Update app.py

app.py

@@ -2,9 +2,9 @@ import os
 import io
 import json
 import hashlib
-import gradio
+import gradio  # 前方参照の解決用（必要なら）
 import gradio as gr
-from typing import Tuple, List, Union
+from typing import Any, List
 
 from pipelines.openai_ingest import (
     extract_text_with_openai,
@@ -14,60 +14,82 @@ from pipelines.openai_ingest import (
 from pipelines.parsing import normalize_resume
 from pipelines.merge import merge_normalized_records
 from pipelines.skills import extract_skills
-from pipelines.anonymize import anonymize_text, render_anonymized_pdf
+
+# ---- anonymize フォールバック（pipelines/anonymize.py が未実装でも動く） ----
+try:
+    from pipelines.anonymize import anonymize_text, render_anonymized_pdf  # type: ignore
+except Exception:
+    import re
+    try:
+        from reportlab.pdfgen import canvas
+        from reportlab.lib.pagesizes import A4
+    except Exception:
+        canvas = None
+        A4 = None
+
+    def anonymize_text(text: str):
+        # 超簡易：メール/電話っぽい所をマスク。氏名は見出し候補を雑にマスク。
+        masked = re.sub(r"([A-Za-z0-9._%+-]+)@([A-Za-z0-9.-]+\.[A-Za-z]{2,})", r"***@\2", text)
+        masked = re.sub(r"(?:\+?\d{1,3}[ -]?)?(?:\(\d{2,4}\)[ -]?)?\d{2,4}[ -]?\d{2,4}[ -]?\d{3,4}", "***-****-****", masked)
+        masked = re.sub(r"(氏名[:：]?\s*)(\S+)", r"\1＊＊＊", masked)
+        return masked, {"fallback": True}
+
+    def render_anonymized_pdf(text: str) -> bytes:
+        if canvas is None:
+            # reportlab が無ければテキストファイルで代替（UIは .pdf 名で返す）
+            return text.encode("utf-8")
+        buf = io.BytesIO()
+        c = canvas.Canvas(buf, pagesize=A4)
+        width, height = A4
+        margin = 40
+        y = height - margin
+        for line in text.splitlines() or ["(no content)"]:
+            if y < margin:
+                c.showPage()
+                y = height - margin
+            c.drawString(margin, y, line[:95])
+            y -= 14
+        c.save()
+        return buf.getvalue()
+# ---------------------------------------------------------------------------
+
 from pipelines.scoring import compute_quality_score
 from pipelines.storage import persist_to_hf
 from pipelines.utils import detect_filetype, load_doc_text
 
 APP_TITLE = "候補者インテーク & レジュメ標準化（OpenAI版）"
 
-# ---- helpers ---------------------------------------------------------------
-
-def _read_file_input(item: Union[str, "gradio.files.TempFile"]) -> Tuple[bytes, str]:
-    """
-    Gradio v4.44 の Files(type='filepath') は str パスを返す。
-    互換のため、パス/ファイルライク双方を許容して (bytes, filename) を返す。
-    """
-    if isinstance(item, str):
-        with open(item, "rb") as rf:
-            data = rf.read()
-        name = os.path.basename(item)
-        return data, name
-    # UploadedFile 等（念のため）
-    if hasattr(item, "read"):
-        data = item.read()
-        name = getattr(item, "name", "uploaded")
-        return data, os.path.basename(name)
-    raise ValueError("Unsupported file input type")
-
 
-def _as_json_code(obj) -> str:
-    return json.dumps(obj, ensure_ascii=False, indent=2)
+def _read_bytes_from_path(path: str) -> bytes:
+    with open(path, "rb") as f:
+        return f.read()
 
 
-# ---- core pipeline ---------------------------------------------------------
-
-def process_resumes(files: List[Union[str, "gradio.files.TempFile"]], candidate_id: str, additional_notes: str = ""):
+def process_resumes(files: List[str], candidate_id: str = "", additional_notes: str = ""):
+    """
+    files: gr.Files(type='filepath') から渡るファイルパスのリスト
+    """
     if not files:
         raise gr.Error("少なくとも1ファイルをアップロードしてください。")
 
     partial_records = []
     raw_texts = []
 
-    for f in files:
-        raw_bytes, fname = _read_file_input(f)
-        filetype = detect_filetype(fname, raw_bytes)
+    for path in files:
+        filename = os.path.basename(path)
+        raw_bytes = _read_bytes_from_path(path)
+        filetype = detect_filetype(filename, raw_bytes)
 
         # 1) テキスト抽出：画像/PDFはOpenAI Vision OCR、docx/txtは生文面＋OpenAI整形
         if filetype in {"pdf", "image"}:
-            text = extract_text_with_openai(raw_bytes, filename=fname, filetype=filetype)
+            text = extract_text_with_openai(raw_bytes, filename=filename, filetype=filetype)
         else:
             base_text = load_doc_text(filetype, raw_bytes)
-            text = extract_text_with_openai(base_text.encode("utf-8"), filename=fname, filetype="txt")
+            text = extract_text_with_openai(base_text.encode("utf-8"), filename=filename, filetype="txt")
 
-        raw_texts.append({"filename": fname, "text": text})
+        raw_texts.append({"filename": filename, "text": text})
 
-        # 2) OpenAIでセクション構造化 → ルール整形
+        # 2) OpenAIでセクション構造化 → ルール正規化
         structured = structure_with_openai(text)
         normalized = normalize_resume({
             "work_experience": structured.get("work_experience_raw", ""),
@@ -76,16 +98,16 @@ def process_resumes(files: List[Union[str, "gradio.files.TempFile"]], candidate_
             "skills": ", ".join(structured.get("skills_list", [])),
         })
         partial_records.append({
-            "source": fname,
+            "source": filename,
             "text": text,
             "structured": structured,
             "normalized": normalized,
         })
 
-    # 3) 統合
+    # 3) 統合（複数ファイル→1候補者）
     merged = merge_normalized_records([r["normalized"] for r in partial_records])
 
-    # 4) スキル抽出
+    # 4) スキル抽出（辞書/正規表現）
     merged_text = "\n\n".join([r["text"] for r in partial_records])
     skills = extract_skills(merged_text, {
         "work_experience": merged.get("raw_sections", {}).get("work_experience", ""),
@@ -101,13 +123,14 @@ def process_resumes(files: List[Union[str, "gradio.files.TempFile"]], candidate_
     # 6) 品質スコア
     score = compute_quality_score(merged_text, merged)
 
-    # 7) 要約
+    # 7) 要約（300/100/1文）
     summaries = summarize_with_openai(merged_text)
 
     # 8) 構造化出力
+    cid = candidate_id or hashlib.sha256(merged_text.encode("utf-8")).hexdigest()[:16]
     result_json = {
-        "candidate_id": candidate_id or hashlib.sha256(merged_text.encode("utf-8")).hexdigest()[:16],
-        "files": [os.path.basename(_read_file_input(f)[1]) if isinstance(f, str) else getattr(f, "name", "uploaded") for f in files],
+        "candidate_id": cid,
+        "files": [os.path.basename(p) for p in files],
         "merged": merged,
         "skills": skills,
         "quality_score": score,
@@ -116,38 +139,39 @@ def process_resumes(files: List[Union[str, "gradio.files.TempFile"]], candidate_
         "notes": additional_notes,
     }
 
-    # 9) HF Datasets 保存
+    # 9) HF Datasets 保存（任意）
     dataset_repo = os.environ.get("DATASET_REPO")
     commit_info = None
     if dataset_repo:
-        file_hash = result_json["candidate_id"]
         commit_info = persist_to_hf(
             dataset_repo=dataset_repo,
             record=result_json,
             anon_pdf_bytes=anon_pdf_bytes,
-            parquet_path=f"candidates/{file_hash}.parquet",
-            json_path=f"candidates/{file_hash}.json",
-            pdf_path=f"candidates/{file_hash}.anon.pdf",
+            parquet_path=f"candidates/{cid}.parquet",
+            json_path=f"candidates/{cid}.json",
+            pdf_path=f"candidates/{cid}.anon.pdf",
         )
 
-    anon_pdf = (result_json["candidate_id"] + ".anon.pdf", anon_pdf_bytes)
+    # UI 向け出力を整形
+    anon_pdf = (f"{cid}.anon.pdf", anon_pdf_bytes)
+    out_json_str = json.dumps(result_json, ensure_ascii=False, indent=2)
+    out_skills_str = json.dumps(skills, ensure_ascii=False, indent=2)  # gr.Code で表示
+    out_score_str = json.dumps(score, ensure_ascii=False, indent=2)
+    out_commit_str = json.dumps(commit_info or {"status": "skipped (DATASET_REPO not set)"}, ensure_ascii=False, indent=2)
 
-    # 返却はすべて文字列 or ファイル
     return (
-        _as_json_code(result_json),
-        _as_json_code(skills),
-        _as_json_code(score),
+        out_json_str,
+        out_skills_str,
+        out_score_str,
         summaries.get("300chars", ""),
         summaries.get("100chars", ""),
         summaries.get("onesent", ""),
         anon_pdf,
-        _as_json_code(commit_info or {"status": "skipped (DATASET_REPO not set)"}),
+        out_commit_str,
     )
 
 
-# ---- UI --------------------------------------------------------------------
-
-with gr.Blocks(title=APP_TITLE, analytics_enabled=False) as demo:
+with gr.Blocks(title=APP_TITLE) as demo:
     gr.Markdown(f"# {APP_TITLE}\n複数ファイルを統合→OpenAIで読み込み/構造化/要約→匿名化→Datasets保存")
 
     with gr.Row():
@@ -155,7 +179,7 @@ with gr.Blocks(title=APP_TITLE, analytics_enabled=False) as demo:
             label="レジュメ類 (PDF/画像/Word/テキスト) 複数可",
             file_count="multiple",
             file_types=[".pdf", ".png", ".jpg", ".jpeg", ".tiff", ".bmp", ".docx", ".txt"],
-            type="filepath",  # ← 重要: 'file' はNG
+            type="filepath",  # ★ 重要：'file' ではなく 'filepath'
         )
         candidate_id = gr.Textbox(label="候補者ID（任意。未入力なら自動生成）")
     notes = gr.Textbox(label="補足メモ（任意）", lines=3)
@@ -166,8 +190,7 @@ with gr.Blocks(title=APP_TITLE, analytics_enabled=False) as demo:
         out_json = gr.Code(label="統合出力 (JSON)")
 
     with gr.Tab("抽出スキル"):
-        # gr.JSON は 4.44 でスキーマ事故発生 → Code に置換
-        out_skills = gr.Code(label="スキル一覧 (JSON)")
+        out_skills = gr.Code(label="スキル一覧 (JSON)")  # ★ JSON -> Code に���更
 
     with gr.Tab("品質スコア"):
         out_score = gr.Code(label="品質評価 (JSON)")
@@ -184,14 +207,12 @@ with gr.Blocks(title=APP_TITLE, analytics_enabled=False) as demo:
         out_commit = gr.Code(label="コミット情報 (JSON)")
 
     run_btn.click(
-        process_resumes,
+        fn=process_resumes,
         inputs=[in_files, candidate_id, notes],
         outputs=[out_json, out_skills, out_score, out_sum_300, out_sum_100, out_sum_1, out_pdf, out_commit],
+        api_name="run",
     )
 
-
 if __name__ == "__main__":
-    # ローカル不可な環境では share=True を強制（環境変数で上書き可）
-    share_env = os.environ.get("GRADIO_SHARE", "true").lower()
-    share_flag = share_env in ("1", "true", "yes", "on")
-    demo.launch(server_name="0.0.0.0", server_port=7860, share=share_flag)
+    # 環境により localhost へ到達できない場合があるため share=True を既定に
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=True)