openlit 1.33.8__py3-none-any.whl → 1.33.9__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (78) hide show
  1. openlit/__helpers.py +83 -0
  2. openlit/__init__.py +1 -1
  3. openlit/instrumentation/ag2/ag2.py +2 -2
  4. openlit/instrumentation/ai21/__init__.py +4 -4
  5. openlit/instrumentation/ai21/ai21.py +370 -319
  6. openlit/instrumentation/ai21/async_ai21.py +371 -319
  7. openlit/instrumentation/anthropic/__init__.py +4 -4
  8. openlit/instrumentation/anthropic/anthropic.py +321 -189
  9. openlit/instrumentation/anthropic/async_anthropic.py +323 -190
  10. openlit/instrumentation/assemblyai/__init__.py +1 -1
  11. openlit/instrumentation/assemblyai/assemblyai.py +59 -43
  12. openlit/instrumentation/astra/astra.py +4 -4
  13. openlit/instrumentation/astra/async_astra.py +4 -4
  14. openlit/instrumentation/azure_ai_inference/__init__.py +4 -4
  15. openlit/instrumentation/azure_ai_inference/async_azure_ai_inference.py +406 -252
  16. openlit/instrumentation/azure_ai_inference/azure_ai_inference.py +406 -252
  17. openlit/instrumentation/bedrock/__init__.py +1 -1
  18. openlit/instrumentation/bedrock/bedrock.py +115 -58
  19. openlit/instrumentation/chroma/chroma.py +4 -4
  20. openlit/instrumentation/cohere/__init__.py +33 -10
  21. openlit/instrumentation/cohere/async_cohere.py +610 -0
  22. openlit/instrumentation/cohere/cohere.py +410 -219
  23. openlit/instrumentation/controlflow/controlflow.py +2 -2
  24. openlit/instrumentation/crawl4ai/async_crawl4ai.py +2 -2
  25. openlit/instrumentation/crawl4ai/crawl4ai.py +2 -2
  26. openlit/instrumentation/crewai/crewai.py +2 -2
  27. openlit/instrumentation/dynamiq/dynamiq.py +2 -2
  28. openlit/instrumentation/elevenlabs/async_elevenlabs.py +73 -47
  29. openlit/instrumentation/elevenlabs/elevenlabs.py +73 -52
  30. openlit/instrumentation/embedchain/embedchain.py +4 -4
  31. openlit/instrumentation/firecrawl/firecrawl.py +2 -2
  32. openlit/instrumentation/google_ai_studio/__init__.py +9 -9
  33. openlit/instrumentation/google_ai_studio/async_google_ai_studio.py +183 -219
  34. openlit/instrumentation/google_ai_studio/google_ai_studio.py +183 -220
  35. openlit/instrumentation/gpt4all/gpt4all.py +17 -17
  36. openlit/instrumentation/groq/async_groq.py +14 -14
  37. openlit/instrumentation/groq/groq.py +14 -14
  38. openlit/instrumentation/haystack/haystack.py +2 -2
  39. openlit/instrumentation/julep/async_julep.py +2 -2
  40. openlit/instrumentation/julep/julep.py +2 -2
  41. openlit/instrumentation/langchain/langchain.py +36 -31
  42. openlit/instrumentation/letta/letta.py +6 -6
  43. openlit/instrumentation/litellm/async_litellm.py +20 -20
  44. openlit/instrumentation/litellm/litellm.py +20 -20
  45. openlit/instrumentation/llamaindex/llamaindex.py +2 -2
  46. openlit/instrumentation/mem0/mem0.py +2 -2
  47. openlit/instrumentation/milvus/milvus.py +4 -4
  48. openlit/instrumentation/mistral/async_mistral.py +18 -18
  49. openlit/instrumentation/mistral/mistral.py +18 -18
  50. openlit/instrumentation/multion/async_multion.py +2 -2
  51. openlit/instrumentation/multion/multion.py +2 -2
  52. openlit/instrumentation/ollama/async_ollama.py +29 -29
  53. openlit/instrumentation/ollama/ollama.py +29 -29
  54. openlit/instrumentation/openai/__init__.py +11 -230
  55. openlit/instrumentation/openai/async_openai.py +434 -409
  56. openlit/instrumentation/openai/openai.py +415 -393
  57. openlit/instrumentation/phidata/phidata.py +2 -2
  58. openlit/instrumentation/pinecone/pinecone.py +4 -4
  59. openlit/instrumentation/premai/premai.py +20 -20
  60. openlit/instrumentation/qdrant/async_qdrant.py +4 -4
  61. openlit/instrumentation/qdrant/qdrant.py +4 -4
  62. openlit/instrumentation/reka/async_reka.py +6 -6
  63. openlit/instrumentation/reka/reka.py +6 -6
  64. openlit/instrumentation/together/async_together.py +18 -18
  65. openlit/instrumentation/together/together.py +18 -18
  66. openlit/instrumentation/transformers/transformers.py +6 -6
  67. openlit/instrumentation/vertexai/async_vertexai.py +53 -53
  68. openlit/instrumentation/vertexai/vertexai.py +53 -53
  69. openlit/instrumentation/vllm/vllm.py +6 -6
  70. openlit/otel/metrics.py +98 -7
  71. openlit/semcov/__init__.py +113 -80
  72. {openlit-1.33.8.dist-info → openlit-1.33.9.dist-info}/METADATA +1 -1
  73. openlit-1.33.9.dist-info/RECORD +121 -0
  74. {openlit-1.33.8.dist-info → openlit-1.33.9.dist-info}/WHEEL +1 -1
  75. openlit/instrumentation/openai/async_azure_openai.py +0 -900
  76. openlit/instrumentation/openai/azure_openai.py +0 -898
  77. openlit-1.33.8.dist-info/RECORD +0 -122
  78. {openlit-1.33.8.dist-info → openlit-1.33.9.dist-info}/LICENSE +0 -0
openlit/instrumentation/google_ai_studio/google_ai_studio.py CHANGED
@@ -1,21 +1,24 @@
1
- # pylint: disable=duplicate-code, broad-exception-caught, too-many-statements, unused-argument, possibly-used-before-assignment, protected-access
2
1
  """
3
2
  Module for monitoring Google AI Studio API calls.
4
3
  """
5
4
 
6
5
  import logging
6
+ import time
7
7
  from opentelemetry.trace import SpanKind, Status, StatusCode
8
- from opentelemetry.sdk.resources import TELEMETRY_SDK_NAME
8
+ from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
9
9
  from openlit.__helpers import (
10
- handle_exception,
11
10
  get_chat_model_cost,
11
+ handle_exception,
12
+ response_as_dict,
13
+ create_metrics_attributes,
14
+ set_server_address_and_port
12
15
  )
13
16
  from openlit.semcov import SemanticConvetion
14
17
 
15
18
  # Initialize logger for logging potential issues and operations
16
19
  logger = logging.getLogger(__name__)
17
20
 
18
- def generate(gen_ai_endpoint, version, environment, application_name,
21
+ def generate(version, environment, application_name,
19
22
  tracer, pricing_info, trace_content, metrics, disable_metrics):
20
23
  """
21
24
  Generates a telemetry wrapper for chat to collect metrics.
@@ -35,230 +38,190 @@ def generate(gen_ai_endpoint, version, environment, application_name,
35
38
 
36
39
  def wrapper(wrapped, instance, args, kwargs):
37
40
  """
38
- Wraps the 'chat' API call to add telemetry.
39
-
41
+ Wraps the 'chat.completions' API call to add telemetry.
42
+
40
43
  This collects metrics such as execution time, cost, and token usage, and handles errors
41
44
  gracefully, adding details to the trace for observability.
42
45
 
43
46
  Args:
44
- wrapped: The original 'chat' method to be wrapped.
47
+ wrapped: The original 'chat.completions' method to be wrapped.
45
48
  instance: The instance of the class where the original method is defined.
46
- args: Positional arguments for the 'chat' method.
47
- kwargs: Keyword arguments for the 'chat' method.
49
+ args: Positional arguments for the 'chat.completions' method.
50
+ kwargs: Keyword arguments for the 'chat.completions' method.
48
51
 
49
52
  Returns:
50
- The response from the original 'chat' method.
53
+ The response from the original 'chat.completions' method.
51
54
  """
52
- # pylint: disable=no-else-return
53
- if kwargs.get("stream", False) is True:
54
- # Special handling for streaming response to accommodate the nature of data flow
55
- def stream_generator():
56
- with tracer.start_as_current_span(gen_ai_endpoint, kind= SpanKind.CLIENT) as span:
57
- # Placeholder for aggregating streaming response
58
- llmresponse = ""
59
-
60
- # Loop through streaming events capturing relevant details
61
- for chunk in wrapped(*args, **kwargs):
62
- # Collect message IDs and aggregated response from events
63
- content = chunk.text
64
- if content:
65
- llmresponse += content
66
-
67
- input_tokens = chunk.usage_metadata.prompt_token_count
68
- output_tokens = chunk.usage_metadata.candidates_token_count
69
- yield chunk
70
-
71
- # Handling exception ensure observability without disrupting operation
72
- try:
73
- prompt = ""
74
- for arg in args:
75
- if isinstance(arg, str):
76
- prompt = f"{prompt}{arg}\n"
77
- elif isinstance(arg, list):
78
- for subarg in arg:
79
- prompt = f"{prompt}{subarg}\n"
80
- if hasattr(instance, "_model_id"):
81
- model = instance._model_id
82
- if hasattr(instance, "_model_name"):
83
- model = instance._model_name.replace("publishers/google/models/", "")
84
- if model.startswith("models/"):
85
- model = model[len("models/"):]
86
-
87
- total_tokens = input_tokens + output_tokens
88
- # Calculate cost of the operation
89
- cost = get_chat_model_cost(model,
90
- pricing_info, input_tokens,
91
- output_tokens)
92
-
93
- # Set Span attributes
94
- span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
95
- span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
96
- SemanticConvetion.GEN_AI_SYSTEM_GOOGLE_AI_STUDIO)
97
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
98
- SemanticConvetion.GEN_AI_TYPE_CHAT)
99
- span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
100
- gen_ai_endpoint)
101
- span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
102
- environment)
103
- span.set_attribute(SemanticConvetion.GEN_AI_APPLICATION_NAME,
104
- application_name)
105
- span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
106
- model)
107
- span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
108
- True)
109
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
110
- input_tokens)
111
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
55
+
56
+ server_address, server_port = set_server_address_and_port(instance, "generativelanguage.googleapis.com", 443)
57
+ request_model = kwargs.get("model", "gemini-2.0-flash")
58
+
59
+ span_name = f"{SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
60
+
61
+ with tracer.start_as_current_span(span_name, kind=SpanKind.CLIENT) as span:
62
+ start_time = time.time()
63
+ response = wrapped(*args, **kwargs)
64
+ end_time = time.time()
65
+
66
+ response_dict = response_as_dict(response)
67
+
68
+ try:
69
+ # Format 'messages' into a single string
70
+ message_prompt = kwargs.get("contents", [])
71
+ formatted_messages = []
72
+
73
+ for content in message_prompt:
74
+ role = content.role
75
+ parts = content.parts
76
+ content_str = []
77
+
78
+ for part in parts:
79
+ # Collect relevant fields and handle each type of data that Part could contain
80
+ if part.text:
81
+ content_str.append(f"text: {part.text}")
82
+ if part.video_metadata:
83
+ content_str.append(f"video_metadata: {part.video_metadata}")
84
+ if part.thought:
85
+ content_str.append(f"thought: {part.thought}")
86
+ if part.code_execution_result:
87
+ content_str.append(f"code_execution_result: {part.code_execution_result}")
88
+ if part.executable_code:
89
+ content_str.append(f"executable_code: {part.executable_code}")
90
+ if part.file_data:
91
+ content_str.append(f"file_data: {part.file_data}")
92
+ if part.function_call:
93
+ content_str.append(f"function_call: {part.function_call}")
94
+ if part.function_response:
95
+ content_str.append(f"function_response: {part.function_response}")
96
+ if part.inline_data:
97
+ content_str.append(f"inline_data: {part.inline_data}")
98
+
99
+ formatted_messages.append(f"{role}: {', '.join(content_str)}")
100
+
101
+ prompt = "\n".join(formatted_messages)
102
+
103
+ input_tokens = response_dict.get('usage_metadata').get('prompt_token_count')
104
+ output_tokens = response_dict.get('usage_metadata').get('candidates_token_count')
105
+
106
+ # Calculate cost of the operation
107
+ cost = get_chat_model_cost(request_model,
108
+ pricing_info, input_tokens,
112
109
  output_tokens)
113
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
114
- total_tokens)
115
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
116
- cost)
117
- if trace_content:
118
- span.add_event(
119
- name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
120
- attributes={
121
- SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
122
- },
123
- )
124
- span.add_event(
125
- name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
126
- attributes={
127
- SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
128
- },
129
- )
130
-
131
- span.set_status(Status(StatusCode.OK))
132
-
133
- if disable_metrics is False:
134
- attributes = {
135
- TELEMETRY_SDK_NAME:
136
- "openlit",
137
- SemanticConvetion.GEN_AI_APPLICATION_NAME:
138
- application_name,
139
- SemanticConvetion.GEN_AI_SYSTEM:
140
- SemanticConvetion.GEN_AI_SYSTEM_GOOGLE_AI_STUDIO,
141
- SemanticConvetion.GEN_AI_ENVIRONMENT:
142
- environment,
143
- SemanticConvetion.GEN_AI_TYPE:
144
- SemanticConvetion.GEN_AI_TYPE_CHAT,
145
- SemanticConvetion.GEN_AI_REQUEST_MODEL:
146
- model
147
- }
148
-
149
- metrics["genai_requests"].add(1, attributes)
150
- metrics["genai_total_tokens"].add(
151
- total_tokens, attributes
152
- )
153
- metrics["genai_completion_tokens"].add(output_tokens, attributes)
154
- metrics["genai_prompt_tokens"].add(input_tokens, attributes)
155
- metrics["genai_cost"].record(cost, attributes)
156
-
157
- except Exception as e:
158
- handle_exception(span, e)
159
- logger.error("Error in trace creation: %s", e)
160
-
161
- return stream_generator()
162
- else:
163
- # pylint: disable=line-too-long
164
- with tracer.start_as_current_span(gen_ai_endpoint, kind= SpanKind.CLIENT) as span:
165
- response = wrapped(*args, **kwargs)
166
-
167
- try:
168
- prompt = ""
169
- for arg in args:
170
- if isinstance(arg, str):
171
- prompt = f"{prompt}{arg}\n"
172
- elif isinstance(arg, list):
173
- for subarg in arg:
174
- prompt = f"{prompt}{subarg}\n"
175
-
176
- if hasattr(instance, "_model_id"):
177
- model = instance._model_id
178
- if hasattr(instance, "_model_name"):
179
- model = instance._model_name.replace("publishers/google/models/", "")
180
- if model.startswith("models/"):
181
- model = model[len("models/"):]
182
-
183
- # Set base span attribues
184
- span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
185
- span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
186
- SemanticConvetion.GEN_AI_SYSTEM_GOOGLE_AI_STUDIO)
187
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
188
- SemanticConvetion.GEN_AI_TYPE_CHAT)
189
- span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
190
- gen_ai_endpoint)
191
- span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
192
- environment)
193
- span.set_attribute(SemanticConvetion.GEN_AI_APPLICATION_NAME,
194
- application_name)
195
- span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
196
- model)
197
- span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
198
- False)
199
-
200
- if trace_content:
201
- span.add_event(
202
- name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
203
- attributes={
204
- SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
205
- },
206
- )
207
- span.add_event(
208
- name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
209
- attributes={
210
- SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.text,
211
- },
212
- )
213
-
214
- prompt_tokens = response.usage_metadata.prompt_token_count
215
- completion_tokens = response.usage_metadata.candidates_token_count
216
- total_tokens = response.usage_metadata.total_token_count
217
- # Calculate cost of the operation
218
- cost = get_chat_model_cost(model,
219
- pricing_info, prompt_tokens, completion_tokens)
220
-
221
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
222
- prompt_tokens)
223
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
224
- completion_tokens)
225
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
226
- total_tokens)
227
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
228
- cost)
229
-
230
- span.set_status(Status(StatusCode.OK))
231
-
232
- if disable_metrics is False:
233
- attributes = {
234
- TELEMETRY_SDK_NAME:
235
- "openlit",
236
- SemanticConvetion.GEN_AI_APPLICATION_NAME:
237
- application_name,
238
- SemanticConvetion.GEN_AI_SYSTEM:
239
- SemanticConvetion.GEN_AI_SYSTEM_GOOGLE_AI_STUDIO,
240
- SemanticConvetion.GEN_AI_ENVIRONMENT:
241
- environment,
242
- SemanticConvetion.GEN_AI_TYPE:
243
- SemanticConvetion.GEN_AI_TYPE_CHAT,
244
- SemanticConvetion.GEN_AI_REQUEST_MODEL:
245
- model
246
- }
247
-
248
- metrics["genai_requests"].add(1, attributes)
249
- metrics["genai_total_tokens"].add(total_tokens, attributes)
250
- metrics["genai_completion_tokens"].add(completion_tokens, attributes)
251
- metrics["genai_prompt_tokens"].add(prompt_tokens, attributes)
252
- metrics["genai_cost"].record(cost, attributes)
253
-
254
- # Return original response
255
- return response
256
-
257
- except Exception as e:
258
- handle_exception(span, e)
259
- logger.error("Error in trace creation: %s", e)
260
-
261
- # Return original response
262
- return response
110
+
111
+ # Set base span attribues (OTel Semconv)
112
+ span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
113
+ span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
114
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
115
+ span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
116
+ SemanticConvetion.GEN_AI_SYSTEM_GEMINI)
117
+ span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
118
+ request_model)
119
+ span.set_attribute(SemanticConvetion.SERVER_PORT,
120
+ server_port)
121
+
122
+ inference_config = kwargs.get('config', {})
123
+
124
+ # List of attributes and their config keys
125
+ attributes = [
126
+ (SemanticConvetion.GEN_AI_REQUEST_FREQUENCY_PENALTY, 'frequency_penalty'),
127
+ (SemanticConvetion.GEN_AI_REQUEST_MAX_TOKENS, 'max_tokens'),
128
+ (SemanticConvetion.GEN_AI_REQUEST_PRESENCE_PENALTY, 'presence_penalty'),
129
+ (SemanticConvetion.GEN_AI_REQUEST_STOP_SEQUENCES, 'stop_sequences'),
130
+ (SemanticConvetion.GEN_AI_REQUEST_TEMPERATURE, 'temperature'),
131
+ (SemanticConvetion.GEN_AI_REQUEST_TOP_P, 'top_p'),
132
+ (SemanticConvetion.GEN_AI_REQUEST_TOP_K, 'top_k'),
133
+ ]
134
+
135
+ # Set each attribute if the corresponding value exists and is not None
136
+ for attribute, key in attributes:
137
+ # Use getattr to get the attribute value from the object
138
+ value = getattr(inference_config, key, None)
139
+ if value is not None:
140
+ span.set_attribute(attribute, value)
141
+
142
+ span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_MODEL,
143
+ response_dict.get('model_version'))
144
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
145
+ input_tokens)
146
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
147
+ output_tokens)
148
+ span.set_attribute(SemanticConvetion.SERVER_ADDRESS,
149
+ server_address)
150
+ span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
151
+ [str(response_dict.get('candidates')[0].get('finish_reason'))])
152
+
153
+ # Set base span attribues (Extras)
154
+ span.set_attribute(DEPLOYMENT_ENVIRONMENT,
155
+ environment)
156
+ span.set_attribute(SERVICE_NAME,
157
+ application_name)
158
+ span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
159
+ False)
160
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
161
+ input_tokens + output_tokens)
162
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
163
+ cost)
164
+ span.set_attribute(SemanticConvetion.GEN_AI_SERVER_TTFT,
165
+ end_time - start_time)
166
+ span.set_attribute(SemanticConvetion.GEN_AI_SDK_VERSION,
167
+ version)
168
+ if trace_content:
169
+ span.add_event(
170
+ name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
171
+ attributes={
172
+ SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
173
+ },
174
+ )
175
+ span.add_event(
176
+ name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
177
+ attributes={
178
+ SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.text,
179
+ },
180
+ )
181
+
182
+ if isinstance(response_dict.get('text'), str):
183
+ span.set_attribute(SemanticConvetion.GEN_AI_OUTPUT_TYPE,
184
+ "text")
185
+ elif response_dict.get('text') is not None:
186
+ span.set_attribute(SemanticConvetion.GEN_AI_OUTPUT_TYPE,
187
+ "json")
188
+
189
+ span.set_status(Status(StatusCode.OK))
190
+
191
+ if disable_metrics is False:
192
+ attributes = create_metrics_attributes(
193
+ service_name=application_name,
194
+ deployment_environment=environment,
195
+ operation=SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
196
+ system=SemanticConvetion.GEN_AI_SYSTEM_GEMINI,
197
+ request_model=request_model,
198
+ server_address=server_address,
199
+ server_port=server_port,
200
+ response_model=response_dict.get('model_version'),
201
+ )
202
+
203
+ metrics["genai_client_usage_tokens"].record(
204
+ input_tokens + output_tokens, attributes
205
+ )
206
+ metrics["genai_client_operation_duration"].record(
207
+ end_time - start_time, attributes
208
+ )
209
+ metrics["genai_server_ttft"].record(
210
+ end_time - start_time, attributes
211
+ )
212
+ metrics["genai_requests"].add(1, attributes)
213
+ metrics["genai_completion_tokens"].add(output_tokens, attributes)
214
+ metrics["genai_prompt_tokens"].add(input_tokens, attributes)
215
+ metrics["genai_cost"].record(cost, attributes)
216
+
217
+ # Return original response
218
+ return response
219
+
220
+ except Exception as e:
221
+ handle_exception(span, e)
222
+ logger.error("Error in trace creation: %s", e)
223
+
224
+ # Return original response
225
+ return response
263
226
 
264
227
  return wrapper
openlit/instrumentation/gpt4all/gpt4all.py CHANGED
@@ -84,8 +84,8 @@ def generate(gen_ai_endpoint, version, environment, application_name,
84
84
  span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
85
85
  span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
86
86
  SemanticConvetion.GEN_AI_SYSTEM_GPT4ALL)
87
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
88
- SemanticConvetion.GEN_AI_TYPE_CHAT)
87
+ span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
88
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
89
89
  span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
90
90
  gen_ai_endpoint)
91
91
  span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
@@ -120,9 +120,9 @@ def generate(gen_ai_endpoint, version, environment, application_name,
120
120
  },
121
121
  )
122
122
 
123
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
123
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
124
124
  prompt_tokens)
125
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
125
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
126
126
  completion_tokens)
127
127
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
128
128
  total_tokens)
@@ -141,8 +141,8 @@ def generate(gen_ai_endpoint, version, environment, application_name,
141
141
  SemanticConvetion.GEN_AI_SYSTEM_GPT4ALL,
142
142
  SemanticConvetion.GEN_AI_ENVIRONMENT:
143
143
  environment,
144
- SemanticConvetion.GEN_AI_TYPE:
145
- SemanticConvetion.GEN_AI_TYPE_CHAT,
144
+ SemanticConvetion.GEN_AI_OPERATION:
145
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
146
146
  SemanticConvetion.GEN_AI_REQUEST_MODEL:
147
147
  model
148
148
  }
@@ -180,8 +180,8 @@ def generate(gen_ai_endpoint, version, environment, application_name,
180
180
  span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
181
181
  span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
182
182
  SemanticConvetion.GEN_AI_SYSTEM_GPT4ALL)
183
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
184
- SemanticConvetion.GEN_AI_TYPE_CHAT)
183
+ span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
184
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
185
185
  span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
186
186
  gen_ai_endpoint)
187
187
  span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
@@ -216,9 +216,9 @@ def generate(gen_ai_endpoint, version, environment, application_name,
216
216
  },
217
217
  )
218
218
 
219
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
219
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
220
220
  prompt_tokens)
221
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
221
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
222
222
  completion_tokens)
223
223
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
224
224
  total_tokens)
@@ -237,8 +237,8 @@ def generate(gen_ai_endpoint, version, environment, application_name,
237
237
  SemanticConvetion.GEN_AI_SYSTEM_GPT4ALL,
238
238
  SemanticConvetion.GEN_AI_ENVIRONMENT:
239
239
  environment,
240
- SemanticConvetion.GEN_AI_TYPE:
241
- SemanticConvetion.GEN_AI_TYPE_CHAT,
240
+ SemanticConvetion.GEN_AI_OPERATION:
241
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
242
242
  SemanticConvetion.GEN_AI_REQUEST_MODEL:
243
243
  model
244
244
  }
@@ -312,8 +312,8 @@ def embed(gen_ai_endpoint, version, environment, application_name,
312
312
  span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
313
313
  span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
314
314
  SemanticConvetion.GEN_AI_SYSTEM_GPT4ALL)
315
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
316
- SemanticConvetion.GEN_AI_TYPE_EMBEDDING)
315
+ span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
316
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_EMBEDDING)
317
317
  span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
318
318
  gen_ai_endpoint)
319
319
  span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
@@ -322,7 +322,7 @@ def embed(gen_ai_endpoint, version, environment, application_name,
322
322
  application_name)
323
323
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
324
324
  model)
325
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
325
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
326
326
  prompt_tokens)
327
327
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
328
328
  prompt_tokens)
@@ -348,8 +348,8 @@ def embed(gen_ai_endpoint, version, environment, application_name,
348
348
  SemanticConvetion.GEN_AI_SYSTEM_GPT4ALL,
349
349
  SemanticConvetion.GEN_AI_ENVIRONMENT:
350
350
  environment,
351
- SemanticConvetion.GEN_AI_TYPE:
352
- SemanticConvetion.GEN_AI_TYPE_EMBEDDING,
351
+ SemanticConvetion.GEN_AI_OPERATION:
352
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_EMBEDDING,
353
353
  SemanticConvetion.GEN_AI_REQUEST_MODEL:
354
354
  model
355
355
  }
openlit/instrumentation/groq/async_groq.py CHANGED
@@ -104,8 +104,8 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
104
104
  span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
105
105
  span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
106
106
  SemanticConvetion.GEN_AI_SYSTEM_GROQ)
107
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
108
- SemanticConvetion.GEN_AI_TYPE_CHAT)
107
+ span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
108
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
109
109
  span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
110
110
  gen_ai_endpoint)
111
111
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_ID,
@@ -132,9 +132,9 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
132
132
  kwargs.get("seed", ""))
133
133
  span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
134
134
  True)
135
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
135
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
136
136
  prompt_tokens)
137
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
137
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
138
138
  completion_tokens)
139
139
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
140
140
  prompt_tokens + completion_tokens)
@@ -166,8 +166,8 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
166
166
  SemanticConvetion.GEN_AI_SYSTEM_GROQ,
167
167
  SemanticConvetion.GEN_AI_ENVIRONMENT:
168
168
  environment,
169
- SemanticConvetion.GEN_AI_TYPE:
170
- SemanticConvetion.GEN_AI_TYPE_CHAT,
169
+ SemanticConvetion.GEN_AI_OPERATION:
170
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
171
171
  SemanticConvetion.GEN_AI_REQUEST_MODEL:
172
172
  kwargs.get("model", "gpt-3.5-turbo")
173
173
  }
@@ -216,8 +216,8 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
216
216
  span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
217
217
  span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
218
218
  SemanticConvetion.GEN_AI_SYSTEM_GROQ)
219
- span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
220
- SemanticConvetion.GEN_AI_TYPE_CHAT)
219
+ span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
220
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
221
221
  span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
222
222
  gen_ai_endpoint)
223
223
  span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_ID,
@@ -259,9 +259,9 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
259
259
  pricing_info, response.usage.prompt_tokens,
260
260
  response.usage.completion_tokens)
261
261
 
262
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
262
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
263
263
  response.usage.prompt_tokens)
264
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
264
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
265
265
  response.usage.completion_tokens)
266
266
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
267
267
  response.usage.total_tokens)
@@ -305,9 +305,9 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
305
305
 
306
306
  span.set_attribute(SemanticConvetion.GEN_AI_CONTENT_COMPLETION,
307
307
  "Function called with tools")
308
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
308
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
309
309
  response.usage.prompt_tokens)
310
- span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
310
+ span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
311
311
  response.usage.completion_tokens)
312
312
  span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
313
313
  response.usage.total_tokens)
@@ -326,8 +326,8 @@ def async_chat(gen_ai_endpoint, version, environment, application_name,
326
326
  SemanticConvetion.GEN_AI_SYSTEM_GROQ,
327
327
  SemanticConvetion.GEN_AI_ENVIRONMENT:
328
328
  environment,
329
- SemanticConvetion.GEN_AI_TYPE:
330
- SemanticConvetion.GEN_AI_TYPE_CHAT,
329
+ SemanticConvetion.GEN_AI_OPERATION:
330
+ SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
331
331
  SemanticConvetion.GEN_AI_REQUEST_MODEL:
332
332
  kwargs.get("model", "gpt-3.5-turbo")
333
333
  }