Publish ML100K leave-one-out samples with index-only plots

.gitattributes

@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+histories.jsonl filter=lfs diff=lfs merge=lfs -text
+train.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,58 @@
+---
+configs:
+  - config_name: default
+    data_files:
+      - split: train
+        path: train.jsonl
+      - split: test
+        path: test.jsonl
+      - split: queries
+        path: test.jsonl
+  - config_name: items
+    data_files:
+      - split: items
+        path: items.jsonl
+---
+
+# MovieLens Leave-One-Out
+
+Five chronological interactions predict the next interaction. One final test target per user; no rating filter. Histories are audit-only, not wholesale model inputs. Actors are supplementary; see actor_sources.json.
+
+```json
+{
+  "schema": "movie-fields-v1",
+  "source": "official MovieLens 100K",
+  "sample_policy": "leave-one-out-windows",
+  "past_order": "oldest-first",
+  "history_length": 5,
+  "stride": 1,
+  "shuffle_seed": 42,
+  "timestamp_policy": "rating timestamp then original u.data row ordinal; ties are not verified watch order",
+  "target_policy": "immediate next interaction, any rating; final interaction held out per user",
+  "catalog_items": 1682,
+  "split_counts": {
+    "train": 94342,
+    "test": 943
+  },
+  "history_rows": 943,
+  "seen_items_filtered": false,
+  "movies_with_actors": 1618,
+  "actor_join_methods": {
+    "unique_official_movielens_title_year_link_imdb_year_checked": 1268,
+    "unique_title_adjacent_release_year_inferred": 56,
+    "unique_imdb_title_year": 268,
+    "unmatched": 31,
+    "explicit_imdb_id_in_original_url": 53,
+    "ambiguous_title_year": 6
+  },
+  "source_events": 100000,
+  "users": 943,
+  "users_too_short": 0,
+  "timestamp_ties": 50561,
+  "held_out_events_in_train": 0
+}
+```
+
+## Index-only plots
+
+Samples and metadata are unchanged. Select `--recommendation-index-content metadata-plot` to index `search_text`; the default still uses metadata-only `text`. Plots never enter histories. Keep plot_source attribution and its content license when distributing. See plot_enrichment.json for coverage and byte-identical files.

histories.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8bbbec4dbeaf05cd0b4a3c4d40fa6004a47e4353a7b5b7c08a9e8277e5918a71
+size 36484649

plot_enrichment.json

@@ -0,0 +1,24 @@
+{
+  "source_dataset": "E:\\Programming\\Github works\\tyrec-codex-updated\\tyrec\\data\\recommendations_loo\\ml_100k",
+  "plot_passages": "E:\\Programming\\Github works\\ml\\query_expansion\\truth_data\\movielens_100k_movie_passages_20260924\\movie_passages.jsonl",
+  "split_counts": {
+    "train": 94342,
+    "test": 943
+  },
+  "catalog_items": 1682,
+  "plot_status_counts": {
+    "with_plot": 1515,
+    "metadata_only_missing_plot": 167
+  },
+  "byte_identical_files": [
+    "train.jsonl",
+    "test.jsonl",
+    "histories.jsonl",
+    "actor_sources.json"
+  ],
+  "metadata_and_history_unchanged": true,
+  "search_text_recipe": "text + newline + plot_text; empty plot leaves text unchanged",
+  "plots_in_model_input": false,
+  "sample_rebuild": false,
+  "new_relevance_labels": false
+}

preparation.json

@@ -0,0 +1,56 @@
+{
+  "schema": "movie-fields-v1",
+  "source": "official MovieLens 100K",
+  "sample_policy": "leave-one-out-windows",
+  "past_order": "oldest-first",
+  "history_length": 5,
+  "stride": 1,
+  "shuffle_seed": 42,
+  "timestamp_policy": "rating timestamp then original u.data row ordinal; ties are not verified watch order",
+  "target_policy": "immediate next interaction, any rating; final interaction held out per user",
+  "catalog_items": 1682,
+  "split_counts": {
+    "train": 94342,
+    "test": 943
+  },
+  "history_rows": 943,
+  "seen_items_filtered": false,
+  "movies_with_actors": 1618,
+  "actor_join_methods": {
+    "unique_official_movielens_title_year_link_imdb_year_checked": 1268,
+    "unique_title_adjacent_release_year_inferred": 56,
+    "unique_imdb_title_year": 268,
+    "unmatched": 31,
+    "explicit_imdb_id_in_original_url": 53,
+    "ambiguous_title_year": 6
+  },
+  "source_events": 100000,
+  "users": 943,
+  "users_too_short": 0,
+  "timestamp_ties": 50561,
+  "held_out_events_in_train": 0,
+  "plot_enrichment": {
+    "source_dataset": "E:\\Programming\\Github works\\tyrec-codex-updated\\tyrec\\data\\recommendations_loo\\ml_100k",
+    "plot_passages": "E:\\Programming\\Github works\\ml\\query_expansion\\truth_data\\movielens_100k_movie_passages_20260924\\movie_passages.jsonl",
+    "split_counts": {
+      "train": 94342,
+      "test": 943
+    },
+    "catalog_items": 1682,
+    "plot_status_counts": {
+      "with_plot": 1515,
+      "metadata_only_missing_plot": 167
+    },
+    "byte_identical_files": [
+      "train.jsonl",
+      "test.jsonl",
+      "histories.jsonl",
+      "actor_sources.json"
+    ],
+    "metadata_and_history_unchanged": true,
+    "search_text_recipe": "text + newline + plot_text; empty plot leaves text unchanged",
+    "plots_in_model_input": false,
+    "sample_rebuild": false,
+    "new_relevance_labels": false
+  }
+}

train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec23d56c101fcd067393b514d6eda1ded8f16ed144c5a45dabc1c53cfaceac06
+size 150900394