SysOps: voice-agy-webbuilder-backup — 2026-06-23 10:04 UTC

docling_routes.py

@@ -0,0 +1,212 @@
+"""Docling API routes for doc-ingest service."""
+from __future__ import annotations
+
+import os
+from pathlib import Path
+from typing import Any, Optional
+
+from fastapi import APIRouter, HTTPException, Query
+from pydantic import BaseModel, Field
+
+from app.docling_service import capabilities, convert_nas_batch, convert_nas_file, get_result, list_history
+
+router = APIRouter(prefix="/docling", tags=["docling"])
+
+NAS_ROOT = Path(os.getenv("NAS_ROOT", "/nas"))
+
+
+class DoclingConvertBody(BaseModel):
+    path: str = Field(..., min_length=1, description="Relatief NAS-pad")
+    ocr: bool = True
+    tables: bool = True
+    page_images: bool = False
+    picture_images: bool = True
+    force_full_page_ocr: bool = False
+    formats: list[str] = Field(default_factory=lambda: ["markdown", "html", "json", "text"])
+
+
+class DoclingBatchBody(BaseModel):
+    paths: list[str] = Field(default_factory=list)
+    limit: int = Field(default=10, ge=1, le=50)
+    ext: Optional[str] = None
+    ocr: bool = True
+    tables: bool = True
+    page_images: bool = False
+    picture_images: bool = True
+    force_full_page_ocr: bool = False
+    formats: list[str] = Field(default_factory=lambda: ["markdown", "json"])
+
+
+def _options(body: DoclingConvertBody | DoclingBatchBody) -> dict[str, Any]:
+    return {
+        "ocr": body.ocr,
+        "tables": body.tables,
+        "page_images": body.page_images,
+        "picture_images": body.picture_images,
+        "force_full_page_ocr": body.force_full_page_ocr,
+        "formats": body.formats,
+    }
+
+
+@router.get("/capabilities")
+def docling_capabilities() -> dict[str, Any]:
+    return {"ok": True, **capabilities()}
+
+
+@router.post("/convert")
+def docling_convert(body: DoclingConvertBody) -> dict[str, Any]:
+    if not NAS_ROOT.is_dir():
+        raise HTTPException(503, f"NAS not mounted at {NAS_ROOT}")
+    caps = capabilities()
+    if not caps.get("installed"):
+        raise HTTPException(503, "Docling not installed on doc-ingest service")
+    out = convert_nas_file(NAS_ROOT, body.path.lstrip("/"), _options(body))
+    if not out.get("ok"):
+        raise HTTPException(502, out.get("error") or "Conversion failed")
+    return out
+
+
+@router.post("/batch")
+def docling_batch(body: DoclingBatchBody) -> dict[str, Any]:
+    if not NAS_ROOT.is_dir():
+        raise HTTPException(503, f"NAS not mounted at {NAS_ROOT}")
+    caps = capabilities()
+    if not caps.get("installed"):
+        raise HTTPException(503, "Docling not installed on doc-ingest service")
+    paths = list(body.paths or [])
+    if not paths:
+        from app.ingest import scan_share_files
+
+        files = scan_share_files(NAS_ROOT)
+        if body.ext:
+            want = body.ext if body.ext.startswith(".") else f".{body.ext}"
+            files = [f for f in files if f.suffix.lower() == want.lower()]
+        paths = [str(f.relative_to(NAS_ROOT)) for f in files[: body.limit]]
+    if not paths:
+        return {"ok": True, "converted": 0, "total": 0, "results": []}
+    return convert_nas_batch(NAS_ROOT, paths, _options(body))
+
+
+@router.get("/history")
+def docling_history(limit: int = Query(default=30, ge=1, le=100)) -> dict[str, Any]:
+    return {"ok": True, "items": list_history(limit)}
+
+
+@router.get("/result")
+def docling_result(
+    path: str = Query(..., min_length=1),
+    file_sig: Optional[str] = None,
+) -> dict[str, Any]:
+    row = get_result(path.lstrip("/"), file_sig)
+    if not row:
+        raise HTTPException(404, "No Docling result for this file")
+    return {"ok": True, "result": row}
+
+
+class NasWriteBody(BaseModel):
+    path: str = Field(..., min_length=1, description="Bronbestand (voor bestandsnaam)")
+    content: str = ""
+    subdir: str = Field(default="Telegram/Exports", max_length=256)
+
+
+@router.get("/nas/diagnostics")
+def docling_nas_diagnostics() -> dict[str, Any]:
+    if not NAS_ROOT.is_dir():
+        return {"ok": False, "nas_mounted": False, "error": f"NAS not at {NAS_ROOT}"}
+    from app.ingest import scan_share_files, SKIP_DIR_NAMES
+
+    all_files = scan_share_files(NAS_ROOT)
+    folder_counts: dict[str, int] = {}
+    for p in all_files:
+        top = p.relative_to(NAS_ROOT).parts[0] if p.relative_to(NAS_ROOT).parts else "."
+        folder_counts[top] = folder_counts.get(top, 0) + 1
+    empty_tops = []
+    for child in NAS_ROOT.iterdir():
+        if not child.is_dir() or child.name.startswith(("#", ".")) or child.name in SKIP_DIR_NAMES:
+            continue
+        if child.name not in folder_counts:
+            empty_tops.append(child.name)
+    return {
+        "ok": True,
+        "nas_mounted": True,
+        "total_files": len(all_files),
+        "by_top_folder": folder_counts,
+        "empty_folder_trees": empty_tops,
+        "hint": (
+            "Mappen zonder zichtbare bestanden = Synology ACL probleem. "
+            "Geef SMB-gebruiker aissa Read/Write op share + submappen in DSM."
+        ) if empty_tops else "",
+    }
+
+
+TEXT_SUFFIXES = {
+    ".txt", ".md", ".csv", ".json", ".xml", ".html", ".htm", ".yaml", ".yml", ".log", ".rtf",
+}
+IMAGE_SUFFIXES = {".png", ".jpg", ".jpeg", ".webp", ".gif", ".bmp", ".tiff", ".tif"}
+PREVIEW_SUFFIXES = TEXT_SUFFIXES | IMAGE_SUFFIXES | {".pdf"}
+
+
+@router.get("/nas/read")
+def docling_nas_read(path: str = Query(..., min_length=1)) -> dict[str, Any]:
+    if not NAS_ROOT.is_dir():
+        raise HTTPException(503, f"NAS not mounted at {NAS_ROOT}")
+    rel = path.lstrip("/")
+    full = (NAS_ROOT / rel).resolve()
+    if not str(full).startswith(str(NAS_ROOT.resolve())) or not full.is_file():
+        raise HTTPException(404, "File not found")
+    ext = full.suffix.lower()
+    if ext not in TEXT_SUFFIXES and ext not in {".csv", ".rtf"}:
+        raise HTTPException(400, f"Not a plain-text file: {ext}")
+    try:
+        content = full.read_text(encoding="utf-8")
+    except UnicodeDecodeError:
+        content = full.read_text(encoding="latin-1", errors="replace")
+    return {
+        "ok": True,
+        "path": rel,
+        "filename": full.name,
+        "extension": ext,
+        "content": content,
+        "bytes": full.stat().st_size,
+        "editable": True,
+    }
+
+
+@router.get("/nas/file")
+def docling_nas_file(path: str = Query(..., min_length=1)):
+    from fastapi.responses import FileResponse
+
+    if not NAS_ROOT.is_dir():
+        raise HTTPException(503, f"NAS not mounted at {NAS_ROOT}")
+    rel = path.lstrip("/")
+    full = (NAS_ROOT / rel).resolve()
+    if not str(full).startswith(str(NAS_ROOT.resolve())) or not full.is_file():
+        raise HTTPException(404, "File not found")
+    media = "application/octet-stream"
+    ext = full.suffix.lower()
+    if ext == ".pdf":
+        media = "application/pdf"
+    elif ext in IMAGE_SUFFIXES:
+        media = f"image/{ext.lstrip('.')}"
+        if ext == ".jpg":
+            media = "image/jpeg"
+    elif ext == ".html":
+        media = "text/html"
+    return FileResponse(full, media_type=media, filename=full.name)
+
+
+@router.post("/nas/write")
+def docling_nas_write(body: NasWriteBody) -> dict[str, Any]:
+    if not NAS_ROOT.is_dir():
+        raise HTTPException(503, f"NAS not mounted at {NAS_ROOT}")
+    subdir = body.subdir.strip().strip("/")
+    out_dir = (NAS_ROOT / subdir).resolve()
+    if not str(out_dir).startswith(str(NAS_ROOT.resolve())):
+        raise HTTPException(400, "Invalid subdir")
+    out_dir.mkdir(parents=True, exist_ok=True)
+    stem = Path(body.path.lstrip("/")).stem
+    fname = f"{stem}-edited.md"
+    full = out_dir / fname
+    full.write_text(body.content or "", encoding="utf-8")
+    rel = str(full.relative_to(NAS_ROOT))
+    return {"ok": True, "path": rel, "filename": fname, "bytes": len(body.content or "")}