6-text-to-3d": "meshy-6",
66
- "meshy-6-image-to-3d": "meshy-6",
67
- "meshy-5-text-to-3d": "meshy-5",
68
- "meshy-5-image-to-3d": "meshy-5",
69
- }
70
-
71
- DEFAULT_MODEL = "meshy-6"
72
-
73
- DONE_STATES = ("succeeded", "success", "completed")
74
- FAILED_STATES = ("failed", "canceled", "cancelled", "expired")
75
-
76
- #: Meshy rejects a target_polycount outside this range with a 400 (terminal).
77
- POLYCOUNT_RANGE = (100, 300_000)
78
-
79
- #: Formats the API can return directly, i.e. keys of `model_urls`.
80
- SUPPORTED_FORMATS = ("glb", "fbx", "obj", "usdz", "stl", "mtl", "3mf")
81
-
82
-
83
- class MeshyModel:
84
- """
85
- Cloud image-to-3D / text-to-3D through Meshy.
86
-
87
- Args:
88
- model_path: [C1] Meshy `ai_model` id ("meshy-6", "meshy-5", "latest"),
89
- or an alias from `MODEL_ALIASES`.
90
- device: [C2] Accepted for interface parity, ignored. "cpu" is fine.
91
- api_key: Explicit key; defaults to `$MESHY_API_KEY`, read at first call.
92
- timeout: [C6] Seconds to wait for one generation task (each stage of
93
- the text-to-3D preview → refine chain gets the full budget).
94
- poll_interval: Seconds between status polls.
95
- max_retries: [C9] Extra attempts for retryable HTTP failures.
96
- cache_dir: [C8] Directory for the response cache. None disables it.
97
- output_format: Any of `SUPPORTED_FORMATS`; picked out of `model_urls`.
98
- low_poly: [B4.3] Sets `model_type="lowpoly"`, i.e. the service produces
99
- a low-poly mesh instead of us decimating one locally.
100
- texture / enable_pbr / topology / should_remesh:
101
- Forwarded to the generation endpoints.
102
- refine: Text-to-3D only. Meshy's preview stage returns an untextured
103
- mesh; the refine stage textures it and costs extra credits.
104
- Ignored when `texture=False`.
105
- api_base: Override the API root (defaults to `$MESHY_API_BASE`).
106
- verbose: Print task progress; off by default (R3.7).
107
- """
108
-
109
- PROVIDER = "meshy"
110
-
111
- def __init__(
112
- self,
113
- model_path: str = DEFAULT_MODEL,
114
- device: str = "cuda",
115
- *,
116
- api_key: Optional[str] = None,
117
- # Measured: one image-to-3D with PBR textures took 283 s here, and the
118
- # Tripo backend took 601 s for the same job. Timing out does not refund
119
- # the credits, so the default errs long; the error carries the task_id.
120
- timeout: int = 1800,
121
- poll_interval: float = 3.0,
122
- max_retries: int = 3,
123
- cache_dir: Optional[str] = None,
124
- output_format: str = "glb",
125
- low_poly: bool = False,
126
- texture: bool = True,
127
- enable_pbr: bool = True,
128
- topology: str = "triangle",
129
- should_remesh: Optional[bool] = None,
130
- refine: bool = True,
131
- api_base: Optional[str] = None,
132
- http_timeout: int = 60,
133
- verbose: bool = False,
134
- **model_specific: Any,
135
- ):
136
- # [CONTRACT-DEVIATION C1] model_path is a version id, not a weight location.
137
- # [CONTRACT-DEVIATION C2] device accepted for interface parity, ignored.
138
- self.model_path = MODEL_ALIASES.get(str(model_path).lower(), model_path)
139
- self.device = device
140
-
141
- # R2.4 — store everything before anything loads.
142
- self.api_key = api_key
143
- self.timeout = timeout
144
- self.poll_interval = poll_interval
145
- self.max_retries = max_retries
146
- self.cache_dir = cache_dir
147
- self.output_format = output_format.lower().lstrip(".")
148
- self.low_poly = low_poly
149
- self.texture = texture
150
- self.enable_pbr = enable_pbr
151
- self.topology = topology
152
- self.should_remesh = should_remesh
153
- self.refine = refine
154
- self.api_base = (api_base or os.environ.get("MESHY_API_BASE")
155
- or DEFAULT_API_BASE)
156
- self.http_timeout = http_timeout
157
- self.verbose = verbose
158
- self.model_specific = model_specific
159
-
160
- if self.output_format not in SUPPORTED_FORMATS:
161
- raise ValueError(
162
- f"output_format={self.output_format!r} is not one of "
163
- f"{SUPPORTED_FORMATS}. Meshy returns these directly in model_urls."
164
- )
165
-
166
- self._client: Optional[cloud_api.CloudAPIClient] = None
167
- self._cache = cloud_api.ResponseCache(cache_dir, self.PROVIDER)
168
-
169
- #: Record of the most recent **real** call — see `TripoModel.last_call_info`.
170
- self.last_call_info: dict[str, Any] = {}
171
-
172
- # ── plumbing ──────────────────────────────────────────────────────────────
173
-
174
- @property
175
- def client(self) -> cloud_api.CloudAPIClient:
176
- """HTTP client, created on first use. [R9.7] The key is resolved here."""
177
- if self._client is None:
178
- key = cloud_api.require_api_key(
179
- self.api_key, "MESHY_API_KEY", SIGNUP_URL, who="MeshyModel")
180
- self._client = cloud_api.CloudAPIClient(
181
- self.api_base, key,
182
- timeout=self.http_timeout,
183
- max_retries=self.max_retries,
184
- )
185
- return self._client
186
-
187
- def unload(self) -> None:
188
- """[C4] Close the HTTP session. Idempotent, and safe before any call."""
189
- if self._client is not None:
190
- self._client.close()
191
- self._client = None
192
-
193
- close = unload
194
-
195
- @staticmethod
196
- def _task_id_of(resp: Any) -> str:
197
- """Meshy answers a creation call with `{"result": "<task id>"}`."""
198
- if isinstance(resp, dict):
199
- for key in ("result", "id", "task_id"):
200
- value = resp.get(key)
201
- if isinstance(value, str) and value:
202
- return value
203
- raise cloud_api.CloudAPIError(
204
- f"Meshy accepted the request but returned no task id: {resp!r}")
205
-
206
- def _fetch_task(self, endpoint: str, task_id: str) -> dict:
207
- task = self.client.request("GET", f"{endpoint}/{task_id}")
208
- if not isinstance(task, dict):
209
- raise cloud_api.CloudAPIError(
210
- f"unexpected task payload for {task_id}: {task!r}", task_id=task_id)
211
- return task
212
-
213
- def _model_url(self, task: dict) -> str:
214
- """URL of the requested format, with an actionable message when absent."""
215
- urls = task.get("model_urls") or {}
216
- url = urls.get(self.output_format)
217
- if isinstance(url, str) and url.startswith("http"):
218
- return url
219
- raise cloud_api.CloudAPIError(
220
- f"task {task.get('id')} finished but has no {self.output_format!r} "
221
- f"URL. Available: {sorted(urls)}. Set output_format to one of those.",
222
- task_id=task.get("id"), payload=task)
223
-
224
- def _clamp_polycount(self, value: Optional[int]) -> Optional[int]:
225
- """
226
- Keep `target_polycount` inside the documented range.
227
-
228
- `Gen3DObjectOperator` forwards TRELLIS.2's default of 1_000_000, which
229
- Meshy rejects with a terminal 400. Clamping (loudly) keeps the operator
230
- interchangeable (R6) instead of making the caller special-case backends.
231
- """
232
- if value is None:
233
- return None
234
- low, high = POLYCOUNT_RANGE
235
- clamped = max(low, min(int(value), high))
236
- if clamped != int(value):
237
- logger.warning("[meshy] target_polycount %s clamped to %s (API range %s-%s)",
238
- value, clamped, low, high)
239
- return clamped
240
-
241
- # ── inference ─────────────────────────────────────────────────────────────
242
-
243
- def infer(
244
- self,
245
- image: Any = None,
246
- seed: int = 42,
247
- decimation_target: Optional[int] = None,
248
- texture_size: Optional[int] = None,
249
- *,
250
- prompt: Optional[str] = None,
251
- negative_prompt: Optional[str] = None,
252
- image_url: Optional[str] = None,
253
- verbose: Optional[bool] = None,
254
- **provider_kwargs: Any,
255
- ) -> bytes:
256
- """
257
- Generate one 3D asset and return the model file bytes.
258
-
259
- Args:
260
- image: Reference image as `PIL.Image.Image`, `np.ndarray` or PNG
261
- `bytes` — sent inline as a base64 data URI. Not a path (R3.2).
262
- seed: [C5] Forwarded for text-to-3D only; image-to-3D has no seed
263
- parameter. Reproducibility is not guaranteed.
264
- decimation_target: Target face count → `target_polycount`, clamped to
265
- the API's 100-300000 range.
266
- texture_size: Accepted for signature parity with `Trellis2Model` and
267
- **ignored** — Meshy exposes no texture resolution parameter.
268
- prompt: Text prompt for text-to-3D. [A4 decision] Same rationale as
269
- `TripoModel`: keeps text-to-3D reachable without a second slot.
270
- negative_prompt: Sent as `texture_prompt`'s counterpart is absent in
271
- the API — rejected explicitly rather than silently dropped.
272
- image_url: Publicly reachable image URL, used instead of the data URI.
273
- verbose: Override the constructor's verbosity for this call.
274
- **provider_kwargs: Merged into the request body verbatim.
275
-
276
- Returns:
277
- The model file as `bytes`, in `output_format`.
278
-
279
- Raises:
280
- CloudAPIAuthError / CloudAPIQuotaError / CloudTaskFailed /
281
- CloudTaskTimeout — see `models/common/cloud_api.py`.
282
- """
283
- verbose = self.verbose if verbose is None else verbose
284
- if image is None and image_url is None and not prompt:
285
- raise ValueError(
286
- "MeshyModel.infer needs an image (`image=` / `image_url=`) or a "
287
- "text `prompt=`; got neither.")
288
- if negative_prompt:
289
- raise ValueError(
290
- "Meshy has no negative_prompt parameter. Fold the constraint into "
291
- "`prompt`, or use TripoModel which supports it.")
292
- if texture_size is not None:
293
- logger.debug("[meshy] texture_size=%s ignored; the API has no equivalent",
294
- texture_size)
295
-
296
- polycount = self._clamp_polycount(decimation_target)
297
- use_image = image is not None or image_url is not None
298
-
299
- payload: dict[str, Any] = {
300
- "ai_model": self.model_path,
301
- "model_type": "lowpoly" if self.low_poly else "standard",
302
- "topology": self.topology,
303
- }
304
- if polycount is not None:
305
- payload["target_polycount"] = polycount
306
- if self.should_remesh is not None:
307
- payload["should_remesh"] = self.should_remesh
308
-
309
- if use_image:
310
- payload["should_texture"] = self.texture
311
- payload["enable_pbr"] = self.enable_pbr
312
- else:
313
- payload["mode"] = "preview"
314
- payload["prompt"] = prompt
315
- # [CONTRACT-DEVIATION C5] accepted by the text endpoint only.
316
- payload["seed"] = int(seed)
317
- payload.update(provider_kwargs)
318
-
319
- # [C8] Everything that changes the output goes into the key; the image is
320
- # represented by its hash, and the API key never appears.
321
- png = cloud_api.image_to_png_bytes(image) if image is not None else None
322
- cache_key = cloud_api.request_hash({
323
- "provider": self.PROVIDER,
324
- "payload": payload,
325
- "image_sha256": cloud_api.bytes_hash(png) if png else None,
326
- "image_url": image_url,
327
- "refine": self.refine and self.texture and not use_image,
328
- "output_format": self.output_format,
329
- })
330
- cached = self._cache.get(cache_key, ext=self.output_format)
331
- if cached is not None:
332
- self.last_call_info = {
333
- "provider": self.PROVIDER, "cached": True, "cache_key": cache_key,
334
- "bytes": len(cached),
335
- **(glb_summary(cached) if self.output_format == "glb" else {}),
336
- }
337
- return cached
338
-
339
- t0 = time.time()
340
- task_ids: list[str] = []
341
-
342
- if use_image:
343
- endpoint = "/openapi/v1/image-to-3d"
344
- payload["image_url"] = (
345
- image_url if png is None else cloud_api.image_to_data_uri(image))
346
- task = self._submit_and_wait(endpoint, payload, verbose=verbose,
347
- task_ids=task_ids)
348
- else:
349
- endpoint = "/openapi/v2/text-to-3d"
350
- task = self._submit_and_wait(endpoint, payload, verbose=verbose,
351
- task_ids=task_ids)
352
- if self.texture and self.refine:
353
- # [C6] Preview geometry is untextured; refine is a second billed
354
- # task that paints it. Skip with texture=False or refine=False.
355
- refine_payload = {
356
- "mode": "refine",
357
- "preview_task_id": task_ids[-1],
358
- "ai_model": self.model_path,
359
- "enable_pbr": self.enable_pbr,
360
- }
361
- task = self._submit_and_wait(endpoint, refine_payload,
362
- verbose=verbose, task_ids=task_ids)
363
-
364
- data = self.client.download(self._model_url(task))
365
- elapsed = time.time() - t0
366
-
367
- info = glb_summary(data) if self.output_format == "glb" else {"bytes": len(data)}
368
- self.last_call_info = {
369
- "provider": self.PROVIDER,
370
- "model": self.model_path,
371
- "task_id": task_ids[-1],
372
- "task_ids": task_ids,
373
- "elapsed_sec": round(elapsed, 2),
374
- "credits": task.get("credits_consumed", task.get("credits")),
375
- "cached": False,
376
- "cache_key": cache_key,
377
- "format": self.output_format,
378
- # The provider's own render of what it produced. Recording the URL
379
- # (rather than downloading it here) keeps the model out of artifact
380
- # placement (R1.2) while giving the caller a free visual check —
381
- # "the asset exists" and "the asset is what was asked for" are
382
- # different claims, and only a picture settles the second one.
383
- "preview_url": task.get("thumbnail_url"),
384
- **info,
385
- }
386
- # R9.8 — one line per billed call.
387
- logger.info(
388
- "[meshy] task=%s model=%s %.1fs credits=%s triangles=%s bytes=%s",
389
- task_ids[-1], self.model_path, elapsed, self.last_call_info.get("credits"),
390
- info.get("triangles"), len(data),
391
- )
392
- if verbose:
393
- print(f"[meshy] {self.last_call_info}")
394
-
395
- self._cache.put(cache_key, data, {
396
- "provider": self.PROVIDER, "payload": payload,
397
- **{k: v for k, v in self.last_call_info.items() if k != "cache_key"},
398
- }, ext=self.output_format)
399
- return data
400
-
401
- def _submit_and_wait(self, endpoint: str, payload: dict, *,
402
- verbose: bool, task_ids: list[str]) -> dict:
403
- """Create one task on `endpoint`, poll it to a terminal state, return it."""
404
- created = self.client.request("POST", endpoint, json_body=payload)
405
- task_id = self._task_id_of(created)
406
- task_ids.append(task_id)
407
- if verbose:
408
- print(f"[meshy] task {task_id} submitted ({endpoint}, "
409
- f"mode={payload.get('mode', 'image')})")
410
- return cloud_api.poll_task(
411
- lambda: self._fetch_task(endpoint, task_id),
412
- task_id=task_id,
413
- timeout=self.timeout,
414
- poll_interval=self.poll_interval,
415
- done_states=DONE_STATES,
416
- failed_states=FAILED_STATES,
417
- verbose=verbose,
418
- )
419
-
420
- def infer_and_save(
421
- self,
422
- image: Any = None,
423
- output_path: Optional[str] = None,
424
- seed: int = 42,
425
- decimation_target: Optional[int] = None,
426
- texture_size: Optional[int] = None,
427
- *,
428
- prompt: Optional[str] = None,
429
- negative_prompt: Optional[str] = None,
430
- image_url: Optional[str] = None,
431
- verbose: Optional[bool] = None,
432
- **provider_kwargs: Any,
433
- ) -> str:
434
- """
435
- [C7] The documented R1.2 exception: run `infer()` and write the result.
436
-
437
- The path comes **in** from the caller and is never derived here.
438
-
439
- Args:
440
- output_path: Destination file. Parent directories are created.
441
- (everything else: see `infer`)
442
-
443
- Returns:
444
- `str(output_path)`.
445
- """
446
- if not output_path:
447
- raise ValueError("output_path is required (model_require.md R1.2)")
448
- data = self.infer(
449
- image, seed, decimation_target, texture_size,
450
- prompt=prompt, negative_prompt=negative_prompt, image_url=image_url,
451
- verbose=verbose, **provider_kwargs,
452
- )
453
- out = Path(output_path)
454
- out.parent.mkdir(parents=True, exist_ok=True)
455
- out.write_bytes(data)
456
- return str(out)
457
-
458
- # ── account ───────────────────────────────────────────────────────────────
459
-
460
- def balance(self) -> dict:
461
- """Remaining credits: ``{"balance": int}``. Cheap and unbilled."""
462
- resp = self.client.request("GET", "/openapi/v1/balance")
463
- return resp if isinstance(resp, dict) else {"balance": resp}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/tripo_model.py DELETED
@@ -1,532 +0,0 @@
1
- """
2
- models/gen_3d_object/tripo_model.py
3
-
4
- TripoModel — wrapper around Tripo3D's cloud 3D-generation API (image-to-3D and
5
- text-to-3D), producing a GLB.
6
-
7
- Reference: https://developers.tripo3d.ai/ (v3 API, base https://openapi.tripo3d.ai/v3)
8
-
9
- CONTRACT DEVIATIONS (model_require.md targets local-weight models; the rules that
10
- replace it live in agent_skills/develop_harness/api_model_require.md):
11
- C1 R2.1/R2.2 → R9.1 `model_path` carries a *model version id*
12
- ("v3.1-20260211"), not a weight location.
13
- C2 R2.3 → R9.2 `device` accepted for interface parity and ignored.
14
- C3 R3.6/R5.5 → R9.3 no @torch.no_grad(); torch is not involved.
15
- C4 R4.1-R4.3 → R9.4 `unload()` is a no-op that closes the HTTP session.
16
- C5 R3.3/R3.4 → R9.5 `seed` is forwarded as model_seed/texture_seed when the
17
- provider accepts it; server-side reproducibility is NOT
18
- guaranteed and is never faked locally.
19
- C6 R3.1 → R9.6 task-based API (submit → poll → download); `infer()`
20
- blocks until completion or `timeout` (default 300 s).
21
- A timeout error carries the `task_id`.
22
- C7 R1.2 uses the documented `infer_and_save(..., output_path)`
23
- exception; the bytes-returning `infer()` exists alongside.
24
- C8 (new) → R9.8 `cache_dir` prevents re-billing an identical request.
25
- C9 (new) → R9.9 `max_retries` + exponential backoff; retryable (5xx/429/
26
- transport) and terminal (400/401/402) failures are split.
27
-
28
- Environment:
29
- TRIPO_API_KEY required at the first infer() call, not at construction.
30
- Get one at https://platform.tripo3d.ai/api-keys
31
- TRIPO_API_BASE optional override of the API root.
32
- `pip install requests` is needed for the API backends only.
33
-
34
- Usage:
35
- from models.gen_3d_object.tripo_model import TripoModel
36
- model = TripoModel(model_path="v3.1-20260211")
37
- path = model.infer_and_save(image, output_path="out/model.glb", seed=42)
38
- print(model.last_call_info) # task_id, elapsed, credits, triangles
39
- """
40
- from __future__ import annotations
41
-
42
- import logging
43
- import os
44
- import time
45
- from pathlib import Path
46
- from typing import Any, Optional
47
-
48
- from models.common import cloud_api
49
- from models.common.glb_utils import glb_summary
50
-
51
- logger = logging.getLogger(__name__)
52
-
53
- #: Where a human gets a key. Quoted verbatim in the fail-fast message (R1.6).
54
- SIGNUP_URL = "https://platform.tripo3d.ai/api-keys"
55
-
56
- DEFAULT_API_BASE = "https://openapi.tripo3d.ai/v3"
57
-
58
- #: Short names accepted for `model_path`, mapped to the ids the API expects.
59
- #: Unknown values are passed through untouched, so a version released after this
60
- #: file was written still works without a code change.
61
- MODEL_ALIASES: dict[str, str] = {
62
- "tripo": "v3.1-20260211",
63
- "tripo-v3.1": "v3.1-20260211",
64
- "v3.1": "v3.1-20260211",
65
- "tripo-v3.0": "v3.0-20250812",
66
- "v3.0": "v3.0-20250812",
67
- "tripo-v2.5": "v2.5-20250123",
68
- "v2.5": "v2.5-20250123",
69
- # P series — natively low-poly, the cheap option for mesh particles (B4.2).
70
- "tripo-p1": "P1-20260311",
71
- "p1": "P1-20260311",
72
- }
73
-
74
- DEFAULT_MODEL = "v3.1-20260211"
75
-
76
- #: Terminal task states. Tripo reports lower-case; compared case-insensitively.
77
- DONE_STATES = ("success", "succeeded", "finished", "completed")
78
- FAILED_STATES = ("failed", "cancelled", "canceled", "banned", "expired", "unknown")
79
-
80
- #: Output fields that have carried the download URL across API versions.
81
- _MODEL_URL_KEYS = ("model_url", "pbr_model", "model", "base_model", "glb", "gltf")
82
-
83
- #: `face_limit` range the service accepts **while `smart_low_poly` is on**.
84
- #: Outside it the request is rejected with a terminal 400:
85
- #: "when smart_low_poly is enabled, face_limit must be between 500 and 20000
86
- #: (triangle mesh), got 30000"
87
- #: Not in the public docs; found by calling it.
88
- LOW_POLY_FACE_LIMIT_RANGE = (500, 20_000)
89
-
90
-
91
- class TripoModel:
92
- """
93
- Cloud image-to-3D / text-to-3D through Tripo3D.
94
-
95
- Interchangeable with `Trellis2Model` in `Gen3DObjectOperator`: the
96
- `infer_and_save(image, output_path, seed, decimation_target, texture_size)`
97
- call signature is positionally identical (model_require.md R6).
98
-
99
- Args:
100
- model_path: [C1] Tripo model version id, or an alias from `MODEL_ALIASES`.
101
- device: [C2] Accepted for interface parity, ignored. "cpu" is fine.
102
- api_key: Explicit key; defaults to `$TRIPO_API_KEY`, read at first call.
103
- timeout: [C6] Seconds to wait for one generation task.
104
- poll_interval: Seconds between status polls.
105
- max_retries: [C9] Extra attempts for retryable HTTP failures.
106
- cache_dir: [C8] Directory for the response cache. None disables it.
107
- output_format: "glb" is produced natively. Any other format needs the
108
- conversion endpoint — see `convert_path`.
109
- convert_path: API path of the format-conversion endpoint. Left as None
110
- because it is not pinned down in the public v3 docs; set it
111
- to enable `output_format="fbx"` without touching this file.
112
- low_poly: [B4.3] Ask the service for clean low-poly topology
113
- (`smart_low_poly`) instead of decimating locally.
114
- texture / pbr / quad / texture_quality / geometry_quality:
115
- Forwarded verbatim to the generation endpoint.
116
- api_base: Override the API root (defaults to `$TRIPO_API_BASE`).
117
- verbose: Print task progress; off by default (R3.7).
118
- """
119
-
120
- #: Provider name, used for cache namespacing and log lines.
121
- PROVIDER = "tripo"
122
-
123
- def __init__(
124
- self,
125
- model_path: str = DEFAULT_MODEL,
126
- device: str = "cuda",
127
- *,
128
- api_key: Optional[str] = None,
129
- # Measured: image-to-3D with smart_low_poly took **601 s**, i.e. it would
130
- # have died one second past a 600 s budget. Timing out does not refund
131
- # the credits — the task finishes server-side and only the download is
132
- # lost — so the default errs long. The error carries the task_id.
133
- timeout: int = 1800,
134
- poll_interval: float = 3.0,
135
- max_retries: int = 3,
136
- cache_dir: Optional[str] = None,
137
- output_format: str = "glb",
138
- convert_path: Optional[str] = None,
139
- low_poly: bool = False,
140
- texture: bool = True,
141
- pbr: bool = True,
142
- quad: bool = False,
143
- texture_quality: str = "standard",
144
- geometry_quality: str = "standard",
145
- api_base: Optional[str] = None,
146
- http_timeout: int = 60,
147
- verbose: bool = False,
148
- **model_specific: Any,
149
- ):
150
- # [CONTRACT-DEVIATION C1] model_path is a version id, not a weight location.
151
- # [CONTRACT-DEVIATION C2] device accepted for interface parity, ignored.
152
- self.model_path = MODEL_ALIASES.get(str(model_path).lower(), model_path)
153
- self.device = device
154
-
155
- # R2.4 — every constructor argument lands on self before anything loads,
156
- # so unload()/reload round-trips without losing configuration.
157
- self.api_key = api_key
158
- self.timeout = timeout
159
- self.poll_interval = poll_interval
160
- self.max_retries = max_retries
161
- self.cache_dir = cache_dir
162
- self.output_format = output_format.lower().lstrip(".")
163
- self.convert_path = convert_path
164
- self.low_poly = low_poly
165
- self.texture = texture
166
- self.pbr = pbr
167
- self.quad = quad
168
- self.texture_quality = texture_quality
169
- self.geometry_quality = geometry_quality
170
- self.api_base = (api_base or os.environ.get("TRIPO_API_BASE")
171
- or DEFAULT_API_BASE)
172
- self.http_timeout = http_timeout
173
- self.verbose = verbose
174
- self.model_specific = model_specific
175
-
176
- self._client: Optional[cloud_api.CloudAPIClient] = None
177
- self._cache = cloud_api.ResponseCache(cache_dir, self.PROVIDER)
178
-
179
- #: Record of the most recent **real** call: task_id, elapsed_sec, credits,
180
- #: triangles, bytes, cached. Operators copy it into meta.json (R9.8).
181
- self.last_call_info: dict[str, Any] = {}
182
-
183
- # ── plumbing ──────────────────────────────────────────────────────────────
184
-
185
- @property
186
- def client(self) -> cloud_api.CloudAPIClient:
187
- """
188
- HTTP client, created on first use.
189
-
190
- [C6/R9.7] The API key is resolved here rather than in `__init__`, so the
191
- wrapper can be imported and constructed on a machine with no credentials
192
- (which is exactly what `test/harness/smoke.py` does).
193
- """
194
- if self._client is None:
195
- key = cloud_api.require_api_key(
196
- self.api_key, "TRIPO_API_KEY", SIGNUP_URL, who="TripoModel")
197
- self._client = cloud_api.CloudAPIClient(
198
- self.api_base, key,
199
- timeout=self.http_timeout,
200
- max_retries=self.max_retries,
201
- )
202
- return self._client
203
-
204
- def unload(self) -> None:
205
- """[C4] Close the HTTP session. Idempotent, and safe before any call."""
206
- if self._client is not None:
207
- self._client.close()
208
- self._client = None
209
-
210
- close = unload
211
-
212
- @staticmethod
213
- def _unwrap(resp: Any) -> dict:
214
- """
215
- Strip Tripo's `{"code": 0, "data": {...}}` envelope.
216
-
217
- A non-zero `code` on an HTTP 200 is a terminal application error — the
218
- same request will fail the same way, so it must not be retried (R9.9).
219
- """
220
- if not isinstance(resp, dict):
221
- raise cloud_api.CloudAPIError(f"unexpected response type {type(resp).__name__}: {resp!r}")
222
- code = resp.get("code")
223
- if code not in (None, 0):
224
- message = resp.get("message") or resp.get("msg") or resp
225
- suggestion = resp.get("suggestion")
226
- raise cloud_api.CloudAPIRequestError(
227
- f"Tripo API error code={code}: {message}"
228
- + (f" — {suggestion}" if suggestion else ""),
229
- payload=resp,
230
- )
231
- data = resp.get("data")
232
- return data if isinstance(data, dict) else resp
233
-
234
- def _upload_image(self, png: bytes) -> str:
235
- """Upload PNG bytes, return the `file_token` used as generation input."""
236
- data = self._unwrap(self.client.request(
237
- "POST", "/files",
238
- files={"file": ("input.png", png, "image/png")},
239
- ))
240
- token = data.get("file_token") or data.get("image_token") or data.get("token")
241
- if not token:
242
- raise cloud_api.CloudAPIError(
243
- f"upload succeeded but no file_token in the response: {data}")
244
- return str(token)
245
-
246
- @staticmethod
247
- def _extract_model_url(task: dict) -> str:
248
- """Find the artifact URL in a finished task, across output field names."""
249
- output = task.get("output") or task.get("result") or {}
250
- for key in _MODEL_URL_KEYS:
251
- value = output.get(key)
252
- if isinstance(value, str) and value.startswith("http"):
253
- return value
254
- if isinstance(value, dict) and isinstance(value.get("url"), str):
255
- return value["url"]
256
- raise cloud_api.CloudAPIError(
257
- f"task {task.get('task_id')} reported success but carries no model URL. "
258
- f"output={output!r}",
259
- task_id=task.get("task_id"), payload=task)
260
-
261
- def _fetch_task(self, task_id: str) -> dict:
262
- return self._unwrap(self.client.request("GET", f"/tasks/{task_id}"))
263
-
264
- # ── payload ───────────────────────────────────────────────────────────────
265
-
266
- def _clamp_face_limit(self, value: int) -> int:
267
- """
268
- Keep `face_limit` inside the range the service accepts.
269
-
270
- `Gen3DObjectOperator` forwards TRELLIS.2's default of 1 000 000, and with
271
- `low_poly=True` anything outside 500-20 000 is a terminal 400. Clamping
272
- (loudly) keeps the backends interchangeable (R6) rather than making the
273
- caller special-case each provider.
274
- """
275
- value = int(value)
276
- if not self.low_poly:
277
- return value
278
- low, high = LOW_POLY_FACE_LIMIT_RANGE
279
- clamped = max(low, min(value, high))
280
- if clamped != value:
281
- logger.warning("[tripo] face_limit %s clamped to %s "
282
- "(smart_low_poly range %s-%s)", value, clamped, low, high)
283
- return clamped
284
-
285
- def _build_payload(
286
- self,
287
- *,
288
- prompt: Optional[str],
289
- negative_prompt: Optional[str],
290
- seed: Optional[int],
291
- decimation_target: Optional[int],
292
- provider_kwargs: dict,
293
- ) -> dict:
294
- """Generation parameters shared by the image and text endpoints."""
295
- payload: dict[str, Any] = {
296
- "model": self.model_path,
297
- "texture": self.texture,
298
- "pbr": self.pbr,
299
- "texture_quality": self.texture_quality,
300
- "geometry_quality": self.geometry_quality,
301
- }
302
- if self.quad:
303
- payload["quad"] = True
304
- if self.low_poly:
305
- # [B4.3] Ask the service for low-poly topology. Decimating a 2M-face
306
- # mesh locally wrecks UVs and normals; generating low-poly does not.
307
- payload["smart_low_poly"] = True
308
- if decimation_target:
309
- payload["face_limit"] = self._clamp_face_limit(decimation_target)
310
- if prompt:
311
- payload["prompt"] = prompt
312
- if negative_prompt:
313
- payload["negative_prompt"] = negative_prompt
314
- if seed is not None:
315
- # [CONTRACT-DEVIATION C5] forwarded, but the service does not promise
316
- # that the same seed reproduces the same mesh. Never faked locally.
317
- payload["model_seed"] = int(seed)
318
- payload["texture_seed"] = int(seed)
319
- payload.update(provider_kwargs)
320
- return payload
321
-
322
- # ── inference ─────────────────────────────────────────────────────────────
323
-
324
- def infer(
325
- self,
326
- image: Any = None,
327
- seed: int = 42,
328
- decimation_target: Optional[int] = None,
329
- texture_size: Optional[int] = None,
330
- *,
331
- prompt: Optional[str] = None,
332
- negative_prompt: Optional[str] = None,
333
- image_url: Optional[str] = None,
334
- verbose: Optional[bool] = None,
335
- **provider_kwargs: Any,
336
- ) -> bytes:
337
- """
338
- Generate one 3D asset and return the model file bytes.
339
-
340
- [C6] Submits the task, polls until it finishes, downloads the artifact —
341
- all inside this call. Typical wall time is 20 s to 4 min.
342
-
343
- Args:
344
- image: Reference image as `PIL.Image.Image`, `np.ndarray` or PNG
345
- `bytes`. Not a path (R3.2). Give `image`, `image_url` or
346
- `prompt`; image wins when several are present.
347
- seed: [C5] Forwarded as `model_seed` / `texture_seed`. Reproducibility
348
- is not guaranteed server-side.
349
- decimation_target: Target face count, forwarded as `face_limit`.
350
- None leaves it to the service.
351
- texture_size: Accepted for signature parity with `Trellis2Model` and
352
- **ignored** — Tripo exposes `texture_quality`, not a
353
- resolution. Pass `texture_quality=` to the constructor.
354
- prompt: Text prompt for text-to-3D. [A4 decision] The operator slot is
355
- image-to-3D today; accepting `prompt` here keeps text-to-3D
356
- available without a second wrapper.
357
- negative_prompt: Text-to-3D only.
358
- image_url: Publicly reachable image URL, used instead of uploading.
359
- verbose: Override the constructor's verbosity for this call.
360
- **provider_kwargs: Merged into the request body verbatim, so a new
361
- API parameter needs no change here.
362
-
363
- Returns:
364
- The model file as `bytes` — GLB unless `output_format`/`convert_path`
365
- say otherwise.
366
-
367
- Raises:
368
- CloudAPIAuthError: no or invalid `TRIPO_API_KEY`.
369
- CloudAPIQuotaError: out of credits (terminal, never retried).
370
- CloudTaskTimeout: not finished within `timeout`; carries `task_id`.
371
- CloudTaskFailed: the service rejected or failed the task.
372
- """
373
- verbose = self.verbose if verbose is None else verbose
374
- if image is None and image_url is None and not prompt:
375
- raise ValueError(
376
- "TripoModel.infer needs an image (`image=` / `image_url=`) or a "
377
- "text `prompt=`; got neither.")
378
- if texture_size is not None:
379
- # [CONTRACT-DEVIATION C2-style] accepted for parity, ignored.
380
- logger.debug("[tripo] texture_size=%s ignored; use texture_quality",
381
- texture_size)
382
- if self.output_format != "glb" and not self.convert_path:
383
- raise NotImplementedError(
384
- f"TripoModel produces GLB; output_format={self.output_format!r} needs "
385
- f"Tripo's conversion endpoint, whose path is not pinned in the public "
386
- f"v3 docs. Either set convert_path=... on the constructor, use "
387
- f"MeshyModel(output_format='fbx') which returns FBX natively, or "
388
- f"convert the GLB with engine_adapters/blender/."
389
- )
390
-
391
- png = cloud_api.image_to_png_bytes(image) if image is not None else None
392
- use_image = png is not None or bool(image_url)
393
- payload = self._build_payload(
394
- # The image endpoint has no prompt field — sending one is a terminal
395
- # 400. An image plus a prompt means "use the image".
396
- prompt=None if use_image else prompt,
397
- negative_prompt=None if use_image else negative_prompt,
398
- seed=seed,
399
- decimation_target=decimation_target, provider_kwargs=provider_kwargs,
400
- )
401
-
402
- # [C8] Cache key covers everything that changes the output. The image is
403
- # represented by its hash, never its bytes, and the key never sees the API key.
404
- cache_key = cloud_api.request_hash({
405
- "provider": self.PROVIDER,
406
- "payload": payload,
407
- "image_sha256": cloud_api.bytes_hash(png) if png else None,
408
- "image_url": image_url,
409
- "output_format": self.output_format,
410
- })
411
- cached = self._cache.get(cache_key, ext=self.output_format)
412
- if cached is not None:
413
- self.last_call_info = {
414
- "provider": self.PROVIDER, "cached": True, "cache_key": cache_key,
415
- "bytes": len(cached), **glb_summary(cached),
416
- }
417
- return cached
418
-
419
- t0 = time.time()
420
- if png is not None:
421
- # `input` takes a file_token, a public URL or a previous task id.
422
- payload["input"] = self._upload_image(png)
423
- endpoint = "/generation/image-to-model"
424
- elif image_url:
425
- payload["input"] = image_url
426
- endpoint = "/generation/image-to-model"
427
- else:
428
- endpoint = "/generation/text-to-model"
429
-
430
- created = self._unwrap(self.client.request("POST", endpoint, json_body=payload))
431
- task_id = str(created.get("task_id") or created.get("id") or "")
432
- if not task_id:
433
- raise cloud_api.CloudAPIError(
434
- f"{endpoint} accepted the request but returned no task_id: {created}")
435
- if verbose:
436
- print(f"[tripo] task {task_id} submitted ({endpoint})")
437
-
438
- # Some responses already carry a finished task; only poll when they do not.
439
- state = str(created.get("status", "")).lower()
440
- if state in DONE_STATES:
441
- task = created
442
- else:
443
- task = cloud_api.poll_task(
444
- lambda: self._fetch_task(task_id),
445
- task_id=task_id,
446
- timeout=self.timeout,
447
- poll_interval=self.poll_interval,
448
- done_states=DONE_STATES,
449
- failed_states=FAILED_STATES,
450
- verbose=verbose,
451
- )
452
-
453
- data = self.client.download(self._extract_model_url(task))
454
- elapsed = time.time() - t0
455
-
456
- info = glb_summary(data) if self.output_format == "glb" else {"bytes": len(data)}
457
- self.last_call_info = {
458
- "provider": self.PROVIDER,
459
- "model": self.model_path,
460
- "task_id": task_id,
461
- "elapsed_sec": round(elapsed, 2),
462
- "credits": task.get("credits_consumed", task.get("credits")),
463
- "cached": False,
464
- "cache_key": cache_key,
465
- # The provider's own render — see the note in meshy_model.py.
466
- "preview_url": (task.get("output") or {}).get("rendered_image_url"),
467
- **info,
468
- }
469
- # R9.8 — one line per billed call: what it cost and what it produced.
470
- logger.info(
471
- "[tripo] task=%s model=%s %.1fs credits=%s triangles=%s bytes=%s",
472
- task_id, self.model_path, elapsed, self.last_call_info.get("credits"),
473
- info.get("triangles"), len(data),
474
- )
475
- if verbose:
476
- print(f"[tripo] {self.last_call_info}")
477
-
478
- self._cache.put(cache_key, data, {
479
- "provider": self.PROVIDER, "payload": payload,
480
- **{k: v for k, v in self.last_call_info.items() if k != "cache_key"},
481
- }, ext=self.output_format)
482
- return data
483
-
484
- def infer_and_save(
485
- self,
486
- image: Any = None,
487
- output_path: Optional[str] = None,
488
- seed: int = 42,
489
- decimation_target: Optional[int] = None,
490
- texture_size: Optional[int] = None,
491
- *,
492
- prompt: Optional[str] = None,
493
- negative_prompt: Optional[str] = None,
494
- image_url: Optional[str] = None,
495
- verbose: Optional[bool] = None,
496
- **provider_kwargs: Any,
497
- ) -> str:
498
- """
499
- [C7] The documented R1.2 exception: run `infer()` and write the result.
500
-
501
- The path comes **in** from the caller and is never derived here — the
502
- operator owns artifact placement.
503
-
504
- Args:
505
- output_path: Destination file. Parent directories are created.
506
- (everything else: see `infer`)
507
-
508
- Returns:
509
- `str(output_path)`.
510
- """
511
- if not output_path:
512
- raise ValueError("output_path is required (model_require.md R1.2)")
513
- data = self.infer(
514
- image, seed, decimation_target, texture_size,
515
- prompt=prompt, negative_prompt=negative_prompt, image_url=image_url,
516
- verbose=verbose, **provider_kwargs,
517
- )
518
- out = Path(output_path)
519
- out.parent.mkdir(parents=True, exist_ok=True)
520
- out.write_bytes(data)
521
- return str(out)
522
-
523
- # ── account ───────────────────────────────────────────────────────────────
524
-
525
- def balance(self) -> dict:
526
- """
527
- Remaining credits: ``{"balance": float, "frozen": float}``.
528
-
529
- Cheap and unbilled — call it before a batch to check the budget (C4 note:
530
- this is the only method that talks to the API without generating).
531
- """
532
- return self._unwrap(self.client.request("GET", "/account/balance"))
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/operators/gen_3d_object/operator.py DELETED
@@ -1,596 +0,0 @@
1
- """
2
- operators/gen_3d_object/operator.py
3
-
4
- Gen3DObjectOperator — accepts a loaded model and processes an input dict into a
5
- 3D output (GLB file on disk).
6
-
7
- The operator is intentionally model-agnostic: you inject any object that
8
- implements the same interface as Trellis2Model. New model wrappers can be
9
- swapped in without touching this file.
10
-
11
- Output layout — two modes, chosen by whether `output_dir` is given:
12
-
13
- * **per-game (default, `output_dir=None`)** — resolved by
14
- `pipeline.common.paths`, so artifacts are grouped by generated game project::
15
-
16
- test_data/outputs/<game_id>/<run_id>/assets/3d_object/<task_id>/
17
- ├── model.glb
18
- └── meta.json
19
-
20
- * **flat (legacy, `output_dir="..."`)** — writes `<output_dir>/<task_id>.glb`,
21
- byte-for-byte the historical behaviour. Existing callers are unaffected.
22
-
23
- Usage:
24
- from models.gen_3d_object.trellis_2_model import Trellis2Model
25
- from operators.gen_3d_object.operator import Gen3DObjectOperator
26
-
27
- model = Trellis2Model(model_path="...")
28
-
29
- # per-game layout
30
- op = Gen3DObjectOperator(model=model, run_id="20260731_1032")
31
-
32
- # legacy flat layout (unchanged)
33
- op = Gen3DObjectOperator(model=model, output_dir="outputs/3d_object")
34
-
35
- result = op.run({
36
- "game_id": "gameA_cyberpunk_shooter", # optional, inferred when omitted
37
- "task_id": "sword_001",
38
- "image_path": "path/to/image.png", # or "image": PIL.Image
39
- "seed": 42,
40
- })
41
- print(result["glb_path"])
42
-
43
- Two optional extras, both additive and off unless asked for:
44
-
45
- * naming an ``asset_id`` in the task hands the finished mesh to the
46
- three.js adapter — staged, oriented, manifest rewritten;
47
- * ``run_art_plan(game_id, image_model)`` generates a whole game's
48
- recorded art plan (``funcs/art_plan.py``) that way in one call.
49
-
50
- ``funcs/asset_pack.py`` covers the other case: three CC0 models,
51
- downloaded and imported in seconds, for when generating is overkill.
52
-
53
- A third route skips inference entirely. Passing ``spec`` instead of an
54
- image builds the mesh from a declarative part list
55
- (``funcs/code_asset.py``), which suits anything exactly describable — a
56
- crate, a sign, a wheel, a rifle. It needs no model and no GPU, states its
57
- own facing and size instead of leaving them to be guessed, and keeps its
58
- parts as named glTF nodes so gameplay can still move them::
59
-
60
- op = Gen3DObjectOperator(model=None, run_id="20260826_1400")
61
- op.run({
62
- "game_id": "gameA_cyberpunk_shooter",
63
- "task_id": "crate_001",
64
- "spec": {
65
- "subject": "wooden supply crate",
66
- "units": "metres", "forward": "+z",
67
- "asset_type": "prop", "height_metres": 0.6,
68
- "parts": [{"id": "body", "kind": "box",
69
- "size": [0.6, 0.6, 0.6], "at": [0, 0.3, 0]}],
70
- },
71
- })
72
-
73
- The two can be mixed, which is usually the best available answer for a
74
- detailed hard-surface asset. A ``mesh`` part reads a GLB off disk — one
75
- component generated by a cloud model — and is placed, measured and gated
76
- exactly like a primitive. What each side contributes is not
77
- interchangeable: the generator supplies shapes no formula states (a
78
- moulded grip's finger swells), and the spec supplies the dimensions and
79
- placement a prompt cannot specify (a rail's 30 mm slot pitch). Spending
80
- generation on something a lathe states exactly buys an approximation at a
81
- hundred times the triangles, and the approximation is the part that can no
82
- longer be adjusted by editing a number — so the ``provenance`` gate
83
- reports what was delegated and at what cost::
84
-
85
- op.run({
86
- "game_id": "gameA_cyberpunk_shooter",
87
- "task_id": "rifle_001",
88
- "spec": {
89
- "subject": "assault rifle",
90
- "units": "metres", "forward": "+z",
91
- "asset_type": "weapon", "height_metres": 0.24,
92
- "parts": [
93
- # exact dimensions: stated
94
- {"id": "receiver", "kind": "box",
95
- "size": [0.04, 0.048, 0.30], "at": [0, 0.139, -0.03]},
96
- # moulded surface: fetched, with a uniform `size` so the
97
- # mesh keeps the proportions it was generated with
98
- {"id": "grip", "kind": "mesh",
99
- "source": "parts/grip.glb",
100
- "size": [0.116, 0.116, 0.116], "at": [0, 0.045, -0.104]},
101
- ],
102
- },
103
- })
104
- """
105
- from __future__ import annotations
106
-
107
- import time
108
- from pathlib import Path
109
- from typing import Any, Optional
110
-
111
- from PIL import Image
112
-
113
- #: Registered task kind for this operator (see pipeline/common/paths.py).
114
- TASK_KIND = "3d_object"
115
-
116
- #: Artifact filename used by the per-game layout.
117
- GLB_FILENAME = "model.glb"
118
-
119
-
120
- class Gen3DObjectOperator:
121
- """
122
- Operator for 3D object generation.
123
-
124
- Args:
125
- model: A loaded model with an `infer_and_save(image, output_path, ...)` method.
126
- Currently supports Trellis2Model; any model with the same interface works.
127
- output_dir (str, optional): **Legacy flat mode.** When given, every GLB is
128
- written as `<output_dir>/<task_id>.glb`, exactly as before. When
129
- omitted (default), the per-game layout under
130
- `test_data/outputs/<game_id>/<run_id>/assets/3d_object/<task_id>/`
131
- is used instead.
132
- run_id (str): Groups all artifacts of one generation run. Per-game mode only.
133
- default_game_id (str, optional): Fallback game project for tasks that carry
134
- no `game_id` and whose input paths reveal none. Per-game mode only.
135
- """
136
-
137
- def __init__(
138
- self,
139
- model: Any,
140
- output_dir: Optional[str] = None,
141
- run_id: str = "default",
142
- default_game_id: Optional[str] = None,
143
- ):
144
- self.model = model
145
- self.run_id = run_id
146
- self.default_game_id = default_game_id
147
- # Legacy flat mode is active only when an explicit output_dir is supplied.
148
- self.output_dir = Path(output_dir) if output_dir else None
149
- if self.output_dir is not None:
150
- self.output_dir.mkdir(parents=True, exist_ok=True)
151
-
152
- # --------------------------------------------------------------------------
153
-
154
- def _load_image(self, inp: dict) -> Image.Image:
155
- """Return a PIL image from an input dict that has 'image_path' or 'image'."""
156
- if "image" in inp and isinstance(inp["image"], Image.Image):
157
- return inp["image"]
158
- if "image_path" in inp:
159
- p = Path(inp["image_path"])
160
- if not p.is_absolute() and not p.exists():
161
- # resolve relative to repo root (codes_official/)
162
- repo_root = Path(__file__).resolve().parents[2]
163
- p = repo_root / inp["image_path"]
164
- return Image.open(str(p)).convert("RGBA")
165
- raise ValueError("Input must contain 'image_path' (str) or 'image' (PIL.Image).")
166
-
167
- # --------------------------------------------------------------------------
168
-
169
- def _artifact_suffix(self) -> str:
170
- """
171
- File extension for the artifact — GLB unless the injected model says
172
- otherwise (the cloud backends can emit FBX / OBJ / USDZ).
173
-
174
- The result dict keeps the ``glb_path`` key whatever the extension is:
175
- renaming a returned key breaks `eval.py` and `test/`.
176
- """
177
- fmt = getattr(self.model, "output_format", None)
178
- return str(fmt).lower().lstrip(".") if fmt else "glb"
179
-
180
- def _resolve_out_path(self, inp: dict, task_id: str) -> tuple[str, Path, Path]:
181
- """
182
- Return ``(game_id, task_dir, glb_path)`` for either output mode.
183
-
184
- Legacy flat mode keeps the historical `<output_dir>/<task_id>.glb`.
185
- """
186
- suffix = self._artifact_suffix()
187
- if self.output_dir is not None:
188
- game_id = str(inp.get("game_id") or inp.get("game") or "")
189
- return game_id, self.output_dir, self.output_dir / f"{task_id}.{suffix}"
190
-
191
- from pipeline.common import paths
192
- game_id = paths.infer_game_id(inp, fallback=self.default_game_id)
193
- task_dir = paths.task_output_dir(game_id, TASK_KIND, task_id, run_id=self.run_id)
194
- filename = GLB_FILENAME if suffix == "glb" else f"model.{suffix}"
195
- return game_id, task_dir, task_dir / filename
196
-
197
- # --------------------------------------------------------------------------
198
-
199
- def run(self, inp: dict) -> dict:
200
- """
201
- Generate a 3D object from an image prompt, or build one from a spec.
202
-
203
- Args:
204
- inp (dict):
205
- - image_path (str): path to the reference / concept image
206
- OR image (PIL.Image): pre-loaded image
207
- OR spec (dict): a part spec, which builds the mesh instead
208
- of inferring it (see `funcs/code_asset.py`)
209
- - compose (dict): combine_avatar arguments, including caller-defined
210
- socket_definitions, grip_templates and optional slot_definitions
211
- - game_id (str, optional): game project this task belongs to;
212
- inferred from `image_path` when omitted
213
- - task_id (str, optional): used to name the output file / directory
214
- - seed (int, optional): random seed (default 42)
215
- - decimation_target (int, optional)
216
- - texture_size (int, optional)
217
-
218
- Returns:
219
- dict:
220
- - task_id (str)
221
- - glb_path (str): absolute path to the saved GLB file
222
- - elapsed_sec (float)
223
- - game_id (str), task_kind (str), output_dir (str) ← additive
224
- """
225
- if inp.get("compose") is not None:
226
- if inp.get("spec") is not None:
227
- raise ValueError("provide compose or spec, not both")
228
- return self._run_compose(inp)
229
- if inp.get("spec") is not None:
230
- return self._run_spec(inp)
231
-
232
- task_id = inp.get("task_id", f"task_{int(time.time())}")
233
- seed = inp.get("seed", 42)
234
- decimation_target = inp.get("decimation_target", 1_000_000)
235
- texture_size = inp.get("texture_size", 4096)
236
-
237
- game_id, task_dir, out_path = self._resolve_out_path(inp, task_id)
238
-
239
- text_mode = inp.get("input_mode") == "text_to_3d"
240
- if text_mode and not str(inp.get("prompt") or "").strip():
241
- raise ValueError("text_to_3d requires a non-empty prompt")
242
- image = None if text_mode else self._load_image(inp)
243
- inference_kwargs = {"prompt": inp["prompt"]} if text_mode else {}
244
-
245
- t0 = time.time()
246
- glb_path = self.model.infer_and_save(
247
- image,
248
- output_path=str(out_path),
249
- seed=seed,
250
- decimation_target=decimation_target,
251
- texture_size=texture_size,
252
- **inference_kwargs,
253
- )
254
- elapsed = time.time() - t0
255
-
256
- result = {
257
- "task_id": task_id,
258
- "glb_path": glb_path,
259
- "elapsed_sec": round(elapsed, 2),
260
- "game_id": game_id,
261
- "task_kind": TASK_KIND,
262
- "output_dir": str(task_dir),
263
- }
264
-
265
- # Image-to-3D infers a floor under a subject that stands on the edge
266
- # of its own crop, and it is never wanted: measured at a fifth of the
267
- # triangle budget, and a grey pancake under the character in the
268
- # scene. Removing it here means every consumer of this task output
269
- # gets a clean mesh, rather than each game discovering the slab.
270
- if inp.get("strip_ground_plate", True):
271
- from .funcs.mesh_cleanup import strip_ground_plate
272
-
273
- cleanup = strip_ground_plate(glb_path)
274
- result["ground_plate"] = cleanup
275
- if cleanup.get("applied"):
276
- result.setdefault("warnings", []).append(
277
- f"removed an inferred floor: {cleanup['plate_count']} "
278
- f"sheets and {cleanup['debris_count']} loose lumps, "
279
- f"{cleanup['removed_fraction']:.0%} of the mesh"
280
- )
281
- elif cleanup.get("reason") and "nothing" not in cleanup["reason"]:
282
- result.setdefault("warnings", []).append(
283
- f"left the mesh alone: {cleanup['reason']}"
284
- )
285
-
286
- # meta.json only in per-game mode, so legacy output dirs stay untouched.
287
- if self.output_dir is None:
288
- from pipeline.common import paths
289
- meta = {
290
- **result,
291
- "run_id": self.run_id,
292
- "prompt": inp.get("prompt"),
293
- "source_image": inp.get("image_path"),
294
- "seed": seed,
295
- "decimation_target": decimation_target,
296
- "texture_size": texture_size,
297
- "model": type(self.model).__name__,
298
- }
299
- for key in ("asset_id", "asset_type", "forward_axis",
300
- "scale_hint_metres", "notes"):
301
- if inp.get(key) is not None:
302
- meta[key] = inp[key]
303
- # Cloud backends expose what the call cost and produced — task id,
304
- # credits, triangle count (api_model_require.md R9.8). Recording it
305
- # here is what makes a billed run reproducible after the fact.
306
- call_info = getattr(self.model, "last_call_info", None)
307
- if call_info:
308
- meta["model_call"] = dict(call_info)
309
- paths.write_task_meta(task_dir, meta)
310
-
311
- # Optional handover to the engine adapter, active only when the task
312
- # names an asset. A generated mesh is not part of a game until the
313
- # adapter has staged it and recorded which way it faces.
314
- if inp.get("asset_id") and self.output_dir is None:
315
- from .funcs.asset_import import import_asset
316
- result.update(
317
- import_asset(
318
- game_id,
319
- task_id,
320
- asset_id=str(inp["asset_id"]),
321
- asset_type=str(inp.get("asset_type") or "prop"),
322
- run_id=self.run_id,
323
- forward_axis=str(inp.get("forward_axis") or ""),
324
- scale_hint_metres=inp.get("scale_hint_metres"),
325
- verified_by=str(inp.get("verified_by") or "heuristic"),
326
- notes=str(inp.get("notes") or ""),
327
- project_hint=str(inp.get("project_hint") or ""),
328
- preview=bool(inp.get("preview", True)),
329
- )
330
- )
331
-
332
- return result
333
-
334
- # --------------------------------------------------------------------------
335
-
336
- def _run_compose(self, inp: dict) -> dict:
337
- """Resolve caller-configured attachments, then build through the spec route."""
338
- import json
339
- from .funcs.code_asset_templates.human_template.combine import (
340
- combine_avatar, write_report,
341
- )
342
-
343
- task_id = inp.get("task_id", f"task_{int(time.time())}")
344
- _, _, out_path = self._resolve_out_path(inp, task_id)
345
- directory = out_path.parent / f"{out_path.stem}_compose"
346
- config = dict(inp["compose"])
347
- config.setdefault("parts_dir", directory / "parts")
348
- started = time.time()
349
- combined = combine_avatar(**config)
350
- result = self._run_spec({**inp, "task_id": task_id, "spec": combined["spec"]})
351
- report_path = write_report(combined["report"], directory / "report.json")
352
- sockets_path = directory / "sockets.json"
353
- sockets_path.write_text(json.dumps(combined["report"]["sockets"], indent=2),
354
- encoding="utf-8")
355
- result["compose"] = {**combined["report"], "report_path": report_path,
356
- "sockets_path": str(sockets_path)}
357
- result["elapsed_sec"] = round(time.time() - started, 2)
358
- return result
359
-
360
- def _run_spec(self, inp: dict) -> dict:
361
- """Build the asset from a declarative spec rather than an image.
362
-
363
- Taken when the task carries a ``spec``. No model is loaded and no
364
- image is read, so this path works with ``model=None`` — a crate
365
- does not need a GPU, and requiring one to build a box is what makes
366
- primitives feel like a defeat rather than a choice.
367
-
368
- Three things come out of the spec that the generated path can only
369
- guess at, which is the reason to prefer it wherever it applies:
370
-
371
- * ``forward_axis`` is *stated*, so the import is recorded as
372
- ``verified_by="spec"`` rather than ``"heuristic"``. Facing is
373
- the one property `orientation_review` exists to establish, and
374
- here there is nothing to establish.
375
- * ``scale_hint_metres`` is stated, so the mesh has a real size
376
- instead of a unit box someone has to scale by eye.
377
- * parts keep their ids as glTF node names, so a wheel can be spun
378
- after import. A reconstructed mesh is one fused body.
379
-
380
- ``strip_ground_plate`` is deliberately not run: it removes a floor
381
- that image-to-3D *invented*, and a spec cannot invent one. Running
382
- it here would risk deleting a plinth that was asked for.
383
-
384
- A spec may also carry ``mesh`` parts, which read a GLB off disk —
385
- typically one component from a cloud model, composed in where no
386
- formula states the shape. Those change one thing here and nothing
387
- else: ``verified_by`` drops to ``"spec+generated"``, because a
388
- generated part's facing was *asserted* by whoever placed it and not
389
- established by anything. Claiming ``"spec"`` for a composition that
390
- is largely fetched geometry would put an unverified facing behind the
391
- one label that is supposed to mean it was not guessed.
392
- """
393
- from .funcs.code_asset import MESH_KIND, build_code_asset
394
-
395
- task_id = inp.get("task_id", f"task_{int(time.time())}")
396
- game_id, task_dir, out_path = self._resolve_out_path(inp, task_id)
397
- # A spec always yields glTF, whatever output_format an injected
398
- # model happens to advertise.
399
- out_path = out_path.with_suffix(".glb")
400
- task_dir.mkdir(parents=True, exist_ok=True)
401
-
402
- t0 = time.time()
403
- report = build_code_asset(
404
- inp["spec"],
405
- str(out_path),
406
- revise=inp.get("revise"),
407
- strict=bool(inp.get("strict", True)),
408
- )
409
- elapsed = time.time() - t0
410
-
411
- # Which parts were fetched rather than derived. Carried on the result
412
- # because after export the two are indistinguishable, and only one of
413
- # them can still be changed by editing a number: a reviewer who wants
414
- # the grip 4 mm thicker needs to know that means regenerating it.
415
- # `long_axis` travels too, since a regenerated part is a new set of
416
- # axes and the rotation that placed the old one will not place it.
417
- generated_parts = [
418
- {
419
- "id": part["id"],
420
- "source": part["source"],
421
- "long_axis": part.get("long_axis"),
422
- }
423
- for part in report["spec"]["parts"]
424
- if part.get("kind") == MESH_KIND
425
- ]
426
-
427
- result: dict[str, Any] = {
428
- "task_id": task_id,
429
- "glb_path": report["glb_path"],
430
- "elapsed_sec": round(elapsed, 2),
431
- "game_id": game_id,
432
- "task_kind": TASK_KIND,
433
- "output_dir": str(task_dir),
434
- "source": "code_asset+generated" if generated_parts else "code_asset",
435
- "spec_ok": report["ok"],
436
- "gates": report["gates"],
437
- "triangles": report["triangles"],
438
- "bounds": report["bounds"],
439
- "part_ids": report["part_ids"],
440
- "generated_parts": generated_parts,
441
- "stop_reason": report["stop_reason"],
442
- "attempts": report["attempts"],
443
- }
444
- if report["warnings"]:
445
- result.setdefault("warnings", []).extend(report["warnings"])
446
- if report["failures"]:
447
- # Surfaced as warnings on the result *and* kept in `gates`: a
448
- # gate failure that only appears in a nested dict is the silent
449
- # early return this whole module is shaped against.
450
- result.setdefault("warnings", []).extend(
451
- f"gate: {message}" for message in report["failures"]
452
- )
453
-
454
- if self.output_dir is None:
455
- from pipeline.common import paths
456
-
457
- meta = {
458
- **result,
459
- "run_id": self.run_id,
460
- "spec": report["spec"],
461
- "forward_axis": report["forward_axis"],
462
- "scale_hint_metres": report["scale_hint_metres"],
463
- "model": "code_asset",
464
- }
465
- for key in ("asset_id", "asset_type", "notes"):
466
- if inp.get(key) is not None:
467
- meta[key] = inp[key]
468
- paths.write_task_meta(task_dir, meta)
469
-
470
- # A failed spec is not handed to the engine. An asset that exists
471
- # gets used, so a gate failure has to stop it becoming an asset.
472
- if inp.get("asset_id") and self.output_dir is None and report["ok"]:
473
- from .funcs.asset_import import import_asset
474
-
475
- result.update(
476
- import_asset(
477
- game_id,
478
- task_id,
479
- asset_id=str(inp["asset_id"]),
480
- asset_type=str(
481
- inp.get("asset_type")
482
- or report["spec"].get("asset_type")
483
- or "prop"
484
- ),
485
- run_id=self.run_id,
486
- # Stated by the spec, not inferred from a view.
487
- forward_axis=report["forward_axis"],
488
- scale_hint_metres=report["scale_hint_metres"],
489
- # "spec" means the facing was stated rather than guessed.
490
- # With generated parts in the composition that is only
491
- # half true: the *placement* is stated, but the mesh
492
- # inside it came from a generator with no facing of its
493
- # own, and whoever wrote the rotation asserted one. Said
494
- # distinctly so a reviewer knows an orientation check is
495
- # still worth a look, instead of trusting a label that
496
- # was earned by a different route.
497
- verified_by="spec+generated" if generated_parts else "spec",
498
- notes=str(inp.get("notes") or "")
499
- or (
500
- f"Built from a spec: {report['subject']}."
501
- + (
502
- f" {len(generated_parts)} generated part(s): "
503
- f"{', '.join(part['id'] for part in generated_parts)}."
504
- if generated_parts else ""
505
- )
506
- ),
507
- project_hint=str(inp.get("project_hint") or ""),
508
- preview=bool(inp.get("preview", True)),
509
- )
510
- )
511
- return result
512
-
513
- # --------------------------------------------------------------------------
514
-
515
- def run_art_plan(
516
- self,
517
- game_id: str,
518
- image_model: Any = None,
519
- plan: Optional[list[dict]] = None,
520
- assets: tuple[str, ...] = (),
521
- ) -> list[dict]:
522
- """
523
- Generate a game's recorded art plan and import each asset.
524
-
525
- The plan lives in `funcs/art_plan.py`; it names a subject, an asset
526
- type and a **height in metres**, because a mesh normalised into a
527
- unit box has no size of its own. Concept images come from
528
- `image_model` unless an entry points at one.
529
-
530
- Facing is recorded as `heuristic`, never as verified: single-image
531
- reconstruction puts the input view on a predictable axis, but a
532
- guess labelled as checked stops anyone checking. Each import leaves
533
- a review sheet for `agent_skills/asset_qa/`.
534
- """
535
- from .funcs.art_plan import concept_image, plan_for_game
536
-
537
- entries = plan if plan is not None else plan_for_game(game_id)
538
- if not entries:
539
- raise ValueError(
540
- f"No art plan for {game_id!r}; add one to "
541
- "operators/gen_3d_object/funcs/art_plan.py"
542
- )
543
-
544
- results: list[dict] = []
545
- for entry in entries:
546
- if assets and entry["asset_id"] not in assets:
547
- continue
548
- image, provenance, framing = concept_image(entry, image_model)
549
- task_dir = Path(
550
- self._resolve_out_path(
551
- {"game_id": game_id}, entry["task_id"]
552
- )[1]
553
- )
554
- task_dir.mkdir(parents=True, exist_ok=True)
555
- # Keep the concept image beside the mesh it produced: a bad
556
- # asset is almost always a bad concept image, and that is not
557
- # diagnosable after the fact if the input was thrown away.
558
- image.save(task_dir / "concept.png")
559
- result = self.run(
560
- {
561
- "game_id": game_id,
562
- "task_id": entry["task_id"],
563
- "image": image,
564
- "prompt": entry["full_prompt"],
565
- "seed": entry.get("seed", 42),
566
- "decimation_target": entry["triangles"],
567
- "texture_size": entry["texture"],
568
- "asset_id": entry["asset_id"],
569
- "asset_type": entry["type"],
570
- "forward_axis": entry["faces"],
571
- "scale_hint_metres": entry["height"],
572
- "verified_by": "heuristic",
573
- "notes": entry["notes"]
574
- or f"Generated from: {provenance}. Facing assumed "
575
- "from the input view.",
576
- }
577
- )
578
- result["asset_id"] = entry["asset_id"]
579
- result["concept_source"] = provenance
580
- result["framing"] = framing
581
- if framing.get("touched_border"):
582
- # The subject ran off the edge of the generated image, so
583
- # part of it was never drawn. Cropping cannot invent it.
584
- result.setdefault("warnings", []).append(
585
- "the concept image was cropped at the border, so the "
586
- "mesh is probably truncated: regenerate with another "
587
- "seed"
588
- )
589
- results.append(result)
590
- return results
591
-
592
- # --------------------------------------------------------------------------
593
-
594
- def run_batch(self, inputs: list[dict]) -> list[dict]:
595
- """Run a list of input dicts sequentially and return results."""
596
- return [self.run(inp) for inp in inputs]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/paths.py DELETED
@@ -1,444 +0,0 @@
1
- """
2
- pipeline/common/paths.py
3
-
4
- Single source of truth for every input / output path in 3AGameFactory.
5
-
6
- The benchmark is authored **per game project**, so outputs are organized the same
7
- way. An artifact is uniquely addressed by ``(game_id, run_id, task_kind, task_id)``::
8
-
9
- test_data/test_samples/<game_id>/<TaskDir>/... # inputs (read-only)
10
- test_data/outputs/<game_id>/<run_id>/<layer>/<task_kind>/<task_id>/ # outputs
11
-
12
- Full output tree::
13
-
14
- test_data/outputs/ # OUTPUT_ROOT ($AAAGF_OUTPUT_ROOT)
15
- └── gameA_cyberpunk_shooter/ # one dir per generated game project
16
- ├── latest -> 20260731_1032/ # symlink to the most recent run
17
- └── <run_id>/ # "default", or a timestamp
18
- ├── run_meta.json # ckpts · seeds · git sha · argv
19
- ├── assets/
20
- │ ├── 3d_object/<task_id>/ # model.glb + meta.json
21
- │ ├── tpose/<task_id>/ # tpose_fg.png + meta.json
22
- │ ├── 3d_scene/<task_id>/
23
- │ ├── motion/<task_id>/
24
- │ ├── cg_video/<task_id>/
25
- │ └── audio/<task_id>/ # dialogue / sound-effect audio + meta.json
26
- ├── mechanic/<task_id>/ # engine project + demo_outputs/ + launch.sh
27
- ├── ui/<task_id>/
28
- ├── pipeline/<task_id>/ # end-to-end game-slice artifacts
29
- └── eval/
30
- ├── <task_kind>/<task_id>/ # metrics.json · build.log · judge_log.json
31
- └── summary.json # aggregate over the whole run
32
-
33
- The root stays next to the test set so any evaluator finds inputs and outputs side
34
- by side. Generated artifacts get large — point ``$AAAGF_OUTPUT_ROOT`` at a scratch
35
- disk to relocate the whole tree without touching code.
36
-
37
- This module has **no third-party dependencies** and imports no models, so it is
38
- safe to import from anywhere (including CPU-only tooling and tests).
39
- """
40
- from __future__ import annotations
41
-
42
- import json
43
- import os
44
- import subprocess
45
- import sys
46
- from datetime import datetime
47
- from pathlib import Path
48
- from typing import Any, Optional
49
-
50
- # ── Roots ─────────────────────────────────────────────────────────────────────
51
-
52
- REPO_ROOT = Path(__file__).resolve().parents[2]
53
-
54
- TEST_DATA_ROOT = REPO_ROOT / "test_data"
55
- TEST_SAMPLES_ROOT = TEST_DATA_ROOT / "test_samples"
56
-
57
- #: Output root. Override with ``export AAAGF_OUTPUT_ROOT=/scratch/aaagf_outputs``.
58
- OUTPUT_ROOT = Path(
59
- os.environ.get("AAAGF_OUTPUT_ROOT", TEST_DATA_ROOT / "outputs")
60
- ).expanduser()
61
-
62
- #: Used when a task carries no game association (ad-hoc single-image demos).
63
- UNASSIGNED_GAME = "_scratch"
64
-
65
- #: Run id used when the caller does not ask for a fresh timestamped one.
66
- DEFAULT_RUN_ID = "default"
67
-
68
- LATEST_LINK = "latest"
69
-
70
-
71
- # ── Task kinds ────────────────────────────────────────────────────────────────
72
-
73
- #: task_kind -> layer directory. Layer A = assets, B = mechanic, C = ui.
74
- TASK_LAYER: dict[str, str] = {
75
- # Layer A — asset generation
76
- "3d_object": "assets",
77
- "tpose": "assets",
78
- "3d_scene": "assets",
79
- "motion": "assets",
80
- "cg_video": "assets",
81
- "audio": "assets",
82
- # Game design (story / chapters / asset & screen plan) — A-layer diagnostic only
83
- "design": "design",
84
- # Layer B / C — code generation
85
- "mechanic": "mechanic",
86
- "ui": "ui",
87
- # End-to-end vertical slice
88
- "pipeline": "pipeline",
89
- # Whitebox-stage game evaluation (stage-1 LLM/Agent whitebox rubric) —
90
- # an eval-only kind: results land under ``eval/whitebox/<task_id>/``
91
- "whitebox": "whitebox",
92
- }
93
-
94
- #: task_kind -> directory name inside ``test_samples/<game_id>/``.
95
- TASK_INPUT_DIR: dict[str, str] = {
96
- "3d_object": "3D_object",
97
- "tpose": "tpose",
98
- "3d_scene": "3D_scene",
99
- "motion": "motion",
100
- "cg_video": "cg_video",
101
- "audio": "audio",
102
- "design": "design",
103
- "mechanic": "mechanic",
104
- "ui": "ui",
105
- "pipeline": "pipeline",
106
- # whitebox case inputs (game_spec/rubric/references) live under
107
- # ``test_bench_example/<case>/``; this dir keeps the registration
108
- # complete for the shared task-input helpers.
109
- "whitebox": "whitebox",
110
- }
111
-
112
- #: task_kind -> per-game task list filename inside ``TASK_INPUT_DIR``.
113
- TASK_JSONL: dict[str, str] = {
114
- "3d_object": "object_tasks.jsonl",
115
- "tpose": "tpose_tasks.jsonl",
116
- "3d_scene": "scene_tasks.jsonl",
117
- "motion": "motion_tasks.jsonl",
118
- "cg_video": "cg_tasks.jsonl",
119
- "audio": "audio_tasks.jsonl",
120
- "design": "design_tasks.jsonl",
121
- "mechanic": "mechanic_tasks.jsonl",
122
- "ui": "ui_tasks.jsonl",
123
- "pipeline": "pipeline_task.jsonl",
124
- "whitebox": "whitebox_tasks.jsonl",
125
- }
126
-
127
- #: task_kind -> cross-game aggregate jsonl under ``test_samples/``.
128
- TASK_COLLECT_JSONL: dict[str, str] = {
129
- "3d_object": "3D_object_gen_collect.jsonl",
130
- "tpose": "tpose_gen_collect.jsonl",
131
- "3d_scene": "3D_scene_gen_collect.jsonl",
132
- "motion": "motion_gen_collect.jsonl",
133
- "cg_video": "cg_video_collect.jsonl",
134
- "audio": "audio_gen_collect.jsonl",
135
- "design": "design_collect.jsonl",
136
- "mechanic": "mechanic_collect.jsonl",
137
- "ui": "ui_collect.jsonl",
138
- "pipeline": "pipeline_collect.jsonl",
139
- "whitebox": "whitebox_collect.jsonl",
140
- }
141
-
142
- #: Layer-A kinds only — what ``pipeline/assets_gen/`` covers.
143
- ASSET_TASK_KINDS = tuple(k for k, v in TASK_LAYER.items() if v == "assets")
144
-
145
-
146
- def check_kind(task_kind: str) -> str:
147
- """Raise a helpful error for an unregistered task kind."""
148
- if task_kind not in TASK_LAYER:
149
- raise ValueError(
150
- f"Unknown task_kind {task_kind!r}. Register it in "
151
- f"pipeline/common/paths.py first. Known: {sorted(TASK_LAYER)}"
152
- )
153
- return task_kind
154
-
155
-
156
- # ── Input paths ───────────────────────────────────────────────────────────────
157
-
158
- def game_input_dir(game_id: str) -> Path:
159
- """``test_data/test_samples/<game_id>/``"""
160
- return TEST_SAMPLES_ROOT / game_id
161
-
162
-
163
- def task_input_dir(game_id: str, task_kind: str) -> Path:
164
- """``test_data/test_samples/<game_id>/<TaskDir>/``"""
165
- return game_input_dir(game_id) / TASK_INPUT_DIR[check_kind(task_kind)]
166
-
167
-
168
- def task_jsonl(game_id: str, task_kind: str) -> Path:
169
- """``test_data/test_samples/<game_id>/<TaskDir>/<kind>_tasks.jsonl``"""
170
- return task_input_dir(game_id, task_kind) / TASK_JSONL[check_kind(task_kind)]
171
-
172
-
173
- def collect_jsonl(task_kind: str) -> Path:
174
- """``test_data/test_samples/<kind>_collect.jsonl`` — cross-game task list."""
175
- return TEST_SAMPLES_ROOT / TASK_COLLECT_JSONL[check_kind(task_kind)]
176
-
177
-
178
- def resolve_tasks_path(task_kind: str, tasks: str | None = None,
179
- game_id: str | None = None) -> Path:
180
- """
181
- Standard task-list resolution shared by every ``run.py`` / ``eval.py``:
182
-
183
- explicit ``--tasks`` → the game's own ``*_tasks.jsonl`` (if non-empty)
184
- → the cross-game ``*_collect.jsonl``.
185
- """
186
- if tasks:
187
- return resolve_input_path(tasks)
188
- if game_id:
189
- per_game = task_jsonl(game_id, task_kind)
190
- if per_game.exists() and per_game.stat().st_size > 0:
191
- return per_game
192
- return collect_jsonl(task_kind)
193
-
194
-
195
- def list_games() -> list[str]:
196
- """Every game project directory present in the test set."""
197
- if not TEST_SAMPLES_ROOT.is_dir():
198
- return []
199
- return sorted(
200
- p.name
201
- for p in TEST_SAMPLES_ROOT.iterdir()
202
- if p.is_dir() and not p.name.startswith((".", "_"))
203
- )
204
-
205
-
206
- def infer_game_id(task: dict, fallback: Optional[str] = None) -> str:
207
- """
208
- Resolve which game project a task belongs to.
209
-
210
- Priority: explicit ``game_id`` / ``game`` field → any path-like field that
211
- sits under ``test_samples/<game_id>/`` → ``fallback`` → ``UNASSIGNED_GAME``.
212
- """
213
- for key in ("game_id", "game"):
214
- if task.get(key):
215
- return str(task[key])
216
-
217
- for value in task.values():
218
- if not isinstance(value, str) or "test_samples" not in value:
219
- continue
220
- parts = Path(value).parts
221
- if "test_samples" in parts:
222
- idx = parts.index("test_samples")
223
- if idx + 1 < len(parts):
224
- return parts[idx + 1]
225
-
226
- return fallback or UNASSIGNED_GAME
227
-
228
-
229
- def iter_tasks(tasks_path: str | Path, game_filter: str | None = None):
230
- """
231
- Yield ``(task_dict, game_id)`` from a jsonl, skipping blanks and comments.
232
-
233
- Every runner should iterate through this so `game_id` resolution and the
234
- `--game` filter behave identically across tasks.
235
- """
236
- with open(tasks_path) as f:
237
- for lineno, line in enumerate(f, 1):
238
- line = line.strip()
239
- if not line or line.startswith("//"):
240
- continue
241
- try:
242
- task = json.loads(line)
243
- except json.JSONDecodeError as e:
244
- raise ValueError(f"{tasks_path}:{lineno} is not valid JSON: {e}") from e
245
- game_id = infer_game_id(task)
246
- if game_filter and game_id != game_filter:
247
- continue
248
- yield task, game_id
249
-
250
-
251
- # ── Output paths ──────────────────────────────────────────────────────────────
252
-
253
- def new_run_id(prefix: str = "") -> str:
254
- """Timestamped run id, e.g. ``20260731_103200`` or ``qwen3b_20260731_103200``."""
255
- stamp = datetime.now().strftime("%Y%m%d_%H%M%S")
256
- return f"{prefix}_{stamp}" if prefix else stamp
257
-
258
-
259
- def game_output_dir(game_id: str) -> Path:
260
- """``test_data/outputs/<game_id>/``"""
261
- return OUTPUT_ROOT / game_id
262
-
263
-
264
- def run_dir(game_id: str, run_id: str = DEFAULT_RUN_ID) -> Path:
265
- """``test_data/outputs/<game_id>/<run_id>/`` — root of one generation run."""
266
- return game_output_dir(game_id) / run_id
267
-
268
-
269
- def task_output_dir(
270
- game_id: str,
271
- task_kind: str,
272
- task_id: str,
273
- run_id: str = DEFAULT_RUN_ID,
274
- create: bool = True,
275
- ) -> Path:
276
- """
277
- ``test_data/outputs/<game_id>/<run_id>/<layer>/[<task_kind>/]<task_id>/``
278
-
279
- One directory per task, so an artifact ships alongside its preview,
280
- intermediates and ``meta.json`` without filename mangling.
281
- """
282
- check_kind(task_kind)
283
- layer = TASK_LAYER[task_kind]
284
- base = run_dir(game_id, run_id) / layer
285
- # assets/ is further split by kind; mechanic/ ui/ pipeline/ *are* the kind.
286
- if layer == "assets":
287
- base = base / task_kind
288
- path = base / task_id
289
- if create:
290
- path.mkdir(parents=True, exist_ok=True)
291
- return path
292
-
293
-
294
- def eval_output_dir(
295
- game_id: str,
296
- task_kind: str,
297
- task_id: str,
298
- run_id: str = DEFAULT_RUN_ID,
299
- create: bool = True,
300
- ) -> Path:
301
- """``test_data/outputs/<game_id>/<run_id>/eval/<task_kind>/<task_id>/``"""
302
- check_kind(task_kind)
303
- path = run_dir(game_id, run_id) / "eval" / task_kind / task_id
304
- if create:
305
- path.mkdir(parents=True, exist_ok=True)
306
- return path
307
-
308
-
309
- def eval_summary_path(game_id: str, run_id: str = DEFAULT_RUN_ID) -> Path:
310
- """``test_data/outputs/<game_id>/<run_id>/eval/summary.json``"""
311
- return run_dir(game_id, run_id) / "eval" / "summary.json"
312
-
313
-
314
- def results_summary_path(game_id: str, task_kind: str,
315
- run_id: str = DEFAULT_RUN_ID) -> Path:
316
- """``test_data/outputs/<game_id>/<run_id>/<task_kind>_results_summary.json``"""
317
- check_kind(task_kind)
318
- return run_dir(game_id, run_id) / f"{task_kind}_results_summary.json"
319
-
320
-
321
- def run_meta_path(game_id: str, run_id: str = DEFAULT_RUN_ID) -> Path:
322
- """``test_data/outputs/<game_id>/<run_id>/run_meta.json``"""
323
- return run_dir(game_id, run_id) / "run_meta.json"
324
-
325
-
326
- # ── Run bookkeeping ───────────────────────────────────────────────────────────
327
-
328
- def _git_sha() -> Optional[str]:
329
- try:
330
- out = subprocess.run(
331
- ["git", "-C", str(REPO_ROOT), "rev-parse", "--short", "HEAD"],
332
- capture_output=True, text=True, timeout=5,
333
- )
334
- return out.stdout.strip() or None
335
- except Exception:
336
- return None
337
-
338
-
339
- def update_latest_link(game_id: str, run_id: str) -> None:
340
- """Repoint ``<game_id>/latest`` at ``run_id`` (best effort, never fatal)."""
341
- if run_id == LATEST_LINK:
342
- return
343
- link = game_output_dir(game_id) / LATEST_LINK
344
- try:
345
- link.parent.mkdir(parents=True, exist_ok=True)
346
- if link.is_symlink() or link.exists():
347
- link.unlink()
348
- link.symlink_to(run_id, target_is_directory=True)
349
- except OSError:
350
- pass # filesystem without symlink support — not fatal
351
-
352
-
353
- def write_run_meta(
354
- game_id: str,
355
- run_id: str = DEFAULT_RUN_ID,
356
- *,
357
- task_kind: Optional[str] = None,
358
- extra: Optional[dict[str, Any]] = None,
359
- ) -> Path:
360
- """Record how a run was produced, so results stay reproducible."""
361
- path = run_meta_path(game_id, run_id)
362
- path.parent.mkdir(parents=True, exist_ok=True)
363
-
364
- meta: dict[str, Any] = {
365
- "game_id": game_id,
366
- "run_id": run_id,
367
- "task_kind": task_kind,
368
- "created_at": datetime.now().isoformat(timespec="seconds"),
369
- "git_sha": _git_sha(),
370
- "argv": sys.argv,
371
- }
372
- if extra:
373
- meta.update(extra)
374
-
375
- if path.exists(): # merge with an earlier task_kind in the same run
376
- try:
377
- prev = json.loads(path.read_text(encoding="utf-8"))
378
- history = prev.pop("history", [])
379
- history.append(prev)
380
- meta["history"] = history
381
- except (json.JSONDecodeError, OSError):
382
- pass
383
-
384
- path.write_text(json.dumps(meta, indent=2, ensure_ascii=False), encoding="utf-8")
385
- return path
386
-
387
-
388
- def write_task_meta(task_dir: str | Path, meta: dict[str, Any]) -> Path:
389
- """Write ``<task_dir>/meta.json`` next to the artifacts."""
390
- d = Path(task_dir)
391
- d.mkdir(parents=True, exist_ok=True)
392
- path = d / "meta.json"
393
- from .record_paths import portable_value
394
- path.write_text(json.dumps(portable_value(meta, d), indent=2, ensure_ascii=False, default=str),
395
- encoding="utf-8")
396
- return path
397
-
398
-
399
- def write_results_summary(results: list[dict], task_kind: str,
400
- run_id: str = DEFAULT_RUN_ID) -> list[Path]:
401
- """
402
- Group ``run()`` results by ``game_id`` and write one summary per game,
403
- refresh ``run_meta.json`` and the ``latest`` symlink.
404
-
405
- Returns the summary paths written.
406
- """
407
- check_kind(task_kind)
408
- grouped: dict[str, list[dict]] = {}
409
- for r in results:
410
- grouped.setdefault(r.get("game_id") or UNASSIGNED_GAME, []).append(r)
411
-
412
- written = []
413
- for game_id, items in grouped.items():
414
- path = results_summary_path(game_id, task_kind, run_id)
415
- path.parent.mkdir(parents=True, exist_ok=True)
416
- path.write_text(json.dumps(items, indent=2, ensure_ascii=False, default=str),
417
- encoding="utf-8")
418
- write_run_meta(game_id, run_id, task_kind=task_kind,
419
- extra={"num_tasks": len(items)})
420
- update_latest_link(game_id, run_id)
421
- written.append(path)
422
- return written
423
-
424
-
425
- # ── Misc ──────────────────────────────────────────────────────────────────────
426
-
427
- def resolve_input_path(path_like: str | Path) -> Path:
428
- """Resolve a possibly repo-relative input path (jsonl paths are repo-relative)."""
429
- if str(path_like).replace("\\", "/").startswith("bench/"):
430
- from .record_paths import resolve_path
431
- return resolve_path(path_like)
432
- p = Path(path_like)
433
- if p.is_absolute() or p.exists():
434
- return p
435
- return REPO_ROOT / p
436
-
437
-
438
- def rel_to_repo(path: str | Path) -> str:
439
- """Repo-relative string if possible, else the absolute path — for logs / meta."""
440
- p = Path(path).resolve()
441
- try:
442
- return str(p.relative_to(REPO_ROOT))
443
- except ValueError:
444
- return str(p)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/record_paths.py DELETED
@@ -1,77 +0,0 @@
1
- """Portable workspace paths for persisted records; resolve only when doing I/O."""
2
- from __future__ import annotations
3
-
4
- from pathlib import Path
5
- import re
6
-
7
- BENCH = Path(__file__).resolve().parents[3]
8
- _PREFIXES = ("test_data/", "GameFactory-3A/", "scripts/", "prompts/", "config/", "reports/")
9
- _ROOT = BENCH.as_posix()
10
- _INLINE_ROOT = re.compile(re.escape(_ROOT).replace("/", r"[\\/]"), re.IGNORECASE)
11
-
12
-
13
- def resolve_path(value: str | Path) -> Path:
14
- text = str(value).replace("\\", "/")
15
- if text == "bench" or text.startswith("bench/"):
16
- path = (BENCH.parent / text).resolve()
17
- path.relative_to(BENCH.resolve())
18
- return path
19
- path = Path(value)
20
- return path if path.is_absolute() else BENCH / path
21
-
22
-
23
- def portable_path(value: str | Path, base: Path | None = None) -> str:
24
- original = str(value)
25
- text = original.replace("\\", "/")
26
- if re.match(r"^[A-Za-z][A-Za-z0-9+.-]*:", text) and not re.match(r"^[A-Za-z]:/", text):
27
- return original # URLs, API URIs and dependency descriptors retain their meaning.
28
- if text.casefold() == _ROOT.casefold() or text.casefold().startswith(_ROOT.casefold() + "/"):
29
- return "bench" + text[len(_ROOT):]
30
- if text == "bench" or text.startswith("bench/"):
31
- path = resolve_path(text)
32
- suffix = path.relative_to(BENCH.resolve()).as_posix()
33
- return "bench" if suffix == "." else "bench/" + suffix
34
- if text.startswith(_PREFIXES):
35
- path = (BENCH / text).resolve()
36
- path.relative_to(BENCH.resolve())
37
- return "bench/" + path.relative_to(BENCH.resolve()).as_posix()
38
- if base is not None and "/" in text and "\n" not in text and not text.startswith("/"):
39
- candidate = base / text
40
- try:
41
- if candidate.exists():
42
- return "bench/" + candidate.resolve().relative_to(BENCH.resolve()).as_posix()
43
- except (OSError, ValueError):
44
- pass
45
- return original
46
-
47
-
48
- def portable_text(value: str) -> str:
49
- """Convert workspace paths embedded in log/error text, leaving URLs intact."""
50
- pattern = re.compile(r"(?:https?://[^\s<>]+)|(" + _INLINE_ROOT.pattern + r"(?:[\\/][^\r\n\"'<>|;,)]*)?)", re.IGNORECASE)
51
- return pattern.sub(lambda match: portable_path(match.group(1)) if match.group(1) else match.group(0), value)
52
-
53
-
54
- def portable_value(value, base: Path | None = None):
55
- if isinstance(value, dict):
56
- return {portable_path(key) if isinstance(key, str) else key: portable_value(item, base)
57
- for key, item in value.items()}
58
- if isinstance(value, (list, tuple)):
59
- return [portable_value(item, base) for item in value]
60
- if isinstance(value, Path):
61
- return portable_path(value, base)
62
- if isinstance(value, str):
63
- result = portable_path(value, base)
64
- return portable_text(result)
65
- return value
66
-
67
-
68
- def runtime_value(value):
69
- """Hydrate a verified frozen plan for runtime use; never use before hashing."""
70
- if isinstance(value, dict):
71
- return {str(resolve_path(key)) if isinstance(key, str) and (key == "bench" or key.startswith("bench/")) else key:
72
- runtime_value(item) for key, item in value.items()}
73
- if isinstance(value, list):
74
- return [runtime_value(item) for item in value]
75
- if isinstance(value, str) and (value == "bench" or value.startswith("bench/")):
76
- return str(resolve_path(value))
77
- return value
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/prompts/stage2_assets.md DELETED
@@ -1,95 +0,0 @@
1
- # GameFactory-3A Stage 2 — asset generation
2
-
3
- 你正在执行同一个游戏测例的第二阶段资产生成。**开始前必须先阅读 `{SETTING_OVERVIEW}`,并严格遵循其中的 asset refinement、asset QA 和 validation 流程。**
4
-
5
- ## 测例输入
6
-
7
- - 游戏测例:`{GAME_ID}`
8
- - 测例目录:`{CASE_ROOT}`
9
- - GameFactory-3A 根目录:`{FRAMEWORK_ROOT}`
10
- - 总入口文档:`{SETTING_OVERVIEW}`
11
- - 目标引擎:`{ENGINE}`
12
- - 完整游戏需求:`{GENERAL_REQUIREMENT}`
13
- - 当前 run 根目录:`{RUN_ROOT}`
14
- - runner 控制面(冻结索引、prompt/packet/log):`{GENERATION_CONTROL_ROOT}`
15
- - 第一阶段白模 run:`{WHITEBOX_RUN_ROOT}`
16
- - 第一阶段白模 workspace:`{WHITEBOX_WORKSPACE}`
17
- - 第一阶段资产接口:`{ASSET_SLOTS}`
18
- - 第一阶段计划/TODO:`{GENERATION_PLAN}`
19
- - 外部冻结的资产任务索引:`{ASSET_TASK_INDEX}`
20
- - 资产输出根目录:`{ASSET_OUTPUT_ROOT}`
21
- - 当前阶段 pipeline task 目录:`{STAGE_WORKSPACE}`
22
- - 资产 QA 证据根:`{RUN_ROOT}/eval/<task_kind>/<task_id>/`
23
-
24
- 请读取游戏完整梗概和玩法,再读取 `{ASSET_TASK_INDEX}` 中列出的**每一条**资产任务、对应的 requirement、JSONL 和参考图片。**每一个资产生成任务都必须使用完整游戏梗概作为上下文**,不能只把孤立的资产 prompt 交给生成流程。
25
-
26
- ## 本阶段任务
27
-
28
- 1. 根据第一阶段的任务拆解、TODO 和 `asset_slots.json` 确认资产用途、生命周期、比例、朝向、挂点和风格;
29
- 2. 按 `setting_overview.md` 路由到对应的 T-pose、3D object、3D scene、Motion、Audio、CG video 或其他资产 Skill;
30
- 3. 逐项生成冻结任务索引中的所有资产,并执行对应的结构/视觉 QA;
31
- 4. 保持第一阶段定义的 asset ID 和游戏/UI/mechanic contract 不变;
32
- 5. 记录资产来源、配置、参考图、QA 和后续导入要求;
33
- 6. 生成完整的资产 manifest,供第三阶段导入使用。
34
-
35
- ## 限制
36
-
37
- - 不修改第一阶段核心游戏机制和需求;
38
- - 不擅自删除、合并或跳过资产任务;
39
- - 资产失败必须记录为失败,不能伪造成功;
40
- - 不将资产生成 QA 当成游戏已经完成;
41
- - 本阶段默认不执行 Agent repair。
42
-
43
- ## 输出位置和 Layer 规则
44
-
45
- 资产必须按框架标准布局写入:
46
-
47
- - `{ASSET_OUTPUT_ROOT}/tpose/<task_id>/`;
48
- - `{ASSET_OUTPUT_ROOT}/3d_object/<task_id>/`;
49
- - `{ASSET_OUTPUT_ROOT}/3d_scene/<task_id>/`;
50
- - `{ASSET_OUTPUT_ROOT}/motion/<task_id>/`;
51
- - `{ASSET_OUTPUT_ROOT}/audio/<task_id>/`;
52
- - `{ASSET_OUTPUT_ROOT}/cg_video/<task_id>/`。
53
-
54
- 每个任务目录至少应有资产文件、`meta.json`、预览(若适用)和 QA 证据。资产 manifest 和阶段产物说明写入 `{ASSET_MANIFEST}`(位于标准 `pipeline/<task_id>/` task 目录);冻结索引只读自 `{ASSET_TASK_INDEX}`,runner 控制文件写入 `{GENERATION_CONTROL_ROOT}`。每个资产任务的 QA 证据写入 `{RUN_ROOT}/eval/<task_kind>/<task_id>/`。至少生成:
55
-
56
- - `asset_manifest.json`;
57
- - `asset_generation_report.json`;
58
- - `asset_task_index.json`(由外部入口生成,不得删改任务);
59
- - 失败任务及其失败原因。
60
-
61
- 不要写入 case 输入目录、不要覆盖白模 run,也不要创建 `pipeline/assets/` stage 目录;真实资产必须按 `assets/<task_kind>/<task_id>/` 放置,不能把所有文件平铺在 pipeline task 目录。
62
-
63
- ## 已确认的优先 3D Object 双服务对比
64
-
65
- 当用户明确选择优先 200 条对象任务的 Meshy / Tripo API 对比时,使用
66
- `scripts/asset/generate_3d_comparison.py` 与 `config/asset_3d_comparison.json`,
67
- 具体命令和费用见 `scripts/asset/README_3d_comparison.md`。这是跨游戏的对象生成实验;
68
- 各游戏只包含 3D Object 子集,不能将其标为六类资产齐全的完整 Stage 2。
69
-
70
- - 输入以 `test_data/priority_gen/3d_object_priority_lowpoly_collect.jsonl` 的全部 200 条为准,
71
- 保留原 task ID、prompt、seed、参考图及最高 8000 三角面的既定政策。
72
- - 完整游戏需求保留在冻结输入包;两家文生调用使用完全相同的短梗概和原资产 prompt,
73
- 合计不超过 800 字符。API 未接收完整正文这一事实必须记录,不能宣称正文已送入模型。
74
- - 两家图生均使用同一冻结参考图,列入 `reference_only` 辅助组;不要为了补文字上下文
75
- 悄悄改图、另画参考图、改变原任务的输入模式或仅给一家添加贴图提示词。
76
- - 每家分别运行全部 200 条。按人物/刚体、文生/图生及上下文条件分组;Hunyuan3D Pro
77
- 仅保留后续接入口,本轮不启用、不消耗第三家额度。
78
- - 固定具体模型版本,不使用 `latest`;Meshy 本轮使用 standard + remesh,
79
- 不依赖将于 2026-10-30 退役的 `model_type=lowpoly`。
80
- - 原始 API mesh 写为 `model.glb`,统一后处理的导入 mesh 写为 `import/model.glb`。
81
- 两家使用同一固定工具、参数与贴���上限;分别记录原始和后处理 QA、耗时及缺陷。
82
- 减面未达到预算应如实记录失败,不增加生成次数、不换模型补齐或挑选 best-of。
83
- - 付费调用前需要用户确认预算并在环境中配置密钥。提交后立刻落盘任务 ID;
84
- 已知远端任务只续查结果,提交结果不确定时停下核对,不自动重发付费任务。
85
- - 本轮按游戏级共享资产目录落盘,在任务目录下加厂商一级:
86
- `test_data/outputs/<game_id>/assets/3d_object/<task_id>/<provider>/`,
87
- `<provider>` 为 `meshy` / `tripo`,不使用 `asset_comparison` 输出目录。
88
- 原始 `model.glb`、`import/model.glb`、`meta.json`、服务端预览均放在厂商目录内;
89
- QA 写入 `test_data/outputs/<game_id>/eval/3d_object/<task_id>/<provider>/`。
90
- manifest/阶段报告按游戏、批次、厂商放在
91
- `test_data/outputs/<game_id>/pipeline/<game_id>_pipeline/<batch_id>/<provider>/`,
92
- 冻结输入、计费回执与跨游戏预览保存在 `test_data/outputs/_asset_batches/<batch_id>/`。
93
- 不覆盖不同批次已有的厂商资产,也不覆盖白模工程。
94
- 结构检查通过不等于视觉 QA 通过;C、Q1/Q2/Q3 的正式评分缺少证据时保持 UNKNOWN。
95
- 第三阶段的资产来源需要另行固定,不能按逐条最高评分混选两家产物。
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/cloud_journal.py DELETED
@@ -1,178 +0,0 @@
1
- """Durable submit receipts for the comparison runner; no automatic POST retry.
2
-
3
- Uses the framework's HTTP plumbing. A lost create response is ambiguous: retain
4
- the credit reservation and require reconciliation instead of buying a new task.
5
- The runner holds one batch lock while using this client.
6
- """
7
- from __future__ import annotations
8
-
9
- import json
10
- import math
11
- import os
12
- from pathlib import Path
13
- from typing import Any
14
-
15
- from models.common import cloud_api
16
- from pipeline.common.record_paths import portable_value, portable_text
17
-
18
-
19
- def read_json(path: Path, default=None):
20
- return json.loads(path.read_text(encoding="utf-8")) if path.is_file() else default
21
-
22
-
23
- def write_json(path: Path, value: Any, *, portable: bool = True) -> None:
24
- path.parent.mkdir(parents=True, exist_ok=True)
25
- temporary = path.with_name(path.name + ".tmp")
26
- with temporary.open("w", encoding="utf-8", newline="\n") as stream:
27
- json.dump(portable_value(value, path.parent) if portable else value,
28
- stream, ensure_ascii=False, indent=2)
29
- stream.write("\n")
30
- stream.flush()
31
- os.fsync(stream.fileno())
32
- os.replace(temporary, path)
33
-
34
-
35
- def safe_error(exc: Exception) -> str:
36
- value = str(exc)
37
- for name in ("MESHY_API_KEY", "TRIPO_API_KEY"):
38
- secret = os.environ.get(name)
39
- if secret:
40
- value = value.replace(secret, "[REDACTED]")
41
- return portable_text(value)[:1600]
42
-
43
-
44
- class BudgetExceeded(RuntimeError):
45
- pass
46
-
47
-
48
- class AmbiguousSubmission(RuntimeError):
49
- pass
50
-
51
-
52
- def reserved_credits(directory: Path) -> int:
53
- return sum(
54
- int(item["reserved_credits"])
55
- for path in directory.glob("*.json")
56
- for item in read_json(path, {}).get("submissions", {}).values()
57
- if item["status"] != "rejected"
58
- )
59
-
60
-
61
- class JournalClient:
62
- """Replay successful create receipts and query the same remote task on resume."""
63
-
64
- def __init__(self, base: str, key: str, journal: Path, cap: int,
65
- prices: dict[str, int], provider: str, transport=None, resume_refused=False):
66
- self.journal = journal
67
- self.cap = cap
68
- self.prices = prices
69
- self.provider = provider
70
- self.resume_refused = resume_refused
71
- # GET is safe to retry. A billed POST gets exactly one transport attempt.
72
- self.reader = transport or cloud_api.CloudAPIClient(base, key, max_retries=3)
73
- self.writer = transport or cloud_api.CloudAPIClient(base, key, max_retries=0)
74
- self.public = None
75
-
76
- def _cost(self, endpoint: str, body: dict) -> int:
77
- if endpoint == "/files":
78
- return 0
79
- if self.provider == "meshy":
80
- if endpoint == "/openapi/v2/text-to-3d":
81
- return self.prices["text_refine" if body.get("mode") == "refine" else "text_preview"]
82
- if endpoint == "/openapi/v1/image-to-3d":
83
- return self.prices["image"]
84
- elif endpoint in ("/generation/text-to-model", "/generation/image-to-model"):
85
- return self.prices["text" if "text-to" in endpoint else "image"]
86
- raise ValueError(f"Unpriced POST endpoint: {endpoint}")
87
-
88
- def request(self, method: str, endpoint: str, **kwargs):
89
- state = read_json(self.journal, {"submissions": {}, "results": {}})
90
- if method != "POST":
91
- result = self.reader.request(method, endpoint, **kwargs)
92
- if method == "GET" and isinstance(result, dict):
93
- state["results"][endpoint] = result
94
- payload = result.get("data", result)
95
- if isinstance(payload, dict):
96
- remote_id = payload.get("task_id") or payload.get("id")
97
- reported = payload.get("credits_consumed", payload.get("credits"))
98
- for record in state["submissions"].values():
99
- if remote_id and record.get("remote_id") == remote_id and isinstance(reported, (int, float)):
100
- record["reported_credits"] = reported
101
- record["reserved_credits"] = max(record["reserved_credits"], math.ceil(reported))
102
- write_json(self.journal, state)
103
- return result
104
-
105
- # Hash binary inputs; never persist image bytes or credentials in receipts.
106
- descriptor = {"endpoint": endpoint, "body": kwargs.get("json_body"),
107
- "files": {name: cloud_api.bytes_hash(value[1])
108
- for name, value in (kwargs.get("files") or {}).items()}}
109
- fingerprint = cloud_api.request_hash(descriptor)
110
- previous = state["submissions"].get(fingerprint)
111
- if previous:
112
- if previous["status"] == "created":
113
- return previous["response"]
114
- if previous["status"] == "rejected" and not (self.resume_refused and previous.get("resource_refusal")):
115
- raise cloud_api.CloudAPIRequestError(previous["error"])
116
- if previous["status"] != "rejected":
117
- raise AmbiguousSubmission(
118
- f"Unconfirmed submission {fingerprint} in {self.journal}. "
119
- "Reconcile its remote task ID; no automatic resubmission.")
120
-
121
- cost = self._cost(endpoint, kwargs.get("json_body") or {})
122
- used = reserved_credits(self.journal.parent)
123
- if used + cost > self.cap:
124
- raise BudgetExceeded(f"{self.provider}: {used} reserved + {cost} exceeds cap {self.cap}")
125
- record = {"endpoint": endpoint, "reserved_credits": cost, "status": "submitting"}
126
- if previous:
127
- record["rejection_history"] = [*previous.get("rejection_history", []),
128
- {k: v for k, v in previous.items() if k != "rejection_history"}]
129
- state["submissions"][fingerprint] = record
130
- write_json(self.journal, state) # commit before the request can incur a charge
131
- try:
132
- response = self.writer.request(method, endpoint, **kwargs)
133
- except cloud_api.CloudAPIError as exc:
134
- # These are explicit refusals. 5xx/timeouts may have created a task.
135
- if exc.status in (400, 401, 402, 403, 404, 405, 413, 415, 422, 429):
136
- record.update(status="rejected", error=safe_error(exc), resource_refusal=isinstance(
137
- exc, (cloud_api.CloudAPIAuthError, cloud_api.CloudAPIQuotaError, cloud_api.CloudAPIRateLimitError)))
138
- else:
139
- record.update(status="uncertain", error=safe_error(exc))
140
- write_json(self.journal, state)
141
- if record["status"] == "uncertain":
142
- raise AmbiguousSubmission(f"Unconfirmed POST {fingerprint}; {safe_error(exc)}") from exc
143
- raise
144
-
145
- rejected = isinstance(response, dict) and response.get("code") not in (None, 0)
146
- envelope = response.get("data", response) if isinstance(response, dict) else {}
147
- remote_id = (envelope.get("task_id") or envelope.get("id") or envelope.get("result")
148
- or envelope.get("file_token") or envelope.get("token")) if isinstance(envelope, dict) else None
149
- if rejected:
150
- rejection = cloud_api.classify_http_error(400, response)
151
- record.update(status="rejected", error=safe_error(rejection), resource_refusal=isinstance(
152
- rejection, (cloud_api.CloudAPIAuthError, cloud_api.CloudAPIQuotaError)))
153
- elif not remote_id:
154
- record.update(status="uncertain", error="Response contained no remote task/file ID")
155
- else:
156
- record.update(status="created", remote_id=remote_id, response=response)
157
- write_json(self.journal, state)
158
- if rejected:
159
- raise rejection
160
- if record["status"] == "uncertain":
161
- raise AmbiguousSubmission(f"Unconfirmed POST {fingerprint}: missing task/file ID")
162
- return response
163
-
164
- def download(self, url: str, **kwargs) -> bytes:
165
- # Signed CDN downloads must not receive the provider's Authorization key.
166
- if not url.startswith("https://"):
167
- raise ValueError("Artifact download must use HTTPS")
168
- if self.public is None:
169
- self.public = cloud_api.CloudAPIClient("", "", max_retries=3)
170
- self.public.session.headers.pop("Authorization", None)
171
- return self.public.download(url, **kwargs)
172
-
173
- def close(self):
174
- self.reader.close()
175
- if self.writer is not self.reader:
176
- self.writer.close()
177
- if self.public is not None:
178
- self.public.close()
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/generate_3d_comparison.py DELETED
@@ -1,603 +0,0 @@
1
- #!/usr/bin/env python3
2
- """Prepare, run and resume the frozen Meshy/Tripo 200-object comparison.
3
-
4
- Prepare/status are offline. Run requires explicit spending caps and API keys.
5
- Native generation outputs remain unmodified. A shared postprocessor produces
6
- separate import meshes; visual acceptance is a separate review.
7
- """
8
- from __future__ import annotations
9
-
10
- import argparse
11
- from collections import Counter, defaultdict
12
- from contextlib import contextmanager
13
- from datetime import datetime, timezone
14
- import hashlib
15
- import io
16
- import json
17
- import math
18
- import os
19
- from pathlib import Path
20
- import re
21
- import shutil
22
- import struct
23
- import subprocess
24
- import sys
25
-
26
- BENCH = Path(__file__).resolve().parents[2]
27
- FRAMEWORK = BENCH / "GameFactory-3A"
28
- sys.path.insert(0, str(FRAMEWORK))
29
- from pipeline.common import paths
30
- from pipeline.common.record_paths import portable_value, resolve_path, runtime_value
31
- from models.common import cloud_api
32
- from models.common.glb_utils import glb_json_chunk, glb_summary
33
- from cloud_journal import (AmbiguousSubmission, BudgetExceeded, JournalClient,
34
- read_json, reserved_credits, safe_error, write_json)
35
-
36
- DEFAULT_CONFIG = BENCH / "config/asset_3d_comparison.json"
37
- IDENTIFIER = re.compile(r"^[A-Za-z0-9][A-Za-z0-9_-]{0,160}$")
38
- RUNTIME_FILES = [Path(__file__), Path(__file__).with_name("cloud_journal.py"),
39
- FRAMEWORK / "pipeline/common/record_paths.py",
40
- FRAMEWORK / "pipeline/common/paths.py",
41
- BENCH / "prompts/stage2_assets.md",
42
- FRAMEWORK / "operators/gen_3d_object/operator.py",
43
- FRAMEWORK / "models/gen_3d_object/meshy_model.py",
44
- FRAMEWORK / "models/gen_3d_object/tripo_model.py",
45
- FRAMEWORK / "models/common/cloud_api.py",
46
- FRAMEWORK / "models/common/glb_utils.py",
47
- Path(__file__).with_name("mesh_tools") / "process_mesh.mjs",
48
- Path(__file__).with_name("mesh_tools") / "package-lock.json"]
49
- MESH_TOOL = Path(__file__).with_name("mesh_tools") / "process_mesh.mjs"
50
-
51
-
52
- def now():
53
- return datetime.now(timezone.utc).isoformat(timespec="seconds")
54
-
55
-
56
- def digest(path: Path):
57
- return hashlib.sha256(path.read_bytes()).hexdigest()
58
-
59
-
60
- def checked_id(value: str):
61
- if not IDENTIFIER.fullmatch(value):
62
- raise ValueError(f"Invalid path identifier: {value!r}")
63
- return value
64
-
65
-
66
- def bench_path(value: str):
67
- result = resolve_path(value).resolve()
68
- result.relative_to(BENCH.resolve())
69
- return result
70
-
71
-
72
- def batch_directory(config: dict, batch_id: str):
73
- return bench_path(config["output_root"]) / "_asset_batches" / checked_id(batch_id)
74
-
75
-
76
- @contextmanager
77
- def exclusive_batch(directory: Path):
78
- """OS lock releases on process exit, including an interrupted run."""
79
- directory.mkdir(parents=True, exist_ok=True)
80
- with (directory / "batch.lock").open("a+b") as stream:
81
- stream.seek(0, 2)
82
- if stream.tell() == 0:
83
- stream.write(b"0")
84
- stream.flush()
85
- stream.seek(0)
86
- if os.name == "nt":
87
- import msvcrt
88
- msvcrt.locking(stream.fileno(), msvcrt.LK_NBLCK, 1)
89
- else:
90
- import fcntl
91
- fcntl.flock(stream, fcntl.LOCK_EX | fcntl.LOCK_NB)
92
- try:
93
- yield
94
- finally:
95
- stream.seek(0)
96
- if os.name == "nt":
97
- msvcrt.locking(stream.fileno(), msvcrt.LK_UNLCK, 1)
98
- else:
99
- fcntl.flock(stream, fcntl.LOCK_UN)
100
-
101
-
102
- def game_directory(plan: dict, game: str):
103
- return Path(plan["output_root"]) / checked_id(game)
104
-
105
-
106
- def pipeline_directory(plan: dict, provider: str, game: str):
107
- return (game_directory(plan, game) / "pipeline" / (game + "_pipeline")
108
- / checked_id(plan["batch_id"]) / checked_id(provider))
109
-
110
-
111
- def generation_index(plan: dict, provider: str, game: str):
112
- return (game_directory(plan, game) / ".generation/3d_object" / checked_id(provider)
113
- / checked_id(plan["batch_id"]) / "asset_task_index.json")
114
-
115
-
116
- def short_context(text: str, limit: int):
117
- """Declared deterministic excerpts, not a claim that the model read the brief."""
118
- title = re.search(r"^title:\s*(.+)$", text, re.M)
119
- lines = [line.strip() for line in text.splitlines() if line.strip()]
120
- title_text = title.group(1) if title else next((s.lstrip("# ") for s in lines if s.startswith("#")), "")
121
- world = next((line for line in lines if re.match(r"-\s*W01\b", line)), "")
122
- loop = next((line for line in lines if any(word in line for word in ("核心循环:", "核心循环:", "核心玩法:", "核心玩法:"))), "")
123
- if not loop:
124
- loop = next((line for line in lines if re.match(r"-\s*S01\b", line)), "")
125
- if not loop:
126
- heading = re.search(r"^##[^\n]*(?:核心循环|一句话定位)[^\n]*\n([^#]+)", text, re.M)
127
- loop = heading.group(1) if heading else ""
128
- pieces = [re.sub(r"\s+", " ", value).strip()[:size]
129
- for value, size in ((title_text, 70), (world, 110), (loop, 110)) if value]
130
- value = " | ".join(pieces)
131
- if not value:
132
- raise ValueError("Cannot extract game context; provide a reviewed short context")
133
- return value[:limit]
134
-
135
-
136
- def frozen_copy(source: Path, destination: Path):
137
- if destination.exists() and digest(destination) != digest(source):
138
- raise ValueError(f"Frozen input changed; use a new batch ID: {destination}")
139
- destination.parent.mkdir(parents=True, exist_ok=True)
140
- if not destination.exists():
141
- shutil.copyfile(source, destination)
142
- return {"path": str(destination.resolve()), "sha256": digest(destination),
143
- "source": str(source.resolve())}
144
-
145
-
146
- def node_process(arguments: list[str], timeout: int):
147
- node = shutil.which("node")
148
- if not node:
149
- raise ValueError("Node.js is required for the fixed mesh pipeline")
150
- environment = os.environ.copy()
151
- for key in ("MESHY_API_KEY", "TRIPO_API_KEY"):
152
- environment.pop(key, None)
153
- if os.name == "nt":
154
- # Conda's DLLs can shadow sharp's libvips dependencies in child processes.
155
- environment["PATH"] = str(Path(node).parent) + os.pathsep + str(Path(os.environ["SystemRoot"]) / "System32")
156
- return subprocess.run([node, str(MESH_TOOL), *arguments], capture_output=True,
157
- text=True, encoding="utf-8", timeout=timeout, env=environment)
158
-
159
-
160
- def mesh_runtime():
161
- result = node_process(["--check"], timeout=60)
162
- if result.returncode:
163
- raise ValueError("Mesh postprocessor unavailable; run npm ci in scripts/asset/mesh_tools. " + result.stderr[-500:])
164
- return json.loads(result.stdout)
165
-
166
-
167
- def subject_group(task: dict):
168
- role = task.get("role", "")
169
- return "character" if task.get("rig", "none") != "none" or role in (
170
- "character", "player", "event_figure", "npc", "creature", "unit_special") or role.startswith("enemy") else "rigid"
171
-
172
-
173
- def prepare(config: dict, batch_id: str):
174
- if config["input_policy"] != "shared_short_context_and_reference_only" or not config["input_policy_confirmed"]:
175
- raise ValueError("The input protocol must be confirmed before preparing this comparison")
176
- directory = batch_directory(config, batch_id)
177
- plan_file = directory / "plan.json"
178
- if plan_file.exists():
179
- raise ValueError(f"Batch is already frozen: {plan_file}. Use status/run or a new batch ID.")
180
- source = bench_path(config["tasks"])
181
- tasks = [json.loads(line) for line in source.read_text(encoding="utf-8-sig").splitlines() if line.strip()]
182
- if len(tasks) != config["expected_tasks"]:
183
- raise ValueError(f"Expected {config['expected_tasks']} tasks; got {len(tasks)}")
184
- for task in tasks:
185
- for provider in config["providers"]:
186
- output = (bench_path(config["output_root"]) / checked_id(task["game_id"])
187
- / "assets/3d_object" / checked_id(task["task_id"]) / checked_id(provider))
188
- if output.is_dir() and any(p.name != ".gitkeep" for p in output.iterdir()):
189
- raise ValueError(f"Provider directory already contains assets; resume its batch: {output}")
190
- seen = set()
191
- frozen_tasks = []
192
- documents = {}
193
- tool_runtime = mesh_runtime()
194
- for task in tasks:
195
- game = checked_id(task["game_id"])
196
- task_id = checked_id(task["task_id"])
197
- identity = (game, task_id)
198
- if identity in seen:
199
- raise ValueError(f"Duplicate task: {identity}")
200
- seen.add(identity)
201
- case = BENCH / "test_data/test_samples" / game
202
- if game not in documents:
203
- requirement = case / "general_requirement.txt"
204
- full_text = requirement.read_text(encoding="utf-8")
205
- doc_root = directory / "inputs" / game
206
- documents[game] = {
207
- "general_requirement": frozen_copy(requirement, doc_root / "general_requirement.txt"),
208
- "object_requirement": frozen_copy(case / "3D_object/requirement.txt", doc_root / "object_requirement.txt"),
209
- "context_excerpt": short_context(full_text, config["context_excerpt_max_chars"]),
210
- "context_method": "deterministic title/world/core-loop excerpts; full brief retained locally",
211
- }
212
- mode = task["input_mode"]
213
- prompt = None
214
- reference = None
215
- if mode == "text_to_3d":
216
- # Never shorten the asset prompt. Both providers receive this exact string.
217
- prompt = f"Game context (excerpt): {documents[game]['context_excerpt']}\nAsset: {task['prompt']}"
218
- if len(prompt) > config["shared_prompt_max_chars"]:
219
- raise ValueError(f"Shared prompt exceeds limit: {task_id} ({len(prompt)} chars)")
220
- elif mode == "image_to_3d":
221
- source_image = bench_path(task["image_path"])
222
- reference = frozen_copy(source_image, directory / "references" / (digest(source_image) + source_image.suffix.lower()))
223
- else:
224
- raise ValueError(f"Unsupported input mode: {mode}")
225
- if not 500 <= task["decimation_target"] <= 8000:
226
- raise ValueError(f"Unexpected triangle target: {task_id}")
227
- frozen_tasks.append({"task": task, "shared_prompt": prompt, "reference": reference,
228
- "subject_group": subject_group(task),
229
- "context_condition": "shared_short_context" if prompt else "reference_only"})
230
- output_root = bench_path(config["output_root"])
231
- plan = {"schema_version": "asset_3d_comparison_plan_v1", "batch_id": batch_id,
232
- "created_at": now(), "scope": "3d_object_only_not_complete_stage2",
233
- "config": config, "output_root": str(output_root), "directory": str(directory),
234
- "source": frozen_copy(source, directory / "inputs/source_tasks.jsonl"),
235
- "method_reference": frozen_copy(bench_path(config["method_reference"]), directory / "inputs/method_selection.md"),
236
- "documents": documents, "tasks": frozen_tasks,
237
- "runtime_sources": {str(p.resolve()): digest(p) for p in RUNTIME_FILES},
238
- "estimated_credits": {}, "recommended_credit_caps": {},
239
- "qa_policy": "basic structural checks + pending visual review; no automatic acceptance",
240
- "texture_policy": "preserve native output; same deterministic downsize for both import meshes",
241
- "postprocess": config["postprocess"], "mesh_runtime": tool_runtime,
242
- "deferred_providers": {"hunyuan3d_pro": {"status": "not_enabled", "adapter_contract":
243
- "infer_and_save(image, output_path, seed, decimation_target, texture_size, *, prompt); requires pinned cloud version and input capability check"}}}
244
- for provider, settings in config["providers"].items():
245
- prices = settings["credits"]
246
- estimate = sum((prices["text_preview"] + prices["text_refine"] if provider == "meshy" else prices["text"])
247
- if item["shared_prompt"] else prices["image"] for item in frozen_tasks)
248
- plan["estimated_credits"][provider] = estimate
249
- plan["recommended_credit_caps"][provider] = (estimate * 11 + 9) // 10
250
- plan["plan_sha256"] = cloud_api.request_hash(portable_value(plan))
251
- write_json(plan_file, plan)
252
- write_json(directory / "asset_task_index.json", {"plan_sha256": plan["plan_sha256"],
253
- "task_count": len(frozen_tasks), "tasks": frozen_tasks})
254
- for provider in config["providers"]:
255
- for game in documents:
256
- write_json(generation_index(plan, provider, game), {
257
- "plan_sha256": plan["plan_sha256"], "game_id": game, "provider": provider,
258
- "scope": plan["scope"], "tasks": [item for item in frozen_tasks if item["task"]["game_id"] == game]})
259
- refresh_reports(plan)
260
- return plan
261
-
262
-
263
- def load_plan(config: dict, batch_id: str):
264
- plan = read_json(batch_directory(config, batch_id) / "plan.json")
265
- if not plan:
266
- raise ValueError("Prepare this batch first")
267
- claimed = plan["plan_sha256"]
268
- if cloud_api.request_hash({k: v for k, v in plan.items() if k != "plan_sha256"}) != claimed:
269
- raise ValueError("Frozen plan was modified")
270
- return runtime_value(plan)
271
-
272
-
273
- def validate_inputs(plan: dict):
274
- if mesh_runtime() != plan["mesh_runtime"]:
275
- raise ValueError("Mesh runtime differs from the frozen version")
276
- for filename, expected in plan["runtime_sources"].items():
277
- if digest(Path(filename)) != expected:
278
- raise ValueError(f"Runtime changed after freeze: {filename}; prepare a new batch")
279
- files = [plan["source"], plan["method_reference"]]
280
- files.extend(doc[field] for doc in plan["documents"].values()
281
- for field in ("general_requirement", "object_requirement"))
282
- files.extend(item["reference"] for item in plan["tasks"] if item["reference"])
283
- for entry in files:
284
- if digest(Path(entry["path"])) != entry["sha256"]:
285
- raise ValueError(f"Frozen input changed: {entry['path']}")
286
-
287
-
288
- def basic_qa(filename: Path, task: dict):
289
- data = filename.read_bytes()
290
- summary = glb_summary(data)
291
- checks = {"parseable_glb": "error" not in summary}
292
- if checks["parseable_glb"]:
293
- doc = glb_json_chunk(data)
294
- checks["glb_v2_exact_length"] = struct.unpack_from("<II", data, 4) == (2, len(data))
295
- checks["positive_triangle_count"] = bool(summary["triangles"])
296
- checks["triangle_budget"] = 0 < summary["triangles"] <= task["decimation_target"]
297
- checks["materials_present"] = summary["materials"] > 0
298
- checks["textures_present"] = summary["textures"] > 0 and summary["images"] > 0
299
- primitives = [p for m in doc.get("meshes", []) for p in m.get("primitives", [])]
300
- checks["uv_present"] = bool(primitives) and all("TEXCOORD_0" in p.get("attributes", {}) for p in primitives)
301
- checks["self_contained"] = all("uri" not in b or b["uri"].startswith("data:") for b in doc.get("buffers", [])) and all(
302
- "bufferView" in image or str(image.get("uri", "")).startswith("data:") for image in doc.get("images", []))
303
- return {"status": "PASS" if all(checks.values()) else "FAIL", "checks": checks,
304
- "summary": summary, "triangle_target": task["decimation_target"],
305
- "requested_texture_size": task.get("texture_size"),
306
- "texture_budget_verified": False, "visual_status": "PENDING",
307
- "limitations": ["No visual, orientation, scale, rigging or engine acceptance yet",
308
- "Native texture resolution retained; task texture_size is not enforced",
309
- "Basic checks do not validate every accessor, topology defect or PBR channel"]}
310
-
311
-
312
- def process_import_mesh(source: Path, task: dict):
313
- destination = source.parent / "import/model.glb"
314
- result = node_process([str(source), str(destination), str(task["decimation_target"]),
315
- str(task["texture_size"])], timeout=300)
316
- if result.returncode:
317
- raise RuntimeError("Postprocessing failed: " + result.stderr[-1200:])
318
- report = json.loads(result.stdout)
319
- return destination, report
320
-
321
-
322
- def task_directories(plan: dict, provider: str, task: dict):
323
- root = game_directory(plan, task["game_id"])
324
- task_id, provider = checked_id(task["task_id"]), checked_id(provider)
325
- output = root / "assets/3d_object" / task_id / provider
326
- evaluation = root / "eval/3d_object" / task_id / provider
327
- return output, evaluation
328
-
329
-
330
- def refresh_reports(plan: dict):
331
- summary = {"batch_id": plan["batch_id"], "plan_sha256": plan["plan_sha256"],
332
- "updated_at": now(), "expected_assets": len(plan["tasks"]) * len(plan["config"]["providers"]),
333
- "estimated_credits": plan["estimated_credits"], "providers": {}}
334
- for provider in plan["config"]["providers"]:
335
- by_game = defaultdict(list)
336
- for item in plan["tasks"]:
337
- task = item["task"]
338
- output, evaluation = task_directories(plan, provider, task)
339
- meta = read_json(output / "meta.json", {})
340
- by_game[task["game_id"]].append({"task_id": task["task_id"], "asset_name": task["asset_name"],
341
- "task_kind": "3d_object", "status": meta.get("status", "prepared"),
342
- "artifact": str(output / "import/model.glb"), "native_artifact": str(output / "model.glb"),
343
- "meta": str(output / "meta.json"), "native_structural_qa": meta.get("native_structural_qa"),
344
- "import_structural_qa": meta.get("import_structural_qa"),
345
- "native_delivered": meta.get("native_delivered", False),
346
- "import_delivered": meta.get("import_delivered", False),
347
- "qa": str(evaluation / "qa.json"), "visual_status": "PENDING",
348
- "context_condition": item["context_condition"], "subject_group": item["subject_group"],
349
- "input_mode": task["input_mode"], "error": meta.get("error")})
350
- all_entries = [entry for entries in by_game.values() for entry in entries]
351
- summary["providers"][provider] = {"task_count": len(all_entries),
352
- "status_counts": dict(Counter(entry["status"] for entry in all_entries)),
353
- "reserved_credits": reserved_credits(Path(plan["directory"]) / "submissions" / provider),
354
- "visual_accepted": 0, "method_group": plan["config"]["method_group"],
355
- "benchmark_metrics": {name: {"status": "UNKNOWN", "value": None}
356
- for name in ("C", "Q1", "Q2", "Q3")},
357
- "subgroups": {key: dict(Counter(e[key] for e in all_entries))
358
- for key in ("subject_group", "input_mode", "context_condition")},
359
- "game_macro_delivery": sum(sum(e["status"] == "delivered" for e in entries) / len(entries)
360
- for entries in by_game.values()) / len(by_game),
361
- "game_macro_native_delivery": sum(sum(e["native_delivered"] for e in entries) / len(entries)
362
- for entries in by_game.values()) / len(by_game),
363
- "native_structural_pass": sum(e["native_structural_qa"] == "PASS" for e in all_entries),
364
- "import_structural_pass": sum(e["import_structural_qa"] == "PASS" for e in all_entries),
365
- "delivery_metric_status": "NOT_RUN" if all(e["status"] == "prepared" for e in all_entries) else "PARTIAL"}
366
- for game, entries in by_game.items():
367
- pipeline = pipeline_directory(plan, provider, game)
368
- manifest = {"game_id": game, "provider": provider, "batch_id": plan["batch_id"],
369
- "plan_sha256": plan["plan_sha256"], "scope": plan["scope"], "assets": entries,
370
- "full_context": plan["documents"][game], "status": "pending_visual_review"}
371
- write_json(pipeline / "asset_manifest.json", manifest)
372
- write_json(pipeline / "asset_generation_report.json", {
373
- "task_count": len(entries), "status_counts": dict(Counter(e["status"] for e in entries)),
374
- "failures": [e for e in entries if e["error"]], "visual_accepted": 0})
375
- paths.write_task_meta(pipeline, {"game_id": game, "task_id": game + "_pipeline",
376
- "task_kind": "pipeline", "run_id": plan["batch_id"], "status": "comparison_only",
377
- "provider": provider, "scope": plan["scope"]})
378
- write_json(Path(plan["directory"]) / "asset_generation_report.json", summary)
379
- return summary
380
-
381
-
382
- def make_model(plan: dict, provider: str, task: dict, cap: int):
383
- from models.gen_3d_object.meshy_model import MeshyModel
384
- from models.gen_3d_object.tripo_model import TripoModel
385
- settings = plan["config"]["providers"][provider]
386
- cls = MeshyModel if provider == "meshy" else TripoModel
387
- control = Path(plan["directory"])
388
- model = cls(model_path=settings["model"], device="cpu", api_base=settings["api_base"],
389
- cache_dir=str(control / "cache" / task["game_id"] / task["task_id"]),
390
- timeout=1800, poll_interval=5, **settings["options"])
391
- key = cloud_api.require_api_key(None, provider.upper() + "_API_KEY", "See the comparison README")
392
- journal = control / "submissions" / provider / (task["game_id"] + "__" + task["task_id"] + ".json")
393
- model._client = JournalClient(settings["api_base"], key, journal, cap, settings["credits"], provider,
394
- resume_refused=True)
395
- return model
396
-
397
-
398
- def download_preview(model, output: Path):
399
- from PIL import Image
400
- url = model.last_call_info.get("preview_url")
401
- if not url:
402
- return {"status": "unavailable"}
403
- try:
404
- data = model.client.download(url)
405
- with Image.open(io.BytesIO(data)) as preview:
406
- preview.verify()
407
- extension = {"PNG": ".png", "JPEG": ".jpg", "WEBP": ".webp"}.get(preview.format)
408
- if not extension:
409
- return {"status": "unsupported_image_format"}
410
- path = output / ("preview" + extension)
411
- path.write_bytes(data)
412
- return {"status": "downloaded", "path": str(path), "visual_review": "PENDING"}
413
- except Exception as exc:
414
- return {"status": "failed", "error": safe_error(exc)}
415
-
416
-
417
- def run(plan: dict, args):
418
- from operators.gen_3d_object.operator import Gen3DObjectOperator
419
-
420
- class SharedStoreOperator(Gen3DObjectOperator):
421
- def _resolve_out_path(self, inp, task_id):
422
- output, _ = task_directories(plan, provider, inp)
423
- output.mkdir(parents=True, exist_ok=True)
424
- return inp["game_id"], output, output / "model.glb"
425
-
426
- if not args.allow_paid:
427
- raise ValueError("Paid generation requires --allow-paid and explicit provider credit caps")
428
- validate_inputs(plan)
429
- selected = list(dict.fromkeys(args.provider or list(plan["config"]["providers"])))
430
- caps = {p: getattr(args, p + "_credit_cap") for p in selected}
431
- for provider in selected:
432
- if not caps[provider] or caps[provider] < reserved_credits(Path(plan["directory"]) / "submissions" / provider):
433
- raise ValueError(f"Supply a positive --{provider}-credit-cap covering existing reservations")
434
- cloud_api.require_api_key(None, provider.upper() + "_API_KEY", "See scripts/asset/README_3d_comparison.md")
435
- selection = set(args.task_id or [])
436
- known = {item["task"]["task_id"] for item in plan["tasks"]}
437
- if selection - known:
438
- raise ValueError(f"Unknown task IDs: {sorted(selection - known)}")
439
- jobs = [item for item in plan["tasks"] if not selection or item["task"]["task_id"] in selection]
440
- if args.limit is not None:
441
- if args.limit < 1:
442
- raise ValueError("--limit must be positive")
443
- jobs = jobs[:args.limit]
444
- write_json(Path(plan["directory"]) / "spending_authorization.json", {
445
- "plan_sha256": plan["plan_sha256"], "credit_caps": caps, "recorded_at": now(),
446
- "selected_task_ids": [item["task"]["task_id"] for item in jobs],
447
- "note": "Caps reserve estimated credits; account-side limits remain authoritative"})
448
- interrupted = False
449
- for provider in selected:
450
- failures = 0
451
- for item in jobs:
452
- task = item["task"]
453
- output, evaluation = task_directories(plan, provider, task)
454
- previous = read_json(output / "meta.json", {})
455
- if previous and previous.get("plan_sha256") != plan["plan_sha256"]:
456
- raise ValueError(f"Another batch owns this provider directory; do not overwrite: {output}")
457
- if not previous and output.is_dir() and any(p.name != ".gitkeep" for p in output.iterdir()):
458
- raise ValueError(f"Untracked provider artifacts must not be overwritten: {output}")
459
- if previous.get("status") == "delivered" and (output / "model.glb").is_file():
460
- imported = output / "import/model.glb"
461
- if previous.get("artifact_sha256") == digest(output / "model.glb") and imported.is_file() and previous.get("import_sha256") == digest(imported):
462
- continue
463
- if previous.get("status") == "failed":
464
- continue # one generation attempt, no cherry-picking on rerun
465
- output.mkdir(parents=True, exist_ok=True)
466
- meta = {"game_id": task["game_id"], "task_id": task["task_id"], "asset_name": task["asset_name"],
467
- "task_kind": "3d_object", "run_id": plan["batch_id"], "provider": provider,
468
- "plan_sha256": plan["plan_sha256"], "context_condition": item["context_condition"],
469
- "full_context": plan["documents"][task["game_id"]], "source_task": task,
470
- "shared_prompt": item["shared_prompt"], "reference": item["reference"],
471
- "status": "running", "updated_at": now(), "visual_status": "PENDING"}
472
- write_json(output / "meta.json", meta)
473
- model = None
474
- stop = False
475
- try:
476
- model = make_model(plan, provider, task, caps[provider])
477
- # Preserve native output for a fair comparison. No automatic mesh cleanup/import.
478
- inp = {**task, "strip_ground_plate": False}
479
- inp.pop("asset_id", None)
480
- if item["reference"]:
481
- inp["image_path"] = item["reference"]["path"]
482
- else:
483
- inp["prompt"] = item["shared_prompt"]
484
- result = SharedStoreOperator(model, run_id=plan["batch_id"]).run(inp)
485
- raw_qa = basic_qa(Path(result["glb_path"]), task)
486
- meta.update(native_delivered=raw_qa["checks"]["parseable_glb"], result=result,
487
- native_structural_qa=raw_qa["status"], model_call=model.last_call_info,
488
- artifact_sha256=digest(output / "model.glb"))
489
- write_json(evaluation / "qa.json", {"native": raw_qa, "import": {"status": "PENDING"},
490
- "visual_status": "PENDING"})
491
- imported, postprocess = process_import_mesh(Path(result["glb_path"]), task)
492
- import_qa = basic_qa(imported, task)
493
- import_qa["texture_budget_verified"] = all(max(t["after"]) <= task["texture_size"] for t in postprocess["textures"])
494
- import_qa["checks"]["texture_budget"] = import_qa["texture_budget_verified"]
495
- import_qa["status"] = "PASS" if all(import_qa["checks"].values()) else "FAIL"
496
- write_json(evaluation / "qa.json", {"native": raw_qa, "import": import_qa,
497
- "postprocess": postprocess, "visual_status": "PENDING"})
498
- meta.update(status="delivered", result=result, model_call=model.last_call_info,
499
- import_delivered=import_qa["checks"]["parseable_glb"],
500
- artifact_sha256=digest(output / "model.glb"), import_sha256=digest(imported),
501
- native_structural_qa=raw_qa["status"], import_structural_qa=import_qa["status"],
502
- postprocess=postprocess,
503
- preview=download_preview(model, output))
504
- failures = 0
505
- except Exception as exc:
506
- status = "failed"
507
- if isinstance(exc, cloud_api.CloudTaskTimeout):
508
- status = "remote_pending"
509
- elif isinstance(exc, AmbiguousSubmission):
510
- status, stop = "submission_uncertain", True
511
- elif isinstance(exc, (BudgetExceeded, cloud_api.CloudAPIAuthError, cloud_api.CloudAPIQuotaError)):
512
- status, stop = "blocked", True
513
- elif isinstance(exc, cloud_api.CloudAPIServerError):
514
- status = "resume_required" # known receipts survive poll/download failures
515
- elif not isinstance(exc, (cloud_api.CloudTaskFailed, cloud_api.CloudAPIRequestError, ValueError)):
516
- status = "resume_required" # retain a downloaded asset after local/tool failures
517
- meta.update(status=status, error=safe_error(exc), error_type=type(exc).__name__)
518
- failures += 1
519
- stop = stop or failures >= 3
520
- finally:
521
- if model is not None:
522
- model.unload()
523
- meta["updated_at"] = now()
524
- journal = read_json(Path(plan["directory"]) / "submissions" / provider /
525
- (task["game_id"] + "__" + task["task_id"] + ".json"), {})
526
- meta["generation_cost"] = [{"endpoint": record["endpoint"], "remote_id": record.get("remote_id"),
527
- "reservation": record["reserved_credits"], "reported_credits": record.get("reported_credits"),
528
- "status": record["status"]} for record in journal.get("submissions", {}).values()]
529
- write_json(output / "meta.json", meta)
530
- refresh_reports(plan)
531
- print(json.dumps({"provider": provider, "task_id": task["task_id"], "status": meta["status"],
532
- "error": meta.get("error")}, ensure_ascii=False), flush=True)
533
- if stop:
534
- interrupted = True
535
- break
536
- if interrupted:
537
- break
538
- incomplete = any(read_json(task_directories(plan, p, item["task"])[0] / "meta.json", {}).get("status") != "delivered"
539
- for p in selected for item in jobs)
540
- return 2 if interrupted or incomplete else 0
541
-
542
-
543
- def recover(plan: dict, args):
544
- """Attach an inspected remote ID to an ambiguous receipt; never submit here."""
545
- checked_id(args.remote_task_id)
546
- matches = [item["task"] for item in plan["tasks"] if item["task"]["task_id"] == args.task_id]
547
- if len(matches) != 1:
548
- raise ValueError("Recover requires exactly one known task")
549
- task = matches[0]
550
- journal = Path(plan["directory"]) / "submissions" / args.provider / (task["game_id"] + "__" + task["task_id"] + ".json")
551
- state = read_json(journal, {})
552
- record = state.get("submissions", {}).get(args.request_hash)
553
- if not record or record["status"] not in ("submitting", "uncertain") or record["endpoint"] == "/files":
554
- raise ValueError("Receipt is not an ambiguous generation submission")
555
- record.update(status="created", remote_id=args.remote_task_id, reconciled_at=now(),
556
- response={"result": args.remote_task_id} if args.provider == "meshy" else
557
- {"code": 0, "data": {"task_id": args.remote_task_id}})
558
- write_json(journal, state)
559
- return {"recovered_task_id": args.remote_task_id, "network_requests": 0}
560
-
561
-
562
- def main(argv=None):
563
- parser = argparse.ArgumentParser(description=__doc__)
564
- parser.add_argument("--config", type=Path, default=DEFAULT_CONFIG)
565
- sub = parser.add_subparsers(dest="command", required=True)
566
- for name in ("prepare", "status", "run", "recover"):
567
- command = sub.add_parser(name)
568
- command.add_argument("--batch-id", required=True)
569
- if name == "run":
570
- command.add_argument("--allow-paid", action="store_true")
571
- command.add_argument("--meshy-credit-cap", type=int)
572
- command.add_argument("--tripo-credit-cap", type=int)
573
- command.add_argument("--provider", action="append", choices=("meshy", "tripo"))
574
- command.add_argument("--task-id", action="append")
575
- command.add_argument("--limit", type=int, help="Limit tasks per selected provider; full inventory is retained")
576
- elif name == "recover":
577
- command.add_argument("--provider", required=True, choices=("meshy", "tripo"))
578
- command.add_argument("--task-id", required=True)
579
- command.add_argument("--request-hash", required=True)
580
- command.add_argument("--remote-task-id", required=True)
581
- args = parser.parse_args(argv)
582
- config = read_json(args.config)
583
- with exclusive_batch(batch_directory(config, args.batch_id)):
584
- if args.command == "prepare":
585
- plan = prepare(config, args.batch_id)
586
- print(json.dumps(portable_value({"plan": str(Path(plan["directory"]) / "plan.json"),
587
- "tasks_per_provider": len(plan["tasks"]), "estimated_credits": plan["estimated_credits"],
588
- "recommended_credit_caps": plan["recommended_credit_caps"]}), indent=2))
589
- else:
590
- plan = load_plan(config, args.batch_id)
591
- if args.command == "run":
592
- return run(plan, args)
593
- value = recover(plan, args) if args.command == "recover" else refresh_reports(plan)
594
- print(json.dumps(portable_value(value), indent=2, ensure_ascii=False))
595
- return 0
596
-
597
-
598
- if __name__ == "__main__":
599
- try:
600
- raise SystemExit(main())
601
- except (ValueError, OSError, cloud_api.CloudAPIError) as exc:
602
- print(safe_error(exc), file=sys.stderr)
603
- raise SystemExit(2)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/package-lock.json DELETED
@@ -1,763 +0,0 @@
1
- {
2
- "name": "bench-asset-comparison-mesh-tools",
3
- "version": "1.0.0",
4
- "lockfileVersion": 3,
5
- "requires": true,
6
- "packages": {
7
- "": {
8
- "name": "bench-asset-comparison-mesh-tools",
9
- "version": "1.0.0",
10
- "dependencies": {
11
- "@gltf-transform/core": "4.2.1",
12
- "@gltf-transform/extensions": "4.2.1",
13
- "@gltf-transform/functions": "4.2.1",
14
- "draco3dgltf": "1.5.7",
15
- "meshoptimizer": "0.24.0",
16
- "sharp": "0.35.5"
17
- }
18
- },
19
- "node_modules/@emnapi/runtime": {
20
- "version": "1.11.3",
21
- "resolved": "https://registry.npmjs.org/@emnapi/runtime/-/runtime-1.11.3.tgz",
22
- "integrity": "sha512-Xz4Tpyki7XyrpbUK1jR1AhdAdaXyhhY4lZ3neLodmhpuWfy2PAQN5B46sAiU4liOXGLkHypn/qU+jvfWSCYYLA==",
23
- "license": "MIT",
24
- "optional": true,
25
- "dependencies": {
26
- "tslib": "^2.4.0"
27
- }
28
- },
29
- "node_modules/@gltf-transform/core": {
30
- "version": "4.2.1",
31
- "resolved": "https://registry.npmjs.org/@gltf-transform/core/-/core-4.2.1.tgz",
32
- "integrity": "sha512-qKhrQ29KJ9K6sz7xGoGRllqPBlROjPJMQPvCLBOfgMPH3S/Ph5E0e6fD5WmwTY5QPyBsDMRqKd1Vl0ncThTXGw==",
33
- "license": "MIT",
34
- "dependencies": {
35
- "property-graph": "^3.0.0"
36
- },
37
- "funding": {
38
- "url": "https://github.com/sponsors/donmccurdy"
39
- }
40
- },
41
- "node_modules/@gltf-transform/extensions": {
42
- "version": "4.2.1",
43
- "resolved": "https://registry.npmjs.org/@gltf-transform/extensions/-/extensions-4.2.1.tgz",
44
- "integrity": "sha512-ieHSJU9qvb3ucWDxgHFcff+C7fLtFPa6G0Wtc/ki7GsN6Rh7o9eoyqi7Ggj2LO4i3ynCwLLPb/onbJKw2fBtsA==",
45
- "license": "MIT",
46
- "dependencies": {
47
- "@gltf-transform/core": "^4.2.1",
48
- "ktx-parse": "^1.0.1"
49
- },
50
- "funding": {
51
- "url": "https://github.com/sponsors/donmccurdy"
52
- }
53
- },
54
- "node_modules/@gltf-transform/functions": {
55
- "version": "4.2.1",
56
- "resolved": "https://registry.npmjs.org/@gltf-transform/functions/-/functions-4.2.1.tgz",
57
- "integrity": "sha512-hFCI80N6zt7yHlqXbzlYEYQbFwK+2OOZNvacrRtrWiCWKI0k1Ad5+nZFsmuIZZV03aXb5XG6UF7JINT3wETS1Q==",
58
- "license": "MIT",
59
- "dependencies": {
60
- "@gltf-transform/core": "^4.2.1",
61
- "@gltf-transform/extensions": "^4.2.1",
62
- "ktx-parse": "^1.0.1",
63
- "ndarray": "^1.0.19",
64
- "ndarray-lanczos": "^0.3.0",
65
- "ndarray-pixels": "^5.0.1"
66
- },
67
- "funding": {
68
- "url": "https://github.com/sponsors/donmccurdy"
69
- }
70
- },
71
- "node_modules/@img/colour": {
72
- "version": "1.1.0",
73
- "resolved": "https://registry.npmjs.org/@img/colour/-/colour-1.1.0.tgz",
74
- "integrity": "sha512-Td76q7j57o/tLVdgS746cYARfSyxk8iEfRxewL9h4OMzYhbW4TAcppl0mT4eyqXddh6L/jwoM75mo7ixa/pCeQ==",
75
- "license": "MIT",
76
- "engines": {
77
- "node": ">=18"
78
- }
79
- },
80
- "node_modules/@img/sharp-darwin-arm64": {
81
- "version": "0.35.5",
82
- "resolved": "https://registry.npmjs.org/@img/sharp-darwin-arm64/-/sharp-darwin-arm64-0.35.5.tgz",
83
- "integrity": "sha512-QRUlFQ0WxvdWyqqG/WtI3iupfD5rBzmCHXSdPsY91sAtVtTo7Q4cb6zOccZ3gqEqkr0f1As1ehLqmEpDsRf+lg==",
84
- "cpu": [
85
- "arm64"
86
- ],
87
- "license": "Apache-2.0",
88
- "optional": true,
89
- "os": [
90
- "darwin"
91
- ],
92
- "engines": {
93
- "node": ">=20.9.0"
94
- },
95
- "funding": {
96
- "url": "https://opencollective.com/libvips"
97
- },
98
- "optionalDependencies": {
99
- "@img/sharp-libvips-darwin-arm64": "1.3.4"
100
- }
101
- },
102
- "node_modules/@img/sharp-darwin-x64": {
103
- "version": "0.35.5",
104
- "resolved": "https://registry.npmjs.org/@img/sharp-darwin-x64/-/sharp-darwin-x64-0.35.5.tgz",
105
- "integrity": "sha512-+BR255RhDlpygUpOc/Jdt1nT6DQ3XG/ERo5wbcdOf5Q320dKtPCKPLR1LJs9VGXRaMa8l1uUa0tkCNOXiAxZUw==",
106
- "cpu": [
107
- "x64"
108
- ],
109
- "license": "Apache-2.0",
110
- "optional": true,
111
- "os": [
112
- "darwin"
113
- ],
114
- "engines": {
115
- "node": ">=20.9.0"
116
- },
117
- "funding": {
118
- "url": "https://opencollective.com/libvips"
119
- },
120
- "optionalDependencies": {
121
- "@img/sharp-libvips-darwin-x64": "1.3.4"
122
- }
123
- },
124
- "node_modules/@img/sharp-freebsd-wasm32": {
125
- "version": "0.35.5",
126
- "resolved": "https://registry.npmjs.org/@img/sharp-freebsd-wasm32/-/sharp-freebsd-wasm32-0.35.5.tgz",
127
- "integrity": "sha512-Y/z91nEZ4uIBX5X3nfTovjU9lHNKFYbL2lpHCLVNmXQK03VIZvXBBt0KxbPGp2SdGSF+2mQU4e+hQaWOt86iAw==",
128
- "license": "Apache-2.0",
129
- "optional": true,
130
- "os": [
131
- "freebsd"
132
- ],
133
- "dependencies": {
134
- "@img/sharp-wasm32": "0.35.5"
135
- },
136
- "engines": {
137
- "node": ">=20.9.0"
138
- },
139
- "funding": {
140
- "url": "https://opencollective.com/libvips"
141
- }
142
- },
143
- "node_modules/@img/sharp-freebsd-wasm32/node_modules/@img/sharp-wasm32": {
144
- "version": "0.35.5",
145
- "resolved": "https://registry.npmjs.org/@img/sharp-wasm32/-/sharp-wasm32-0.35.5.tgz",
146
- "integrity": "sha512-Ptsga1su4tQx+LLF1ECS9U6nz5kmrXKo6XVbtR48Ke3ZRxxgaWBu7IDtEe1quo8hiupwm6WFqxVlXaSf7IINGQ==",
147
- "license": "Apache-2.0 AND LGPL-3.0-or-later AND MIT",
148
- "optional": true,
149
- "dependencies": {
150
- "@emnapi/runtime": "^1.11.3"
151
- },
152
- "engines": {
153
- "node": ">=20.9.0"
154
- },
155
- "funding": {
156
- "url": "https://opencollective.com/libvips"
157
- }
158
- },
159
- "node_modules/@img/sharp-libvips-darwin-arm64": {
160
- "version": "1.3.4",
161
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-darwin-arm64/-/sharp-libvips-darwin-arm64-1.3.4.tgz",
162
- "integrity": "sha512-5R89nBYiRdUlSWJxPhO+GVtaXzXSxKnRu/xqMn3KTA3L9EB9Oy/P+Nn2f2vlhPuUdy/Zusb2DarbyTpGCfEDuw==",
163
- "cpu": [
164
- "arm64"
165
- ],
166
- "license": "LGPL-3.0-or-later",
167
- "optional": true,
168
- "os": [
169
- "darwin"
170
- ],
171
- "funding": {
172
- "url": "https://opencollective.com/libvips"
173
- }
174
- },
175
- "node_modules/@img/sharp-libvips-darwin-x64": {
176
- "version": "1.3.4",
177
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-darwin-x64/-/sharp-libvips-darwin-x64-1.3.4.tgz",
178
- "integrity": "sha512-iR2OKH80yi0U+dUplyh3/xdpFvps6YkCwsXenIJxqxR1v9o+xtKTGbS9H7cps+2Vxjc8B1j96p75NmTGjIhtpQ==",
179
- "cpu": [
180
- "x64"
181
- ],
182
- "license": "LGPL-3.0-or-later",
183
- "optional": true,
184
- "os": [
185
- "darwin"
186
- ],
187
- "funding": {
188
- "url": "https://opencollective.com/libvips"
189
- }
190
- },
191
- "node_modules/@img/sharp-libvips-linux-arm": {
192
- "version": "1.3.4",
193
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linux-arm/-/sharp-libvips-linux-arm-1.3.4.tgz",
194
- "integrity": "sha512-LmRtTsOHuvM2+wlO2Db37dx5MiZhB0FvSunciw48YjdOkZz9KAiRbm8ujeMOA1INqmei5NapFxYEK1D1ZSidmw==",
195
- "cpu": [
196
- "arm"
197
- ],
198
- "license": "LGPL-3.0-or-later",
199
- "optional": true,
200
- "os": [
201
- "linux"
202
- ],
203
- "funding": {
204
- "url": "https://opencollective.com/libvips"
205
- }
206
- },
207
- "node_modules/@img/sharp-libvips-linux-arm64": {
208
- "version": "1.3.4",
209
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linux-arm64/-/sharp-libvips-linux-arm64-1.3.4.tgz",
210
- "integrity": "sha512-Y3dgX/6lE2QhQb+Gxy0WZxfg9MEm/JBjamZpS2IklP7xIQoKN4hzAm7KcMVGtaVDt3neE9OKBC7vAfonA/Lr1A==",
211
- "cpu": [
212
- "arm64"
213
- ],
214
- "license": "LGPL-3.0-or-later",
215
- "optional": true,
216
- "os": [
217
- "linux"
218
- ],
219
- "funding": {
220
- "url": "https://opencollective.com/libvips"
221
- }
222
- },
223
- "node_modules/@img/sharp-libvips-linux-ppc64": {
224
- "version": "1.3.4",
225
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linux-ppc64/-/sharp-libvips-linux-ppc64-1.3.4.tgz",
226
- "integrity": "sha512-Le6boB8Tai0Nis+gIxIpKx68UDVVIqdR8Tin5Yf1z2LJJQLDJvCDRqRu+jC2qCoD+eIomonmOwB4smBRxfVpYQ==",
227
- "cpu": [
228
- "ppc64"
229
- ],
230
- "license": "LGPL-3.0-or-later",
231
- "optional": true,
232
- "os": [
233
- "linux"
234
- ],
235
- "funding": {
236
- "url": "https://opencollective.com/libvips"
237
- }
238
- },
239
- "node_modules/@img/sharp-libvips-linux-riscv64": {
240
- "version": "1.3.4",
241
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linux-riscv64/-/sharp-libvips-linux-riscv64-1.3.4.tgz",
242
- "integrity": "sha512-aHkkIEHPRdQEegJN20MLmGtxYD9R2wQr3Cwpddnu5+YKMt6Uzax7S9h5gpZTo8wyrGuZSlfQ63OevL5mTyOC7Q==",
243
- "cpu": [
244
- "riscv64"
245
- ],
246
- "license": "LGPL-3.0-or-later",
247
- "optional": true,
248
- "os": [
249
- "linux"
250
- ],
251
- "funding": {
252
- "url": "https://opencollective.com/libvips"
253
- }
254
- },
255
- "node_modules/@img/sharp-libvips-linux-s390x": {
256
- "version": "1.3.4",
257
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linux-s390x/-/sharp-libvips-linux-s390x-1.3.4.tgz",
258
- "integrity": "sha512-ra/mB6MikESDUO7Yg+Mi95bFBb9GsObURuhnOv3OqknjGe9sZrG8tCe9q0xSIGrtLgvgw0gKnFWcK4blSgQOuQ==",
259
- "cpu": [
260
- "s390x"
261
- ],
262
- "license": "LGPL-3.0-or-later",
263
- "optional": true,
264
- "os": [
265
- "linux"
266
- ],
267
- "funding": {
268
- "url": "https://opencollective.com/libvips"
269
- }
270
- },
271
- "node_modules/@img/sharp-libvips-linux-x64": {
272
- "version": "1.3.4",
273
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linux-x64/-/sharp-libvips-linux-x64-1.3.4.tgz",
274
- "integrity": "sha512-GJ//SSXbnwSDes02umB3nDJLFcQzw8a18V8fyhqr6tV515tOEMdImjjxj1AoafMRz56F3PHgftnj1QEKSU1zkw==",
275
- "cpu": [
276
- "x64"
277
- ],
278
- "license": "LGPL-3.0-or-later",
279
- "optional": true,
280
- "os": [
281
- "linux"
282
- ],
283
- "funding": {
284
- "url": "https://opencollective.com/libvips"
285
- }
286
- },
287
- "node_modules/@img/sharp-libvips-linuxmusl-arm64": {
288
- "version": "1.3.4",
289
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linuxmusl-arm64/-/sharp-libvips-linuxmusl-arm64-1.3.4.tgz",
290
- "integrity": "sha512-hvulFwtjUcagsis6BBxHwGFwWoNZjgYmULGVrZcyfNbjA8hKILbRxGg15/7w5HDyXHXUos/j6baAWqnCyQ2DWA==",
291
- "cpu": [
292
- "arm64"
293
- ],
294
- "license": "LGPL-3.0-or-later",
295
- "optional": true,
296
- "os": [
297
- "linux"
298
- ],
299
- "funding": {
300
- "url": "https://opencollective.com/libvips"
301
- }
302
- },
303
- "node_modules/@img/sharp-libvips-linuxmusl-x64": {
304
- "version": "1.3.4",
305
- "resolved": "https://registry.npmjs.org/@img/sharp-libvips-linuxmusl-x64/-/sharp-libvips-linuxmusl-x64-1.3.4.tgz",
306
- "integrity": "sha512-6zXKeE/p39I1AmA3cJG35eyBGNqNddLnUXjhwBnsGjFPWqf5VKkDBEqaEkPDoTEtkxwi2vv8Tcr2mDyP4So7Fg==",
307
- "cpu": [
308
- "x64"
309
- ],
310
- "license": "LGPL-3.0-or-later",
311
- "optional": true,
312
- "os": [
313
- "linux"
314
- ],
315
- "funding": {
316
- "url": "https://opencollective.com/libvips"
317
- }
318
- },
319
- "node_modules/@img/sharp-linux-arm": {
320
- "version": "0.35.5",
321
- "resolved": "https://registry.npmjs.org/@img/sharp-linux-arm/-/sharp-linux-arm-0.35.5.tgz",
322
- "integrity": "sha512-LEaXK2WdXVK5ykcw0buWyPMsmLLL2vpHLD6yrNSW+JGEL3BZPA4tpKN6iaMc4AxTTAoaX/sU1rOL51lcIz48ZQ==",
323
- "cpu": [
324
- "arm"
325
- ],
326
- "license": "Apache-2.0",
327
- "optional": true,
328
- "os": [
329
- "linux"
330
- ],
331
- "engines": {
332
- "node": ">=20.9.0"
333
- },
334
- "funding": {
335
- "url": "https://opencollective.com/libvips"
336
- },
337
- "optionalDependencies": {
338
- "@img/sharp-libvips-linux-arm": "1.3.4"
339
- }
340
- },
341
- "node_modules/@img/sharp-linux-arm64": {
342
- "version": "0.35.5",
343
- "resolved": "https://registry.npmjs.org/@img/sharp-linux-arm64/-/sharp-linux-arm64-0.35.5.tgz",
344
- "integrity": "sha512-LYVx5JTsOM2CBzmxreh+nl64/3H6Xb09iSLknqH47z2T2DFFxDeFLP5y4dJwe6H7uGQlHPyEEtIqyo3DYsRwdQ==",
345
- "cpu": [
346
- "arm64"
347
- ],
348
- "license": "Apache-2.0",
349
- "optional": true,
350
- "os": [
351
- "linux"
352
- ],
353
- "engines": {
354
- "node": ">=20.9.0"
355
- },
356
- "funding": {
357
- "url": "https://opencollective.com/libvips"
358
- },
359
- "optionalDependencies": {
360
- "@img/sharp-libvips-linux-arm64": "1.3.4"
361
- }
362
- },
363
- "node_modules/@img/sharp-linux-ppc64": {
364
- "version": "0.35.5",
365
- "resolved": "https://registry.npmjs.org/@img/sharp-linux-ppc64/-/sharp-linux-ppc64-0.35.5.tgz",
366
- "integrity": "sha512-QVxAAq8evVRI9ia2vqgwrmWucn5Dfv+JdWzj75pD8omHLPSP7f8p20O8jxzjCcuCEQEOtYOZUmX1hkiZ0kdevA==",
367
- "cpu": [
368
- "ppc64"
369
- ],
370
- "license": "Apache-2.0",
371
- "optional": true,
372
- "os": [
373
- "linux"
374
- ],
375
- "engines": {
376
- "node": ">=20.9.0"
377
- },
378
- "funding": {
379
- "url": "https://opencollective.com/libvips"
380
- },
381
- "optionalDependencies": {
382
- "@img/sharp-libvips-linux-ppc64": "1.3.4"
383
- }
384
- },
385
- "node_modules/@img/sharp-linux-riscv64": {
386
- "version": "0.35.5",
387
- "resolved": "https://registry.npmjs.org/@img/sharp-linux-riscv64/-/sharp-linux-riscv64-0.35.5.tgz",
388
- "integrity": "sha512-LtdreXguaavKODPIfzJ4kffx7UNt1omwtK0rch4EBbbSTXPnxWmYSayXdLJw0fJzQ97kHt1gL/yh4tvU+nCyRQ==",
389
- "cpu": [
390
- "riscv64"
391
- ],
392
- "license": "Apache-2.0",
393
- "optional": true,
394
- "os": [
395
- "linux"
396
- ],
397
- "engines": {
398
- "node": ">=20.9.0"
399
- },
400
- "funding": {
401
- "url": "https://opencollective.com/libvips"
402
- },
403
- "optionalDependencies": {
404
- "@img/sharp-libvips-linux-riscv64": "1.3.4"
405
- }
406
- },
407
- "node_modules/@img/sharp-linux-s390x": {
408
- "version": "0.35.5",
409
- "resolved": "https://registry.npmjs.org/@img/sharp-linux-s390x/-/sharp-linux-s390x-0.35.5.tgz",
410
- "integrity": "sha512-UZasTOFiYzotTsGOCu42BfUzP6Tu6Do/947iRm1RsLKvlllxwGcn4RN27LibGWceix4Y+Pmw3jsnTcCQIgWjqA==",
411
- "cpu": [
412
- "s390x"
413
- ],
414
- "license": "Apache-2.0",
415
- "optional": true,
416
- "os": [
417
- "linux"
418
- ],
419
- "engines": {
420
- "node": ">=20.9.0"
421
- },
422
- "funding": {
423
- "url": "https://opencollective.com/libvips"
424
- },
425
- "optionalDependencies": {
426
- "@img/sharp-libvips-linux-s390x": "1.3.4"
427
- }
428
- },
429
- "node_modules/@img/sharp-linux-x64": {
430
- "version": "0.35.5",
431
- "resolved": "https://registry.npmjs.org/@img/sharp-linux-x64/-/sharp-linux-x64-0.35.5.tgz",
432
- "integrity": "sha512-SxFtLTeJInhAA9Q836kux2vZNeOBQEx658qvbboZScr0wIARym3IcGmW7KpVD5sbVg0Ojy+udFQdayYIZyoNog==",
433
- "cpu": [
434
- "x64"
435
- ],
436
- "license": "Apache-2.0",
437
- "optional": true,
438
- "os": [
439
- "linux"
440
- ],
441
- "engines": {
442
- "node": ">=20.9.0"
443
- },
444
- "funding": {
445
- "url": "https://opencollective.com/libvips"
446
- },
447
- "optionalDependencies": {
448
- "@img/sharp-libvips-linux-x64": "1.3.4"
449
- }
450
- },
451
- "node_modules/@img/sharp-linuxmusl-arm64": {
452
- "version": "0.35.5",
453
- "resolved": "https://registry.npmjs.org/@img/sharp-linuxmusl-arm64/-/sharp-linuxmusl-arm64-0.35.5.tgz",
454
- "integrity": "sha512-9HbMclmI1zlNkFRs3z9/eBtDjfD0sGlrX1z6b1qwmiFY5ElDLh4BC0LPBdVp7z1DXFiKlIcznf+ZlsuZzLxQqg==",
455
- "cpu": [
456
- "arm64"
457
- ],
458
- "license": "Apache-2.0",
459
- "optional": true,
460
- "os": [
461
- "linux"
462
- ],
463
- "engines": {
464
- "node": ">=20.9.0"
465
- },
466
- "funding": {
467
- "url": "https://opencollective.com/libvips"
468
- },
469
- "optionalDependencies": {
470
- "@img/sharp-libvips-linuxmusl-arm64": "1.3.4"
471
- }
472
- },
473
- "node_modules/@img/sharp-linuxmusl-x64": {
474
- "version": "0.35.5",
475
- "resolved": "https://registry.npmjs.org/@img/sharp-linuxmusl-x64/-/sharp-linuxmusl-x64-0.35.5.tgz",
476
- "integrity": "sha512-4KOphqB035HrVdqLZfCgMzzERrQkkzOwRhl4OAkRO1YCldbaFjySXMaK534Mo0V+LndnlJk+sbUyLeU0ULyD1A==",
477
- "cpu": [
478
- "x64"
479
- ],
480
- "license": "Apache-2.0",
481
- "optional": true,
482
- "os": [
483
- "linux"
484
- ],
485
- "engines": {
486
- "node": ">=20.9.0"
487
- },
488
- "funding": {
489
- "url": "https://opencollective.com/libvips"
490
- },
491
- "optionalDependencies": {
492
- "@img/sharp-libvips-linuxmusl-x64": "1.3.4"
493
- }
494
- },
495
- "node_modules/@img/sharp-webcontainers-wasm32": {
496
- "version": "0.35.5",
497
- "resolved": "https://registry.npmjs.org/@img/sharp-webcontainers-wasm32/-/sharp-webcontainers-wasm32-0.35.5.tgz",
498
- "integrity": "sha512-hfhF/FmoQyTUkA0bIKFOtw536BQSeBMe6BF6QyWlrPxT754+TFLaZ7sKKTfvvM0yJgKgaYTwnFCIZ/GuDw5SUA==",
499
- "cpu": [
500
- "wasm32"
501
- ],
502
- "license": "Apache-2.0",
503
- "optional": true,
504
- "dependencies": {
505
- "@img/sharp-wasm32": "0.35.5"
506
- },
507
- "engines": {
508
- "node": ">=20.9.0"
509
- },
510
- "funding": {
511
- "url": "https://opencollective.com/libvips"
512
- }
513
- },
514
- "node_modules/@img/sharp-webcontainers-wasm32/node_modules/@img/sharp-wasm32": {
515
- "version": "0.35.5",
516
- "resolved": "https://registry.npmjs.org/@img/sharp-wasm32/-/sharp-wasm32-0.35.5.tgz",
517
- "integrity": "sha512-Ptsga1su4tQx+LLF1ECS9U6nz5kmrXKo6XVbtR48Ke3ZRxxgaWBu7IDtEe1quo8hiupwm6WFqxVlXaSf7IINGQ==",
518
- "license": "Apache-2.0 AND LGPL-3.0-or-later AND MIT",
519
- "optional": true,
520
- "dependencies": {
521
- "@emnapi/runtime": "^1.11.3"
522
- },
523
- "engines": {
524
- "node": ">=20.9.0"
525
- },
526
- "funding": {
527
- "url": "https://opencollective.com/libvips"
528
- }
529
- },
530
- "node_modules/@img/sharp-win32-arm64": {
531
- "version": "0.35.5",
532
- "resolved": "https://registry.npmjs.org/@img/sharp-win32-arm64/-/sharp-win32-arm64-0.35.5.tgz",
533
- "integrity": "sha512-X4t7g+7ZA5DKblCBEXGjUqqemj4vczING/5viFwAL8h4N3qYeyjwdCvRLHi4EdOUI+2Z7UFlp1VM+p/AuEtm6Q==",
534
- "cpu": [
535
- "arm64"
536
- ],
537
- "license": "Apache-2.0 AND LGPL-3.0-or-later",
538
- "optional": true,
539
- "os": [
540
- "win32"
541
- ],
542
- "engines": {
543
- "node": ">=20.9.0"
544
- },
545
- "funding": {
546
- "url": "https://opencollective.com/libvips"
547
- }
548
- },
549
- "node_modules/@img/sharp-win32-ia32": {
550
- "version": "0.35.5",
551
- "resolved": "https://registry.npmjs.org/@img/sharp-win32-ia32/-/sharp-win32-ia32-0.35.5.tgz",
552
- "integrity": "sha512-5Zm82LoBc43nhwNybZlG7Y1KO//Zhsn306fQl29ZOuStHLGTo3BWL83q3cznX0poxSAMuYL1On/BHBxkBeKr6A==",
553
- "cpu": [
554
- "ia32"
555
- ],
556
- "license": "Apache-2.0 AND LGPL-3.0-or-later",
557
- "optional": true,
558
- "os": [
559
- "win32"
560
- ],
561
- "engines": {
562
- "node": "^20.9.0"
563
- },
564
- "funding": {
565
- "url": "https://opencollective.com/libvips"
566
- }
567
- },
568
- "node_modules/@img/sharp-win32-x64": {
569
- "version": "0.35.5",
570
- "resolved": "https://registry.npmjs.org/@img/sharp-win32-x64/-/sharp-win32-x64-0.35.5.tgz",
571
- "integrity": "sha512-x76eH0vEiHlcMQu8Y8IenntaACtddpT6W0wmXtWrnKcnKI7ME5DdgqhAD6SEWOEl1v2zDvkZDhFA9KnURwpfqg==",
572
- "cpu": [
573
- "x64"
574
- ],
575
- "license": "Apache-2.0 AND LGPL-3.0-or-later",
576
- "optional": true,
577
- "os": [
578
- "win32"
579
- ],
580
- "engines": {
581
- "node": ">=20.9.0"
582
- },
583
- "funding": {
584
- "url": "https://opencollective.com/libvips"
585
- }
586
- },
587
- "node_modules/@types/ndarray": {
588
- "version": "1.1.0",
589
- "resolved": "https://registry.npmjs.org/@types/ndarray/-/ndarray-1.1.0.tgz",
590
- "integrity": "sha512-go+C6vHBUuO9jH6Tj6qEZ63JEGoMSxFdmd+6QAtthGD6xumBfIz3RbNDDs2aoRVLf9naH8lk5jnWy2UHHKgYQQ==",
591
- "license": "MIT"
592
- },
593
- "node_modules/cwise-compiler": {
594
- "version": "1.1.3",
595
- "resolved": "https://registry.npmjs.org/cwise-compiler/-/cwise-compiler-1.1.3.tgz",
596
- "integrity": "sha512-WXlK/m+Di8DMMcCjcWr4i+XzcQra9eCdXIJrgh4TUgh0pIS/yJduLxS9JgefsHJ/YVLdgPtXm9r62W92MvanEQ==",
597
- "license": "MIT",
598
- "dependencies": {
599
- "uniq": "^1.0.0"
600
- }
601
- },
602
- "node_modules/detect-libc": {
603
- "version": "2.1.2",
604
- "resolved": "https://registry.npmjs.org/detect-libc/-/detect-libc-2.1.2.tgz",
605
- "integrity": "sha512-Btj2BOOO83o3WyH59e8MgXsxEQVcarkUOpEYrubB0urwnN10yQ364rsiByU11nZlqWYZm05i/of7io4mzihBtQ==",
606
- "license": "Apache-2.0",
607
- "engines": {
608
- "node": ">=8"
609
- }
610
- },
611
- "node_modules/draco3dgltf": {
612
- "version": "1.5.7",
613
- "resolved": "https://registry.npmjs.org/draco3dgltf/-/draco3dgltf-1.5.7.tgz",
614
- "integrity": "sha512-LeqcpmoHIyYUi0z70/H3tMkGj8QhqVxq6FJGPjlzR24BNkQ6jyMheMvFKJBI0dzGZrEOUyQEmZ8axM1xRrbRiw==",
615
- "license": "Apache-2.0"
616
- },
617
- "node_modules/iota-array": {
618
- "version": "1.0.0",
619
- "resolved": "https://registry.npmjs.org/iota-array/-/iota-array-1.0.0.tgz",
620
- "integrity": "sha512-pZ2xT+LOHckCatGQ3DcG/a+QuEqvoxqkiL7tvE8nn3uuu+f6i1TtpB5/FtWFbxUuVr5PZCx8KskuGatbJDXOWA==",
621
- "license": "MIT"
622
- },
623
- "node_modules/is-buffer": {
624
- "version": "1.1.6",
625
- "resolved": "https://registry.npmjs.org/is-buffer/-/is-buffer-1.1.6.tgz",
626
- "integrity": "sha512-NcdALwpXkTm5Zvvbk7owOUSvVvBKDgKP5/ewfXEznmQFfs4ZRmanOeKBTjRVjka3QFoN6XJ+9F3USqfHqTaU5w==",
627
- "license": "MIT"
628
- },
629
- "node_modules/ktx-parse": {
630
- "version": "1.1.0",
631
- "resolved": "https://registry.npmjs.org/ktx-parse/-/ktx-parse-1.1.0.tgz",
632
- "integrity": "sha512-mKp3y+FaYgR7mXWAbyyzpa/r1zDWeaunH+INJO4fou3hb45XuNSwar+7llrRyvpMWafxSIi99RNFJ05MHedaJQ==",
633
- "license": "MIT"
634
- },
635
- "node_modules/meshoptimizer": {
636
- "version": "0.24.0",
637
- "resolved": "https://registry.npmjs.org/meshoptimizer/-/meshoptimizer-0.24.0.tgz",
638
- "integrity": "sha512-Iq/8cM5cK2B0P0QdKHexr30zEJn5sQxjXKtQJGp6LO7r1VUPwhgxau18pvybVDnEDsKE49GdlJYwqPX9A/P9Sw==",
639
- "license": "MIT"
640
- },
641
- "node_modules/ndarray": {
642
- "version": "1.1.1",
643
- "resolved": "https://registry.npmjs.org/ndarray/-/ndarray-1.1.1.tgz",
644
- "integrity": "sha512-AsE8w/M2MS+0pD3nP5qbRRtJFeJJJaJ+qXZaZr0XjVGXhy6LIj2TERTlkpdK2AI3WufoOP+pwPk3q2ilAJcZ9g==",
645
- "license": "MIT",
646
- "dependencies": {
647
- "iota-array": "^1.0.0",
648
- "is-buffer": "^1.0.2"
649
- }
650
- },
651
- "node_modules/ndarray-lanczos": {
652
- "version": "0.3.0",
653
- "resolved": "https://registry.npmjs.org/ndarray-lanczos/-/ndarray-lanczos-0.3.0.tgz",
654
- "integrity": "sha512-5kBmmG3Zvyj77qxIAC4QFLKuYdDIBJwCG+DukT6jQHNa1Ft74/hPH1z5mbQXeHBt8yvGPBGVrr3wEOdJPYYZYg==",
655
- "license": "MIT",
656
- "dependencies": {
657
- "@types/ndarray": "^1.0.11",
658
- "ndarray": "^1.0.19"
659
- }
660
- },
661
- "node_modules/ndarray-ops": {
662
- "version": "1.2.2",
663
- "resolved": "https://registry.npmjs.org/ndarray-ops/-/ndarray-ops-1.2.2.tgz",
664
- "integrity": "sha512-BppWAFRjMYF7N/r6Ie51q6D4fs0iiGmeXIACKY66fLpnwIui3Wc3CXiD/30mgLbDjPpSLrsqcp3Z62+IcHZsDw==",
665
- "license": "MIT",
666
- "dependencies": {
667
- "cwise-compiler": "^1.0.0"
668
- }
669
- },
670
- "node_modules/ndarray-pixels": {
671
- "version": "5.2.0",
672
- "resolved": "https://registry.npmjs.org/ndarray-pixels/-/ndarray-pixels-5.2.0.tgz",
673
- "integrity": "sha512-lTh4tFKziAatVTa9crIsidUyn+lqujVOQpzfdBWvdFu2wo9Uo6z261lVX7SgMyP89xGmj3TMTPbbxl9YDnV4SA==",
674
- "license": "MIT",
675
- "dependencies": {
676
- "@types/ndarray": "^1.0.14",
677
- "ndarray": "^1.0.19",
678
- "ndarray-ops": "^1.2.2",
679
- "sharp": "^0.35.0"
680
- }
681
- },
682
- "node_modules/property-graph": {
683
- "version": "3.0.0",
684
- "resolved": "https://registry.npmjs.org/property-graph/-/property-graph-3.0.0.tgz",
685
- "integrity": "sha512-TnzxUsttmGtw+OiU0LDw+0FlMbJ8vV8pOjyDI7+Kdni4Tj0hW5BFh7TatQu7Y68hcvvFmiFOHilKShsA4R82fA==",
686
- "license": "MIT"
687
- },
688
- "node_modules/semver": {
689
- "version": "7.8.5",
690
- "resolved": "https://registry.npmjs.org/semver/-/semver-7.8.5.tgz",
691
- "integrity": "sha512-Y7/KDsb8LjooZpwaqGyulO6DQlksgCncchHGk+sZIY4SBvUocMBEFH5Ur1fI4dV+Jvl0w6cjvucaIi40puRioA==",
692
- "license": "ISC",
693
- "bin": {
694
- "semver": "bin/semver.js"
695
- },
696
- "engines": {
697
- "node": ">=10"
698
- }
699
- },
700
- "node_modules/sharp": {
701
- "version": "0.35.5",
702
- "resolved": "https://registry.npmjs.org/sharp/-/sharp-0.35.5.tgz",
703
- "integrity": "sha512-Ywn4OnzGukp7CDMrp08RQ50YKmuwG47brZgIVPTvBaaAfQlRlygrRqSrxdCiL9M+LlzLBiJ68IR1QqvzHyjC7g==",
704
- "license": "Apache-2.0",
705
- "dependencies": {
706
- "@img/colour": "^1.1.0",
707
- "detect-libc": "^2.1.2",
708
- "semver": "^7.8.5"
709
- },
710
- "engines": {
711
- "node": ">=20.9.0"
712
- },
713
- "funding": {
714
- "url": "https://opencollective.com/libvips"
715
- },
716
- "optionalDependencies": {
717
- "@img/sharp-darwin-arm64": "0.35.5",
718
- "@img/sharp-darwin-x64": "0.35.5",
719
- "@img/sharp-freebsd-wasm32": "0.35.5",
720
- "@img/sharp-libvips-darwin-arm64": "1.3.4",
721
- "@img/sharp-libvips-darwin-x64": "1.3.4",
722
- "@img/sharp-libvips-linux-arm": "1.3.4",
723
- "@img/sharp-libvips-linux-arm64": "1.3.4",
724
- "@img/sharp-libvips-linux-ppc64": "1.3.4",
725
- "@img/sharp-libvips-linux-riscv64": "1.3.4",
726
- "@img/sharp-libvips-linux-s390x": "1.3.4",
727
- "@img/sharp-libvips-linux-x64": "1.3.4",
728
- "@img/sharp-libvips-linuxmusl-arm64": "1.3.4",
729
- "@img/sharp-libvips-linuxmusl-x64": "1.3.4",
730
- "@img/sharp-linux-arm": "0.35.5",
731
- "@img/sharp-linux-arm64": "0.35.5",
732
- "@img/sharp-linux-ppc64": "0.35.5",
733
- "@img/sharp-linux-riscv64": "0.35.5",
734
- "@img/sharp-linux-s390x": "0.35.5",
735
- "@img/sharp-linux-x64": "0.35.5",
736
- "@img/sharp-linuxmusl-arm64": "0.35.5",
737
- "@img/sharp-linuxmusl-x64": "0.35.5",
738
- "@img/sharp-webcontainers-wasm32": "0.35.5",
739
- "@img/sharp-win32-arm64": "0.35.5",
740
- "@img/sharp-win32-ia32": "0.35.5",
741
- "@img/sharp-win32-x64": "0.35.5"
742
- },
743
- "peerDependenciesMeta": {
744
- "@types/node": {
745
- "optional": true
746
- }
747
- }
748
- },
749
- "node_modules/tslib": {
750
- "version": "2.8.1",
751
- "resolved": "https://registry.npmjs.org/tslib/-/tslib-2.8.1.tgz",
752
- "integrity": "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==",
753
- "license": "0BSD",
754
- "optional": true
755
- },
756
- "node_modules/uniq": {
757
- "version": "1.0.1",
758
- "resolved": "https://registry.npmjs.org/uniq/-/uniq-1.0.1.tgz",
759
- "integrity": "sha512-Gw+zz50YNKPDKXs+9d+aKAjVwpjNwqzvNpLigIruT4HA9lMZNdMqs9x07kKHB/L9WRzqp4+DlTU5s4wG2esdoA==",
760
- "license": "MIT"
761
- }
762
- }
763
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/process_mesh.mjs DELETED
@@ -1,77 +0,0 @@
1
- // One fixed pipeline for both services; keep the original GLB separately.
2
- import {readFileSync, mkdirSync} from 'node:fs';
3
- import {dirname} from 'node:path';
4
- import {performance} from 'node:perf_hooks';
5
- import {NodeIO, Logger} from '@gltf-transform/core';
6
- import {ALL_EXTENSIONS} from '@gltf-transform/extensions';
7
- import {simplify, weld} from '@gltf-transform/functions';
8
- import {MeshoptDecoder, MeshoptSimplifier} from 'meshoptimizer';
9
- import draco from 'draco3dgltf';
10
- import sharp from 'sharp';
11
-
12
- const settings = JSON.parse(readFileSync(new URL('./package.json', import.meta.url)));
13
- await Promise.all([MeshoptDecoder.ready, MeshoptSimplifier.ready]);
14
- const runtime = {node: process.versions.node, dependencies: settings.dependencies};
15
- if (process.argv[2] === '--check') {
16
- console.log(JSON.stringify({status: 'ready', ...runtime}));
17
- } else {
18
- const [source, destination, triangleArg, textureArg] = process.argv.slice(2);
19
- const triangleTarget = Number(triangleArg), textureTarget = Number(textureArg);
20
- if (!source || !destination || !Number.isInteger(triangleTarget) || triangleTarget <= 0
21
- || !Number.isInteger(textureTarget) || textureTarget <= 0) {
22
- throw new Error('Expected: source.glb destination.glb triangle_target texture_size');
23
- }
24
- const started = performance.now();
25
- const io = new NodeIO().registerExtensions(ALL_EXTENSIONS).registerDependencies({
26
- 'draco3d.decoder': await draco.createDecoderModule(),
27
- 'meshopt.decoder': MeshoptDecoder,
28
- });
29
- const document = await io.read(source);
30
- document.setLogger(new Logger(Logger.Verbosity.SILENT));
31
- // The decoded import GLB is not dependent on the provider's mesh codec.
32
- for (const extension of document.getRoot().listExtensionsUsed()) {
33
- if (['KHR_draco_mesh_compression', 'EXT_meshopt_compression'].includes(extension.extensionName)) {
34
- extension.dispose();
35
- }
36
- }
37
- function triangleCount() {
38
- let count = 0;
39
- for (const mesh of document.getRoot().listMeshes()) {
40
- for (const primitive of mesh.listPrimitives()) {
41
- if (primitive.getMode() !== 4) throw new Error('Only triangle primitives are supported');
42
- const n = (primitive.getIndices() || primitive.getAttribute('POSITION')).getCount();
43
- count += n / 3;
44
- }
45
- }
46
- return count;
47
- }
48
- const before = triangleCount();
49
- if (before > triangleTarget) {
50
- await document.transform(weld(), simplify({simplifier: MeshoptSimplifier,
51
- ratio: triangleTarget / before, error: 0.01, lockBorder: false}));
52
- }
53
- const textures = [];
54
- for (const texture of document.getRoot().listTextures()) {
55
- const image = texture.getImage();
56
- if (!image) throw new Error('Texture has no decoded image');
57
- const input = Buffer.from(image);
58
- const info = await sharp(input).metadata();
59
- let sizeAfter = [info.width, info.height];
60
- if (Math.max(...sizeAfter) > textureTarget) {
61
- const {data, info: resized} = await sharp(input).resize({width: textureTarget, height: textureTarget,
62
- fit: 'inside', withoutEnlargement: true, kernel: 'lanczos3'}).png({compressionLevel: 9}).toBuffer({resolveWithObject: true});
63
- texture.setImage(new Uint8Array(data)).setMimeType('image/png').setURI('');
64
- sizeAfter = [resized.width, resized.height];
65
- }
66
- textures.push({name: texture.getName(), before: [info.width, info.height], after: sizeAfter});
67
- }
68
- const after = triangleCount();
69
- mkdirSync(dirname(destination), {recursive: true});
70
- await io.write(destination, document);
71
- console.log(JSON.stringify({status: after <= triangleTarget ? 'PASS' : 'FAIL',
72
- triangles_before: before, triangles_after: after, triangle_target: triangleTarget,
73
- texture_target: textureTarget, textures, runtime,
74
- policy: 'weld; one simplify pass at error 0.01 when over budget; downsize textures only; preserve UV/material/node data',
75
- elapsed_sec: Math.round((performance.now() - started) / 10) / 100,
76
- visual_review: 'PENDING'}));
77
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/sources_index.json DELETED
@@ -1,178 +0,0 @@
1
- {
2
- "files": [
3
- {
4
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/general_requirement.txt",
5
- "sha256": "285a181c4ed0275c4c96a9cd71c6e22abfd90212018ef6c284f55315037589a3",
6
- "bytes": 59694
7
- },
8
- {
9
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/object_requirement.txt",
10
- "sha256": "8b8e455ca2bb791b3718602c902186b576bfd8a7e172262e9e4ff3ba25c51173",
11
- "bytes": 947
12
- },
13
- {
14
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
15
- "sha256": "ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55",
16
- "bytes": 686650
17
- },
18
- {
19
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
20
- "sha256": "b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d",
21
- "bytes": 4872
22
- },
23
- {
24
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/inputs/method_selection.md",
25
- "sha256": "d9b0471f4c0e0f1e70c76e2bba676d3275d10530451435dc625972dbcea9abbe",
26
- "bytes": 11846
27
- },
28
- {
29
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/generate_3d_comparison.py",
30
- "sha256": "70f8d79c8596908f7edcd7ff68120f558cea6881f23dc2f0e29ef62f2bab75c4",
31
- "bytes": 33749
32
- },
33
- {
34
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/cloud_journal.py",
35
- "sha256": "92d075ef9308e0d6496a901d85b7847b8f602f88de8e19caeaf85ac893081253",
36
- "bytes": 8458
37
- },
38
- {
39
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/record_paths.py",
40
- "sha256": "93c42e454d60d03c30b12e58e8e75453c7caa8fc92ff30bbc20aa645b3fc23dd",
41
- "bytes": 3390
42
- },
43
- {
44
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/pipeline/common/paths.py",
45
- "sha256": "ed1cf1105f95c76b4893e98a84b3e37ef7d1f87c873c93c8e58413457104d28b",
46
- "bytes": 17054
47
- },
48
- {
49
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/prompts/stage2_assets.md",
50
- "sha256": "172339b9f6c6d834ef0d782c0d3dd779b62a12bd34acddc746b80d7cd99f63c2",
51
- "bytes": 6384
52
- },
53
- {
54
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/operators/gen_3d_object/operator.py",
55
- "sha256": "5f88c457ee58f73de5ddbd4d96fa91e474c29bd3ecefbb02f3faf18b93c4365a",
56
- "bytes": 27332
57
- },
58
- {
59
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/meshy_model.py",
60
- "sha256": "ef437ff79554469390c0d6c71f11788431fe4a1d1e51e92feae3adf296cebaec",
61
- "bytes": 20927
62
- },
63
- {
64
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/gen_3d_object/tripo_model.py",
65
- "sha256": "29ab3944260d62585a991f384eee4917368f97c4476abc3eda85299bc1a8089e",
66
- "bytes": 24255
67
- },
68
- {
69
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/cloud_api.py",
70
- "sha256": "a3c363a3439b3864e05e782340eb3059380c4b1c28eccb2524e9cbf9131fe996",
71
- "bytes": 21838
72
- },
73
- {
74
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/GameFactory-3A/models/common/glb_utils.py",
75
- "sha256": "6f3ee0147b345415cb6d8b1b1433fa557d0d543bb92276840dc85b43cd7a66ad",
76
- "bytes": 23765
77
- },
78
- {
79
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/process_mesh.mjs",
80
- "sha256": "b1b6ada273a576083bb167772ec03fee6fe3275470180563722e56687688995c",
81
- "bytes": 3939
82
- },
83
- {
84
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/runtime/bench/scripts/asset/mesh_tools/package-lock.json",
85
- "sha256": "0d9c3ad5898a78a0db22f361caa7242b48ff86586e898f5ae5c1bb4a7134631a",
86
- "bytes": 26253
87
- },
88
- {
89
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/provenance/frozen_parent_plan_original.zip",
90
- "sha256": "a72e49a29033d161d0ec0abd27dc003f34c8b7a190201c7e8b514e0afc4bce45",
91
- "bytes": 60760,
92
- "kind": "immutable_frozen_parent_plan_archive"
93
- },
94
- {
95
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_001.json",
96
- "sha256": "5894be33a2c6f03e1f5f03d8cca375cc7ca0ac2b487d4550091d5a5116a9398a",
97
- "bytes": 8260
98
- },
99
- {
100
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/c4784de99ad02419f6e080c08b3388e1e8700ed26b6d9e46e7dbb72cafa7311f.meta.json",
101
- "sha256": "6b3e4607c12b5fa01d292a14763ac5c8535cdd91b25fbe63843c7514aea9100d",
102
- "bytes": 1097324
103
- },
104
- {
105
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_002.json",
106
- "sha256": "e916997a4a508619798a3185f42e5e4728f2ba2b4fb33b3d981b13f54a029002",
107
- "bytes": 8260
108
- },
109
- {
110
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/440b4733d80d11bc6c02e030f10f3e14a40728326f88b899eb77291c6944107d.meta.json",
111
- "sha256": "12e1f67885fe0de9561c160b9f170393131ab70ef0fe51f2bf9b3b9db71d1dc6",
112
- "bytes": 10631
113
- },
114
- {
115
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_003.json",
116
- "sha256": "fd1b9e78e2184db4c1cd116d4ecd028b5c79b03c43f81ff289abb4752796a827",
117
- "bytes": 13007
118
- },
119
- {
120
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/2fe8790006de37b67087226226b42b450fd22be3f3c30f7b8d0ddbaf50246d62.meta.json",
121
- "sha256": "733a4dfff0e27e876acfefc2ba669366abb2bba58b1acc21e22c1698246cc103",
122
- "bytes": 1927
123
- },
124
- {
125
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/meshy/G055_inner_court_investigation__G055_inner_court_investigation_obj_004.json",
126
- "sha256": "8d5798763592494ea43a8cdb3444e0d8434405077a714df072c70a385c060a22",
127
- "bytes": 13023
128
- },
129
- {
130
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/meshy/8fc448662209e1e2fdf9c5ebb0e8e8a012c7533282406ec50f103a9b5b99f7b6.meta.json",
131
- "sha256": "24fe21a314f083f1f9f082301fa50af947d971ae5f1a5be2a80b26b2d797877c",
132
- "bytes": 1936
133
- },
134
- {
135
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_001.json",
136
- "sha256": "b5cf0b889ac83023fc8a59341f90e1d9330551e4917942988599bdd1aaf6ad5f",
137
- "bytes": 2551
138
- },
139
- {
140
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/ecb7e508f9dc367e4d5c4b91e9c7b37b68dbb47a2d9173ce1b5c30f01813d8a9.meta.json",
141
- "sha256": "aeec1d4543021901196c813bf83f17909af22ab14b7420afbe00840d0ed6fbf7",
142
- "bytes": 973
143
- },
144
- {
145
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_002.json",
146
- "sha256": "a1e9ac277155c21a86abac106f1eafdee5ee2451d93777fa203748c8c375d19e",
147
- "bytes": 2551
148
- },
149
- {
150
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/072537beaa31697bd206f8b13b746d96ac409e06dffd39e613fddaa8e2f755f0.meta.json",
151
- "sha256": "cfc119bdd9f2aa784c3304a8a9b377edecfe89073abd41782c30283e54d3bc0f",
152
- "bytes": 973
153
- },
154
- {
155
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_003.json",
156
- "sha256": "d545acfd66a3673d8a3fa23de2754648bf81cd01ac459c532d93657e182bfeb6",
157
- "bytes": 2790
158
- },
159
- {
160
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/c0f6d2616b8bb8de62f37adee4acaa3039af81bbe0a507b57e9e150aaa700dad.meta.json",
161
- "sha256": "f327271326166ed8e341a5a5279684f4d7d209d74c58c4e929704c8974e6d1d7",
162
- "bytes": 1536
163
- },
164
- {
165
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/receipts/tripo/G055_inner_court_investigation__G055_inner_court_investigation_obj_004.json",
166
- "sha256": "8af678f1954427efb2de7f1c6fa5ff009dfeb8dc69b9632f25fa4ce0ebac8d4c",
167
- "bytes": 2798
168
- },
169
- {
170
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/cache_metadata/tripo/2ce49f964f8b611e965e5f39d37edc4b84a43acb9238f0caaabd9215699fbabf.meta.json",
171
- "sha256": "c250939ec0997d42af67675d4151f320e9b67c315c5dc710622e030f875f4933",
172
- "bytes": 1544
173
- }
174
- ],
175
- "receipt_bytes_preserved": true,
176
- "frozen_plan_unchanged": true,
177
- "parent_plan_sha256": "5c287b8399f00fbb0dddf94c9e3f5ffcbcc2e679ff76882bc8fc369d6eae840d"
178
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/.generation/3d_object/tripo/obj200_meshy_tripo_20261015_v2/asset_task_index.json DELETED
@@ -1,111 +0,0 @@
1
- {
2
- "plan_sha256": "5c287b8399f00fbb0dddf94c9e3f5ffcbcc2e679ff76882bc8fc369d6eae840d",
3
- "game_id": "G055_inner_court_investigation",
4
- "provider": "tripo",
5
- "scope": "3d_object_only_not_complete_stage2",
6
- "tasks": [
7
- {
8
- "task": {
9
- "game_id": "G055_inner_court_investigation",
10
- "task_id": "G055_inner_court_investigation_obj_001",
11
- "asset_name": "shen_zhiwei",
12
- "name_zh": "沈知微",
13
- "role": "player",
14
- "input_mode": "image_to_3d",
15
- "prompt": "Original young court investigator Shen Zhiwei: understated dark-teal commoner robe with faint embroidery, hair in a plain bun with wooden pin, cloth shoes, sharp calm gaze, stylized semi-realistic ancient court style, vermilion-and-indigo palette, T-pose, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
16
- "seed": 5501,
17
- "decimation_target": 8000,
18
- "texture_size": 2048,
19
- "rig": "humanoid",
20
- "animation_strategy": "skeletal",
21
- "spec_ref": "P01,W06",
22
- "image_path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
23
- "poly_policy": "low_poly_priority",
24
- "l_tier_repair_target": true
25
- },
26
- "shared_prompt": null,
27
- "reference": {
28
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png",
29
- "sha256": "ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55",
30
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/ca09a6abbf7f0b51f14f52e78d6165f579a7423d7de6f82eac8699975a640e55.png"
31
- },
32
- "subject_group": "character",
33
- "context_condition": "reference_only"
34
- },
35
- {
36
- "task": {
37
- "game_id": "G055_inner_court_investigation",
38
- "task_id": "G055_inner_court_investigation_obj_002",
39
- "asset_name": "evidence_pouch_ash",
40
- "name_zh": "旧香囊残灰",
41
- "role": "evidence",
42
- "input_mode": "image_to_3d",
43
- "prompt": "Small scorched silk incense pouch with charred rim and faint residual ash, tied with a cut thread, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
44
- "seed": 1552,
45
- "decimation_target": 4000,
46
- "texture_size": 1024,
47
- "rig": "none",
48
- "animation_strategy": "procedural",
49
- "spec_ref": "P20.E01",
50
- "image_path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
51
- "reference_image_role": "single_object_visual_anchor",
52
- "poly_policy": "low_poly_priority",
53
- "l_tier_repair_target": true
54
- },
55
- "shared_prompt": null,
56
- "reference": {
57
- "path": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png",
58
- "sha256": "b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d",
59
- "source": "bench/test_data/outputs/G055_inner_court_investigation/.generation/3d_object/obj200_meshy_tripo_20261015_v2/delivery/references/b5a13889899eea16de82e49b034a09b90d0475c86c32d50da1ee3416319f134d.png"
60
- },
61
- "subject_group": "rigid",
62
- "context_condition": "reference_only"
63
- },
64
- {
65
- "task": {
66
- "game_id": "G055_inner_court_investigation",
67
- "task_id": "G055_inner_court_investigation_obj_003",
68
- "asset_name": "rota_manuscript",
69
- "name_zh": "排班抄本",
70
- "role": "evidence",
71
- "input_mode": "text_to_3d",
72
- "prompt": "Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
73
- "seed": 1553,
74
- "decimation_target": 5000,
75
- "texture_size": 1024,
76
- "rig": "none",
77
- "animation_strategy": "procedural",
78
- "spec_ref": "P20.E03",
79
- "poly_policy": "low_poly_priority",
80
- "l_tier_repair_target": true
81
- },
82
- "shared_prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、��令、\nAsset: Hand-bound duty-rota manuscript copy with thread binding holes, brush annotations and half-page tear, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
83
- "reference": null,
84
- "subject_group": "rigid",
85
- "context_condition": "shared_short_context"
86
- },
87
- {
88
- "task": {
89
- "game_id": "G055_inner_court_investigation",
90
- "task_id": "G055_inner_court_investigation_obj_004",
91
- "asset_name": "seal_wax_imprint",
92
- "name_zh": "火漆拓印",
93
- "role": "evidence",
94
- "input_mode": "text_to_3d",
95
- "prompt": "Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
96
- "seed": 1554,
97
- "decimation_target": 4000,
98
- "texture_size": 1024,
99
- "rig": "none",
100
- "animation_strategy": "procedural",
101
- "spec_ref": "P20.E06",
102
- "poly_policy": "low_poly_priority",
103
- "l_tier_repair_target": true
104
- },
105
- "shared_prompt": "Game context (excerpt): 宫墙影语 | 1. 抵达区域:查看区域平面、门禁和当日作息摘要。 2. 观察环境:标记守卫、闲谈、可攀爬点、遮蔽物、器具与异常物。 3. 提出假设:在笔记本写下“谁、何时、凭什么、可能路线”。 4. 选定身份与路线:制服、差事、口令、\nAsset: Wax seal imprint fragment on cracked paper with a clear directional fissure pattern and vermilion stamp edge, ancient court stylized, game-ready, PBR textures, low-poly style, minimal triangle count, flat faceted surfaces, clean silhouette, no fine surface detail",
106
- "reference": null,
107
- "subject_group": "rigid",
108
- "context_condition": "shared_short_context"
109
- }
110
- ]
111
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
outputs/G055_inner_court_investigation/assets/3d_object/.gitattributes ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ G055_inner_court_investigation_obj_001/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
2
+ G055_inner_court_investigation_obj_001/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
3
+ G055_inner_court_investigation_obj_001/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
4
+ G055_inner_court_investigation_obj_001/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
5
+ G055_inner_court_investigation_obj_002/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
6
+ G055_inner_court_investigation_obj_002/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
7
+ G055_inner_court_investigation_obj_002/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
8
+ G055_inner_court_investigation_obj_002/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
9
+ G055_inner_court_investigation_obj_003/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
10
+ G055_inner_court_investigation_obj_003/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
11
+ G055_inner_court_investigation_obj_003/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
12
+ G055_inner_court_investigation_obj_003/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
13
+ G055_inner_court_investigation_obj_004/meshy/import/model.glb filter=lfs diff=lfs merge=lfs -text
14
+ G055_inner_court_investigation_obj_004/meshy/model.glb filter=lfs diff=lfs merge=lfs -text
15
+ G055_inner_court_investigation_obj_004/tripo/import/model.glb filter=lfs diff=lfs merge=lfs -text
16
+ G055_inner_court_investigation_obj_004/tripo/model.glb filter=lfs diff=lfs merge=lfs -text
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/contact_sheet.jpg DELETED

Git LFS Details

  • SHA256: baca169a34981c1b88b6fd4d8a6ac6d6651ae33ddf06a397376f3626705157ce
  • Pointer size: 130 Bytes
  • Size of remote file: 63.4 kB
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/iso.png DELETED

Git LFS Details

  • SHA256: 5dc7cb3902500760e442c39a96ca2cbcae3abef2556696f018539112f67ff586
  • Pointer size: 130 Bytes
  • Size of remote file: 57.1 kB
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/neg_x.png DELETED

Git LFS Details

  • SHA256: bb9bb501e9112977e8b1c7ffecadf36bdc83e43046558ac2907b8f36e35ea05a
  • Pointer size: 130 Bytes
  • Size of remote file: 38.4 kB
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/neg_z.png DELETED

Git LFS Details

  • SHA256: c150fe924698b3847d0c3e7f28762352544fb99b6e79fd76fba64e8d3773c226
  • Pointer size: 130 Bytes
  • Size of remote file: 52 kB
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/pos_x.png DELETED

Git LFS Details

  • SHA256: d544168f0ebed51f010c6b58ed280d9518df446afa872281628f1f88a4d09a4b
  • Pointer size: 130 Bytes
  • Size of remote file: 39.6 kB
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/pos_y.png DELETED

Git LFS Details

  • SHA256: 6a4c624851a4a7cde2b69ce3208668c665b3cc4845d2158965202b7e3e25a90f
  • Pointer size: 130 Bytes
  • Size of remote file: 29 kB
outputs/G055_inner_court_investigation/eval/3d_object/G055_inner_court_investigation_obj_001/meshy/previews/import/pos_z.png DELETED

Git LFS Details

  • SHA256: 009b52494d7e188cd62771373407cd3f571fcc61ad8d658a2297697388be18a3
  • Pointer size: 130 Bytes
  • Size of remote file: 64.3 kB