feat: export pipeline outputs (HTML copy + upload archiving) via EXPORT_* envs

README.md

@@ -287,6 +287,9 @@ curl -s -X POST http://localhost:26000/paperless/sync \
 - `PAPERLESS_BASE_URL`, `PAPERLESS_TOKEN`(선택): Paperless API 연동 시 사용
 - `PAPERLESS_VERIFY_SSL`(기본 `true`): Paperless HTTPS 검증 비활성화는 `false`
 - `PAPERLESS_CA_BUNDLE`(선택): 신뢰할 CA 번들 경로 지정 시 해당 번들로 검증
+- `OUTPUT_DIR`(기본 `outputs`): 파이프라인 산출물(HTML) 저장 루트
+- `EXPORT_HTML_DIR`(선택): HTML 산출물 사본을 내보낼 디렉터리(예: 시놀로지 공유 폴더)
+- `EXPORT_UPLOAD_DIR`(선택): 업로드 원본 파일 보관 디렉터리
 - `API_KEY`(선택): 설정 시 모든 민감 엔드포인트 호출에 `X-API-Key` 헤더 필요
 - `CORS_ORIGINS`(선택): CORS 허용 오리진(쉼표 구분), 미설정 시 `*`
 

server/config.py

@@ -13,6 +13,11 @@ class Settings:
     english_ratio_threshold: float = float(os.getenv("ENGLISH_RATIO_THRESHOLD", "0.65"))
     embedding_model: str = os.getenv("EMBEDDING_MODEL", "nomic-embed-text")
     index_path: str = os.getenv("INDEX_PATH", "data/index.jsonl")
+    output_dir: str = os.getenv("OUTPUT_DIR", "outputs")
+
+    # Optional export targets (e.g., Synology NAS shares)
+    export_html_dir: str = os.getenv("EXPORT_HTML_DIR", "")
+    export_upload_dir: str = os.getenv("EXPORT_UPLOAD_DIR", "")
 
     # Paperless (user will provide API details)
     paperless_base_url: str = os.getenv("PAPERLESS_BASE_URL", "")

server/main.py

@@ -4,6 +4,8 @@ from fastapi import FastAPI, HTTPException, Depends, UploadFile, File, Form
 from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
 from typing import List, Dict, Any
+import shutil
+from pathlib import Path
 
 from .config import settings
 from .ollama_client import OllamaClient
@@ -29,7 +31,7 @@ app.add_middleware(
 )
 ollama = OllamaClient(settings.ollama_host)
 index = JsonlIndex(settings.index_path)
-pipeline = DocumentPipeline(ollama, settings.embedding_model, settings.boost_model)
+pipeline = DocumentPipeline(ollama, settings.embedding_model, settings.boost_model, output_dir=settings.output_dir)
 
 
 class ChatRequest(BaseModel):
@@ -178,7 +180,13 @@ def pipeline_ingest(req: PipelineIngestRequest, _: None = Depends(require_api_ke
         summary_sentences=req.summary_sentences,
         summary_language=req.summary_language,
     )
-    return {"status": "ok", "doc_id": result.doc_id, "added": result.added_chunks, "chunks": result.chunks, "html_path": result.html_path}
+    exported_html: str | None = None
+    if result.html_path and settings.export_html_dir:
+        Path(settings.export_html_dir).mkdir(parents=True, exist_ok=True)
+        dst = str(Path(settings.export_html_dir) / Path(result.html_path).name)
+        shutil.copyfile(result.html_path, dst)
+        exported_html = dst
+    return {"status": "ok", "doc_id": result.doc_id, "added": result.added_chunks, "chunks": result.chunks, "html_path": result.html_path, "exported_html": exported_html}
 
 
 @app.post("/pipeline/ingest_file")
@@ -226,7 +234,18 @@ async def pipeline_ingest_file(
         translate=translate,
         target_language=target_language,
     )
-    return {"status": "ok", "doc_id": result.doc_id, "added": result.added_chunks, "chunks": result.chunks, "html_path": result.html_path}
+    exported_html: str | None = None
+    if result.html_path and settings.export_html_dir:
+        Path(settings.export_html_dir).mkdir(parents=True, exist_ok=True)
+        dst = str(Path(settings.export_html_dir) / Path(result.html_path).name)
+        shutil.copyfile(result.html_path, dst)
+        exported_html = dst
+    if settings.export_upload_dir:
+        Path(settings.export_upload_dir).mkdir(parents=True, exist_ok=True)
+        orig_name = f"{doc_id}__{file.filename}"
+        with open(str(Path(settings.export_upload_dir) / orig_name), "wb") as f:
+            f.write(raw)
+    return {"status": "ok", "doc_id": result.doc_id, "added": result.added_chunks, "chunks": result.chunks, "html_path": result.html_path, "exported_html": exported_html}
 
 
 # Paperless webhook placeholder (to be wired with user-provided details)