Create app.py

app.py

@@ -0,0 +1,173 @@
+import os
+import io
+import json
+import hashlib
+import gradio as gr
+
+from pipelines.openai_ingest import (
+    extract_text_with_openai,
+    structure_with_openai,
+    summarize_with_openai,
+)
+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
+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）"
+
+
+def _read_file_bytes(path: str) -> bytes:
+    with open(path, "rb") as f:
+        return f.read()
+
+
+def process_resumes(filepaths, candidate_id: str, additional_notes: str = ""):
+    if not filepaths:
+        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 p in filepaths:
+        raw_bytes = _read_file_bytes(p)
+        fname = os.path.basename(p)
+        filetype = detect_filetype(fname, 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)
+        else:
+            base_text = load_doc_text(filetype, raw_bytes)
+            text = extract_text_with_openai(base_text.encode("utf-8"), filename=fname, filetype="txt")
+
+        raw_texts.append({"filename": fname, "text": text})
+
+        # 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", ""),
+            "education": structured.get("education_raw", ""),
+            "certifications": structured.get("certifications_raw", ""),
+            "skills": ", ".join(structured.get("skills_list", [])),
+        })
+        partial_records.append({
+            "source": fname,
+            "text": text,
+            "structured": structured,
+            "normalized": normalized,
+        })
+
+    # 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
+    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", ""),
+        "education": merged.get("raw_sections", {}).get("education", ""),
+        "certifications": merged.get("raw_sections", {}).get("certifications", ""),
+        "skills": ", ".join(merged.get("skills", [])),
+    })
+
+    # 5) 匿名化
+    anonymized_text, anon_map = anonymize_text(merged_text)
+    anon_pdf_bytes = render_anonymized_pdf(anonymized_text)
+
+    # 6) 品質スコア
+    score = compute_quality_score(merged_text, merged)
+
+    # 7) 要約（300/100/1文）
+    summaries = summarize_with_openai(merged_text)
+
+    # 8) 構造化出力
+    result_json = {
+        "candidate_id": candidate_id or hashlib.sha256(merged_text.encode("utf-8")).hexdigest()[:16],
+        "files": [os.path.basename(p) for p in filepaths],
+        "merged": merged,
+        "skills": skills,
+        "quality_score": score,
+        "summaries": summaries,
+        "anonymization_map": anon_map,
+        "notes": additional_notes,
+    }
+
+    # 9) HF Datasets 保存
+    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",
+        )
+
+    anon_pdf = (result_json["candidate_id"] + ".anon.pdf", anon_pdf_bytes)
+
+    # 出力はすべて文字列/ファイルに統一（GradioのAPI情報生成で安全）
+    return (
+        json.dumps(result_json, ensure_ascii=False, indent=2),
+        json.dumps(skills, ensure_ascii=False, indent=2),
+        json.dumps(score, ensure_ascii=False, indent=2),
+        summaries.get("300chars", ""),
+        summaries.get("100chars", ""),
+        summaries.get("onesent", ""),
+        anon_pdf,
+        json.dumps(commit_info or {"status": "skipped (DATASET_REPO not set)"}, ensure_ascii=False, indent=2),
+    )
+
+
+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():
+        in_files = gr.Files(
+            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：filepath に統一
+        )
+        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)
+
+    run_btn = gr.Button("実行")
+
+    with gr.Tab("構造化JSON"):
+        out_json = gr.Code(label="統合出力 (JSON)")
+
+    with gr.Tab("抽出スキル"):
+        out_skills = gr.Code(label="スキル一覧 (JSON)")
+
+    with gr.Tab("品質スコア"):
+        out_score = gr.Code(label="品質評価 (JSON)")
+
+    with gr.Tab("要約 (300/100/1文)"):
+        out_sum_300 = gr.Textbox(label="300字要約")
+        out_sum_100 = gr.Textbox(label="100字要約")
+        out_sum_1 = gr.Textbox(label="1文要約")
+
+    with gr.Tab("匿名PDF"):
+        out_pdf = gr.File(label="匿名PDFダウンロード")
+
+    with gr.Tab("Datasets 保存ログ"):
+        out_commit = gr.Code(label="コミット情報")
+
+    run_btn.click(
+        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],
+    )
+
+
+if __name__ == "__main__":
+    # Spaces でもローカルでも OK
+    demo.launch(server_name=os.getenv("GRADIO_SERVER_NAME", "0.0.0.0"),
+                server_port=int(os.getenv("GRADIO_SERVER_PORT", "7860")),
+                share=False)