fix(document): catch httpx timeout from gateway OCR backend (#892 r3)

Round-3 review on PR #892 found a real bug: the gateway backend's httpx.Timeout
raises httpx.ReadTimeout (a httpx.TimeoutException, NOT a builtin TimeoutError),
so the `except TimeoutError` added in r2 only covered the Mistral
(anyio.fail_after) path — gateway timeouts still fell through to
reason="error". Catch both (TimeoutError, httpx.TimeoutException) so either
backend's timeout lands in the dedicated parse_failed_reason="timeout" bucket.
Add an end-to-end test driving a gateway httpx.ReadTimeout through the
processor.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

nextcloud_mcp_server/document_processors/ocr.py

@@ -264,11 +264,13 @@ class OcrProcessor(DocumentProcessor):
             text, boundaries = await backend.ocr(
                 content, content_type.split(";")[0].strip().lower()
             )
-        except TimeoutError:
-            # anyio.fail_after / httpx read-timeout raise TimeoutError with an
-            # empty message; give it its own reason bucket and a useful log so a
-            # too-low DOCUMENT_OCR_TIMEOUT_SECONDS is distinguishable from a
-            # provider that's actually erroring.
+        except (TimeoutError, httpx.TimeoutException):
+            # Two timeout shapes reach here: the Mistral backend's
+            # anyio.fail_after raises the builtin TimeoutError, while the gateway
+            # backend's httpx.Timeout raises httpx.ReadTimeout (a
+            # httpx.TimeoutException, NOT a TimeoutError). Catch both so a
+            # too-low DOCUMENT_OCR_TIMEOUT_SECONDS lands in its own reason bucket
+            # rather than being conflated with provider errors.
             timeout = settings.document_ocr_timeout_seconds
             logger.warning(
                 "OCR timed out for %s after %.1fs", filename or "<bytes>", timeout

tests/unit/test_ocr_processor.py

@@ -149,6 +149,25 @@ async def test_processor_timeout_returns_timeout_reason(monkeypatch):
     assert "timed out" in r.error
 
 
+async def test_gateway_httpx_timeout_maps_to_timeout_reason(monkeypatch):
+    """A gateway httpx.ReadTimeout (not a builtin TimeoutError) must still map to
+    parse_failed_reason='timeout', not 'error'."""
+    import httpx
+
+    class _HttpxTimeoutBackend:
+        async def ocr(self, content, mime_type):
+            raise httpx.ReadTimeout("read timed out")
+
+    monkeypatch.setattr(
+        ocr, "get_settings", lambda: _settings(document_ocr_timeout_seconds=5.0)
+    )
+    monkeypatch.setattr(ocr, "build_ocr_backend", lambda s: _HttpxTimeoutBackend())
+    r = await ocr.OcrProcessor().process(b"%PDF-1.7", "application/pdf")
+    assert r.success is False
+    assert r.metadata["parse_failed_reason"] == "timeout"
+    assert "timed out" in r.error
+
+
 async def test_gateway_backend_uses_configured_timeout(mocker, monkeypatch):
     """The gateway OCR call must use DOCUMENT_OCR_TIMEOUT_SECONDS (resolved per
     call), not the old hardcoded 180s constant."""