wikikg-fact-phd / src /analysis /label_distribution.py
minhy112's picture
Add files using upload-large-folder tool
715cc5a verified
Raw
History Blame Contribute Delete
2.05 kB
from __future__ import annotations
import argparse
from collections import Counter, defaultdict
from pathlib import Path
from typing import Any
from src.data.io_utils import read_jsonl, write_csv
DEFAULT_LABEL_ORDER = ["SUPPORTS", "REFUTES", "NEI", "CONFLICTING"]
def infer_split(path: Path, rows: list[dict[str, Any]]) -> str:
splits = sorted({str(row.get("split")) for row in rows if row.get("split")})
if len(splits) == 1:
return splits[0]
stem = path.stem
if stem.startswith("claims_"):
return stem.removeprefix("claims_")
return stem
def distribution_rows(input_path: Path, label_order: list[str]) -> list[dict[str, Any]]:
rows = read_jsonl(input_path)
grouped: dict[str, list[dict[str, Any]]] = defaultdict(list)
fallback_split = infer_split(input_path, rows)
for row in rows:
grouped[str(row.get("split") or fallback_split)].append(row)
output_rows: list[dict[str, Any]] = []
for split, split_rows in sorted(grouped.items()):
counts = Counter(str(row.get("label") or "UNLABELED") for row in split_rows)
labels = list(label_order)
labels.extend(label for label in sorted(counts) if label not in labels)
total = len(split_rows)
out: dict[str, Any] = {"Split": split, "Total": total}
for label in labels:
out[label] = counts.get(label, 0)
for label in labels:
out[f"{label}_pct"] = round(counts.get(label, 0) / max(1, total) * 100, 4)
output_rows.append(out)
return output_rows
def main() -> None:
parser = argparse.ArgumentParser()
parser.add_argument("--input", type=Path, required=True)
parser.add_argument("--output", type=Path, required=True)
parser.add_argument("--label-order", nargs="*", default=DEFAULT_LABEL_ORDER)
args = parser.parse_args()
rows = distribution_rows(args.input, args.label_order)
write_csv(args.output, rows)
print(f"Wrote {len(rows)} label distribution rows to {args.output}")
if __name__ == "__main__":
main()