data_factory / fill_answer_from_text.py
marcuskwan's picture
Add files using upload-large-folder tool
d18d6b1 verified
#!/usr/bin/env python3
"""Extract answer markers from text and populate the `answer` field."""
from __future__ import annotations
import argparse
import json
import re
from pathlib import Path
ANSWER_PATTERN = re.compile(
r"\*\*\s*answer\s*:\s*(.+?)\s*\*\*",
flags=re.IGNORECASE | re.DOTALL,
)
def build_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(
description=(
"Extract `**Answer:x**` from each item's `text` and fill `answer`."
)
)
parser.add_argument(
"--input",
default="video_dataset.json",
help="Input JSON path (default: video_dataset.json).",
)
parser.add_argument(
"--output",
default="",
help="Output JSON path (default: overwrite --input in place).",
)
parser.add_argument(
"--overwrite",
action="store_true",
help="Overwrite non-empty existing `answer` values.",
)
parser.add_argument(
"--dry-run",
action="store_true",
help="Only report stats; do not write output.",
)
parser.add_argument(
"--ensure-ascii",
action="store_true",
help="Write escaped unicode (default keeps unicode characters).",
)
return parser
def extract_answer(text: str) -> str:
match = ANSWER_PATTERN.search(text or "")
if not match:
return ""
raw = match.group(1).strip()
return re.sub(r"\s+", " ", raw)
def main() -> None:
args = build_parser().parse_args()
input_path = Path(args.input)
output_path = Path(args.output) if args.output else input_path
with input_path.open("r", encoding="utf-8") as f:
data = json.load(f)
if not isinstance(data, list):
raise ValueError("Input JSON must be a list of objects.")
total = 0
with_marker = 0
updated = 0
skipped_existing = 0
no_marker = 0
for item in data:
if not isinstance(item, dict):
continue
total += 1
text = "" if item.get("text") is None else str(item.get("text"))
extracted = extract_answer(text)
if not extracted:
no_marker += 1
continue
with_marker += 1
current = "" if item.get("answer") is None else str(item.get("answer")).strip()
if current and not args.overwrite:
skipped_existing += 1
continue
if current != extracted:
item["answer"] = extracted
updated += 1
if not args.dry_run:
with output_path.open("w", encoding="utf-8") as f:
json.dump(data, f, ensure_ascii=args.ensure_ascii, indent=2)
f.write("\n")
print(
"Processed "
f"{total} items | markers: {with_marker} | updated: {updated} | "
f"skipped_existing: {skipped_existing} | no_marker: {no_marker} | "
f"output: {output_path if not args.dry_run else '(dry-run)'}"
)
if __name__ == "__main__":
main()