Skip to content

Python SDK for AI agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks including CrewAI, Agno, OpenAI Agents SDK, Langchain, Autogen, AG2, and CamelAI

rename prompt and completion (#131)

Shawn Qiu4018ac78dbc0
diff --git a/agentops/event.py b/agentops/event.py
index b4b10d0e..8e1ab163 100644
--- a/agentops/event.py
+++ b/agentops/event.py
@@ -76,9 +76,9 @@ class LLMEvent(Event):
agent_id(UUID, optional): The unique identifier of the agent that triggered the event.
thread_id(UUID, optional): The unique identifier of the contextual thread that a message pertains to.
- prompt_message(str, list, optional): The message or messages that were used to prompt the LLM. Preferably in ChatML format which is more fully supported by AgentOps.
+ prompt(str, list, optional): The message or messages that were used to prompt the LLM. Preferably in ChatML format which is more fully supported by AgentOps.
prompt_tokens(int, optional): The number of tokens in the prompt message.
- completion_message(str, object, optional): The message or returned by the LLM. Preferably in ChatML format which is more fully supported by AgentOps.
+ completion(str, object, optional): The message or returned by the LLM. Preferably in ChatML format which is more fully supported by AgentOps.
completion_tokens(int, optional): The number of tokens in the completion message.
model(Models, str, optional): LLM model e.g. "gpt-4". Models defined in enums.Models are more fully supported by AgentOps e.g. extra features in dashboard.
@@ -87,9 +87,9 @@ class LLMEvent(Event):
event_type: str = EventType.LLM.value
agent_id: Optional[UUID] = None
thread_id: Optional[UUID] = None
- prompt_message: str | List = None
+ prompt: str | List = None
prompt_tokens: Optional[int] = None
- completion_message: str | object = None
+ completion: str | object = None
completion_tokens: Optional[int] = None
model: Optional[Models | str] = None
diff --git a/agentops/langchain_callback_handler.py b/agentops/langchain_callback_handler.py
index daca246d..a79d3ea9 100644
--- a/agentops/langchain_callback_handler.py
+++ b/agentops/langchain_callback_handler.py
@@ -64,7 +64,7 @@ class LangchainCallbackHandler(BaseCallbackHandler):
**({} if metadata is None else metadata),
**kwargs}, # TODO: params is inconsistent, in ToolEvent we put it in logs
model=kwargs['invocation_params']['model'],
- prompt_message=prompts[0]
+ prompt=prompts[0]
# tags=tags # TODO
)
@@ -99,7 +99,7 @@ class LangchainCallbackHandler(BaseCallbackHandler):
}
llm_event.end_timestamp = get_ISO_time()
if response.llm_output is not None:
- llm_event.completion_message = response.generations[0][0].message.content # TODO
+ llm_event.completion = response.generations[0][0].message.content # TODO
llm_event.prompt_tokens = response.llm_output['token_usage']['prompt_tokens']
llm_event.completion_tokens = response.llm_output['token_usage']['completion_tokens']
self.ao_client.record(llm_event)
@@ -363,7 +363,7 @@ class AsyncLangchainCallbackHandler(AsyncCallbackHandler):
**({} if metadata is None else metadata),
**kwargs}, # TODO: params is inconsistent, in ToolEvent we put it in logs
model=kwargs['invocation_params']['model'],
- prompt_message=prompts[0]
+ prompt=prompts[0]
)
@debug_print_function_params
@@ -424,7 +424,7 @@ class AsyncLangchainCallbackHandler(AsyncCallbackHandler):
}
llm_event.end_timestamp = get_ISO_time()
if response.llm_output is not None:
- llm_event.completion_message = response.generations[0][0].message.content # TODO
+ llm_event.completion = response.generations[0][0].message.content # TODO
llm_event.prompt_tokens = response.llm_output['token_usage']['prompt_tokens']
llm_event.completion_tokens = response.llm_output['token_usage']['completion_tokens']
self.ao_client.record(llm_event)
diff --git a/agentops/llm_tracker.py b/agentops/llm_tracker.py
index 6b03c840..efb32a3c 100644
--- a/agentops/llm_tracker.py
+++ b/agentops/llm_tracker.py
@@ -50,8 +50,8 @@ class LlmTracker:
if finish_reason:
self.llm_event.agent_id = check_call_stack_for_agent_id()
- self.llm_event.prompt_message = kwargs["messages"]
- self.llm_event.completion_message = {"role": "assistant", "content": self.completion}
+ self.llm_event.prompt = kwargs["messages"]
+ self.llm_event.completion = {"role": "assistant", "content": self.completion}
self.llm_event.returns = {"finish_reason": finish_reason, "content": self.completion}
self.llm_event.model = model
self.llm_event.end_timestamp = get_ISO_time()
@@ -87,9 +87,9 @@ class LlmTracker:
# v0.0.0 responses are dicts
try:
self.llm_event.agent_id = check_call_stack_for_agent_id()
- self.llm_event.prompt_message = kwargs["messages"]
+ self.llm_event.prompt = kwargs["messages"]
self.llm_event.prompt_tokens = response['usage']['prompt_tokens']
- self.llm_event.completion_message = response['choices'][0]['message']['content']
+ self.llm_event.completion = response['choices'][0]['message']['content']
self.llm_event.completion_tokens = response['usage']['completion_tokens']
self.llm_event.returns = {"content": response['choices'][0]['message']['content']}
self.llm_event.model = response["model"]
@@ -134,8 +134,8 @@ class LlmTracker:
if finish_reason:
self.llm_event.agent_id = check_call_stack_for_agent_id()
- self.llm_event.prompt_message = kwargs["messages"]
- self.llm_event.completion_message = {"role": "assistant", "content": self.completion}
+ self.llm_event.prompt = kwargs["messages"]
+ self.llm_event.completion = {"role": "assistant", "content": self.completion}
self.llm_event.returns = {"finish_reason": finish_reason, "content": self.completion,
"function_call": function_call, "tool_calls": tool_calls, "role": role}
self.llm_event.model = model
@@ -179,9 +179,9 @@ class LlmTracker:
# v1.0.0+ responses are objects
try:
self.llm_event.agent_id = check_call_stack_for_agent_id()
- self.llm_event.prompt_message = kwargs["messages"]
+ self.llm_event.prompt = kwargs["messages"]
self.llm_event.prompt_tokens = response.usage.prompt_tokens
- self.llm_event.completion_message = response.choices[0].message.model_dump().get('content')
+ self.llm_event.completion = response.choices[0].message.model_dump().get('content')
self.llm_event.completion_tokens = response.usage.completion_tokens
self.llm_event.returns = response.model_dump()
self.llm_event.model = response.model