jiagengliu02 commited on
Commit
8536af5
·
verified ·
1 Parent(s): ce8d37d

Replace with Croissant-friendly questions table

Browse files
data/questions.jsonl CHANGED
The diff for this file is too large to render. See raw diff
 
scripts/export_hf_questions.py CHANGED
@@ -2,7 +2,6 @@
2
  from __future__ import annotations
3
 
4
  import argparse
5
- import hashlib
6
  import json
7
  from pathlib import Path
8
  from typing import Any
@@ -84,12 +83,6 @@ def as_list(value: Any) -> list[Any]:
84
  return [value]
85
 
86
 
87
- def stable_id(source_rel: str) -> str:
88
- digest = hashlib.sha1(source_rel.encode("utf-8")).hexdigest()[:12]
89
- readable = source_rel.removesuffix(".json").replace("/", "__").replace(" ", "_").replace("&", "and")
90
- return f"{readable}__{digest}"
91
-
92
-
93
  def normalize_answer(raw: Any) -> str | int | float | bool | list[Any] | dict[str, Any] | None:
94
  if raw is None:
95
  return None
@@ -233,7 +226,7 @@ def compact_metadata(payload: dict[str, Any]) -> dict[str, Any]:
233
  return metadata
234
 
235
 
236
- def record_for_file(path: Path, json_root: Path) -> dict[str, Any]:
237
  rel = path.relative_to(json_root)
238
  parts = rel.parts
239
  if len(parts) < 5:
@@ -242,12 +235,11 @@ def record_for_file(path: Path, json_root: Path) -> dict[str, Any]:
242
  big_task, small_task, scene_from_path, room_from_path = parts[:4]
243
  payload = json.loads(path.read_text(encoding="utf-8"))
244
  question, answer, options = extract_question_answer(payload)
245
- source_rel = rel.as_posix()
246
  scene = text(payload.get("scene")) or scene_from_path
247
  room = text(payload.get("room")) or room_from_path
248
 
249
  return {
250
- "id": stable_id(source_rel),
251
  "big_task": big_task,
252
  "small_task": small_task,
253
  "runner_task": RUNNER_TASK_BY_SMALL.get(small_task),
@@ -270,7 +262,11 @@ def main() -> int:
270
 
271
  json_root = args.json_root
272
  question_files = sorted(p for p in json_root.rglob("*.json") if p.parent != json_root)
273
- records = [record_for_file(path, json_root) for path in question_files]
 
 
 
 
274
 
275
  args.output.parent.mkdir(parents=True, exist_ok=True)
276
  with args.output.open("w", encoding="utf-8") as f:
 
2
  from __future__ import annotations
3
 
4
  import argparse
 
5
  import json
6
  from pathlib import Path
7
  from typing import Any
 
83
  return [value]
84
 
85
 
 
 
 
 
 
 
86
  def normalize_answer(raw: Any) -> str | int | float | bool | list[Any] | dict[str, Any] | None:
87
  if raw is None:
88
  return None
 
226
  return metadata
227
 
228
 
229
+ def record_for_file(path: Path, json_root: Path, row_id: str) -> dict[str, Any]:
230
  rel = path.relative_to(json_root)
231
  parts = rel.parts
232
  if len(parts) < 5:
 
235
  big_task, small_task, scene_from_path, room_from_path = parts[:4]
236
  payload = json.loads(path.read_text(encoding="utf-8"))
237
  question, answer, options = extract_question_answer(payload)
 
238
  scene = text(payload.get("scene")) or scene_from_path
239
  room = text(payload.get("room")) or room_from_path
240
 
241
  return {
242
+ "id": row_id,
243
  "big_task": big_task,
244
  "small_task": small_task,
245
  "runner_task": RUNNER_TASK_BY_SMALL.get(small_task),
 
262
 
263
  json_root = args.json_root
264
  question_files = sorted(p for p in json_root.rglob("*.json") if p.parent != json_root)
265
+ id_width = max(4, len(str(len(question_files))))
266
+ records = [
267
+ record_for_file(path, json_root, f"{index:0{id_width}d}")
268
+ for index, path in enumerate(question_files, start=1)
269
+ ]
270
 
271
  args.output.parent.mkdir(parents=True, exist_ok=True)
272
  with args.output.open("w", encoding="utf-8") as f:
scripts/upload_hf_dataset.py ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ from __future__ import annotations
3
+
4
+ import argparse
5
+ import subprocess
6
+ import sys
7
+ from pathlib import Path
8
+
9
+ from huggingface_hub import HfApi
10
+
11
+
12
+ def run_export(repo_root: Path) -> None:
13
+ subprocess.run(
14
+ [sys.executable, "scripts/export_hf_questions.py"],
15
+ cwd=repo_root,
16
+ check=True,
17
+ )
18
+
19
+
20
+ def main() -> int:
21
+ parser = argparse.ArgumentParser(description="Upload the Croissant-friendly ESI-Bench table to Hugging Face.")
22
+ parser.add_argument("--repo-id", default="ESI-Bench/esi-bench")
23
+ parser.add_argument("--repo-root", type=Path, default=Path(__file__).resolve().parents[1])
24
+ parser.add_argument("--private", action="store_true", help="Create the dataset repo as private if it does not exist.")
25
+ parser.add_argument("--skip-export", action="store_true", help="Upload existing data/questions.jsonl without regenerating it.")
26
+ parser.add_argument("--commit-message", default="Replace with Croissant-friendly questions table")
27
+ args = parser.parse_args()
28
+
29
+ repo_root = args.repo_root.resolve()
30
+ if not args.skip_export:
31
+ run_export(repo_root)
32
+
33
+ api = HfApi()
34
+ api.create_repo(repo_id=args.repo_id, repo_type="dataset", private=args.private, exist_ok=True)
35
+ api.upload_folder(
36
+ folder_path=repo_root,
37
+ repo_id=args.repo_id,
38
+ repo_type="dataset",
39
+ allow_patterns=[
40
+ "README.md",
41
+ "data/questions.jsonl",
42
+ "scripts/export_hf_questions.py",
43
+ "scripts/upload_hf_dataset.py",
44
+ ],
45
+ delete_patterns=[
46
+ "data/**",
47
+ "dataset/**",
48
+ "scripts/**",
49
+ ],
50
+ commit_message=args.commit_message,
51
+ )
52
+ print(f"Uploaded {repo_root} to https://huggingface.co/datasets/{args.repo_id}")
53
+ return 0
54
+
55
+
56
+ if __name__ == "__main__":
57
+ raise SystemExit(main())