feat: document pipeline (embedding->Korean translation->HTML). Add /pipeline/ingest endpoint

server/main.py

@@ -11,6 +11,7 @@ from .index_store import JsonlIndex
 from .security import require_api_key
 from .paperless_client import PaperlessClient
 from .utils import chunk_text
+from .pipeline import DocumentPipeline
 
 
 app = FastAPI(title="Local AI Server", version="0.2.1")
@@ -28,6 +29,7 @@ app.add_middleware(
 )
 ollama = OllamaClient(settings.ollama_host)
 index = JsonlIndex(settings.index_path)
+pipeline = DocumentPipeline(ollama, settings.embedding_model, settings.boost_model)
 
 
 class ChatRequest(BaseModel):
@@ -55,6 +57,12 @@ class UpsertRequest(BaseModel):
     batch: int = 16
 
 
+class PipelineIngestRequest(BaseModel):
+    doc_id: str
+    text: str
+    generate_html: bool = True
+
+
 @app.get("/health")
 def health() -> Dict[str, Any]:
     return {
@@ -152,6 +160,12 @@ def index_reload() -> Dict[str, Any]:
     return {"total": total}
 
 
+@app.post("/pipeline/ingest")
+def pipeline_ingest(req: PipelineIngestRequest, _: None = Depends(require_api_key)) -> Dict[str, Any]:
+    result = pipeline.process(doc_id=req.doc_id, text=req.text, index=index, generate_html=req.generate_html)
+    return {"status": "ok", "doc_id": result.doc_id, "added": result.added_chunks, "chunks": result.chunks, "html_path": result.html_path}
+
+
 # Paperless webhook placeholder (to be wired with user-provided details)
 class PaperlessHook(BaseModel):
     document_id: int