Merge branch 'development' of github.com:Scale3-Labs/langtrace-python-sdk into development

Author: rohit-kadhe

pyproject.toml

@@ -18,10 +18,10 @@ classifiers=[
     "Operating System :: OS Independent",
 ]
 dependencies = [
-  'trace-attributes==7.0.0',
+  'trace-attributes==7.0.1',
   'opentelemetry-api>=1.25.0',
   'opentelemetry-sdk>=1.25.0',
-  'opentelemetry-instrumentation>=0.46b0',
+  'opentelemetry-instrumentation>=0.47b0',
   'opentelemetry-instrumentation-sqlalchemy>=0.46b0',
   'opentelemetry-exporter-otlp-proto-http>=1.25.0',
   'opentelemetry-exporter-otlp-proto-grpc>=1.25.0',

src/langtrace_python_sdk/constants/instrumentation/common.py

@@ -5,6 +5,7 @@
     "gpt-4-1106-preview": "cl100k_base",
     "gpt-4-1106-vision-preview": "cl100k_base",
     "gpt-4o": "0200k_base",
+    "gpt-4o-mini": "0200k_base",
 }
 
 SERVICE_PROVIDERS = {

src/langtrace_python_sdk/instrumentation/gemini/instrumentation.py

@@ -14,7 +14,7 @@ def instrumentation_dependencies(self) -> Collection[str]:
     def _instrument(self, **kwargs):
         trace_provider = kwargs.get("tracer_provider")
         tracer = get_tracer(__name__, "", trace_provider)
-        version = v("google-cloud-aiplatform")
+        version = v("google-generativeai")
 
         for _, api_config in APIS.items():
             module = api_config.get("module")

src/langtrace_python_sdk/instrumentation/ollama/patch.py

@@ -167,20 +167,21 @@ def _handle_streaming_response(span, response, api):
     accumulated_tokens = None
     if api == "chat":
         accumulated_tokens = {"message": {"content": "", "role": ""}}
-    if api == "completion":
+    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, chunk.get("response") or chunk.get("message").get("content")
-            )
+            set_event_completion_chunk(span, content)
 
         _set_response_attributes(span, chunk | accumulated_tokens)
     finally:
@@ -196,19 +197,22 @@ async def _ahandle_streaming_response(span, response, api):
     accumulated_tokens = None
     if api == "chat":
         accumulated_tokens = {"message": {"content": "", "role": ""}}
-    if api == "completion":
+    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, chunk)
+            set_event_completion_chunk(span, content)
         _set_response_attributes(span, chunk | accumulated_tokens)
     finally:
         # Finalize span after processing all chunks

src/langtrace_python_sdk/version.py

@@ -1 +1 @@
-__version__ = "2.2.12"
+__version__ = "2.2.17"