Same agent Smaller bill. The tools and infrastructure to get more work from your AI. Fewer tokens. Less compute. Lower cost.
TypeScript agent-harnessagent-harness-engineeringai-engineeringai-observabilityamd-gpuclaude-codeclickhousecodexcoding-agentcoding-agentscursorgenaigpu-monitoringllm-observabilityllmopsnvidia-smiobservabilityopen-sourceopenaiopentelemetry
[Feat]: Refactor ElevenLabs and AssemblyAI OpenTelemetry Monitoring Instrumentation (#772)
Patcher35425467217b
diff --git a/sdk/python/pyproject.toml b/sdk/python/pyproject.tomlindex 4a131781..749b3509 100644--- a/sdk/python/pyproject.toml+++ b/sdk/python/pyproject.toml@@ -1,6 +1,6 @@[tool.poetry]name = "openlit"-version = "1.34.8"+version = "1.34.9"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/assemblyai/__init__.py b/sdk/python/src/openlit/instrumentation/assemblyai/__init__.pyindex d17a01b2..178c45df 100644--- a/sdk/python/src/openlit/instrumentation/assemblyai/__init__.py+++ b/sdk/python/src/openlit/instrumentation/assemblyai/__init__.py@@ -5,39 +5,35 @@ import importlib.metadatafrom opentelemetry.instrumentation.instrumentor import BaseInstrumentorfrom wrapt import wrap_function_wrapper-from openlit.instrumentation.assemblyai.assemblyai import (- transcribe-)+from openlit.instrumentation.assemblyai.assemblyai import transcribe-_instruments = ('assemblyai >= 0.35.1',)+_instruments = ("assemblyai >= 0.35.1",)class AssemblyAIInstrumentor(BaseInstrumentor):"""- An instrumentor for AssemblyAI's client library.+ An instrumentor for AssemblyAI client library."""def instrumentation_dependencies(self) -> Collection[str]:return _instrumentsdef _instrument(self, **kwargs):- application_name = kwargs.get('application_name', 'default')- environment = kwargs.get('environment', 'default')- tracer = kwargs.get('tracer')- event_provider = kwargs.get('event_provider')- metrics = kwargs.get('metrics_dict')- 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('assemblyai')+ 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", False)+ disable_metrics = kwargs.get("disable_metrics")+ version = importlib.metadata.version("assemblyai")# sync transcribewrap_function_wrapper(- 'assemblyai.transcriber',- 'Transcriber.transcribe',+ "assemblyai.transcriber",+ "Transcriber.transcribe",transcribe(version, environment, application_name,- tracer, event_provider, pricing_info, capture_message_content, metrics, disable_metrics),+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),)def _uninstrument(self, **kwargs):- # Proper uninstrumentation logic to revert patched methodspassdiff --git a/sdk/python/src/openlit/instrumentation/assemblyai/assemblyai.py b/sdk/python/src/openlit/instrumentation/assemblyai/assemblyai.pyindex 5c3e6cbb..902d8191 100644--- a/sdk/python/src/openlit/instrumentation/assemblyai/assemblyai.py+++ b/sdk/python/src/openlit/instrumentation/assemblyai/assemblyai.py@@ -1,150 +1,59 @@"""-Module for monitoring Assembly AI API calls.+Module for monitoring AssemblyAI API calls."""import loggingimport time-from opentelemetry.trace import SpanKind, Status, StatusCode-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT-from openlit.__helpers import (- get_audio_model_cost,- handle_exception,- create_metrics_attributes,- set_server_address_and_port,- otel_event-)+from opentelemetry.trace import SpanKind+from openlit.__helpers import handle_exception, set_server_address_and_port+from openlit.instrumentation.assemblyai.utils import process_audio_responsefrom openlit.semcov import SemanticConvention# Initialize logger for logging potential issues and operationslogger = logging.getLogger(__name__)def transcribe(version, environment, application_name,- tracer, event_provider, pricing_info, capture_message_content, metrics, disable_metrics):+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):"""- Generates a telemetry wrapper for GenAI function call+ Generates a telemetry wrapper for AssemblyAI transcribe function call"""-def wrapper(wrapped, instance, args, kwargs):"""- Wraps the GenAI function call.+ Wraps the AssemblyAI transcribe function call."""- server_address, server_port = set_server_address_and_port(instance, 'api.assemblyai.com', 443)- request_model = kwargs.get('speech_model', 'best')+ server_address, server_port = set_server_address_and_port(instance, "api.assemblyai.com", 443)+ request_model = kwargs.get("speech_model", "best")- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO} {request_model}'+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO} {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()try:- # Calculate cost of the operation- cost = get_audio_model_cost(request_model,- pricing_info, None, response.audio_duration)-- # Set Span attributes (OTel Semconv)- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,- SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO)- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,- SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI)- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,- request_model)- span.set_attribute(SemanticConvention.SERVER_ADDRESS,- server_address)- span.set_attribute(SemanticConvention.SERVER_PORT,- server_port)- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,- request_model)- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,- 'text')-- # Set Span attributes (Extras)- span.set_attribute(DEPLOYMENT_ENVIRONMENT,- environment)- span.set_attribute(SERVICE_NAME,- application_name)- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_DURATION,- response.audio_duration)- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,- cost)- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,- version)-- # To be removed one the change to log events (from span events) is complete- if capture_message_content:- span.add_event(- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,- attributes={- SemanticConvention.GEN_AI_CONTENT_PROMPT: response.audio_url,- },- )- span.add_event(- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,- attributes={- SemanticConvention.GEN_AI_CONTENT_COMPLETION: response.text,- },- )-- input_event = otel_event(- name=SemanticConvention.GEN_AI_USER_MESSAGE,- attributes={- SemanticConvention.GEN_AI_SYSTEM: SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI- },- body={- **({'content': response.audio_url} if capture_message_content else {}),- 'role': 'user'- }+ response = wrapped(*args, **kwargs)++ response = process_audio_response(+ response=response,+ gen_ai_endpoint="assemblyai.transcribe",+ 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)- event_provider.emit(input_event)-- output_event = otel_event(- name=SemanticConvention.GEN_AI_CHOICE,- attributes={- SemanticConvention.GEN_AI_SYSTEM: SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI- },- body={- 'finish_reason': 'stop',- 'index': 0,- 'message': {- **({'content': response.text} if capture_message_content else {}),- 'role': 'assistant'- }- }- )- event_provider.emit(output_event)-- 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_AUDIO,- system=SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI,- request_model=request_model,- server_address=server_address,- server_port=server_port,- response_model=request_model,- )-- metrics['genai_client_operation_duration'].record(- end_time - start_time, attributes- )- metrics['genai_requests'].add(1, attributes)- metrics['genai_cost'].record(cost, attributes)-- # Return original response- return responseexcept Exception as e:handle_exception(span, e)- logger.error('Error in trace creation: %s', e)- # Return original response- return response+ return responsereturn wrapperdiff --git a/sdk/python/src/openlit/instrumentation/assemblyai/utils.py b/sdk/python/src/openlit/instrumentation/assemblyai/utils.pynew file mode 100644index 00000000..1a412a93--- /dev/null+++ b/sdk/python/src/openlit/instrumentation/assemblyai/utils.py@@ -0,0 +1,142 @@+"""+AssemblyAI OpenTelemetry instrumentation utility functions+"""+import time++from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT+from opentelemetry.trace import Status, StatusCode++from openlit.__helpers import (+ get_audio_model_cost,+ create_metrics_attributes,+)+from openlit.semcov import SemanticConvention++def format_audio_url(audio_url):+ """+ Process audio URL input to extract content.+ """+ return str(audio_url) if audio_url else ""++def common_span_attributes(scope, gen_ai_operation, gen_ai_system, server_address, server_port,+ request_model, response_model, environment, application_name, is_stream, tbt, ttft, version):+ """+ Set common span attributes for both chat and RAG operations.+ """++ scope._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")+ scope._span.set_attribute(SemanticConvention.GEN_AI_OPERATION, gen_ai_operation)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM, gen_ai_system)+ scope._span.set_attribute(SemanticConvention.SERVER_ADDRESS, server_address)+ scope._span.set_attribute(SemanticConvention.SERVER_PORT, server_port)+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL, request_model)+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL, scope._response_model)+ scope._span.set_attribute(DEPLOYMENT_ENVIRONMENT, environment)+ scope._span.set_attribute(SERVICE_NAME, application_name)+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM, is_stream)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT, scope._tbt)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT, scope._ttft)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION, version)++def record_audio_metrics(metrics, gen_ai_operation, gen_ai_system, server_address, server_port,+ request_model, response_model, environment, application_name, start_time, end_time, cost):+ """+ Record audio metrics for the operation.+ """++ attributes = create_metrics_attributes(+ operation=gen_ai_operation,+ system=gen_ai_system,+ server_address=server_address,+ server_port=server_port,+ request_model=request_model,+ response_model=response_model,+ service_name=application_name,+ deployment_environment=environment,+ )+ metrics["genai_client_operation_duration"].record(end_time - start_time, attributes)+ metrics["genai_requests"].add(1, attributes)+ metrics["genai_cost"].record(cost, attributes)++def common_audio_logic(scope, gen_ai_endpoint, pricing_info, environment, application_name,+ metrics, capture_message_content, disable_metrics, version):+ """+ Process audio transcription request and generate Telemetry+ """++ prompt = scope._response.audio_url+ request_model = scope._kwargs.get("speech_model", "best")+ is_stream = False++ # Calculate cost based on audio duration+ cost = get_audio_model_cost(request_model, pricing_info, prompt, scope._response.audio_duration)++ # Common Span Attributes+ common_span_attributes(scope,+ SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO, SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI,+ scope._server_address, scope._server_port, request_model, request_model,+ environment, application_name, is_stream, scope._tbt, scope._ttft, version)++ # Span Attributes for Response parameters+ scope._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE, "text")+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID, scope._response.id)++ # Span Attributes for Cost+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST, cost)++ # Audio-specific span attributes+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_DURATION, scope._response.audio_duration)++ # 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._response.text)++ # 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._response.text,+ },+ )++ scope._span.set_status(Status(StatusCode.OK))++ # Metrics+ if not disable_metrics:+ record_audio_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO,+ SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI, scope._server_address, scope._server_port,+ request_model, request_model, environment, application_name, scope._start_time,+ scope._end_time, cost)++def process_audio_response(response, gen_ai_endpoint, 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 audio transcription request and generate Telemetry+ """++ scope = type("GenericScope", (), {})()++ scope._start_time = start_time+ scope._end_time = time.time()+ scope._span = span+ scope._server_address, scope._server_port = server_address, server_port+ scope._kwargs = kwargs+ scope._response = response++ # Initialize streaming and timing values for AssemblyAI transcription+ scope._response_model = kwargs.get("speech_model", "best")+ scope._tbt = 0.0+ scope._ttft = scope._end_time - scope._start_time++ common_audio_logic(scope, gen_ai_endpoint, pricing_info, environment, application_name,+ metrics, capture_message_content, disable_metrics, version)++ return responsediff --git a/sdk/python/src/openlit/instrumentation/elevenlabs/__init__.py b/sdk/python/src/openlit/instrumentation/elevenlabs/__init__.pyindex ba08ed6a..ebb898a2 100644--- a/sdk/python/src/openlit/instrumentation/elevenlabs/__init__.py+++ b/sdk/python/src/openlit/instrumentation/elevenlabs/__init__.py@@ -1,4 +1,3 @@-# pylint: disable=useless-return, bad-staticmethod-argument, disable=duplicate-code"""Initializer of Auto Instrumentation of ElevenLabs Functions"""from typing import Collection@@ -6,18 +5,14 @@ import importlib.metadatafrom opentelemetry.instrumentation.instrumentor import BaseInstrumentorfrom wrapt import wrap_function_wrapper-from openlit.instrumentation.elevenlabs.elevenlabs import (- generate-)-from openlit.instrumentation.elevenlabs.async_elevenlabs import (- async_generate-)+from openlit.instrumentation.elevenlabs.elevenlabs import generate+from openlit.instrumentation.elevenlabs.async_elevenlabs import async_generate_instruments = ("elevenlabs >= 1.4.0",)class ElevenLabsInstrumentor(BaseInstrumentor):"""- An instrumentor for ElevenLabs's client library.+ An instrumentor for ElevenLabs client library."""def instrumentation_dependencies(self) -> Collection[str]:@@ -33,14 +28,6 @@ class ElevenLabsInstrumentor(BaseInstrumentor):disable_metrics = kwargs.get("disable_metrics")version = importlib.metadata.version("elevenlabs")- # sync generate- wrap_function_wrapper(- "elevenlabs.client",- "ElevenLabs.generate",- generate("elevenlabs.generate", version, environment, application_name,- tracer, pricing_info, capture_message_content, metrics, disable_metrics),- )-# sync text_to_speech.convertwrap_function_wrapper("elevenlabs.text_to_speech.client",@@ -49,22 +36,13 @@ class ElevenLabsInstrumentor(BaseInstrumentor):tracer, pricing_info, capture_message_content, metrics, disable_metrics),)- # async generate- wrap_function_wrapper(- "elevenlabs.client",- "AsyncElevenLabs.generate",- async_generate("elevenlabs.generate", version, environment, application_name,- tracer, pricing_info, capture_message_content, metrics, disable_metrics),- )-- # sync text_to_speech.convert+ # async text_to_speech.convertwrap_function_wrapper("elevenlabs.text_to_speech.client","AsyncTextToSpeechClient.convert",- generate("elevenlabs.text_to_speech", version, environment, application_name,+ async_generate("elevenlabs.text_to_speech", version, environment, application_name,tracer, pricing_info, capture_message_content, metrics, disable_metrics),)def _uninstrument(self, **kwargs):- # Proper uninstrumentation logic to revert patched methodspassdiff --git a/sdk/python/src/openlit/instrumentation/elevenlabs/async_elevenlabs.py b/sdk/python/src/openlit/instrumentation/elevenlabs/async_elevenlabs.pyindex 6297cccb..a8229bd4 100644--- a/sdk/python/src/openlit/instrumentation/elevenlabs/async_elevenlabs.py+++ b/sdk/python/src/openlit/instrumentation/elevenlabs/async_elevenlabs.py@@ -1,145 +1,55 @@"""-Module for monitoring Ollama API calls.+Module for monitoring ElevenLabs API calls."""-import loggingimport time-from opentelemetry.trace import SpanKind, Status, StatusCode-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT-from openlit.__helpers import (- get_audio_model_cost,- handle_exception,- create_metrics_attributes,-)+from opentelemetry.trace import SpanKind+from openlit.__helpers import handle_exception+from openlit.instrumentation.elevenlabs.utils import process_audio_responsefrom openlit.semcov import SemanticConvention-# Initialize logger for logging potential issues and operations-logger = logging.getLogger(__name__)-def async_generate(gen_ai_endpoint, version, environment, application_name,- tracer, pricing_info, capture_message_content, metrics, disable_metrics):+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):"""- Generates a telemetry wrapper for creating speech audio 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 ElevenLabs API.- tracer: OpenTelemetry tracer for creating spans.- pricing_info: Information used for calculating the cost of generating speech audio.- capture_message_content: Flag indicating whether to trace the input text and generated audio.-- Returns:- A function that wraps the speech audio creation method to add telemetry.+ Generates a telemetry wrapper for GenAI function call"""async def wrapper(wrapped, instance, args, kwargs):"""- Wraps the 'generate' API call to add telemetry.-- This collects metrics such as execution time, cost, and handles errors- gracefully, adding details to the trace for observability.-- Args:- wrapped: The original 'generate' method to be wrapped.- instance: The instance of the class where the original method is defined.- args: Positional arguments for the 'generate' method.- kwargs: Keyword arguments for the 'generate' method.-- Returns:- The response from the original 'generate' method.+ Wraps the GenAI function call."""server_address, server_port = "api.elevenlabs.io", 443- request_model = kwargs.get('model', kwargs.get('model_id', 'eleven_multilingual_v2'))+ request_model = kwargs.get("model", kwargs.get("model_id", "eleven_multilingual_v2"))- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO} {request_model}'+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO} {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 = await wrapped(*args, **kwargs)- end_time = time.time()+ response = wrapped(*args, **kwargs)try:- # Calculate cost of the operation- cost = get_audio_model_cost(request_model,- pricing_info, kwargs.get('text', ''))-- # Set Span attributes- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,- SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO)- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,- SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI)- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,- request_model)- span.set_attribute(SemanticConvention.SERVER_ADDRESS,- server_address)- span.set_attribute(SemanticConvention.SERVER_PORT,- server_port)- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,- request_model)- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,- request_model)- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,- 'audio')-- # Set Span attributes (Extras)- if gen_ai_endpoint == 'elevenlabs.generate':- if isinstance(kwargs.get('voice', 'Rachel'), str):- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_VOICE,- kwargs.get('voice', 'Rachel'))- else:- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_VOICE,- kwargs.get('voice_id', ''))- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_RESPONSE_FORMAT,- kwargs.get('output_format', 'mp3'))- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_SETTINGS,- str(kwargs.get('voice_settings', '')))- span.set_attribute(DEPLOYMENT_ENVIRONMENT,- environment)- span.set_attribute(SERVICE_NAME,- application_name)- 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('text', '')),- },- )-- 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_AUDIO,- system=SemanticConvention.GEN_AI_SYSTEM_ELEVENLABS,- request_model=request_model,- server_address=server_address,- server_port=server_port,- response_model=request_model,- )-- metrics['genai_client_operation_duration'].record(- end_time - start_time, attributes- )- metrics['genai_requests'].add(1, attributes)- metrics['genai_cost'].record(cost, attributes)-- # Return original response- return response+ response = process_audio_response(+ response=response,+ gen_ai_endpoint=gen_ai_endpoint,+ 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,+ args=args,+ kwargs=kwargs,+ capture_message_content=capture_message_content,+ disable_metrics=disable_metrics,+ version=version+ )except Exception as e:handle_exception(span, e)- logger.error('Error in trace creation: %s', e)- # Return original response- return response+ return responsereturn wrapperdiff --git a/sdk/python/src/openlit/instrumentation/elevenlabs/elevenlabs.py b/sdk/python/src/openlit/instrumentation/elevenlabs/elevenlabs.pyindex 84f9c99b..4963c89c 100644--- a/sdk/python/src/openlit/instrumentation/elevenlabs/elevenlabs.py+++ b/sdk/python/src/openlit/instrumentation/elevenlabs/elevenlabs.py@@ -1,145 +1,55 @@"""-Module for monitoring Ollama API calls.+Module for monitoring ElevenLabs API calls."""-import loggingimport time-from opentelemetry.trace import SpanKind, Status, StatusCode-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT-from openlit.__helpers import (- get_audio_model_cost,- handle_exception,- create_metrics_attributes,-)+from opentelemetry.trace import SpanKind+from openlit.__helpers import handle_exception+from openlit.instrumentation.elevenlabs.utils import process_audio_responsefrom openlit.semcov import SemanticConvention-# Initialize logger for logging potential issues and operations-logger = logging.getLogger(__name__)-def generate(gen_ai_endpoint, version, environment, application_name,- tracer, pricing_info, capture_message_content, metrics, disable_metrics):+ tracer, pricing_info, capture_message_content, metrics, disable_metrics):"""- Generates a telemetry wrapper for creating speech audio 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 ElevenLabs API.- tracer: OpenTelemetry tracer for creating spans.- pricing_info: Information used for calculating the cost of generating speech audio.- capture_message_content: Flag indicating whether to trace the input text and generated audio.-- Returns:- A function that wraps the speech audio creation method to add telemetry.+ Generates a telemetry wrapper for GenAI function call"""def wrapper(wrapped, instance, args, kwargs):"""- Wraps the 'generate' API call to add telemetry.-- This collects metrics such as execution time, cost, and handles errors- gracefully, adding details to the trace for observability.-- Args:- wrapped: The original 'generate' method to be wrapped.- instance: The instance of the class where the original method is defined.- args: Positional arguments for the 'generate' method.- kwargs: Keyword arguments for the 'generate' method.-- Returns:- The response from the original 'generate' method.+ Wraps the GenAI function call."""server_address, server_port = "api.elevenlabs.io", 443- request_model = kwargs.get('model', kwargs.get('model_id', 'eleven_multilingual_v2'))+ request_model = kwargs.get("model", kwargs.get("model_id", "eleven_multilingual_v2"))- span_name = f'{SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO} {request_model}'+ span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO} {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()try:- # Calculate cost of the operation- cost = get_audio_model_cost(request_model,- pricing_info, kwargs.get('text', ''))-- # Set Span attributes- span.set_attribute(TELEMETRY_SDK_NAME, 'openlit')- span.set_attribute(SemanticConvention.GEN_AI_OPERATION,- SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO)- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,- SemanticConvention.GEN_AI_SYSTEM_ASSEMBLYAI)- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,- request_model)- span.set_attribute(SemanticConvention.SERVER_ADDRESS,- server_address)- span.set_attribute(SemanticConvention.SERVER_PORT,- server_port)- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,- request_model)- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,- request_model)- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,- 'audio')-- # Set Span attributes (Extras)- if gen_ai_endpoint == 'elevenlabs.generate':- if isinstance(kwargs.get('voice', 'Rachel'), str):- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_VOICE,- kwargs.get('voice', 'Rachel'))- else:- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_VOICE,- kwargs.get('voice_id', ''))- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_RESPONSE_FORMAT,- kwargs.get('output_format', 'mp3'))- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_SETTINGS,- str(kwargs.get('voice_settings', '')))- span.set_attribute(DEPLOYMENT_ENVIRONMENT,- environment)- span.set_attribute(SERVICE_NAME,- application_name)- 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('text', '')),- },- )-- 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_AUDIO,- system=SemanticConvention.GEN_AI_SYSTEM_ELEVENLABS,- request_model=request_model,- server_address=server_address,- server_port=server_port,- response_model=request_model,- )-- metrics['genai_client_operation_duration'].record(- end_time - start_time, attributes- )- metrics['genai_requests'].add(1, attributes)- metrics['genai_cost'].record(cost, attributes)-- # Return original response- return response+ response = process_audio_response(+ response=response,+ gen_ai_endpoint=gen_ai_endpoint,+ 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,+ args=args,+ kwargs=kwargs,+ capture_message_content=capture_message_content,+ disable_metrics=disable_metrics,+ version=version+ )except Exception as e:handle_exception(span, e)- logger.error('Error in trace creation: %s', e)- # Return original response- return response+ return responsereturn wrapperdiff --git a/sdk/python/src/openlit/instrumentation/elevenlabs/utils.py b/sdk/python/src/openlit/instrumentation/elevenlabs/utils.pynew file mode 100644index 00000000..2a83d952--- /dev/null+++ b/sdk/python/src/openlit/instrumentation/elevenlabs/utils.py@@ -0,0 +1,133 @@+"""+ElevenLabs OpenTelemetry instrumentation utility functions+"""+import time++from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT+from opentelemetry.trace import Status, StatusCode++from openlit.__helpers import (+ get_audio_model_cost,+ create_metrics_attributes,+)+from openlit.semcov import SemanticConvention++def format_content(text):+ """+ Process text input to extract content.+ """+ return str(text) if text else ""++def common_span_attributes(scope, gen_ai_operation, gen_ai_system, server_address, server_port,+ request_model, response_model, environment, application_name, is_stream, tbt, ttft, version):+ """+ Set common span attributes for both chat and RAG operations.+ """++ scope._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")+ scope._span.set_attribute(SemanticConvention.GEN_AI_OPERATION, gen_ai_operation)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM, gen_ai_system)+ scope._span.set_attribute(SemanticConvention.SERVER_ADDRESS, server_address)+ scope._span.set_attribute(SemanticConvention.SERVER_PORT, server_port)+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL, request_model)+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL, scope._response_model)+ scope._span.set_attribute(DEPLOYMENT_ENVIRONMENT, environment)+ scope._span.set_attribute(SERVICE_NAME, application_name)+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM, is_stream)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT, scope._tbt)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT, scope._ttft)+ scope._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION, version)++def record_audio_metrics(metrics, gen_ai_operation, gen_ai_system, server_address, server_port,+ request_model, response_model, environment, application_name, start_time, end_time, cost):+ """+ Record audio generation metrics for the operation.+ """++ attributes = create_metrics_attributes(+ operation=gen_ai_operation,+ system=gen_ai_system,+ server_address=server_address,+ server_port=server_port,+ request_model=request_model,+ response_model=response_model,+ service_name=application_name,+ deployment_environment=environment,+ )+ metrics["genai_client_operation_duration"].record(end_time - start_time, attributes)+ metrics["genai_requests"].add(1, attributes)+ metrics["genai_cost"].record(cost, attributes)++def common_audio_logic(scope, gen_ai_endpoint, pricing_info, environment, application_name,+ metrics, capture_message_content, disable_metrics, version):+ """+ Process audio generation request and generate Telemetry+ """++ text = format_content(scope._kwargs.get("text", ""))+ request_model = scope._kwargs.get("model", scope._kwargs.get("model_id", "eleven_multilingual_v2"))+ is_stream = False # ElevenLabs audio generation is not streaming++ cost = get_audio_model_cost(request_model, pricing_info, text)++ # Common Span Attributes+ common_span_attributes(scope,+ SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO, SemanticConvention.GEN_AI_SYSTEM_ELEVENLABS,+ scope._server_address, scope._server_port, request_model, request_model,+ environment, application_name, is_stream, scope._tbt, scope._ttft, version)++ # Span Attributes for Cost and Tokens+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST, cost)++ # Span Attributes for Response parameters+ scope._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE, scope._kwargs.get("output_format", "mp3_44100_128"))++ # Audio-specific span attributes+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_VOICE, scope._kwargs.get("voice_id", ""))+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_AUDIO_SETTINGS, str(scope._kwargs.get("voice_settings", "")))++ # Span Attributes for Content+ if capture_message_content:+ scope._span.set_attribute(SemanticConvention.GEN_AI_CONTENT_PROMPT, text)++ # 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: text,+ },+ )++ scope._span.set_status(Status(StatusCode.OK))++ # Metrics+ if not disable_metrics:+ record_audio_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_AUDIO, SemanticConvention.GEN_AI_SYSTEM_ELEVENLABS,+ scope._server_address, scope._server_port, request_model, request_model, environment,+ application_name, scope._start_time, scope._end_time, cost)++def process_audio_response(response, gen_ai_endpoint, pricing_info, server_port, server_address,+ environment, application_name, metrics, start_time, span, args, kwargs, capture_message_content=False,+ disable_metrics=False, version="1.0.0"):+ """+ Process audio generation request and generate Telemetry+ """++ scope = type("GenericScope", (), {})()++ scope._start_time = start_time+ scope._end_time = time.time()+ scope._span = span+ scope._server_address, scope._server_port = server_address, server_port+ scope._kwargs = kwargs+ scope._args = args++ # Initialize streaming and timing values for ElevenLabs audio generation+ scope._response_model = kwargs.get("model", kwargs.get("model_id", "eleven_multilingual_v2"))+ scope._tbt = 0.0+ scope._ttft = scope._end_time - scope._start_time++ common_audio_logic(scope, gen_ai_endpoint, pricing_info, environment, application_name,+ metrics, capture_message_content, disable_metrics, version)++ return responsediff --git a/sdk/python/src/openlit/instrumentation/gpt4all/utils.py b/sdk/python/src/openlit/instrumentation/gpt4all/utils.pyindex 1dfc380e..56481687 100644--- a/sdk/python/src/openlit/instrumentation/gpt4all/utils.py+++ b/sdk/python/src/openlit/instrumentation/gpt4all/utils.py@@ -109,7 +109,7 @@ def record_embedding_metrics(metrics, gen_ai_operation, gen_ai_system, server_admetrics["genai_prompt_tokens"].add(input_tokens, attributes)metrics["genai_cost"].record(cost, attributes)-def common_generate_logic(scope, pricing_info, environment, application_name, metrics,+def common_t2s_logic(scope, pricing_info, environment, application_name, metrics,capture_message_content, disable_metrics, version, is_stream):"""Process generate request and generate Telemetry@@ -228,7 +228,7 @@ def process_streaming_generate_response(scope, pricing_info, environment, applic"""Process generate request and generate Telemetry"""- common_generate_logic(scope, pricing_info, environment, application_name, metrics,+ common_t2s_logic(scope, pricing_info, environment, application_name, metrics,capture_message_content, disable_metrics, version, is_stream=True)def process_generate_response(response, request_model, pricing_info, server_port, server_address,@@ -252,7 +252,7 @@ def process_generate_response(response, request_model, pricing_info, server_portscope._args = argsscope._tools = None- common_generate_logic(scope, pricing_info, environment, application_name, metrics,+ common_t2s_logic(scope, pricing_info, environment, application_name, metrics,capture_message_content, disable_metrics, version, is_stream=False)return responsediff --git a/sdk/python/tests/test_elevenlabs.py.hold b/sdk/python/tests/test_elevenlabs.pysimilarity index 55%rename from sdk/python/tests/test_elevenlabs.py.holdrename to sdk/python/tests/test_elevenlabs.pyindex 0ef7c21c..76bf4b68 100644--- a/sdk/python/tests/test_elevenlabs.py.hold+++ b/sdk/python/tests/test_elevenlabs.py@@ -28,36 +28,6 @@ openlit.init(application_name="openlit-python-elevenlabs-test")-def test_sync_elevenlabs_generate():- """- Tests synchronous generate text-to-speech with the 'eleven_multilingual_v2' model.-- Raises:- AssertionError: If the generate text-to-speech response object is not as expected.- """-- try:- audio = sync_client.generate(- text="Say LLM Monitoring",- voice="Sarah",- model="eleven_multilingual_v2"- )- assert isinstance(audio, types.GeneratorType)-- audio = sync_client.generate(- text="Say LLM Observability!",- voice="Sarah",- model="eleven_multilingual_v2",- stream=True- )- assert isinstance(audio, types.GeneratorType)-- # pylint: disable=broad-exception-caught- except Exception as e:- if "rate limit" in str(e).lower():- print("Rate Limited:", e)- else:- raisedef test_sync_elevenlabs_t2s():"""@@ -83,38 +53,6 @@ def test_sync_elevenlabs_t2s():else:raise-async def test_async_elevenlabs_generate():- """- Tests synchronous generate text-to-speech with the 'eleven_multilingual_v2' model.-- Raises:- AssertionError: If the generate text-to-speech response object is not as expected.- """-- try:- audio = await async_client.generate(- text="Say LLM Monitoring",- voice="Sarah",- model="eleven_multilingual_v2"- )- assert isinstance(audio, types.AsyncGeneratorType)-- audio = await async_client.generate(- text="Say LLM Observability!",- voice="Sarah",- model="eleven_multilingual_v2",- stream=True- )- assert isinstance(audio, types.AsyncGeneratorType)-- # pylint: disable=broad-exception-caught- except Exception as e:- if "rate limit" in str(e).lower():- print("Rate Limited:", e)- else:- raise-@pytest.mark.asyncioasync def test_async_elevenlabs_t2s():"""