ragaai-catalyst 2.1.7.2b0__py3-none-any.whl → 2.1.7.3__py3-none-any.whl

ragaai_catalyst/tracers/upload_traces.py

@@ -45,6 +45,9 @@ class UploadTraces:
                 else:
                     SCHEMA_MAPPING_NEW[key] = {"columnType": key, "parentColumn": "response"}
 
+        if "error" in additional_metadata_keys and additional_metadata_keys["error"]:
+            SCHEMA_MAPPING_NEW["error"] = {"columnType": "metadata"}
+
         if additional_pipeline_keys:
             for key in additional_pipeline_keys:
                 SCHEMA_MAPPING_NEW[key] = {"columnType": "pipeline"}

ragaai_catalyst/tracers/utils/convert_langchain_callbacks_output.py

@@ -58,4 +58,5 @@ def convert_langchain_callbacks_output(result, project_name="", metadata="", pip
 
     initial_struc[0]["traces"] = traces_data
 
+    initial_struc[0]["error"] = result["error"]
     return initial_struc

ragaai_catalyst/tracers/utils/rag_trace_json_converter.py

@@ -44,7 +44,7 @@ def rag_trace_json_converter(input_trace, custom_model_cost, trace_id, user_deta
                                             if isinstance(message, str):
                                                 human_index = message.find("Human:")
                                                 if human_index != -1:
-                                                    human_message = message[human_index:]
+                                                    human_message = message[human_index:].replace("Human:", "")
                                                     break
                                         return human_message if human_message else value
                                 except Exception as e:
@@ -164,10 +164,30 @@ def rag_trace_json_converter(input_trace, custom_model_cost, trace_id, user_deta
         except Exception as e:
             logger.error(f"Error while extracting context from trace: {str(e)}")
             return ""
-        
+    
+    def get_span_errors(input_trace):
+        try:
+            if tracer_type == "langchain":
+                span_errors = {}
+                for span in input_trace:
+                    try:
+                        if "status" in span.keys() and span.get("status", {}).get("status_code", "").lower() == "error":
+                            span_errors[f"{span['name']}"] = span["status"]
+                    except:
+                        logger.error(f"Error fetching status from span")
+                return span_errors
+        except:
+            logger.error(f"Error in get_span_errors")
+            return None
+
+
+
+
+
     prompt = get_prompt(input_trace)
     response = get_response(input_trace)
     context = get_context(input_trace)
+    error = get_span_errors(input_trace)
     
     if tracer_type == "langchain":
         trace_aggregate["tracer_type"] = "langchain"
@@ -183,6 +203,7 @@ def rag_trace_json_converter(input_trace, custom_model_cost, trace_id, user_deta
     trace_aggregate["data"]["prompt"] = prompt
     trace_aggregate["data"]["response"] = response
     trace_aggregate["data"]["context"] = context
+    trace_aggregate["error"] = error
     
     if tracer_type == "langchain":
         additional_metadata = get_additional_metadata(input_trace, custom_model_cost, model_cost, prompt, response)
@@ -191,6 +212,9 @@ def rag_trace_json_converter(input_trace, custom_model_cost, trace_id, user_deta
     
     trace_aggregate["metadata"] = user_details.get("trace_user_detail", {}).get("metadata")
     trace_aggregate["metadata"].update(additional_metadata)
+    trace_aggregate["metadata"]["error"] = f"{error}"
+    additional_metadata["error"] = error if error else None
+
     additional_metadata.pop("total_cost")
     additional_metadata.pop("total_latency")
     return trace_aggregate, additional_metadata
@@ -212,14 +236,14 @@ def get_additional_metadata(spans, custom_model_cost, model_cost_dict, prompt=""
                     additional_metadata["tokens"]["prompt"] = span["attributes"]["llm.token_count.prompt"]
 
                 except:
-                    logger.warning("Warning: prompt token not found. using fallback strategies to get tokens.")
+                    logger.debug("Warning: prompt token not found. using fallback strategies to get tokens.")
                     try:
                         additional_metadata["tokens"]["prompt"] = num_tokens_from_messages(
                             model=additional_metadata["model_name"],
                             message=prompt
                         )
                     except Exception as e:
-                        logger.warning(f"Failed to count prompt tokens: {str(e)}. Using 'gpt-4o-mini' model count as fallback.")
+                        logger.debug(f"Failed to count prompt tokens: {str(e)}. Using 'gpt-4o-mini' model count as fallback.")
                         additional_metadata["tokens"]["prompt"] = num_tokens_from_messages(
                             model="gpt-4o-mini",
                             message=prompt
@@ -228,14 +252,14 @@ def get_additional_metadata(spans, custom_model_cost, model_cost_dict, prompt=""
                 try:
                     additional_metadata["tokens"]["completion"] = span["attributes"]["llm.token_count.completion"]
                 except:
-                    logger.warning("Warning: completion token not found. using fallback strategies to get tokens.")
+                    logger.debug("Warning: completion token not found. using fallback strategies to get tokens.")
                     try:
                         additional_metadata["tokens"]["completion"] = num_tokens_from_messages(
                             model=additional_metadata["model_name"],
                             message=response
                         )
                     except Exception as e:
-                        logger.warning(f"Failed to count completion tokens: {str(e)}. Using 'gpt-4o-mini' model count as fallback.")
+                        logger.debug(f"Failed to count completion tokens: {str(e)}. Using 'gpt-4o-mini' model count as fallback.")
                         additional_metadata["tokens"]["completion"] = num_tokens_from_messages(
                             model="gpt-4o-mini",
                             message=response
@@ -299,96 +323,47 @@ def get_additional_metadata(spans, custom_model_cost, model_cost_dict, prompt=""
 
 def num_tokens_from_messages(model, message):
     try:
-        # Handle None or empty message
         if not message:
-            logger.warning("Empty or None message provided to token counter")
+            logger.error("Empty or None message provided to token counter")
             return 0
-            
-        # GPT models
-        if re.match(r'^gpt-', model):
-            """Check if the model is any GPT model (pattern: ^gpt-)
-            This matches any model name that starts with 'gpt-'
-            """
-            def num_tokens_from_string(string: str, encoding_name: str) -> int:
-                """Returns the number of tokens in a text string."""
+        
+        def num_tokens_from_string(text_content: str, encoding_name: str) -> int:
+            """Returns the number of tokens in a text string."""
+            if isinstance(text_content, list):
+                list_str = str(text_content[0]) if text_content else ""
+                pattern = r"content=\'(.*?)\'(?:\s+additional_kwargs=|$)"
+                match = re.search(pattern, list_str, re.DOTALL)
+                if match:
+                    text_content = match.group(1)  # Extract content and process it for tokens
+                else:
+                    text_content = list_str 
+            try:
+                encoding = tiktoken.get_encoding(encoding_name)
+                return len(encoding.encode(text_content))
+            except Exception as e:
+                logger.warning(f"Error encoding with {encoding_name}: {str(e)}")
                 try:
-                    encoding = tiktoken.get_encoding(encoding_name)
-                    num_tokens = len(encoding.encode(string))
-                    return num_tokens
-                except Exception as e:
-                    logger.warning(f"Error encoding with {encoding_name}: {str(e)}")
-                    # Fallback to a different encoding if the requested one fails
-                    try:
-                        fallback_encoding = tiktoken.get_encoding("cl100k_base")
-                        return len(fallback_encoding.encode(string))
-                    except:
-                        logger.error("Failed to use fallback encoding")
-                        return 0
-            
-            if re.match(r'^gpt-4o.*', model):
-                """Check for GPT-4 Optimized models (pattern: ^gpt-4o.*)
-                Examples that match:
-                - gpt-4o
-                - gpt-4o-mini
-                - gpt-4o-2024-08-06
-                The .* allows for any characters after 'gpt-4o'
-                """
+                    fallback_encoding = tiktoken.get_encoding("cl100k_base")
+                    return len(fallback_encoding.encode(text_content))
+                except:
+                    logger.debug("Failed to use fallback encoding")
+                    return 0
+        
+        # Determine which encoding to use based on model name
+        encoding_name = "o200k_base"  
+        
+        if re.match(r'^gpt-', model):
+            if re.match(r'^gpt-(4o|4\.1).*', model):
+                # GPT-4o and GPT-4.1 models
                 encoding_name = "o200k_base"
-                return num_tokens_from_string(message, encoding_name)
-            
             elif re.match(r'^gpt-(4|3\.5).*', model):
-                """Check for GPT-4 and GPT-3.5 models (pattern: ^gpt-(4|3\.5).*)
-                Uses cl100k_base encoding for GPT-4 and GPT-3.5 models
-                Examples that match:
-                - gpt-4
-                - gpt-4-turbo
-                - gpt-4-2024-08-06
-                - gpt-3.5-turbo
-                - gpt-3.5-turbo-16k
-                """
+                # GPT-4 and GPT-3.5 models
                 encoding_name = "cl100k_base"
-                return num_tokens_from_string(message, encoding_name)
+        else:
+            logger.debug(f"Using default token counter for: {model}.")
             
-            else:
-                """Default case for any other GPT models
-                Uses o200k_base encoding as the default tokenizer
-                """
-                return num_tokens_from_string(message, encoding_name="o200k_base")
+        return num_tokens_from_string(message, encoding_name)
             
-
-        # Gemini models 
-        elif re.match(r'^gemini-', model):
-            try:
-                GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
-                if not GOOGLE_API_KEY:
-                    logger.warning("GOOGLE_API_KEY not found in environment variables")
-                    return 0
-                    
-                import google.generativeai as genai
-                client = genai.Client(api_key=GOOGLE_API_KEY)
-
-                response = client.models.count_tokens(
-                        model=model,
-                        contents=message,
-                    )
-                return response.total_tokens
-            except ImportError:
-                logger.warning("google.generativeai module not found. Install with pip install google-generativeai")
-                return 0
-            except Exception as e:
-                logger.warning(f"Error counting tokens for Gemini model: {str(e)}")
-                return 0
-        
-        # Default case for unknown models
-        else:
-            logger.warning(f"Unknown model type: {model}. Using default token counter.")
-            try:
-                # Use cl100k_base as a fallback for unknown models
-                encoding = tiktoken.get_encoding("cl100k_base")
-                return len(encoding.encode(message))
-            except:
-                logger.error("Failed to use fallback encoding for unknown model")
-                return 0
     except Exception as e:
         logger.error(f"Unexpected error in token counting: {str(e)}")
         return 0

{ragaai_catalyst-2.1.7.2b0.dist-info → ragaai_catalyst-2.1.7.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ragaai_catalyst
-Version: 2.1.7.2b0
+Version: 2.1.7.3
 Summary: RAGA AI CATALYST
 Author-email: Kiran Scaria <kiran.scaria@raga.ai>, Kedar Gaikwad <kedar.gaikwad@raga.ai>, Dushyant Mahajan <dushyant.mahajan@raga.ai>, Siddhartha Kosti <siddhartha.kosti@raga.ai>, Ritika Goel <ritika.goel@raga.ai>, Vijay Chaurasia <vijay.chaurasia@raga.ai>, Tushar Kumar <tushar.kumar@raga.ai>
 Requires-Python: <=3.13.2,>=3.10

{ragaai_catalyst-2.1.7.2b0.dist-info → ragaai_catalyst-2.1.7.3.dist-info}/RECORD

@@ -32,7 +32,7 @@ ragaai_catalyst/tracers/langchain_callback.py,sha256=CB75zzG3-DkYTELj0vI1MOHQTY0
 ragaai_catalyst/tracers/llamaindex_callback.py,sha256=ZY0BJrrlz-P9Mg2dX-ZkVKG3gSvzwqBtk7JL_05MiYA,14028
 ragaai_catalyst/tracers/llamaindex_instrumentation.py,sha256=Ys_jLkvVqo12bKgXDmkp4TxJu9HkBATrFE8cIcTYxWw,14329
 ragaai_catalyst/tracers/tracer.py,sha256=LHRd35rTFgYovxsI4tyXHgsCCdccO1_O1txFLFyLUKc,39069
-ragaai_catalyst/tracers/upload_traces.py,sha256=PEE_JhAmOAMKyb-pl4ZoFWhIePxJm1zs93crrk94iEg,5887
+ragaai_catalyst/tracers/upload_traces.py,sha256=xVDciX7iYtxasehKISYsXmKR_-pnB-fRGlszSnBkqtc,6043
 ragaai_catalyst/tracers/agentic_tracing/README.md,sha256=X4QwLb7-Jg7GQMIXj-SerZIgDETfw-7VgYlczOR8ZeQ,4508
 ragaai_catalyst/tracers/agentic_tracing/__init__.py,sha256=yf6SKvOPSpH-9LiKaoLKXwqj5sez8F_5wkOb91yp0oE,260
 ragaai_catalyst/tracers/agentic_tracing/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -80,16 +80,16 @@ ragaai_catalyst/tracers/exporters/raga_exporter.py,sha256=6xvjWXyh8XPkHKSLLmAZUQ
 ragaai_catalyst/tracers/exporters/ragaai_trace_exporter.py,sha256=BzL5u7yOcWHHtPueoJclXcvexBQo1SFu5lHY8ANnzP8,12998
 ragaai_catalyst/tracers/instrumentators/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 ragaai_catalyst/tracers/utils/__init__.py,sha256=KeMaZtYaTojilpLv65qH08QmpYclfpacDA0U3wg6Ybw,64
-ragaai_catalyst/tracers/utils/convert_langchain_callbacks_output.py,sha256=e0URuRWCdzpxuBLfL82FOTMjbRuDAkW8aIRi7s7Nocc,1655
+ragaai_catalyst/tracers/utils/convert_langchain_callbacks_output.py,sha256=SehrD7q8ytAiUYoWr406b4mWs3Lk0Rcy6Ekkihh22TI,1703
 ragaai_catalyst/tracers/utils/convert_llama_instru_callback.py,sha256=8qLo7x4Zsn3dhJfSv9gviB60YXZ2TOsWEouucJmBM0c,1724
 ragaai_catalyst/tracers/utils/extraction_logic_llama_index.py,sha256=ZhPs0YhVtB82-Pq9o1BvCinKE_WPvVxPTEcZjlJbFYM,2371
 ragaai_catalyst/tracers/utils/langchain_tracer_extraction_logic.py,sha256=XS2_x2qneqEx9oAighLg-LRiueWcESLwIC2r7eJT-Ww,3117
 ragaai_catalyst/tracers/utils/model_prices_and_context_window_backup.json,sha256=C3uwkibJ08C9sOX-54kulZYmJlIpZ-SQpfE6HNGrjbM,343502
-ragaai_catalyst/tracers/utils/rag_trace_json_converter.py,sha256=Y1xWS9aPBUt4cg1Rsgg6m3K6B9LuWBp2BXPliCmGk3E,20615
+ragaai_catalyst/tracers/utils/rag_trace_json_converter.py,sha256=adCKk7Nj8307XYYg2sB-QT-66OShOs2iTGwNVwqbHig,19373
 ragaai_catalyst/tracers/utils/trace_json_converter.py,sha256=E0_QfciQMMpCtQYrNB4l8HJhlaFalr5bkMqkVRgQahY,14073
 ragaai_catalyst/tracers/utils/utils.py,sha256=ViygfJ7vZ7U0CTSA1lbxVloHp4NSlmfDzBRNCJuMhis,2374
-ragaai_catalyst-2.1.7.2b0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-ragaai_catalyst-2.1.7.2b0.dist-info/METADATA,sha256=gZkHYlK9kJaM3_kNJtTQzFbIYV1T18LokNMGtqdwHzQ,17607
-ragaai_catalyst-2.1.7.2b0.dist-info/WHEEL,sha256=pxyMxgL8-pra_rKaQ4drOZAegBVuX-G_4nRHjjgWbmo,91
-ragaai_catalyst-2.1.7.2b0.dist-info/top_level.txt,sha256=HpgsdRgEJMk8nqrU6qdCYk3di7MJkDL0B19lkc7dLfM,16
-ragaai_catalyst-2.1.7.2b0.dist-info/RECORD,,
+ragaai_catalyst-2.1.7.3.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+ragaai_catalyst-2.1.7.3.dist-info/METADATA,sha256=mQ05LSLjmZkiOkRHe2SmSL0SmysTeZWOECduUWhQPuQ,17605
+ragaai_catalyst-2.1.7.3.dist-info/WHEEL,sha256=ck4Vq1_RXyvS4Jt6SI0Vz6fyVs4GWg7AINwpsaGEgPE,91
+ragaai_catalyst-2.1.7.3.dist-info/top_level.txt,sha256=HpgsdRgEJMk8nqrU6qdCYk3di7MJkDL0B19lkc7dLfM,16
+ragaai_catalyst-2.1.7.3.dist-info/RECORD,,

{ragaai_catalyst-2.1.7.2b0.dist-info → ragaai_catalyst-2.1.7.3.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (79.0.0)
+Generator: setuptools (80.0.0)
 Root-Is-Purelib: true
 Tag: py3-none-any