Skip to content

Same agent Smaller bill. The tools and infrastructure to get more work from your AI. Fewer tokens. Less compute. Lower cost.

[Feat]: Refactor Mistral AI OpenTelemetry Monitoring Instrumentation (#780)

Patcher11a83949cf37
diff --git a/sdk/python/pyproject.toml b/sdk/python/pyproject.toml
index 0a15583b..baf30a30 100644
--- a/sdk/python/pyproject.toml
+++ b/sdk/python/pyproject.toml
@@ -1,6 +1,6 @@
[tool.poetry]
name = "openlit"
-version = "1.34.16"
+version = "1.34.17"
description = "OpenTelemetry-native Auto instrumentation library for monitoring LLM Applications and GPUs, facilitating the integration of observability into your GenAI-driven projects"
authors = ["OpenLIT"]
license = "Apache-2.0"
diff --git a/sdk/python/src/openlit/instrumentation/mistral/__init__.py b/sdk/python/src/openlit/instrumentation/mistral/__init__.py
index e9b88b64..a9f20822 100644
--- a/sdk/python/src/openlit/instrumentation/mistral/__init__.py
+++ b/sdk/python/src/openlit/instrumentation/mistral/__init__.py
@@ -1,80 +1,88 @@
-# pylint: disable=useless-return, bad-staticmethod-argument, disable=duplicate-code
"""Initializer of Auto Instrumentation of Mistral Functions"""
+
from typing import Collection
import importlib.metadata
from opentelemetry.instrumentation.instrumentor import BaseInstrumentor
from wrapt import wrap_function_wrapper
-from openlit.instrumentation.mistral.mistral import chat, chat_stream, embeddings
-from openlit.instrumentation.mistral.async_mistral import async_chat, async_chat_stream
-from openlit.instrumentation.mistral.async_mistral import async_embeddings
+from openlit.instrumentation.mistral.mistral import (
+ complete,
+ stream,
+ embed
+)
+from openlit.instrumentation.mistral.async_mistral import (
+ async_complete,
+ async_stream,
+ async_embed
+)
_instruments = ("mistralai >= 1.0.0",)
class MistralInstrumentor(BaseInstrumentor):
- """An instrumentor for Mistral's client library."""
+ """
+ An instrumentor for Mistral client library.
+ """
def instrumentation_dependencies(self) -> Collection[str]:
return _instruments
def _instrument(self, **kwargs):
- application_name = kwargs.get("application_name")
- environment = kwargs.get("environment")
+ application_name = kwargs.get("application_name", "default")
+ environment = kwargs.get("environment", "default")
tracer = kwargs.get("tracer")
metrics = kwargs.get("metrics_dict")
- pricing_info = kwargs.get("pricing_info")
- capture_message_content = kwargs.get("capture_message_content")
+ pricing_info = kwargs.get("pricing_info", {})
+ capture_message_content = kwargs.get("capture_message_content", False)
disable_metrics = kwargs.get("disable_metrics")
version = importlib.metadata.version("mistralai")
- # sync
+ # sync chat completions
wrap_function_wrapper(
- "mistralai.chat",
- "Chat.complete",
- chat(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ "mistralai.chat",
+ "Chat.complete",
+ complete(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- # sync
+ # sync chat streaming
wrap_function_wrapper(
- "mistralai.chat",
- "Chat.stream",
- chat_stream(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ "mistralai.chat",
+ "Chat.stream",
+ stream(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- # sync
+ # sync embeddings
wrap_function_wrapper(
- "mistralai.embeddings",
- "Embeddings.create",
- embeddings(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ "mistralai.embeddings",
+ "Embeddings.create",
+ embed(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- # Async
+ # async chat completions
wrap_function_wrapper(
- "mistralai.chat",
- "Chat.complete_async",
- async_chat(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ "mistralai.chat",
+ "Chat.complete_async",
+ async_complete(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- # Async
+ # async chat streaming
wrap_function_wrapper(
- "mistralai.chat",
- "Chat.stream_async",
- async_chat_stream(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ "mistralai.chat",
+ "Chat.stream_async",
+ async_stream(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- #sync
+ # async embeddings
wrap_function_wrapper(
- "mistralai.embeddings",
- "Embeddings.create_async",
- async_embeddings(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ "mistralai.embeddings",
+ "Embeddings.create_async",
+ async_embed(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- @staticmethod
def _uninstrument(self, **kwargs):
pass
diff --git a/sdk/python/src/openlit/instrumentation/mistral/async_mistral.py b/sdk/python/src/openlit/instrumentation/mistral/async_mistral.py
index 1ec656fb..a441c0a4 100644
--- a/sdk/python/src/openlit/instrumentation/mistral/async_mistral.py
+++ b/sdk/python/src/openlit/instrumentation/mistral/async_mistral.py
@@ -1,611 +1,200 @@
"""
-Module for monitoring Mistral API calls.
+Module for monitoring Mistral API calls (async version).
"""
-import logging
import time
-from opentelemetry.trace import SpanKind, Status, StatusCode
-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
+from opentelemetry.trace import SpanKind
from openlit.__helpers import (
- get_chat_model_cost,
- get_embed_model_cost,
handle_exception,
- response_as_dict,
- calculate_ttft,
- calculate_tbt,
- create_metrics_attributes,
- set_server_address_and_port
+ set_server_address_and_port,
+)
+from openlit.instrumentation.mistral.utils import (
+ process_chunk,
+ process_chat_response,
+ process_streaming_chat_response,
+ process_embedding_response,
)
from openlit.semcov import SemanticConvention
-# Initialize logger for logging potential issues and operations
-logger = logging.getLogger(__name__)
-
-def async_chat(version, environment, application_name, tracer,
- pricing_info, capture_message_content, metrics, disable_metrics):
+def async_complete(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for chat to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Mistral API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Mistral usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the chat method to add telemetry.
+ Generates a telemetry wrapper for GenAI complete function call
"""
async def wrapper(wrapped, instance, args, kwargs):
"""
- Wraps the 'chat' API call to add telemetry.
-
- This collects metrics such as execution time, cost, and token usage, and handles errors
- gracefully, adding details to the trace for observability.
-
- Args:
- wrapped: The original 'chat' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'chat' method.
- kwargs: Keyword arguments for the 'chat' method.
-
- Returns:
- The response from the original 'chat' method.
+ Wraps the GenAI complete function call.
"""
- server_address, server_port = set_server_address_and_port(instance, 'api.mistral.ai', 443)
- request_model = kwargs.get('model', 'mistral-small-latest')
+ server_address, server_port = set_server_address_and_port(instance, "api.mistral.ai", 443)
+ request_model = kwargs.get("model", "mistral-small-latest")
- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}'
+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
with tracer.start_as_current_span(span_name, kind=SpanKind.CLIENT) as span:
start_time = time.time()
response = await wrapped(*args, **kwargs)
- end_time = time.time()
-
- response_dict = response_as_dict(response)
-
- try:
- # Format 'messages' into a single string
- message_prompt = kwargs.get('messages', '')
- formatted_messages = []
- for message in message_prompt:
- role = message['role']
- content = message['content']
-
- if isinstance(content, list):
- content_str = ", ".join(
- f'{item["type"]}: {item["text"] if "text" in item else item["image_url"]}'
- if "type" in item else f'text: {item["text"]}'
- for item in content
- )
- formatted_messages.append(f'{role}: {content_str}')
- else:
- formatted_messages.append(f'{role}: {content}')
- prompt = '\n'.join(formatted_messages)
-
- input_tokens = response_dict.get('usage').get('prompt_tokens')
- output_tokens = response_dict.get('usage').get('completion_tokens')
-
- # Calculate cost of the operation
- cost = get_chat_model_cost(request_model,
- pricing_info, input_tokens,
- output_tokens)
-
- # Set base span attribues (OTel Semconv)
- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')
- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_MISTRAL)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- kwargs.get('seed', ''))
- span.set_attribute(SemanticConvention.SERVER_PORT,
- server_port)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- kwargs.get('frequency_penalty', 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- kwargs.get('max_tokens', -1))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- kwargs.get('presence_penalty', 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- kwargs.get('stop', []))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- kwargs.get('temperature', 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- kwargs.get('top_p', 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- response_dict.get('id'))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- response_dict.get('model'))
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- input_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- output_tokens)
- span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- server_address)
-
- # Set base span attribues (Extras)
- span.set_attribute(DEPLOYMENT_ENVIRONMENT,
- environment)
- span.set_attribute(SERVICE_NAME,
- application_name)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- False)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- input_tokens + output_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- end_time - start_time)
- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
-
- for i in range(kwargs.get('n',1)):
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [response_dict.get('choices')[i].get('finish_reason')])
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- # pylint: disable=line-too-long
- SemanticConvention.GEN_AI_CONTENT_COMPLETION: str(response_dict.get('choices')[i].get('message').get('content')),
- },
- )
- if kwargs.get('tools'):
- span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALLS,
- str(response_dict.get('choices')[i].get('message').get('tool_calls')))
-
- if isinstance(response_dict.get('choices')[i].get('message').get('content'), str):
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'text')
- elif response_dict.get('choices')[i].get('message').get('content') is not None:
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'json')
-
- span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
- request_model=request_model,
- server_address=server_address,
- server_port=server_port,
- response_model=response_dict.get('model'),
- )
-
- metrics['genai_client_usage_tokens'].record(
- input_tokens + output_tokens, attributes
- )
- metrics['genai_client_operation_duration'].record(
- end_time - start_time, attributes
- )
- metrics['genai_server_ttft'].record(
- end_time - start_time, attributes
- )
- metrics['genai_requests'].add(1, attributes)
- metrics['genai_completion_tokens'].add(output_tokens, attributes)
- metrics['genai_prompt_tokens'].add(input_tokens, attributes)
- metrics['genai_cost'].record(cost, attributes)
-
- # Return original response
- return response
-
- except Exception as e:
- handle_exception(span, e)
- logger.error('Error in trace creation: %s', e)
-
- # Return original response
- return response
+ response = process_chat_response(
+ response=response,
+ request_model=request_model,
+ pricing_info=pricing_info,
+ server_port=server_port,
+ server_address=server_address,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ start_time=start_time,
+ span=span,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version,
+ **kwargs
+ )
+
+ return response
return wrapper
-def async_chat_stream(version, environment, application_name,
+def async_stream(version, environment, application_name,
tracer, pricing_info, capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for chat_stream to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Mistral API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Mistral usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the chat method to add telemetry.
+ Generates a telemetry wrapper for GenAI stream function call
"""
- async def wrapper(wrapped, instance, args, kwargs):
+ class TracedAsyncStream:
"""
- Wraps the 'chat_stream' API call to add telemetry.
-
- This collects metrics such as execution time, cost, and token usage, and handles errors
- gracefully, adding details to the trace for observability.
-
- Args:
- wrapped: The original 'chat_stream' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'chat_stream' method.
- kwargs: Keyword arguments for the 'chat_stream' method.
-
- Returns:
- The response from the original 'chat_stream' method.
+ Wrapper for async streaming responses to collect telemetry.
"""
- class TracedAsyncStream:
- """
- Wrapper for streaming responses to collect metrics and trace data.
- Wraps the 'mistral.syncStream' response to collect message IDs and aggregated response.
-
- This class implements the '__aiter__' and '__anext__' methods that
- handle asynchronous streaming responses.
-
- This class also implements '__aenter__' and '__aexit__' methods that
- handle asynchronous context management protocol.
- """
- def __init__(
- self,
- wrapped,
- span,
- kwargs,
- server_address,
- server_port,
- **args,
- ):
- self.__wrapped__ = wrapped
- self._span = span
- # Placeholder for aggregating streaming response
- self._llmresponse = ''
- self._response_id = ''
- self._response_model = ''
- self._finish_reason = ''
- self._input_tokens = ''
- self._output_tokens = ''
-
- self._args = args
- self._kwargs = kwargs
- self._start_time = time.time()
- self._end_time = None
- self._timestamps = []
- self._ttft = 0
- self._tbt = 0
- self._server_address = server_address
- self._server_port = server_port
-
- async def __aenter__(self):
- await self.__wrapped__.__aenter__()
- return self
-
- async def __aexit__(self, exc_type, exc_value, traceback):
- await self.__wrapped__.__aexit__(exc_type, exc_value, traceback)
-
- def __aiter__(self):
- return self
-
- async def __getattr__(self, name):
- """Delegate attribute access to the wrapped object."""
- return getattr(await self.__wrapped__, name)
-
- async def __anext__(self):
+ def __init__(
+ self,
+ wrapped,
+ span,
+ span_name,
+ kwargs,
+ server_address,
+ server_port,
+ **args,
+ ):
+ self.__wrapped__ = wrapped
+ self._span = span
+ self._span_name = span_name
+ self._llmresponse = ""
+ self._response_id = ""
+ self._response_model = ""
+ self._finish_reason = ""
+ self._tools = None
+ self._input_tokens = 0
+ self._output_tokens = 0
+
+ self._args = args
+ self._kwargs = kwargs
+ self._start_time = time.time()
+ self._end_time = None
+ self._timestamps = []
+ self._ttft = 0
+ self._tbt = 0
+ self._server_address = server_address
+ self._server_port = server_port
+
+ async def __aenter__(self):
+ await self.__wrapped__.__aenter__()
+ return self
+
+ async def __aexit__(self, exc_type, exc_value, traceback):
+ await self.__wrapped__.__aexit__(exc_type, exc_value, traceback)
+
+ def __aiter__(self):
+ return self
+
+ async def __getattr__(self, name):
+ """Delegate attribute access to the wrapped object."""
+ return getattr(await self.__wrapped__, name)
+
+ async def __anext__(self):
+ try:
+ chunk = await self.__wrapped__.__anext__()
+ process_chunk(self, chunk)
+ return chunk
+ except StopAsyncIteration:
try:
- chunk = await self.__wrapped__.__anext__()
- end_time = time.time()
- # Record the timestamp for the current chunk
- self._timestamps.append(end_time)
-
- if len(self._timestamps) == 1:
- # Calculate time to first chunk
- self._ttft = calculate_ttft(self._timestamps, self._start_time)
-
- chunked = response_as_dict(chunk)
-
- self._llmresponse += chunked.get('data').get('choices')[0].get('delta').get('content')
- if chunked.get('data').get('usage') is not None:
- self._response_id = chunked.get('data').get('id')
- self._response_model = chunked.get('data').get('model')
- self._input_tokens = chunked.get('data').get('usage').get('prompt_tokens')
- self._output_tokens = chunked.get('data').get('usage').get('completion_tokens')
- self._finish_reason = chunked.get('data').get('choices')[0].get('finish_reason')
-
- return chunk
- except StopAsyncIteration:
- # Handling exception ensure observability without disrupting operation
- try:
- self._end_time = time.time()
- if len(self._timestamps) > 1:
- self._tbt = calculate_tbt(self._timestamps)
-
- # Format 'messages' into a single string
- message_prompt = self._kwargs.get('messages', '')
- formatted_messages = []
- for message in message_prompt:
- role = message['role']
- content = message['content']
-
- if isinstance(content, list):
- content_str_list = []
- for item in content:
- if item['type'] == 'text':
- content_str_list.append(f'text: {item["text"]}')
- elif (item['type'] == 'image_url' and
- not item['image_url']['url'].startswith('data:')):
- content_str_list.append(f'image_url: {item["image_url"]["url"]}')
- content_str = ", ".join(content_str_list)
- formatted_messages.append(f'{role}: {content_str}')
- else:
- formatted_messages.append(f'{role}: {content}')
- prompt = '\n'.join(formatted_messages)
-
- request_model = self._kwargs.get('model', 'mistral-small-latest')
-
- # Calculate cost of the operation
- cost = get_chat_model_cost(request_model,
- pricing_info, self._input_tokens,
- self._output_tokens)
-
- # Set Span attributes (OTel Semconv)
- self._span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')
- self._span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- self._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_MISTRAL)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- self._kwargs.get('seed', ''))
- self._span.set_attribute(SemanticConvention.SERVER_PORT,
- self._server_port)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- self._kwargs.get('frequency_penalty', 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- self._kwargs.get('max_tokens', -1))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- self._kwargs.get('presence_penalty', 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- self._kwargs.get('stop_sequences', []))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- self._kwargs.get('temperature', 0.3))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_K,
- self._kwargs.get('k', 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- self._kwargs.get('p', 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [self._finish_reason])
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- self._response_id)
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- self._response_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- self._input_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- self._output_tokens)
- self._span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- self._server_address)
-
- if isinstance(self._llmresponse, str):
- self._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'text')
- else:
- self._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'json')
-
- # Set Span attributes (Extra)
- self._span.set_attribute(DEPLOYMENT_ENVIRONMENT,
- environment)
- self._span.set_attribute(SERVICE_NAME,
- application_name)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- True)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- self._input_tokens + self._output_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT,
- self._tbt)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- self._ttft)
- self._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_COMPLETION: self._llmresponse,
- },
- )
- self._span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
- request_model=request_model,
- server_address=self._server_address,
- server_port=self._server_port,
- response_model=self._response_model,
- )
-
- metrics['genai_client_usage_tokens'].record(
- self._input_tokens + self._output_tokens, attributes
- )
- metrics['genai_client_operation_duration'].record(
- self._end_time - self._start_time, attributes
- )
- metrics['genai_server_tbt'].record(
- self._tbt, attributes
- )
- metrics['genai_server_ttft'].record(
- self._ttft, attributes
- )
- metrics['genai_requests'].add(1, attributes)
- metrics['genai_completion_tokens'].add(self._output_tokens, attributes)
- metrics['genai_prompt_tokens'].add(self._input_tokens, attributes)
- metrics['genai_cost'].record(cost, attributes)
-
- except Exception as e:
- handle_exception(self._span, e)
- logger.error('Error in trace creation: %s', e)
- finally:
- self._span.end()
- raise
-
- server_address, server_port = set_server_address_and_port(instance, 'api.mistral.ai', 443)
- request_model = kwargs.get('model', 'mistral-small-latest')
-
- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}'
+ with tracer.start_as_current_span(self._span_name, kind= SpanKind.CLIENT) as self._span:
+ process_streaming_chat_response(
+ self,
+ pricing_info=pricing_info,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version
+ )
+
+ except Exception as e:
+ handle_exception(self._span, e)
+
+ raise
+ async def wrapper(wrapped, instance, args, kwargs):
+ """
+ Wraps the GenAI stream function call.
+ """
+
+ server_address, server_port = set_server_address_and_port(instance, "api.mistral.ai", 443)
+ request_model = kwargs.get("model", "mistral-small-latest")
+
+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
+
+ # Stream endpoint is always streaming
awaited_wrapped = await wrapped(*args, **kwargs)
span = tracer.start_span(span_name, kind=SpanKind.CLIENT)
- return TracedAsyncStream(awaited_wrapped, span, kwargs, server_address, server_port)
+
+ return TracedAsyncStream(awaited_wrapped, span, span_name, kwargs, server_address, server_port)
return wrapper
-def async_embeddings(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics):
+def async_embed(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for embeddings to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Mistral API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Mistral usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the embeddings method to add telemetry.
+ Generates a telemetry wrapper for GenAI embedding function call
"""
- def wrapper(wrapped, instance, args, kwargs):
+ async def wrapper(wrapped, instance, args, kwargs):
"""
- Wraps the 'embeddings' API call to add telemetry.
-
- This collects metrics such as execution time, cost, and token usage, and handles errors
- gracefully, adding details to the trace for observability.
-
- Args:
- wrapped: The original 'embeddings' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'embeddings' method.
- kwargs: Keyword arguments for the 'embeddings' method.
-
- Returns:
- The response from the original 'embeddings' method.
+ Wraps the GenAI embedding function call.
"""
- server_address, server_port = set_server_address_and_port(instance, 'api.mistral.ai', 443)
- request_model = kwargs.get('model', 'mistral-embed')
+ server_address, server_port = set_server_address_and_port(instance, "api.mistral.ai", 443)
+ request_model = kwargs.get("model", "mistral-embed")
- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING} {request_model}'
+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING} {request_model}"
- with tracer.start_as_current_span(span_name, kind= SpanKind.CLIENT) as span:
+ with tracer.start_as_current_span(span_name, kind=SpanKind.CLIENT) as span:
start_time = time.time()
- response = wrapped(*args, **kwargs)
- end_time = time.time()
+ response = await wrapped(*args, **kwargs)
- response_dict = response_as_dict(response)
try:
- input_tokens = response_dict.get('usage').get('prompt_tokens')
-
- # Calculate cost of the operation
- cost = get_embed_model_cost(request_model,
- pricing_info, input_tokens)
-
- # Set Span attributes (OTel Semconv)
- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')
- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING)
- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_MISTRAL)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_ENCODING_FORMATS,
- [kwargs.get('encoding_format', 'float')])
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- response_dict.get('model'))
- span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- server_address)
- span.set_attribute(SemanticConvention.SERVER_PORT,
- server_port)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- input_tokens)
-
- # Set Span attributes (Extras)
- span.set_attribute(DEPLOYMENT_ENVIRONMENT,
- environment)
- span.set_attribute(SERVICE_NAME,
- application_name)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- input_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
-
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: str(kwargs.get('inputs', '')),
- },
- )
-
- span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING,
- system=SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
- request_model=request_model,
- server_address=server_address,
- server_port=server_port,
- response_model=response_dict.get('model'),
- )
- metrics['genai_client_usage_tokens'].record(
- input_tokens, attributes
- )
- metrics['genai_client_operation_duration'].record(
- end_time - start_time, attributes
- )
- metrics['genai_requests'].add(1, attributes)
- metrics['genai_prompt_tokens'].add(input_tokens, attributes)
- metrics['genai_cost'].record(cost, attributes)
-
- # Return original response
- return response
+ response = process_embedding_response(
+ response=response,
+ request_model=request_model,
+ pricing_info=pricing_info,
+ server_port=server_port,
+ server_address=server_address,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ start_time=start_time,
+ span=span,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version,
+ **kwargs
+ )
except Exception as e:
handle_exception(span, e)
- logger.error('Error in trace creation: %s', e)
- # Return original response
- return response
+ return response
return wrapper
diff --git a/sdk/python/src/openlit/instrumentation/mistral/mistral.py b/sdk/python/src/openlit/instrumentation/mistral/mistral.py
index 923bd839..9acfd206 100644
--- a/sdk/python/src/openlit/instrumentation/mistral/mistral.py
+++ b/sdk/python/src/openlit/instrumentation/mistral/mistral.py
@@ -2,610 +2,199 @@
Module for monitoring Mistral API calls.
"""
-import logging
import time
-from opentelemetry.trace import SpanKind, Status, StatusCode
-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
+from opentelemetry.trace import SpanKind
from openlit.__helpers import (
- get_chat_model_cost,
- get_embed_model_cost,
handle_exception,
- response_as_dict,
- calculate_ttft,
- calculate_tbt,
- create_metrics_attributes,
- set_server_address_and_port
+ set_server_address_and_port,
+)
+from openlit.instrumentation.mistral.utils import (
+ process_chunk,
+ process_chat_response,
+ process_streaming_chat_response,
+ process_embedding_response,
)
from openlit.semcov import SemanticConvention
-# Initialize logger for logging potential issues and operations
-logger = logging.getLogger(__name__)
-
-def chat(version, environment, application_name, tracer,
- pricing_info, capture_message_content, metrics, disable_metrics):
+def complete(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for chat to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Mistral API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Mistral usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the chat method to add telemetry.
+ Generates a telemetry wrapper for GenAI complete function call
"""
def wrapper(wrapped, instance, args, kwargs):
"""
- Wraps the 'chat' API call to add telemetry.
-
- This collects metrics such as execution time, cost, and token usage, and handles errors
- gracefully, adding details to the trace for observability.
-
- Args:
- wrapped: The original 'chat' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'chat' method.
- kwargs: Keyword arguments for the 'chat' method.
-
- Returns:
- The response from the original 'chat' method.
+ Wraps the GenAI complete function call.
"""
- server_address, server_port = set_server_address_and_port(instance, 'api.mistral.ai', 443)
- request_model = kwargs.get('model', 'mistral-small-latest')
+ server_address, server_port = set_server_address_and_port(instance, "api.mistral.ai", 443)
+ request_model = kwargs.get("model", "mistral-small-latest")
- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}'
+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
with tracer.start_as_current_span(span_name, kind=SpanKind.CLIENT) as span:
start_time = time.time()
response = wrapped(*args, **kwargs)
- end_time = time.time()
+ response = process_chat_response(
+ response=response,
+ request_model=request_model,
+ pricing_info=pricing_info,
+ server_port=server_port,
+ server_address=server_address,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ start_time=start_time,
+ span=span,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version,
+ **kwargs
+ )
+
+ return response
- response_dict = response_as_dict(response)
+ return wrapper
- try:
- # Format 'messages' into a single string
- message_prompt = kwargs.get('messages', '')
- formatted_messages = []
- for message in message_prompt:
- role = message['role']
- content = message['content']
-
- if isinstance(content, list):
- content_str = ", ".join(
- f'{item["type"]}: {item["text"] if "text" in item else item["image_url"]}'
- if "type" in item else f'text: {item["text"]}'
- for item in content
- )
- formatted_messages.append(f'{role}: {content_str}')
- else:
- formatted_messages.append(f'{role}: {content}')
- prompt = '\n'.join(formatted_messages)
-
- input_tokens = response_dict.get('usage').get('prompt_tokens')
- output_tokens = response_dict.get('usage').get('completion_tokens')
-
- # Calculate cost of the operation
- cost = get_chat_model_cost(request_model,
- pricing_info, input_tokens,
- output_tokens)
-
- # Set base span attribues (OTel Semconv)
- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')
- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_MISTRAL)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- kwargs.get('seed', ''))
- span.set_attribute(SemanticConvention.SERVER_PORT,
- server_port)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- kwargs.get('frequency_penalty', 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- kwargs.get('max_tokens', -1))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- kwargs.get('presence_penalty', 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- kwargs.get('stop', []))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- kwargs.get('temperature', 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- kwargs.get('top_p', 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- response_dict.get('id'))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- response_dict.get('model'))
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- input_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- output_tokens)
- span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- server_address)
-
- # Set base span attribues (Extras)
- span.set_attribute(DEPLOYMENT_ENVIRONMENT,
- environment)
- span.set_attribute(SERVICE_NAME,
- application_name)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- False)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- input_tokens + output_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- end_time - start_time)
- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
-
- for i in range(kwargs.get('n',1)):
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [response_dict.get('choices')[i].get('finish_reason')])
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- # pylint: disable=line-too-long
- SemanticConvention.GEN_AI_CONTENT_COMPLETION: str(response_dict.get('choices')[i].get('message').get('content')),
- },
- )
- if kwargs.get('tools'):
- span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALLS,
- str(response_dict.get('choices')[i].get('message').get('tool_calls')))
-
- if isinstance(response_dict.get('choices')[i].get('message').get('content'), str):
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'text')
- elif response_dict.get('choices')[i].get('message').get('content') is not None:
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'json')
-
- span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
- request_model=request_model,
- server_address=server_address,
- server_port=server_port,
- response_model=response_dict.get('model'),
- )
-
- metrics['genai_client_usage_tokens'].record(
- input_tokens + output_tokens, attributes
- )
- metrics['genai_client_operation_duration'].record(
- end_time - start_time, attributes
- )
- metrics['genai_server_ttft'].record(
- end_time - start_time, attributes
- )
- metrics['genai_requests'].add(1, attributes)
- metrics['genai_completion_tokens'].add(output_tokens, attributes)
- metrics['genai_prompt_tokens'].add(input_tokens, attributes)
- metrics['genai_cost'].record(cost, attributes)
-
- # Return original response
- return response
+def stream(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):
+ """
+ Generates a telemetry wrapper for GenAI stream function call
+ """
- except Exception as e:
- handle_exception(span, e)
- logger.error('Error in trace creation: %s', e)
+ class TracedSyncStream:
+ """
+ Wrapper for streaming responses to collect telemetry.
+ """
- # Return original response
- return response
+ def __init__(
+ self,
+ wrapped,
+ span,
+ span_name,
+ kwargs,
+ server_address,
+ server_port,
+ **args,
+ ):
+ self.__wrapped__ = wrapped
+ self._span = span
+ self._span_name = span_name
+ self._llmresponse = ""
+ self._response_id = ""
+ self._response_model = ""
+ self._finish_reason = ""
+ self._tools = None
+ self._input_tokens = 0
+ self._output_tokens = 0
+
+ self._args = args
+ self._kwargs = kwargs
+ self._start_time = time.time()
+ self._end_time = None
+ self._timestamps = []
+ self._ttft = 0
+ self._tbt = 0
+ self._server_address = server_address
+ self._server_port = server_port
+
+ def __enter__(self):
+ self.__wrapped__.__enter__()
+ return self
+
+ def __exit__(self, exc_type, exc_value, traceback):
+ self.__wrapped__.__exit__(exc_type, exc_value, traceback)
+
+ def __iter__(self):
+ return self
+
+ def __getattr__(self, name):
+ """Delegate attribute access to the wrapped object."""
+ return getattr(self.__wrapped__, name)
+
+ def __next__(self):
+ try:
+ chunk = self.__wrapped__.__next__()
+ process_chunk(self, chunk)
+ return chunk
+ except StopIteration:
+ try:
+ with tracer.start_as_current_span(self._span_name, kind= SpanKind.CLIENT) as self._span:
+ process_streaming_chat_response(
+ self,
+ pricing_info=pricing_info,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version
+ )
- return wrapper
+ except Exception as e:
+ handle_exception(self._span, e)
-def chat_stream(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics):
- """
- Generates a telemetry wrapper for chat_stream to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Mistral API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Mistral usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the chat method to add telemetry.
- """
+ raise
def wrapper(wrapped, instance, args, kwargs):
"""
- Wraps the 'chat_stream' API call to add telemetry.
-
- This collects metrics such as execution time, cost, and token usage, and handles errors
- gracefully, adding details to the trace for observability.
-
- Args:
- wrapped: The original 'chat_stream' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'chat_stream' method.
- kwargs: Keyword arguments for the 'chat_stream' method.
-
- Returns:
- The response from the original 'chat_stream' method.
+ Wraps the GenAI stream function call.
"""
- class TracedSyncStream:
- """
- Wrapper for streaming responses to collect metrics and trace data.
- Wraps the 'mistral.syncStream' response to collect message IDs and aggregated response.
-
- This class implements the '__aiter__' and '__anext__' methods that
- handle asynchronous streaming responses.
-
- This class also implements '__aenter__' and '__aexit__' methods that
- handle asynchronous context management protocol.
- """
- def __init__(
- self,
- wrapped,
- span,
- kwargs,
- server_address,
- server_port,
- **args,
- ):
- self.__wrapped__ = wrapped
- self._span = span
- # Placeholder for aggregating streaming response
- self._llmresponse = ''
- self._response_id = ''
- self._response_model = ''
- self._finish_reason = ''
- self._input_tokens = ''
- self._output_tokens = ''
-
- self._args = args
- self._kwargs = kwargs
- self._start_time = time.time()
- self._end_time = None
- self._timestamps = []
- self._ttft = 0
- self._tbt = 0
- self._server_address = server_address
- self._server_port = server_port
-
- def __enter__(self):
- self.__wrapped__.__enter__()
- return self
-
- def __exit__(self, exc_type, exc_value, traceback):
- self.__wrapped__.__exit__(exc_type, exc_value, traceback)
-
- def __iter__(self):
- return self
-
- def __getattr__(self, name):
- """Delegate attribute access to the wrapped object."""
- return getattr(self.__wrapped__, name)
-
- def __next__(self):
- try:
- chunk = self.__wrapped__.__next__()
- end_time = time.time()
- # Record the timestamp for the current chunk
- self._timestamps.append(end_time)
-
- if len(self._timestamps) == 1:
- # Calculate time to first chunk
- self._ttft = calculate_ttft(self._timestamps, self._start_time)
-
- chunked = response_as_dict(chunk)
-
- self._llmresponse += chunked.get('data').get('choices')[0].get('delta').get('content')
- if chunked.get('data').get('usage') is not None:
- self._response_id = chunked.get('data').get('id')
- self._response_model = chunked.get('data').get('model')
- self._input_tokens = chunked.get('data').get('usage').get('prompt_tokens')
- self._output_tokens = chunked.get('data').get('usage').get('completion_tokens')
- self._finish_reason = chunked.get('data').get('choices')[0].get('finish_reason')
-
- return chunk
- except StopIteration:
- # Handling exception ensure observability without disrupting operation
- try:
- self._end_time = time.time()
- if len(self._timestamps) > 1:
- self._tbt = calculate_tbt(self._timestamps)
-
- # Format 'messages' into a single string
- message_prompt = self._kwargs.get('messages', '')
- formatted_messages = []
- for message in message_prompt:
- role = message['role']
- content = message['content']
-
- if isinstance(content, list):
- content_str_list = []
- for item in content:
- if item['type'] == 'text':
- content_str_list.append(f'text: {item["text"]}')
- elif (item['type'] == 'image_url' and
- not item['image_url']['url'].startswith('data:')):
- content_str_list.append(f'image_url: {item["image_url"]["url"]}')
- content_str = ", ".join(content_str_list)
- formatted_messages.append(f'{role}: {content_str}')
- else:
- formatted_messages.append(f'{role}: {content}')
- prompt = '\n'.join(formatted_messages)
-
- request_model = self._kwargs.get('model', 'mistral-small-latest')
-
- # Calculate cost of the operation
- cost = get_chat_model_cost(request_model,
- pricing_info, self._input_tokens,
- self._output_tokens)
-
- # Set Span attributes (OTel Semconv)
- self._span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')
- self._span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- self._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_MISTRAL)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- self._kwargs.get('seed', ''))
- self._span.set_attribute(SemanticConvention.SERVER_PORT,
- self._server_port)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- self._kwargs.get('frequency_penalty', 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- self._kwargs.get('max_tokens', -1))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- self._kwargs.get('presence_penalty', 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- self._kwargs.get('stop_sequences', []))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- self._kwargs.get('temperature', 0.3))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_K,
- self._kwargs.get('k', 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- self._kwargs.get('p', 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [self._finish_reason])
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- self._response_id)
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- self._response_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- self._input_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- self._output_tokens)
- self._span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- self._server_address)
-
- if isinstance(self._llmresponse, str):
- self._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'text')
- else:
- self._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- 'json')
-
- # Set Span attributes (Extra)
- self._span.set_attribute(DEPLOYMENT_ENVIRONMENT,
- environment)
- self._span.set_attribute(SERVICE_NAME,
- application_name)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- True)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- self._input_tokens + self._output_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT,
- self._tbt)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- self._ttft)
- self._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_COMPLETION: self._llmresponse,
- },
- )
- self._span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
- request_model=request_model,
- server_address=self._server_address,
- server_port=self._server_port,
- response_model=self._response_model,
- )
-
- metrics['genai_client_usage_tokens'].record(
- self._input_tokens + self._output_tokens, attributes
- )
- metrics['genai_client_operation_duration'].record(
- self._end_time - self._start_time, attributes
- )
- metrics['genai_server_tbt'].record(
- self._tbt, attributes
- )
- metrics['genai_server_ttft'].record(
- self._ttft, attributes
- )
- metrics['genai_requests'].add(1, attributes)
- metrics['genai_completion_tokens'].add(self._output_tokens, attributes)
- metrics['genai_prompt_tokens'].add(self._input_tokens, attributes)
- metrics['genai_cost'].record(cost, attributes)
-
- except Exception as e:
- handle_exception(self._span, e)
- logger.error('Error in trace creation: %s', e)
- finally:
- self._span.end()
- raise
-
- server_address, server_port = set_server_address_and_port(instance, 'api.mistral.ai', 443)
- request_model = kwargs.get('model', 'mistral-small-latest')
-
- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}'
+ server_address, server_port = set_server_address_and_port(instance, "api.mistral.ai", 443)
+ request_model = kwargs.get("model", "mistral-small-latest")
+
+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
+ # Stream endpoint is always streaming
awaited_wrapped = wrapped(*args, **kwargs)
span = tracer.start_span(span_name, kind=SpanKind.CLIENT)
- return TracedSyncStream(awaited_wrapped, span, kwargs, server_address, server_port)
+
+ return TracedSyncStream(awaited_wrapped, span, span_name, kwargs, server_address, server_port)
return wrapper
-def embeddings(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics):
+def embed(version, environment, application_name,
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for embeddings to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Mistral API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Mistral usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the embeddings method to add telemetry.
+ Generates a telemetry wrapper for GenAI embedding function call
"""
def wrapper(wrapped, instance, args, kwargs):
"""
- Wraps the 'embeddings' API call to add telemetry.
-
- This collects metrics such as execution time, cost, and token usage, and handles errors
- gracefully, adding details to the trace for observability.
-
- Args:
- wrapped: The original 'embeddings' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'embeddings' method.
- kwargs: Keyword arguments for the 'embeddings' method.
-
- Returns:
- The response from the original 'embeddings' method.
+ Wraps the GenAI embedding function call.
"""
- server_address, server_port = set_server_address_and_port(instance, 'api.mistral.ai', 443)
- request_model = kwargs.get('model', 'mistral-embed')
+ server_address, server_port = set_server_address_and_port(instance, "api.mistral.ai", 443)
+ request_model = kwargs.get("model", "mistral-embed")
- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING} {request_model}'
+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING} {request_model}"
- with tracer.start_as_current_span(span_name, kind= SpanKind.CLIENT) as span:
+ with tracer.start_as_current_span(span_name, kind=SpanKind.CLIENT) as span:
start_time = time.time()
response = wrapped(*args, **kwargs)
- end_time = time.time()
- response_dict = response_as_dict(response)
try:
- input_tokens = response_dict.get('usage').get('prompt_tokens')
-
- # Calculate cost of the operation
- cost = get_embed_model_cost(request_model,
- pricing_info, input_tokens)
-
- # Set Span attributes (OTel Semconv)
- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')
- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING)
- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_MISTRAL)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_ENCODING_FORMATS,
- [kwargs.get('encoding_format', 'float')])
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- response_dict.get('model'))
- span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- server_address)
- span.set_attribute(SemanticConvention.SERVER_PORT,
- server_port)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- input_tokens)
-
- # Set Span attributes (Extras)
- span.set_attribute(DEPLOYMENT_ENVIRONMENT,
- environment)
- span.set_attribute(SERVICE_NAME,
- application_name)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- input_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
-
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: str(kwargs.get('inputs', '')),
- },
- )
-
- span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING,
- system=SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
- request_model=request_model,
- server_address=server_address,
- server_port=server_port,
- response_model=response_dict.get('model'),
- )
- metrics['genai_client_usage_tokens'].record(
- input_tokens, attributes
- )
- metrics['genai_client_operation_duration'].record(
- end_time - start_time, attributes
- )
- metrics['genai_requests'].add(1, attributes)
- metrics['genai_prompt_tokens'].add(input_tokens, attributes)
- metrics['genai_cost'].record(cost, attributes)
-
- # Return original response
- return response
+ response = process_embedding_response(
+ response=response,
+ request_model=request_model,
+ pricing_info=pricing_info,
+ server_port=server_port,
+ server_address=server_address,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ start_time=start_time,
+ span=span,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version,
+ **kwargs
+ )
except Exception as e:
handle_exception(span, e)
- logger.error('Error in trace creation: %s', e)
- # Return original response
- return response
+ return response
return wrapper
diff --git a/sdk/python/src/openlit/instrumentation/mistral/utils.py b/sdk/python/src/openlit/instrumentation/mistral/utils.py
new file mode 100644
index 00000000..54863d7d
--- /dev/null
+++ b/sdk/python/src/openlit/instrumentation/mistral/utils.py
@@ -0,0 +1,298 @@
+"""
+Mistral OpenTelemetry instrumentation utility functions
+"""
+import time
+
+from opentelemetry.trace import Status, StatusCode
+
+from openlit.__helpers import (
+ calculate_ttft,
+ response_as_dict,
+ calculate_tbt,
+ get_chat_model_cost,
+ get_embed_model_cost,
+ common_span_attributes,
+ record_completion_metrics,
+ record_embedding_metrics,
+)
+from openlit.semcov import SemanticConvention
+
+def format_content(messages):
+ """
+ Process a list of messages to extract content.
+ """
+
+ formatted_messages = []
+ for message in messages:
+ role = message.get("role", "user")
+ content = message.get("content", "")
+
+ if isinstance(content, list):
+ content_str = ", ".join(
+ f'{item["type"]}: {item["text"] if "text" in item else item.get("image_url", "")}'
+ if "type" in item else f'text: {item.get("text", "")}'
+ for item in content
+ )
+ formatted_messages.append(f"{role}: {content_str}")
+ else:
+ formatted_messages.append(f"{role}: {content}")
+
+ return "\n".join(formatted_messages)
+
+def process_chunk(scope, chunk):
+ """
+ Process a chunk of response data and update state.
+ """
+
+ end_time = time.time()
+ # Record the timestamp for the current chunk
+ scope._timestamps.append(end_time)
+
+ if len(scope._timestamps) == 1:
+ # Calculate time to first chunk
+ scope._ttft = calculate_ttft(scope._timestamps, scope._start_time)
+
+ chunked = response_as_dict(chunk)
+
+ # Collect message IDs and aggregated response from events
+ if chunked.get("data"):
+ data = chunked.get("data")
+ choices = data.get("choices", [])
+
+ if choices and "delta" in choices[0]:
+ delta = choices[0]["delta"]
+ content = delta.get("content")
+ if content:
+ scope._llmresponse += content
+
+ # Handle tool calls in streaming - optimized
+ delta_tools = delta.get("tool_calls")
+ if delta_tools:
+ scope._tools = scope._tools or []
+
+ for tool in delta_tools:
+ idx = tool.get("index", 0)
+
+ # Extend list if needed
+ scope._tools.extend([{}] * (idx + 1 - len(scope._tools)))
+
+ if tool.get("id"): # New tool (id exists)
+ func = tool.get("function", {})
+ scope._tools[idx] = {
+ "id": tool["id"],
+ "function": {"name": func.get("name", ""), "arguments": func.get("arguments", "")},
+ "type": tool.get("type", "function")
+ }
+ elif scope._tools[idx] and "function" in tool: # Append args (id is None)
+ scope._tools[idx]["function"]["arguments"] += tool["function"].get("arguments", "")
+
+ # Handle usage information (typically only in final chunk)
+ if data.get("usage"):
+ scope._input_tokens = data.get("usage").get("prompt_tokens", 0)
+ scope._output_tokens = data.get("usage").get("completion_tokens", 0)
+ scope._response_id = data.get("id")
+ scope._response_model = data.get("model")
+ scope._finish_reason = choices[0].get("finish_reason", "") if choices else ""
+ scope._end_time = time.time()
+
+def common_chat_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version, is_stream):
+ """
+ Process chat request and generate Telemetry
+ """
+
+ if len(scope._timestamps) > 1:
+ scope._tbt = calculate_tbt(scope._timestamps)
+
+ prompt = format_content(scope._kwargs.get("messages", []))
+ request_model = scope._kwargs.get("model", "mistral-small-latest")
+
+ cost = get_chat_model_cost(request_model, pricing_info, scope._input_tokens, scope._output_tokens)
+
+ # Common Span Attributes
+ common_span_attributes(scope,
+ SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT, SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
+ scope._server_address, scope._server_port, request_model, scope._response_model,
+ environment, application_name, is_stream, scope._tbt, scope._ttft, version)
+
+ # Span Attributes for Request parameters
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED, scope._kwargs.get("seed", ""))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY, scope._kwargs.get("frequency_penalty", 0.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS, scope._kwargs.get("max_tokens", -1))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY, scope._kwargs.get("presence_penalty", 0.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES, scope._kwargs.get("stop_sequences", []))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE, scope._kwargs.get("temperature", 0.3))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_K, scope._kwargs.get("k", 1.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P, scope._kwargs.get("p", 1.0))
+
+ # Span Attributes for Response parameters
+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID, scope._response_id)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON, [scope._finish_reason])
+ scope._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE, "text" if isinstance(scope._llmresponse, str) else "json")
+
+ # Span Attributes for Cost and Tokens
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS, scope._input_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS, scope._output_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CLIENT_TOKEN_USAGE, scope._input_tokens + scope._output_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST, cost)
+
+ # Span Attributes for Tools - optimized
+ if scope._tools:
+ tools = scope._tools if isinstance(scope._tools, list) else [scope._tools]
+
+ names, ids, args = zip(*[
+ (t.get("function", {}).get("name", ""),
+ str(t.get("id", "")),
+ str(t.get("function", {}).get("arguments", "")))
+ for t in tools if isinstance(t, dict) and t
+ ]) if tools else ([], [], [])
+
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_NAME, ", ".join(filter(None, names)))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALL_ID, ", ".join(filter(None, ids)))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_ARGS, ", ".join(filter(None, args)))
+
+ # Span Attributes for Content
+ if capture_message_content:
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CONTENT_PROMPT, prompt)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CONTENT_COMPLETION, scope._llmresponse)
+
+ # To be removed once the change to span_attributes (from span events) is complete
+ scope._span.add_event(
+ name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
+ attributes={
+ SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
+ },
+ )
+ scope._span.add_event(
+ name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
+ attributes={
+ SemanticConvention.GEN_AI_CONTENT_COMPLETION: scope._llmresponse,
+ },
+ )
+
+ scope._span.set_status(Status(StatusCode.OK))
+
+ # Metrics
+ if not disable_metrics:
+ record_completion_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT, SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
+ scope._server_address, scope._server_port, request_model, scope._response_model, environment,
+ application_name, scope._start_time, scope._end_time, scope._input_tokens, scope._output_tokens,
+ cost, scope._tbt, scope._ttft)
+
+def process_streaming_chat_response(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content=False, disable_metrics=False, version=""):
+ """
+ Process streaming chat request and generate Telemetry
+ """
+
+ common_chat_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version, is_stream=True)
+
+def process_chat_response(response, request_model, pricing_info, server_port, server_address,
+ environment, application_name, metrics, start_time, span, capture_message_content=False,
+ disable_metrics=False, version="1.0.0", **kwargs):
+ """
+ Process chat request and generate Telemetry
+ """
+
+ # Create scope object
+ scope = type("GenericScope", (), {})()
+ response_dict = response_as_dict(response)
+
+ scope._start_time = start_time
+ scope._end_time = time.time()
+ scope._span = span
+ scope._llmresponse = " ".join(
+ (choice.get("message", {}).get("content") or "")
+ for choice in response_dict.get("choices", [])
+ )
+ scope._response_id = response_dict.get("id")
+ scope._response_model = response_dict.get("model")
+ scope._input_tokens = response_dict.get("usage", {}).get("prompt_tokens", 0)
+ scope._output_tokens = response_dict.get("usage", {}).get("completion_tokens", 0)
+ scope._timestamps = []
+ scope._ttft, scope._tbt = scope._end_time - scope._start_time, 0
+ scope._server_address, scope._server_port = server_address, server_port
+ scope._kwargs = kwargs
+ scope._finish_reason = str(response_dict.get("choices", [])[0].get("finish_reason", "")) if response_dict.get("choices") else ""
+
+ # Handle tool calls
+ if kwargs.get("tools"):
+ scope._tools = response_dict.get("choices", [{}])[0].get("message", {}).get("tool_calls")
+ else:
+ scope._tools = None
+
+ common_chat_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version, is_stream=False)
+
+ return response
+
+def common_embedding_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version):
+ """
+ Process embedding request and generate Telemetry
+ """
+
+ request_model = scope._kwargs.get("model", "mistral-embed")
+ inputs = scope._kwargs.get("inputs", [])
+
+ cost = get_embed_model_cost(request_model, pricing_info, scope._input_tokens)
+
+ # Common Span Attributes
+ common_span_attributes(scope,
+ SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING, SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
+ scope._server_address, scope._server_port, request_model, scope._response_model,
+ environment, application_name, False, 0, scope._ttft, version)
+
+ # Span Attributes for Request parameters
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_ENCODING_FORMATS, [scope._kwargs.get("encoding_format", "float")])
+
+ # Span Attributes for Cost and Tokens
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS, scope._input_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CLIENT_TOKEN_USAGE, scope._input_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST, cost)
+
+ # Span Attributes for Content
+ if capture_message_content:
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CONTENT_PROMPT, str(inputs))
+
+ # To be removed once the change to span_attributes (from span events) is complete
+ scope._span.add_event(
+ name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
+ attributes={
+ SemanticConvention.GEN_AI_CONTENT_PROMPT: str(inputs),
+ },
+ )
+
+ scope._span.set_status(Status(StatusCode.OK))
+
+ # Metrics
+ if not disable_metrics:
+ record_embedding_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_EMBEDDING, SemanticConvention.GEN_AI_SYSTEM_MISTRAL,
+ scope._server_address, scope._server_port, request_model, scope._response_model, environment,
+ application_name, scope._start_time, scope._end_time, scope._input_tokens, cost)
+
+def process_embedding_response(response, request_model, pricing_info, server_port, server_address,
+ environment, application_name, metrics, start_time, span, capture_message_content=False,
+ disable_metrics=False, version="1.0.0", **kwargs):
+ """
+ Process embedding request and generate Telemetry
+ """
+
+ # Create scope object
+ scope = type("GenericScope", (), {})()
+ response_dict = response_as_dict(response)
+
+ scope._start_time = start_time
+ scope._end_time = time.time()
+ scope._span = span
+ scope._response_model = response_dict.get("model")
+ scope._input_tokens = response_dict.get("usage", {}).get("prompt_tokens", 0)
+ scope._ttft = scope._end_time - scope._start_time
+ scope._server_address, scope._server_port = server_address, server_port
+ scope._kwargs = kwargs
+
+ common_embedding_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version)
+
+ return response