fix: add document_timeout to PdfPipelineOptions (C1)

Docling's native document_timeout is the only mechanism that can
interrupt processing inside a blocked thread (OCR, table extraction).
Without it, asyncio.wait_for cannot stop a frozen conversion.

Configurable via DOCUMENT_TIMEOUT env var (default: 120s).

Closes #57 (C1)

document-parser/infra/local_converter.py

@@ -42,6 +42,7 @@ from domain.value_objects import (
     PageElement,
 )
 from infra.bbox import to_topleft_list
+from infra.settings import settings
 
 logger = logging.getLogger(__name__)
 
@@ -102,6 +103,7 @@ def _build_docling_converter(options: ConversionOptions) -> DoclingConverter:
         generate_page_images=options.generate_page_images,
         generate_picture_images=options.generate_picture_images,
         images_scale=options.images_scale,
+        document_timeout=settings.document_timeout,
     )
 
     return DoclingConverter(

document-parser/infra/settings.py

@@ -14,6 +14,7 @@ class Settings:
     docling_serve_url: str = "http://localhost:5001"
     docling_serve_api_key: str | None = None
     conversion_timeout: int = 900
+    document_timeout: float = 120.0  # Docling-level per-document timeout (seconds)
     max_concurrent_analyses: int = 3
     max_page_count: int = 0  # 0 = unlimited
     upload_dir: str = "./uploads"
@@ -33,6 +34,7 @@ class Settings:
             docling_serve_url=os.environ.get("DOCLING_SERVE_URL", "http://localhost:5001"),
             docling_serve_api_key=os.environ.get("DOCLING_SERVE_API_KEY"),
             conversion_timeout=int(os.environ.get("CONVERSION_TIMEOUT", "900")),
+            document_timeout=float(os.environ.get("DOCUMENT_TIMEOUT", "120.0")),
             max_concurrent_analyses=int(os.environ.get("MAX_CONCURRENT_ANALYSES", "3")),
             max_page_count=int(os.environ.get("MAX_PAGE_COUNT", "0")),
             upload_dir=os.environ.get("UPLOAD_DIR", "./uploads"),

document-parser/tests/test_pipeline_options.py

@@ -53,6 +53,7 @@ class TestBuildConverter:
         assert opts.generate_page_images is False
         assert opts.generate_picture_images is False
         assert opts.images_scale == 1.0
+        assert opts.document_timeout is not None
 
     def test_ocr_disabled(self):
         conv = build_converter(ConversionOptions(do_ocr=False))

document-parser/tests/test_settings.py

@@ -14,6 +14,7 @@ class TestSettingsDefaults:
         assert s.docling_serve_url == "http://localhost:5001"
         assert s.docling_serve_api_key is None
         assert s.conversion_timeout == 900
+        assert s.document_timeout == 120.0
         assert s.max_page_count == 0
         assert s.upload_dir == "./uploads"
         assert s.db_path == "./data/docling_studio.db"
@@ -36,6 +37,7 @@ class TestSettingsFromEnv:
         monkeypatch.setenv("DOCLING_SERVE_URL", "http://serve:9000")
         monkeypatch.setenv("DOCLING_SERVE_API_KEY", "secret-key")
         monkeypatch.setenv("CONVERSION_TIMEOUT", "120")
+        monkeypatch.setenv("DOCUMENT_TIMEOUT", "60.0")
         monkeypatch.setenv("MAX_PAGE_COUNT", "20")
         monkeypatch.setenv("UPLOAD_DIR", "/data/uploads")
         monkeypatch.setenv("DB_PATH", "/data/test.db")
@@ -49,6 +51,7 @@ class TestSettingsFromEnv:
         assert s.docling_serve_url == "http://serve:9000"
         assert s.docling_serve_api_key == "secret-key"
         assert s.conversion_timeout == 120
+        assert s.document_timeout == 60.0
         assert s.max_page_count == 20
         assert s.upload_dir == "/data/uploads"
         assert s.db_path == "/data/test.db"