import argparse import csv import hashlib from collections import Counter from datetime import datetime, timedelta, timezone from pathlib import Path CORE_METALS_HIGH = {"\u94dc", "\u94dd", "\u91d1", "\u94f6", "\u9502", "\u7a00\u571f"} CORE_METALS_MEDIUM = {"\u9521", "\u954d", "\u94b4", "\u950c", "\u94c5", "\u94a8", "\u9511", "\u94bc", "\u9511"} def read_csv(path): with path.open("r", encoding="utf-8-sig", newline="") as handle: return list(csv.DictReader(handle)) def write_csv(path, fieldnames, rows): path.parent.mkdir(parents=True, exist_ok=True) with path.open("w", encoding="utf-8-sig", newline="") as handle: writer = csv.DictWriter(handle, fieldnames=fieldnames) writer.writeheader() writer.writerows(rows) def sha256_file(path): return hashlib.sha256(path.read_bytes()).hexdigest() def split_tags(value): return [part.strip() for part in (value or "").split("|") if part.strip()] def priority(metals, status, has_value, located): metal_set = set(metals) if metal_set & CORE_METALS_HIGH and has_value and located: return "HIGH" if status == "NEEDS_COMPANY_OFFICIAL_SOURCE" and metal_set & CORE_METALS_HIGH: return "HIGH" if metal_set & CORE_METALS_MEDIUM or has_value: return "MEDIUM" return "LOW" def source_type(status, themes): theme_set = set(themes) if status == "NEEDS_PRICE_INVENTORY_EXTERNAL_SOURCE": if "inventory" in theme_set: return "exchange_inventory_or_industry_inventory" return "exchange_or_SMM_or_industry_price" if status == "NEEDS_COMPANY_OFFICIAL_SOURCE": if "project" in theme_set: return "annual_report_exchange_filing_or_project_announcement" return "annual_report_exchange_filing_or_company_ir" return "manual_source_classification" def question(status, metals, themes, value_candidates): metal_text = "|".join(metals) or "UNKNOWN_METAL" theme_text = "|".join(themes) or "UNKNOWN_THEME" if status == "NEEDS_PRICE_INVENTORY_EXTERNAL_SOURCE": return f"\u4e3a {metal_text} / {theme_text} \u8865\u5145\u53ef\u590d\u6838\u7684\u4ef7\u683c\u3001\u5e93\u5b58\u6216\u4ea4\u6613\u6240/\u884c\u4e1a\u6765\u6e90\uff0c\u6838\u5bf9\u5019\u9009\u6570\u503c {value_candidates or 'NA'}" if status == "NEEDS_COMPANY_OFFICIAL_SOURCE": return f"\u4e3a {metal_text} / {theme_text} \u8865\u5145\u5e74\u62a5\u3001\u4ea4\u6613\u6240\u516c\u544a\u6216\u516c\u53f8 IR \u6765\u6e90\uff0c\u6838\u5bf9\u4ea7\u80fd/\u9879\u76ee/\u4f9b\u7ed9\u8868\u8ff0" return f"\u4eba\u5de5\u5224\u65ad {metal_text} / {theme_text} \u6240\u9700\u5916\u90e8\u6765\u6e90" def main(): parser = argparse.ArgumentParser() parser.add_argument("--project-root", default=".") parser.add_argument("--crosscheck", required=True) parser.add_argument("--cards", required=True) parser.add_argument("--locations", required=True) parser.add_argument("--output", required=True) parser.add_argument("--manifest", required=True) parser.add_argument("--summary", required=True) parser.add_argument("--run-id", default="RUN-ANA-YS-GAP-PRIORITY-007") args = parser.parse_args() project_root = Path(args.project_root).resolve() crosscheck = read_csv(project_root / args.crosscheck) cards = {row.get("evidence_card_id"): row for row in read_csv(project_root / args.cards)} locations = {row.get("evidence_card_id"): row for row in read_csv(project_root / args.locations)} targets = [ row for row in crosscheck if row.get("evidence_strength_draft") == "GAP_REVIEW" or row.get("cross_check_status", "").startswith("NEEDS_") ] created_at = datetime.now(timezone(timedelta(hours=8))).isoformat(timespec="seconds") rows = [] for item in targets: card = cards.get(item.get("evidence_card_id"), {}) loc = locations.get(item.get("evidence_card_id"), {}) metals = split_tags(item.get("metal_tags", "")) themes = split_tags(item.get("theme_tags", "")) has_value = bool(item.get("value_unit_candidates") or card.get("value_unit_candidates")) located = loc.get("location_status") == "LOCATED_IN_CONVERTED_TEXT" status = item.get("cross_check_status", "") pr = priority(metals, status, has_value, located) stype = source_type(status, themes) rows.append( { "gap_priority_id": f"YS-GAP-PRIORITY-007-{len(rows) + 1:04d}", "case_id": item.get("case_id", ""), "batch_id": item.get("batch_id", ""), "run_id": args.run_id, "evidence_card_id": item.get("evidence_card_id", ""), "fact_id": item.get("fact_id", ""), "doc_id": item.get("doc_id", ""), "metal_tags": item.get("metal_tags", ""), "theme_tags": item.get("theme_tags", ""), "cross_check_status": status, "gap_type": "PRICE_INVENTORY_SOURCE_GAP" if "PRICE" in status else "COMPANY_OFFICIAL_SOURCE_GAP" if "COMPANY" in status else "MANUAL_SOURCE_GAP", "priority": pr, "preferred_source_type": stype, "supplement_question": question(status, metals, themes, item.get("value_unit_candidates", "")), "located_in_converted_text": "YES" if located else "NO", "precheck_page_no": loc.get("precheck_page_no", ""), "precheck_location": loc.get("precheck_location", ""), "value_unit_candidates": item.get("value_unit_candidates", ""), "date_candidates": item.get("date_candidates", ""), "next_action": "supplement_now" if pr == "HIGH" else "queue_for_next_pass" if pr == "MEDIUM" else "defer_low_priority", "review_status": "DRAFT_FOR_REVIEW", "created_at": created_at, } ) output_path = project_root / args.output fields = list(rows[0].keys()) if rows else ["gap_priority_id", "case_id", "run_id", "review_status"] write_csv(output_path, fields, rows) output_sha = sha256_file(output_path) manifest_path = project_root / args.manifest write_csv( manifest_path, ["case_id", "batch_id", "run_id", "artifact_type", "artifact_path", "row_count", "sha256", "review_status", "created_at"], [ { "case_id": "ANA-YS-INDUSTRY-001", "batch_id": "BATCH-001+BATCH-003", "run_id": args.run_id, "artifact_type": "key_fact_gap_priority", "artifact_path": output_path.relative_to(project_root).as_posix(), "row_count": str(len(rows)), "sha256": output_sha, "review_status": "DRAFT_FOR_REVIEW", "created_at": created_at, } ], ) priority_counts = Counter(row["priority"] for row in rows) gap_counts = Counter(row["gap_type"] for row in rows) action_counts = Counter(row["next_action"] for row in rows) summary_path = project_root / args.summary lines = [ "# \u5173\u952e\u4e8b\u5b9e\u5916\u90e8\u6765\u6e90\u7f3a\u53e3\u4f18\u5148\u7ea7 PASS-007 \u6458\u8981", "", "\u72b6\u6001\uff1aDRAFT_FOR_REVIEW", f"\u751f\u6210\u65f6\u95f4\uff1a{created_at}", "", "## \u8f93\u51fa", "", f"- \u7f3a\u53e3\u961f\u5217\uff1a`{output_path.relative_to(project_root).as_posix()}`", f"- manifest\uff1a`{manifest_path.relative_to(project_root).as_posix()}`", f"- \u8bb0\u5f55\u6570\uff1a{len(rows)}", f"- sha256\uff1a`{output_sha}`", "", "## \u4f18\u5148\u7ea7", "", "| \u4f18\u5148\u7ea7 | \u6570\u91cf |", "|---|---:|", ] for key, count in priority_counts.most_common(): lines.append(f"| {key} | {count} |") lines.extend(["", "## \u7f3a\u53e3\u7c7b\u578b", "", "| \u7c7b\u578b | \u6570\u91cf |", "|---|---:|"]) for key, count in gap_counts.most_common(): lines.append(f"| {key} | {count} |") lines.extend(["", "## \u4e0b\u4e00\u6b65", "", "| \u52a8\u4f5c | \u6570\u91cf |", "|---|---:|"]) for key, count in action_counts.most_common(): lines.append(f"| {key} | {count} |") lines.extend( [ "", "## \u8fb9\u754c", "", "\u672c\u8f6e\u53ea\u662f\u5bf9 PASS-006 \u7f3a\u53e3\u505a\u4f18\u5148\u7ea7\u62c6\u5206\uff0c\u4e0d\u8868\u793a\u7f3a\u53e3\u5df2\u8865\u9f50\uff0c\u4e0d\u4f5c\u4e3a\u6b63\u5f0f\u8bc1\u636e\u6216\u6b63\u5f0f\u6307\u6807\u3002", ] ) summary_path.parent.mkdir(parents=True, exist_ok=True) summary_path.write_text("\n".join(lines) + "\n", encoding="utf-8") print(f"rows={len(rows)}") print(dict(priority_counts)) print(dict(gap_counts)) print(f"output={output_path.relative_to(project_root).as_posix()}") print(f"manifest={manifest_path.relative_to(project_root).as_posix()}") print(f"summary={summary_path.relative_to(project_root).as_posix()}") if __name__ == "__main__": main()