fix: raise botocore exceptions (#193)

Motivation:
Old implementation is swallowing botocore exceptions and translating it
into `ValueError` exception which complicates handling different errors,
some botocore exceptions could be just retried in the user code, whereas
for others the user wants to just trigger alert and re-raise or raise
some specific exception. Raising `ValueError` for all botocore
exceptions just make it impossible.

libs/aws/langchain_aws/chat_models/bedrock_converse.py

@@ -705,7 +705,7 @@ def _messages_to_bedrock(
             )
             bedrock_messages.append(curr)
         else:
-            raise ValueError()
+            raise ValueError(f"Unsupported message type {type(msg)}")
     return bedrock_messages, bedrock_system
 
 

libs/aws/langchain_aws/embeddings/bedrock.py

@@ -1,5 +1,6 @@
 import asyncio
 import json
+import logging
 import os
 from typing import Any, Dict, List, Optional
 
@@ -154,7 +155,8 @@ def _embedding_func(self, text: str) -> List[float]:
                 return response_body.get("embedding")
 
         except Exception as e:
-            raise ValueError(f"Error raised by inference endpoint: {e}")
+            logging.error(f"Error raised by inference endpoint: {e}")
+            raise e
 
     def _normalize_vector(self, embeddings: List[float]) -> List[float]:
         """Normalize the embedding to a unit vector."""

libs/aws/langchain_aws/llms/bedrock.py

@@ -1,5 +1,6 @@
 import asyncio
 import json
+import logging
 import os
 import warnings
 from abc import ABC
@@ -788,7 +789,10 @@ def _prepare_input_and_invoke(
             ) = LLMInputOutputAdapter.prepare_output(provider, response).values()
 
         except Exception as e:
-            raise ValueError(f"Error raised by bedrock service: {e}")
+            logging.error(f"Error raised by bedrock service: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
         if stop is not None:
             text = enforce_stop_tokens(text, stop)
@@ -908,7 +912,10 @@ def _prepare_input_and_invoke_stream(
             response = self.client.invoke_model_with_response_stream(**request_options)
 
         except Exception as e:
-            raise ValueError(f"Error raised by bedrock service: {e}")
+            logging.error(f"Error raised by bedrock service: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
         for chunk in LLMInputOutputAdapter.prepare_output_stream(
             provider,

libs/aws/langchain_aws/llms/sagemaker_endpoint.py

@@ -1,6 +1,7 @@
 """Sagemaker InvokeEndpoint API."""
 
 import io
+import logging
 import re
 from abc import abstractmethod
 from typing import Any, Dict, Generic, Iterator, List, Mapping, Optional, TypeVar, Union
@@ -336,7 +337,10 @@ def _stream(
                         run_manager.on_llm_new_token(chunk.text)
 
         except Exception as e:
-            raise ValueError(f"Error raised by streaming inference endpoint: {e}")
+            logging.error(f"Error raised by streaming inference endpoint: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
     def _call(
         self,
@@ -382,7 +386,10 @@ def _call(
                 **_endpoint_kwargs,
             )
         except Exception as e:
-            raise ValueError(f"Error raised by inference endpoint: {e}")
+            logging.error(f"Error raised by inference endpoint: {e}")
+            if run_manager is not None:
+                run_manager.on_llm_error(e)
+            raise e
 
         text = self.content_handler.transform_output(response["Body"])
         if stop is not None: