Publish Habr IR retrieval release

Release fingerprint: 0a18ee0a5ee2373a1b861f1d08d325052ccaeec94fb69110d90d16c2630d96e2

README.md

@@ -11,26 +11,16 @@ tags:
 - russian
 - synthetic-queries
 configs:
-- config_name: pairs
+- config_name: default
   data_files:
   - split: train
-    path: data/pairs/train.parquet
-  - split: validation
-    path: data/pairs/validation.parquet
+    path: data/train-*.parquet
+  - split: dev
+    path: data/dev-*.parquet
   - split: test
-    path: data/pairs/test.parquet
-- config_name: corpus-100k
-  data_files:
-  - split: train
-    path: data/corpus-100k/corpus.parquet
-- config_name: qrels
-  data_files:
-  - split: train
-    path: data/qrels/train.parquet
-  - split: validation
-    path: data/qrels/validation.parquet
-  - split: test
-    path: data/qrels/test.parquet
+    path: data/test-*.parquet
+  - split: corpus
+    path: data/corpus-*.parquet
 ---
 
 # Habr IR
@@ -43,18 +33,19 @@ Queries and short answers were generated from passages and passed deterministic
 and competitor checks. The generation process is synthetic: these examples are not a substitute for fully
 human-authored relevance judgments. The local pilot was manually reviewed before scale generation.
 
-## Configurations
+## Splits
 
-- `pairs`: queries, answers, positive passages, intent labels, and generation provenance.
-- `corpus-100k`: the complete retrieval corpus; `is_positive` marks passages referenced by accepted pairs.
-- `qrels`: one positive relevance judgment per query.
+- `train`, `dev`, and `test` contain query-positive rows.
+- `corpus` contains candidate documents as `doc_id, content`.
+- All splits share one schema and one `default` Dataset Viewer configuration.
+- Raw qrels, the audit sheet, and release provenance remain downloadable repository artifacts.
 
 ```python
 from datasets import load_dataset
 
-pairs = load_dataset("justatom/habr-ir", "pairs")
-corpus = load_dataset("justatom/habr-ir", "corpus-100k", split="train")
-qrels = load_dataset("justatom/habr-ir", "qrels")
+dataset = load_dataset("justatom/habr-ir")
+print(dataset["train"][0]["query"])
+print(dataset["corpus"][0]["content"])
 ```
 
 ## Source and limitations

data/{pairs/train.parquet → corpus-00000-of-00001.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:42fc497302c80acd1718d4edad4f2a48659e42a04ef337adea281b7f520f52a0
-size 10993763
+oid sha256:933c5b9efadb1714933eb69206f1331a33b92c6ca35a95cc26cba766e7ba9523
+size 92678880

data/{pairs/validation.parquet → dev-00000-of-00001.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:377a259710919e7304e6c992bdd874c48f85e1abe329d193b7b74881f2bcf719
-size 1378297
+oid sha256:298571b6fcc119c39d25f62bc59558c4bec32c78f312d8c08ab00ec7d43355da
+size 1379703

data/{pairs/test.parquet → test-00000-of-00001.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:58d7097b597009b73d3979bcc99d7db4450d537713d38c78593a10eea4c5d8aa
-size 1412293
+oid sha256:241106c7571e072629905e19c5559217d5ae06927c4afdddb6e27464fea9a93a
+size 1413736

data/{corpus-100k/corpus.parquet → train-00000-of-00001.parquet}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:740f1ccf27880da6b7899ad776cec43812487c1587f5e93937ee6bb4e5d6153f
-size 172563126
+oid sha256:1f1bbf9d161fb6df7ad0a4d7a3d6fe42b3b916e0a95c310d162781a820af6e55
+size 10995248

data/manifests/release-manifest.json → manifest.json

@@ -1,67 +1,81 @@
 {
   "artifacts": [
     {
-      "bytes": 2182,
+      "bytes": 1889,
       "path": "README.md",
-      "sha256": "6fe6f923fb9897d382c2b8b74933ea52c93cf35a1ec901118cf62e3861d6d38a"
+      "sha256": "65d882d5ae99d5e926cad5e10cc161a3280e6d8178504fa4a3caac91a584072b"
     },
     {
-      "bytes": 125476479,
-      "path": "audit/pilot-review.csv",
-      "sha256": "bb0d7b9763bf8e6937299ed08163b730f865ff8c2daa70803270401a515dec66"
+      "bytes": 73377,
+      "path": "artifacts/qrels/dev.parquet",
+      "sha256": "e65c75b20b3ef8e0f62942e3efb283a89cd84626a912a0aa710884045281b1b2"
     },
     {
-      "bytes": 172563126,
-      "path": "data/corpus-100k/corpus.parquet",
-      "sha256": "740f1ccf27880da6b7899ad776cec43812487c1587f5e93937ee6bb4e5d6153f"
+      "bytes": 73693,
+      "path": "artifacts/qrels/test.parquet",
+      "sha256": "e172bdca77be49becb27a81dcd0578b6eafeb453cc145147df3532dbe359a8ec"
     },
     {
-      "bytes": 1412293,
-      "path": "data/pairs/test.parquet",
-      "sha256": "58d7097b597009b73d3979bcc99d7db4450d537713d38c78593a10eea4c5d8aa"
+      "bytes": 581271,
+      "path": "artifacts/qrels/train.parquet",
+      "sha256": "76afe4a65cadad3bd812767c384a5c11d35c28aa1909840fabbd7e33ff255cc0"
     },
     {
-      "bytes": 10993763,
-      "path": "data/pairs/train.parquet",
-      "sha256": "42fc497302c80acd1718d4edad4f2a48659e42a04ef337adea281b7f520f52a0"
+      "bytes": 125476479,
+      "path": "audit/pilot-review.csv",
+      "sha256": "bb0d7b9763bf8e6937299ed08163b730f865ff8c2daa70803270401a515dec66"
     },
     {
-      "bytes": 1378297,
-      "path": "data/pairs/validation.parquet",
-      "sha256": "377a259710919e7304e6c992bdd874c48f85e1abe329d193b7b74881f2bcf719"
+      "bytes": 92678880,
+      "path": "data/corpus-00000-of-00001.parquet",
+      "sha256": "933c5b9efadb1714933eb69206f1331a33b92c6ca35a95cc26cba766e7ba9523"
     },
     {
-      "bytes": 73693,
-      "path": "data/qrels/test.parquet",
-      "sha256": "e172bdca77be49becb27a81dcd0578b6eafeb453cc145147df3532dbe359a8ec"
+      "bytes": 1379703,
+      "path": "data/dev-00000-of-00001.parquet",
+      "sha256": "298571b6fcc119c39d25f62bc59558c4bec32c78f312d8c08ab00ec7d43355da"
     },
     {
-      "bytes": 581271,
-      "path": "data/qrels/train.parquet",
-      "sha256": "76afe4a65cadad3bd812767c384a5c11d35c28aa1909840fabbd7e33ff255cc0"
+      "bytes": 1413736,
+      "path": "data/test-00000-of-00001.parquet",
+      "sha256": "241106c7571e072629905e19c5559217d5ae06927c4afdddb6e27464fea9a93a"
     },
     {
-      "bytes": 73377,
-      "path": "data/qrels/validation.parquet",
-      "sha256": "e65c75b20b3ef8e0f62942e3efb283a89cd84626a912a0aa710884045281b1b2"
+      "bytes": 10995248,
+      "path": "data/train-00000-of-00001.parquet",
+      "sha256": "1f1bbf9d161fb6df7ad0a4d7a3d6fe42b3b916e0a95c310d162781a820af6e55"
     }
   ],
-  "contract": "habr-ir-combined-release-v1",
+  "contract": "habr-ir-retrieval-release-v1",
   "counts": {
     "audit_rows": 11000,
     "corpus": 100000,
-    "pair_splits": {
-      "test": 1029,
-      "train": 8215,
-      "validation": 1016
-    },
     "pairs": 10260,
-    "qrels": 10260
+    "qrels": 10260,
+    "splits": {
+      "corpus": 100000,
+      "dev": 1016,
+      "test": 1029,
+      "train": 8215
+    }
   },
-  "fingerprint": "66465c0f1c1ac414fdd7317464f105ae0ed54e06685a269db2fe2130a3daf4eb",
+  "fingerprint": "0a18ee0a5ee2373a1b861f1d08d325052ccaeec94fb69110d90d16c2630d96e2",
   "git": {
     "dirty": false,
-    "sha": "2cbc50c499f2cb532b37515e196642a972068978"
+    "sha": "87f075a6ca85c7170821331721afd1f4c2cc8090"
+  },
+  "release": {
+    "config_name": "default",
+    "include_audit": true,
+    "include_qrels_artifacts": true,
+    "layout": "retrieval",
+    "repo_id": "justatom/habr-ir",
+    "split_map": {
+      "corpus": "corpus",
+      "test": "test",
+      "train": "train",
+      "validation": "dev"
+    }
   },
   "releases": [
     {
@@ -227,41 +241,29 @@
       "title": "String",
       "url": "String"
     },
-    "corpus": {
-      "article_id": "String",
-      "char_count": "Int64",
-      "content": "String",
-      "corpus_rank": "UInt32",
-      "flows": "List(String)",
-      "hubs": "List(String)",
-      "is_positive": "Boolean",
-      "passage_id": "String",
-      "section": "String",
-      "serialized_passage": "String",
-      "source_hash": "String",
-      "tags": "List(String)",
-      "title": "String",
-      "token_count": "Int64",
-      "url": "String"
-    },
-    "pairs": {
+    "public": {
       "actual_intent": "String",
       "answer": "String",
       "article_id": "String",
+      "bucket": "String",
+      "content": "String",
+      "doc_id": "String",
       "evidence": "String",
       "generation_attempt": "Int64",
       "generation_batch_id": "String",
       "generation_custom_id": "String",
       "generator_model": "String",
       "generator_prompt_hash": "String",
+      "is_positive": "Boolean",
+      "negative_doc_id": "String",
       "pair_id": "String",
-      "positive_passage": "String",
-      "positive_passage_id": "String",
+      "positive": "String",
+      "positive_doc_id": "String",
       "query": "String",
       "query_id": "String",
       "requested_intent": "String",
       "section": "String",
-      "split": "String",
+      "source": "String",
       "tags": "List(String)",
       "title": "String",
       "topic_flows": "List(String)",