feat(03-02): replace exception handlers in embeddings and chunker

- Catch openai.APIError in embedding_dimension property
- Use generic message for UnexpectedEmbeddingError
- Catch VLM APITimeoutError and APIError separately in chunker
- Use logger.exception for unexpected errors in contextualize_chunks

Author: paultranvan

openrag/components/indexer/chunker/chunker.py

@@ -68,18 +68,20 @@ async def _generate_context(
                 ]
                 output = await self.context_generator.ainvoke(messages)
                 return output.content
+
             except openai.APITimeoutError:
-                logger.warning(
-                    f"OpenAI API timeout contextualizing chunk after {CONTEXTUALIZATION_TIMEOUT}s",
-                    filename=filename,
-                )
+                # VLM timeout - graceful degradation
+                logger.warning("VLM context generation timeout", timeout=CONTEXTUALIZATION_TIMEOUT)
+                return ""
+
+            except openai.APIError as e:
+                # Other VLM API errors - log but don't fail chunking
+                logger.error("VLM context generation failed", error=str(e))
                 return ""
+
             except Exception as e:
-                logger.warning(
-                    "Error contextualizing chunk of document",
-                    filename=filename,
-                    error=str(e),
-                )
+                # Unexpected errors - log but still gracefully degrade
+                logger.exception("Unexpected error during context generation")
                 return ""
 
     async def contextualize_chunks(
@@ -131,7 +133,7 @@ async def contextualize_chunks(
             ]
 
         except Exception as e:
-            logger.warning(f"Error contextualizing chunks from `{filename}`: {e}")
+            logger.exception("Error contextualizing chunks", filename=filename)
             return chunks
 
 

openrag/components/indexer/embeddings/openai.py

@@ -23,8 +23,18 @@ def embedding_dimension(self) -> int:
             # Test call to get embedding dimension
             output = self.embed_documents([Document(page_content="test")])
             return len(output[0])
-        except Exception:
-            raise
+
+        except openai.APIError as e:
+            logger.error("Failed to get embedding dimension", error=str(e))
+            raise EmbeddingAPIError(f"API error: {e}", model_name=self.embedding_model)
+
+        except (IndexError, AttributeError) as e:
+            logger.error("Invalid embedding response format", error=str(e))
+            raise EmbeddingResponseError("Unexpected response format", error=str(e))
+
+        except Exception as e:
+            logger.exception("Unexpected error getting embedding dimension")
+            raise UnexpectedEmbeddingError("An unexpected error occurred")
 
     def embed_documents(self, texts: list[str | Document]) -> list[list[float]]:
         """
@@ -62,7 +72,7 @@ def embed_documents(self, texts: list[str | Document]) -> list[list[float]]:
         except Exception as e:
             logger.exception("Unexpected error while embedding documents", error=str(e))
             raise UnexpectedEmbeddingError(
-                f"Failed to embed documents: {e!s}",
+                "An unexpected error occurred during document embedding",
                 model_name=self.embedding_model,
                 base_url=self.base_url,
                 error=str(e),