explicit conversion timeout and handling

python/packages/officeconvert/src/officeconvert/conversion.py

@@ -57,13 +57,19 @@ def convert_pptx_to_pdf(pptx_path: Path, pdf_path: Path, *, timeout_s: int = 120
         str(output_dir),
         str(pptx_path.resolve()),
     ]
-    completed = subprocess.run(
-        command,
-        check=False,
-        capture_output=True,
-        text=True,
-        timeout=timeout_s,
-    )
+    try:
+        completed = subprocess.run(
+            command,
+            check=False,
+            capture_output=True,
+            text=True,
+            timeout=timeout_s,
+        )
+    except subprocess.TimeoutExpired as exc:
+        raise RuntimeError(
+            "LibreOffice conversion timed out after "
+            f"{timeout_s} seconds while rendering {pptx_path.name}"
+        ) from exc
     if completed.returncode != 0:
         raise RuntimeError(
             f"LibreOffice conversion failed: {completed.stderr.strip() or completed.stdout.strip()}"
@@ -116,13 +122,20 @@ def render_pdf_to_images(
         str(pdf_path.resolve()),
         str(prefix_path),
     ]
-    completed = subprocess.run(
-        command,
-        check=False,
-        capture_output=True,
-        text=True,
-        timeout=timeout_s,
-    )
+    try:
+        completed = subprocess.run(
+            command,
+            check=False,
+            capture_output=True,
+            text=True,
+            timeout=timeout_s,
+        )
+    except subprocess.TimeoutExpired as exc:
+        raise RuntimeError(
+            "Poppler rasterization timed out after "
+            f"{timeout_s} seconds while rendering {pdf_path.name}; "
+            "increase conversion PDF render timeout or lower image DPI"
+        ) from exc
     if completed.returncode != 0:
         raise RuntimeError(
             f"Poppler rasterization failed: {completed.stderr.strip() or completed.stdout.strip()}"
@@ -165,6 +178,8 @@ def convert_pptx_to_slidedeck(
     *,
     dpi: int = 180,
     image_format: str = "png",
+    pptx_to_pdf_timeout_s: int = 180,
+    pdf_to_images_timeout_s: int = 600,
 ) -> SlideDeckResult:
     """Convert a PPTX into rendered images and extracted notes.
 
@@ -177,6 +192,8 @@ def convert_pptx_to_slidedeck(
         work_dir: Scratch directory for generated outputs.
         dpi: Rasterization DPI for output slide images.
         image_format: Output image format accepted by `pdftoppm`.
+        pptx_to_pdf_timeout_s: Timeout in seconds for the LibreOffice subprocess.
+        pdf_to_images_timeout_s: Timeout in seconds for the Poppler subprocess.
 
     Returns:
         Fully materialized `SlideDeckResult` with local image paths.
@@ -189,12 +206,13 @@ def convert_pptx_to_slidedeck(
     pdf_path = work_dir / f"{pptx_path.stem}.pdf"
     image_dir = work_dir / "slides"
 
-    convert_pptx_to_pdf(pptx_path, pdf_path)
+    convert_pptx_to_pdf(pptx_path, pdf_path, timeout_s=pptx_to_pdf_timeout_s)
     image_paths = render_pdf_to_images(
         pdf_path,
         image_dir,
         dpi=dpi,
         image_format=image_format,
+        timeout_s=pdf_to_images_timeout_s,
     )
     notes = extract_slide_notes(pptx_path)
 

python/packages/server/src/officeconvert_server/config.py

@@ -16,6 +16,9 @@ class ServerConfig:
     s3_secure: bool
     s3_public_endpoint: str
     s3_session_ttl_seconds: int
+    conversion_image_dpi: int
+    conversion_pptx_to_pdf_timeout_seconds: int
+    conversion_pdf_to_images_timeout_seconds: int
     conversion_cleanup_delay_seconds: int
 
 
@@ -28,6 +31,13 @@ def load_server_config() -> ServerConfig:
         s3_secure=os.getenv("S3_USE_SSL", "false").lower() == "true",
         s3_public_endpoint=os.getenv("S3_PUBLIC_ENDPOINT", "localhost:8333"),
         s3_session_ttl_seconds=int(os.getenv("S3_SESSION_TTL_SECONDS", "3600")),
+        conversion_image_dpi=int(os.getenv("CONVERSION_IMAGE_DPI", "150")),
+        conversion_pptx_to_pdf_timeout_seconds=int(
+            os.getenv("CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS", "180")
+        ),
+        conversion_pdf_to_images_timeout_seconds=int(
+            os.getenv("CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS", "600")
+        ),
         conversion_cleanup_delay_seconds=int(
             os.getenv("CONVERSION_CLEANUP_DELAY_SECONDS", "3600")
         ),

python/packages/server/src/officeconvert_server/service.py

@@ -182,6 +182,9 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
                 convert_pptx_to_slidedeck,
                 source_path,
                 work_dir,
+                dpi=self._config.conversion_image_dpi,
+                pptx_to_pdf_timeout_s=self._config.conversion_pptx_to_pdf_timeout_seconds,
+                pdf_to_images_timeout_s=self._config.conversion_pdf_to_images_timeout_seconds,
             )
             session.slide_deck = await asyncio.to_thread(
                 self._upload_and_build_slide_deck,