Add Support for CrewAI using Ollama

src/langtrace_python_sdk/instrumentation/ollama/patch.py

@@ -1,6 +1,7 @@
 from langtrace_python_sdk.constants.instrumentation.ollama import APIS
 from langtrace_python_sdk.utils import set_span_attribute
 from langtrace_python_sdk.utils.llm import (
+    StreamWrapper,
     get_extra_attributes,
     get_langtrace_attributes,
     get_llm_request_attributes,
@@ -16,9 +17,10 @@
 import json
 from opentelemetry.trace.status import Status, StatusCode
 from langtrace.trace_attributes import SpanAttributes
+from opentelemetry.trace import Tracer
 
 
-def generic_patch(operation_name, version, tracer):
+def generic_patch(operation_name, version, tracer: Tracer):
     def traced_method(wrapped, instance, args, kwargs):
         api = APIS[operation_name]
         service_provider = SERVICE_PROVIDERS["OLLAMA"]
@@ -35,36 +37,29 @@ def traced_method(wrapped, instance, args, kwargs):
         }
 
         attributes = LLMSpanAttributes(**span_attributes)
-        with tracer.start_as_current_span(
-            name=get_span_name(f'ollama.{api["METHOD"]}'), kind=SpanKind.CLIENT
-        ) as span:
-            _set_input_attributes(span, kwargs, attributes)
-
-            try:
-                result = wrapped(*args, **kwargs)
-                if result:
-                    if span.is_recording():
-
-                        if kwargs.get("stream"):
-                            return _handle_streaming_response(
-                                span, result, api["METHOD"]
-                            )
 
-                        _set_response_attributes(span, result)
-                        span.set_status(Status(StatusCode.OK))
+        span = tracer.start_span(
+            name=get_span_name(f'ollama.{api["METHOD"]}'), kind=SpanKind.CLIENT
+        )
+        _set_input_attributes(span, kwargs, attributes)
 
-                span.end()
-                return result
+        try:
+            result = wrapped(*args, **kwargs)
+            if kwargs.get("stream"):
+                return StreamWrapper(result, span)
+            else:
+                _set_response_attributes(span, result)
+            return result
 
-            except Exception as err:
-                # Record the exception in the span
-                span.record_exception(err)
+        except Exception as err:
+            # Record the exception in the span
+            span.record_exception(err)
 
-                # Set the span status to indicate an error
-                span.set_status(Status(StatusCode.ERROR, str(err)))
+            # Set the span status to indicate an error
+            span.set_status(Status(StatusCode.ERROR, str(err)))
 
-                # Reraise the exception to ensure it's not swallowed
-                raise
+            # Reraise the exception to ensure it's not swallowed
+            raise
 
     return traced_method
 
@@ -82,30 +77,28 @@ async def traced_method(wrapped, instance, args, kwargs):
             **get_extra_attributes(),
         }
         attributes = LLMSpanAttributes(**span_attributes)
-        with tracer.start_as_current_span(api["METHOD"], kind=SpanKind.CLIENT) as span:
-            _set_input_attributes(span, kwargs, attributes)
-            try:
-                result = await wrapped(*args, **kwargs)
-                if result:
-                    if span.is_recording():
-                        if kwargs.get("stream"):
-                            return _ahandle_streaming_response(
-                                span, result, api["METHOD"]
-                            )
-
-                        _set_response_attributes(span, result)
-                        span.set_status(Status(StatusCode.OK))
-                span.end()
-                return result
-            except Exception as err:
-                # Record the exception in the span
-                span.record_exception(err)
-
-                # Set the span status to indicate an error
-                span.set_status(Status(StatusCode.ERROR, str(err)))
-
-                # Reraise the exception to ensure it's not swallowed
-                raise
+        span = tracer.start_span(
+            name=get_span_name(f'ollama.{api["METHOD"]}'), kind=SpanKind.CLIENT
+        )
+
+        _set_input_attributes(span, kwargs, attributes)
+        try:
+            result = await wrapped(*args, **kwargs)
+            if kwargs.get("stream"):
+                return StreamWrapper(span, result)
+            else:
+                _set_response_attributes(span, result)
+            span.end()
+            return result
+        except Exception as err:
+            # Record the exception in the span
+            span.record_exception(err)
+
+            # Set the span status to indicate an error
+            span.set_status(Status(StatusCode.ERROR, str(err)))
+
+            # Reraise the exception to ensure it's not swallowed
+            raise
 
     return traced_method
 
@@ -162,63 +155,3 @@ def _set_input_attributes(span, kwargs, attributes):
             SpanAttributes.LLM_PRESENCE_PENALTY,
             options.get("presence_penalty"),
         )
-
-
-def _handle_streaming_response(span, response, api):
-    accumulated_tokens = None
-    if api == "chat":
-        accumulated_tokens = {"message": {"content": "", "role": ""}}
-    if api == "completion" or api == "generate":
-        accumulated_tokens = {"response": ""}
-    span.add_event(Event.STREAM_START.value)
-    try:
-        for chunk in response:
-            content = None
-            if api == "chat":
-                content = chunk["message"]["content"]
-                accumulated_tokens["message"]["content"] += chunk["message"]["content"]
-                accumulated_tokens["message"]["role"] = chunk["message"]["role"]
-            if api == "generate":
-                content = chunk["response"]
-                accumulated_tokens["response"] += chunk["response"]
-
-            set_event_completion_chunk(span, content)
-
-        _set_response_attributes(span, chunk | accumulated_tokens)
-    finally:
-        # Finalize span after processing all chunks
-        span.add_event(Event.STREAM_END.value)
-        span.set_status(StatusCode.OK)
-        span.end()
-
-    return response
-
-
-async def _ahandle_streaming_response(span, response, api):
-    accumulated_tokens = None
-    if api == "chat":
-        accumulated_tokens = {"message": {"content": "", "role": ""}}
-    if api == "completion" or api == "generate":
-        accumulated_tokens = {"response": ""}
-
-    span.add_event(Event.STREAM_START.value)
-    try:
-        async for chunk in response:
-            content = None
-            if api == "chat":
-                content = chunk["message"]["content"]
-                accumulated_tokens["message"]["content"] += chunk["message"]["content"]
-                accumulated_tokens["message"]["role"] = chunk["message"]["role"]
-            if api == "generate":
-                content = chunk["response"]
-                accumulated_tokens["response"] += chunk["response"]
-
-            set_event_completion_chunk(span, content)
-        _set_response_attributes(span, chunk | accumulated_tokens)
-    finally:
-        # Finalize span after processing all chunks
-        span.add_event(Event.STREAM_END.value)
-        span.set_status(StatusCode.OK)
-        span.end()
-
-    return response

src/langtrace_python_sdk/utils/llm.py

@@ -237,7 +237,6 @@ class StreamWrapper:
     def __init__(
         self, stream, span, prompt_tokens=0, function_call=False, tool_calls=False
     ):
-
         self.stream = stream
         self.span = span
         self.prompt_tokens = prompt_tokens
@@ -284,7 +283,6 @@ def cleanup(self):
                     }
                 ],
             )
-
             self.span.set_status(StatusCode.OK)
             self.span.end()
             self._span_started = False
@@ -377,6 +375,10 @@ def build_streaming_response(self, chunk):
         if hasattr(chunk, "delta") and chunk.delta is not None:
             content = [chunk.delta.text] if hasattr(chunk.delta, "text") else []
 
+        if isinstance(chunk, dict):
+            if "message" in chunk:
+                if "content" in chunk["message"]:
+                    content = [chunk["message"]["content"]]
         if content:
             self.result_content.append(content[0])
 
@@ -401,6 +403,13 @@ def set_usage_attributes(self, chunk):
             self.completion_tokens = chunk.usage_metadata.candidates_token_count
             self.prompt_tokens = chunk.usage_metadata.prompt_token_count
 
+        # Ollama
+        if isinstance(chunk, dict):
+            if "prompt_eval_count" in chunk:
+                self.prompt_tokens = chunk["prompt_eval_count"]
+            if "eval_count" in chunk:
+                self.completion_tokens = chunk["eval_count"]
+
     def process_chunk(self, chunk):
         self.set_response_model(chunk=chunk)
         self.build_streaming_response(chunk=chunk)

src/langtrace_python_sdk/version.py

@@ -1 +1 @@
-__version__ = "2.2.21"
+__version__ = "2.2.22"