Python SDK for agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks like CrewAI, Langchain, and Autogen
Prevents an issue with the check in stop_instrumenting().
To properly test this, I also needed to add a parameter to override the behavior of singleton.
๐ฅ Pull Request
๐ Description
Initialize llm_tracker to None when instrument_llm_calls is False.
Add "allow_multiple_instances" kwargs to suppress singleton decorator so the client may be tested.
๐ Related Issue (if applicable)
256
๐ฏ Goal
Fix an issue that may occur when Crew and Autogen packages exist but you aren't using them.
(In my case, it popped up in a secondary file's main section)
Prevents an issue with the check in stop_instrumenting(). To properly test this, I also needed to add a parameter to override the behavior of singleton.
๐ฅ Pull Request
๐ Description Initialize llm_tracker to None when instrument_llm_calls is False. Add "allow_multiple_instances" kwargs to suppress singleton decorator so the client may be tested.
๐ Related Issue (if applicable)
256
๐ฏ Goal Fix an issue that may occur when Crew and Autogen packages exist but you aren't using them. (In my case, it popped up in a secondary file's main section)
๐งช Testing Added a unit test; passed on tox.