Upload backend/scripts/export_intent_backlog.py with huggingface_hub

backend/scripts/export_intent_backlog.py

@@ -0,0 +1,79 @@
+#!/usr/bin/env python3
+"""
+Parse backend/logs/intent/low_confidence.csv and export a Markdown backlog.
+"""
+from __future__ import annotations
+
+import argparse
+import csv
+from collections import defaultdict
+from datetime import datetime
+from pathlib import Path
+
+
+def parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(description="Export low-confidence intent queries.")
+    parser.add_argument(
+        "--output-date",
+        default=datetime.utcnow().strftime("%Y-%m-%d"),
+        help="Ngày (YYYY-MM-DD) dùng cho thư mục báo cáo.",
+    )
+    return parser.parse_args()
+
+
+def load_backlog(log_path: Path) -> list[dict[str, str]]:
+    if not log_path.exists():
+        return []
+    rows: list[dict[str, str]] = []
+    with log_path.open("r", encoding="utf-8") as fp:
+        reader = csv.DictReader(fp)
+        for row in reader:
+            rows.append(row)
+    return rows
+
+
+def render_markdown(rows: list[dict[str, str]]) -> str:
+    lines = [
+        "<!-- Auto-generated by export_intent_backlog.py -->",
+        "# Backlog truy vấn low-confidence",
+        "",
+        f"Tổng số mẫu: {len(rows)}",
+        "",
+    ]
+    by_intent: dict[str, list[dict[str, str]]] = defaultdict(list)
+    for row in rows:
+        by_intent[row.get("intent", "unknown")].append(row)
+
+    for intent, samples in sorted(by_intent.items()):
+        lines.append(f"## Intent: {intent} ({len(samples)} mẫu)")
+        lines.append("")
+        lines.append("| Thời gian | Confidence | Route | Query |")
+        lines.append("| --- | --- | --- | --- |")
+        for sample in samples[:50]:
+            lines.append(
+                f"| {sample.get('timestamp','')} | {sample.get('confidence','')} "
+                f"| {sample.get('route','')} | {sample.get('query','').strip()} |"
+            )
+        lines.append("")
+    if not by_intent:
+        lines.append("_Chưa có dữ liệu._")
+    return "\n".join(lines).strip() + "\n"
+
+
+def main() -> None:
+    repo_root = Path(__file__).resolve().parents[2]
+    args = parse_args()
+    log_path = repo_root / "backend" / "logs" / "intent" / "low_confidence.csv"
+    rows = load_backlog(log_path)
+
+    output_dir = repo_root / "tài nguyên" / "báo cáo" / args.output_date / "backend"
+    output_dir.mkdir(parents=True, exist_ok=True)
+    output_path = output_dir / "intent_backlog.md"
+    markdown = render_markdown(rows)
+    output_path.write_text(markdown, encoding="utf-8")
+    print(f"✅ Wrote {len(rows)} entries to {output_path}")
+
+
+if __name__ == "__main__":
+    main()
+