Improve README with curated examples and usage guidance

README.md

@@ -14,26 +14,119 @@ tags:
 - scene-graph
 ---
 
-# ChartGalaxy++
+<h1 align="center">ChartGalaxy++</h1>
 
-Scene graph annotations for **217,195 infographic charts**, containing **18,897,483 nodes** and **39,989,150 relationships**.
+<p align="center"><strong>A Richly Annotated Dataset for Chart Understanding and Generation</strong></p>
 
-[Project and application data](https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus) · [Image2JSON model](https://huggingface.co/ChartGalaxyPP/ChartGalaxyPlusPlus-Image2JSON) · [Data format](DATA_FORMAT.md) · [Loading guide](USAGE.md)
+<p align="center"><a href="https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus">Project</a> &nbsp; · &nbsp; <a href="https://huggingface.co/ChartGalaxyPP/ChartGalaxyPlusPlus-Image2JSON">Image2JSON model</a> &nbsp; · &nbsp; <a href="https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus/releases/tag/v1.0">Application data</a> &nbsp; · &nbsp; <a href="USAGE.md">Loading guide</a></p>
 
-| Chart source | Train | Test | Download contents |
-| --- | ---: | ---: | --- |
-| Real | 56,244 | 500 | URLs and annotation JSON; no image binaries |
-| Synthetic | 159,951 | 500 | PNG and annotation JSON |
-| Total | 216,195 | 1,000 | |
+**ChartGalaxy++ connects what is in an infographic chart with how it is organized.** Each chart is paired with a scene graph: visual elements and semantic groups form the nodes, while hierarchical and spatial relationships connect them. These annotations support chart structure prediction, visual question answering, and evaluation of image generation.
 
-The 926 downloadable `tar.gz` shards contain **one JSON file per chart** and a same-stem PNG for each synthetic chart. `sample_index.jsonl.gz` maps sample IDs to shard and member names. The original split membership and IDs are retained. This repository contains the main dataset only; the separate QA package includes its original evaluated images.
+| Infographic charts | Annotated nodes | Relationships |
+| :---: | :---: | :---: |
+| **217,195** | **18.90 million** | **39.99 million** |
 
-There are **15,591,656 visual elements**, **3,088,632 groups**, and **217,195 implicit roots**. Relationships comprise **18,680,288 parent links** and **21,308,862 stored spatial records**. Each chart stores at most 100 spatial records. Proximity is Boolean `is_near`; overlapping boxes are not near, and both the absolute and relative distance thresholds must hold.
+## Explore the dataset
 
-Layout illustrations have received generated replacement assets. Their boxes, appearance descriptions, affected group attributes, and stored spatial relations were updated with the released PNGs. Icons are retained. Historical benchmark packages identify their evaluated inputs by hash; changed images must not be substituted for those inputs.
+![Eight synthetic infographic charts with varied chart structures, editorial layouts, typography, and illustrations](assets/gallery.png)
 
-Annotations are model-produced (`human_gold: false`). Structural and content-integrity checks are not a claim that every annotation was verified by a human. Available real-image URLs and archive references are supplied without checking their current availability.
+*Selected synthetic examples from the released PNGs, shown in full. The gallery spans compact comparisons, layered circular charts, dense radial marks, and illustrated narratives. [View at full resolution](assets/gallery.png) · [Sample identities](assets/sources.json)*
 
-Contributed annotations use **CC BY-NC 4.0**. Third-party chart content and required upstream license attributions retain their respective rights. See [license and sources](LICENSE_AND_SOURCES.md). Annotation, training, evaluation, and review pipeline code is outside this release.
+## What is annotated?
 
-![Synthetic example](examples/01-layout/image.png)
+![An infographic chart and an excerpt of its actual scene graph, linking a sector label, icon, and two years of bar marks](assets/annotation.png)
+
+*One data item connects the FINANCE label and icon to both years of marks. The excerpt preserves the released node IDs and parent links; other branches are omitted. [Enlarge](assets/annotation.png)*
+
+| Layer | Annotation content |
+| --- | --- |
+| **Visual elements** | Text, images, and shapes with bounding boxes, semantic roles, text content, and appearance attributes |
+| **Semantic groups** | Charts, axes, legends, legend items, data items, series, and panels |
+| **Hierarchy** | Parent–child links connecting elements to groups and the chart composition |
+| **Spatial relationships** | Relative position, alignment, overlap, and Boolean proximity (`is_near`) |
+
+See the [annotation guide](ANNOTATION_GUIDE.md), [JSON schema](scene_graph.schema.json), and [data format](DATA_FORMAT.md) for all fields and coordinate conventions.
+
+## Dataset at a glance
+
+| Split | Real charts | Synthetic charts | Total |
+| --- | ---: | ---: | ---: |
+| Train | 56,244 | 159,951 | 216,195 |
+| Test | 500 | 500 | 1,000 |
+| **Total** | **56,744** | **160,451** | **217,195** |
+
+**Real charts: URLs + annotation JSON. Synthetic charts: PNG + annotation JSON.** The main dataset is distributed in 926 independently extractable `tar.gz` shards. `sample_index.jsonl.gz` maps each sample ID to its shard and member files. The separate QA package includes its original images.
+
+
+## Get started
+
+Install the download client with `pip install huggingface_hub`. Start with one standalone example before downloading shards:
+
+```python
+import json
+from huggingface_hub import HfApi, hf_hub_download
+
+repo = "ChartGalaxyPP/ChartGalaxyPlusPlus"
+revision = HfApi().dataset_info(repo).sha
+def download(name):
+    return hf_hub_download(repo, name, repo_type="dataset", revision=revision)
+
+image_path = download("examples/01-layout/image.png")
+graph_path = download("examples/01-layout/scene_graph.json")
+spatial_path = download("examples/01-layout/spatial_relations.json")
+with open(graph_path, encoding="utf-8") as f:
+    graph = json.load(f)
+nodes = graph["compositional_deconstruction"]["nodes"]
+print(f"{len(nodes)} explicit nodes", image_path)
+```
+
+The example is an existing training record, not an additional sample. To load any chart, use the [sample index](sample_index.jsonl.gz) and the [shard-loading example](USAGE.md). Bounding boxes in dataset annotations use **`[y0, x0, y1, x1]`, normalized to 0–1000**.
+
+<details>
+<summary>Download the complete dataset</summary>
+
+```python
+from huggingface_hub import snapshot_download
+
+snapshot_download(
+    repo_id=repo,
+    repo_type="dataset",
+    revision=revision,
+    local_dir="chartgalaxy-plus-plus",
+)
+```
+
+The data shards total approximately 105 GB. The release includes shard hashes in [manifest.json](manifest.json). See [USAGE.md](USAGE.md) for extraction and integrity checks.
+
+</details>
+
+## What can you do with it?
+
+| Application | What the scene graph enables | Released resource |
+| --- | --- | --- |
+| **Image-to-scene-graph prediction** | Recover elements, semantic groups, and chart hierarchy | [Image2JSON model](https://huggingface.co/ChartGalaxyPP/ChartGalaxyPlusPlus-Image2JSON); [1,000-chart benchmark](https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus/tree/main/applications/image_to_scene_graph) |
+| **Infographic question answering** | Associate text and marks through explicit groups and relationships | [1,266 questions, reference answers, and original images](https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus/tree/main/applications/qa) |
+| **Scene graph preservation** | Evaluate structural fidelity in generated infographic charts | [11-model benchmark and generated outputs](https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus/tree/main/applications/scene_graph_preservation) |
+
+See the [project page](https://github.com/ChartGalaxyPP/ChartGalaxyPlusPlus#three-applications) for qualitative results and links to all three application packages. Under the paper's evaluation protocol, the released Image2JSON model achieves **89.4% node F1, 85.1% hierarchy F1, and 88.0% spatial F1**.
+
+## Annotation and release notes
+
+<details>
+<summary>Exact annotation counts and release notes</summary>
+
+- **Nodes:** 15,591,656 visual elements + 3,088,632 explicit groups + 217,195 implicit roots = 18,897,483.
+- **Relationships:** 18,680,288 parent links + 21,308,862 stored spatial records = 39,989,150. Each chart stores at most 100 spatial records; unrecorded pairs are not negative labels.
+- `is_near` is Boolean. Positive-area overlaps and containment are excluded; both absolute and relative distance thresholds must hold. The format guide specifies the rule.
+- Annotations are model-produced (`human_gold: false`); the release does not claim that every chart has been independently verified by a human.
+- Generated replacement illustrations and their affected annotations are included in the main dataset. Historical benchmark packages retain the identities of their evaluated inputs; use their supplied references when inspecting reported results.
+- Real-image references are supplied without checking their current availability. Source pages, direct image URLs, and archive references are distinguished in the format guide.
+
+</details>
+
+
+This release contains images or image references, scene graph annotations, and spatial records. Underlying data-table files and annotation/training/evaluation pipeline code are not included. The gallery is a curated visual preview, not a random sample or annotation-quality evaluation.
+
+## License
+
+Contributed annotations are licensed under **CC BY-NC 4.0**. Third-party chart content and required upstream attributions retain their respective rights; see [LICENSE](LICENSE) and [license and sources](LICENSE_AND_SOURCES.md).

SHA256SUMS

@@ -3,8 +3,11 @@
 6de7772ea60a4a144f5a1aa7ea734700f69bc4b97227c225682fc4b21c4d402e  DATA_FORMAT.md
 cc97a2c4ff7723a1f25179dc7c18f676fa84191bbb36ea7d0061acaf3c102559  LICENSE
 81c88405668ceb8b2aba512be68db631e75a509cdfcd0267fa6e143680a97e50  LICENSE_AND_SOURCES.md
-53cbf27200afbe3b4d3b58320de52f4ad6d987f5ad3039627a4aceea7203c4f4  README.md
+efd3c9ce1c51e4c8db5398cffbfe414bcdb4e2820ddd455263d4da51c0f976b6  README.md
 c8c494822b72a8b9790b7b941e430e6977b2ad2f5534634251c0297f7d9fab9b  USAGE.md
+43b4cf4d63e13b94c2be7e42810c6c98be74954dd54c600ed5d5b1d9a2c6447b  assets/annotation.png
+3f67c1b2d4b551dadde581d92a04725cdde151e17767e6f5678f0b7ed8f0b0f6  assets/gallery.png
+04fea58165003bf611cbb0f5ce508ab823258abfca9b07d95a986ad2c9d3d3bc  assets/sources.json
 d0852894bbda4c3e7dff2a4e863af4ce683c2ea9430a85b09e1793a69fda02c1  data/real/test/part-00000.tar.gz
 ec6f21c29bdab7a69f11fcc9238f427c73d0d86cfc5fbb661042dddc1d35ed4b  data/real/train/part-00116.tar.gz
 f5970041c9dc754b6e220613e293783bb72d51e666d3deed857a83d68264ba99  data/real/train/part-00117.tar.gz

assets/sources.json

@@ -0,0 +1,192 @@
+{
+  "gallery": [
+    {
+      "position": 1,
+      "caption": "Dumbbell / range",
+      "index": 67,
+      "sample_id": "nb_bb2220eb07d52c0f4167b07d",
+      "benchmark_id": "nb2_16739d6064307d33b328",
+      "source_type": "synthetic",
+      "track": "nb2",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00002.tar.gz",
+      "annotation_file": "16739d6064307d33b328e5ea8b6b749c.json",
+      "image_file": "16739d6064307d33b328e5ea8b6b749c.png",
+      "image_sha256": "db9ca2a3385b7ea950edf1df4288ee9facf500482a439047f3b91deb095105db",
+      "annotation_sha256": "e1073435aab0148e5e73a8f197f637b7037b12bea286b18d223338b916b4dea9",
+      "scene_graph_sha256": "098bd137c4b1bdf8cb8a6f5076b354348f435f332adecb00ad0ed7350ab07b46"
+    },
+    {
+      "position": 2,
+      "caption": "Radial bars",
+      "index": 35,
+      "sample_id": "d3-js_circular_grouped_bar_chart_circular_grouped_bar_plain_chart_01/sample_34497",
+      "benchmark_id": "layout_89a243a4e8328ebbf21a",
+      "source_type": "synthetic",
+      "track": "layout",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00001.tar.gz",
+      "annotation_file": "89a243a4e8328ebbf21af9214ffb735b.json",
+      "image_file": "89a243a4e8328ebbf21af9214ffb735b.png",
+      "image_sha256": "6c8a6259bba8db954af0f415e96373d585bf16a2597e078e166793dd4a7a4657",
+      "annotation_sha256": "f45e902b57243e50a4b113ce71956d564faf424f006fc8c6c6f47620640f4ed5",
+      "scene_graph_sha256": "32bfb81da3d4c60201a2b01037b59150b5c699a97af6975955bd650185a767ba"
+    },
+    {
+      "position": 3,
+      "caption": "Interval marks",
+      "index": 103,
+      "sample_id": "nb_e885643f9025fb53201bacef",
+      "benchmark_id": "nb2_a7c1c15db8c85a280655",
+      "source_type": "synthetic",
+      "track": "nb2",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00004.tar.gz",
+      "annotation_file": "a7c1c15db8c85a280655da6ea8149de9.json",
+      "image_file": "a7c1c15db8c85a280655da6ea8149de9.png",
+      "image_sha256": "b1b1b2905b8cfb36f24cc06286cddb9a9433bded431d7e9bdaaa7a62ae5f43ed",
+      "annotation_sha256": "90b3ee3ca19a0f80cd231c7b2d3a58219afa959a9f0d0d4f053f227b742903b7",
+      "scene_graph_sha256": "126a85aef31a831a25c568274aad4c70929bc5d8effeea44dc0d79069e104928"
+    },
+    {
+      "position": 4,
+      "caption": "Circular layers",
+      "index": 106,
+      "sample_id": "nb_e3eefe82f2f150b05faa34aa",
+      "benchmark_id": "nb2_b1b73c72d9cf543dd871",
+      "source_type": "synthetic",
+      "track": "nb2",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00004.tar.gz",
+      "annotation_file": "b1b73c72d9cf543dd8719327d64628aa.json",
+      "image_file": "b1b73c72d9cf543dd8719327d64628aa.png",
+      "image_sha256": "c0aade03cf54487d165ebb58191288029bcc95d8f8b6df5a807f485e66028fd8",
+      "annotation_sha256": "ceccbbea2e7aba73f334f08e42796ee3f8929d74710e493adc9c0422741ab43e",
+      "scene_graph_sha256": "990e41bb59342b306c23d10a76669b819e40973f9d775526120370abcc9b5915"
+    },
+    {
+      "position": 5,
+      "caption": "Diverging bars",
+      "index": 20,
+      "sample_id": "d3-js_diverging_bar_chart_diverging_bar_plain_chart_03/sample_5364",
+      "benchmark_id": "layout_400ffffeb13d992a35ac",
+      "source_type": "synthetic",
+      "track": "layout",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00000.tar.gz",
+      "annotation_file": "400ffffeb13d992a35ac0090ccdd3f4d.json",
+      "image_file": "400ffffeb13d992a35ac0090ccdd3f4d.png",
+      "image_sha256": "cb2901ae06da12e8a171ada83a298f592fe0c9f58a27fc4836730919a726e6a4",
+      "annotation_sha256": "6e8b6ae8a8d58102538467bcfefc245f7d0697b1bc917b1a4ab34db85a571022",
+      "scene_graph_sha256": "ba0529dc354b736ce0fe44b8e545d09aa3b8c438fffbd6c7b44b330ab026e911"
+    },
+    {
+      "position": 6,
+      "caption": "Stacked bars",
+      "index": 124,
+      "sample_id": "nb_cce8b07b058c5ee16ba1835d",
+      "benchmark_id": "nb2_fa0cc382ebd8c5dd2d6c",
+      "source_type": "synthetic",
+      "track": "nb2",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00004.tar.gz",
+      "annotation_file": "fa0cc382ebd8c5dd2d6cf8e8dd1372cc.json",
+      "image_file": "fa0cc382ebd8c5dd2d6cf8e8dd1372cc.png",
+      "image_sha256": "7257aab8c526e73d982a9e8d6382d2919cdb52565073cc5e1b25bafd4cccda37",
+      "annotation_sha256": "ca80feb8d3c9cce797fe057a8fe5e395fb3697ac82708cfec027317fc44a74c2",
+      "scene_graph_sha256": "3f1828d1ca4fe0ab44dc3d2e918d41cae58d5beadb69164be9809eed6013be4c"
+    },
+    {
+      "position": 7,
+      "caption": "Time series",
+      "index": 71,
+      "sample_id": "nb_4c7fa3d3284baebd7748d089",
+      "benchmark_id": "nb2_332b024d1902d001e278",
+      "source_type": "synthetic",
+      "track": "nb2",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00002.tar.gz",
+      "annotation_file": "332b024d1902d001e278687fc5e990c7.json",
+      "image_file": "332b024d1902d001e278687fc5e990c7.png",
+      "image_sha256": "e70d5cc908fbf1058ba549d0abc01b3c04cc1e934b79d69eac2660849fddfc27",
+      "annotation_sha256": "c08f57b5b2643f8ee21c7f386c9aded432bbdaff88577b60e5e6781973460664",
+      "scene_graph_sha256": "d945be0c1f81af3adea881861a1e4012bd2bc33c784b6983e66fa624d1d3eb07"
+    },
+    {
+      "position": 8,
+      "caption": "Multiple lines",
+      "index": 112,
+      "sample_id": "nb_bd3466f097f2219fb2487e6b",
+      "benchmark_id": "nb2_cd147ce88da91fa3d28d",
+      "source_type": "synthetic",
+      "track": "nb2",
+      "split": "test",
+      "shard": "data/synthetic/test/part-00004.tar.gz",
+      "annotation_file": "cd147ce88da91fa3d28daf463014fb6f.json",
+      "image_file": "cd147ce88da91fa3d28daf463014fb6f.png",
+      "image_sha256": "021b5a4a11557e7f058e58f87caf01dc252c2748eebc9085b6d37290e5afc38e",
+      "annotation_sha256": "19ddc281f6544ac527fd02481c83594ac4be5699936b99d38467fa22c907819b",
+      "scene_graph_sha256": "e8463816043041524a01a43ec7d488892d75d0b80cdf1a7b18701a89bc8bf013"
+    }
+  ],
+  "annotation": {
+    "index": 20,
+    "sample_id": "d3-js_diverging_bar_chart_diverging_bar_plain_chart_03/sample_5364",
+    "benchmark_id": "layout_400ffffeb13d992a35ac",
+    "source_type": "synthetic",
+    "track": "layout",
+    "split": "test",
+    "shard": "data/synthetic/test/part-00000.tar.gz",
+    "annotation_file": "400ffffeb13d992a35ac0090ccdd3f4d.json",
+    "image_file": "400ffffeb13d992a35ac0090ccdd3f4d.png",
+    "image_sha256": "cb2901ae06da12e8a171ada83a298f592fe0c9f58a27fc4836730919a726e6a4",
+    "annotation_sha256": "6e8b6ae8a8d58102538467bcfefc245f7d0697b1bc917b1a4ab34db85a571022",
+    "scene_graph_sha256": "ba0529dc354b736ce0fe44b8e545d09aa3b8c438fffbd6c7b44b330ab026e911",
+    "shown_parent_links": [
+      [
+        "<chart 5>",
+        "ROOT"
+      ],
+      [
+        "<data_item 9>",
+        "<chart 5>"
+      ],
+      [
+        "<data_item 10>",
+        "<data_item 9>"
+      ],
+      [
+        "<data_item 13>",
+        "<data_item 9>"
+      ],
+      [
+        "<image 16>",
+        "<data_item 9>"
+      ],
+      [
+        "<text 17>",
+        "<data_item 9>"
+      ],
+      [
+        "<text 11>",
+        "<data_item 10>"
+      ],
+      [
+        "<shape 12>",
+        "<data_item 10>"
+      ],
+      [
+        "<text 14>",
+        "<data_item 13>"
+      ],
+      [
+        "<shape 15>",
+        "<data_item 13>"
+      ]
+    ]
+  },
+  "paper_figures": {
+    "image2json-results.png": "Qualitative comparison of image-to-scene-graph predictions (paper figure).",
+    "generation-results.png": "Qualitative scene graph preservation comparison (paper figure)."
+  }
+}