v0.1.0a24

Increase max character limit

v0.1.0a23

Instantiate OpenAI monitor with inference pipeline id directly to avoid API request overhead.

v0.1.0a22

Update ground truth merging to be generic merging

v0.1.0a21

Add cost per token

v0.1.0a20

Raises an error if all LLM runner rows throw exceptions & adds those exceptions to the returned dataframe if they only represent a subset

v0.1.0a19

Send prompt with production data

v0.1.0a18

Add data stream method and use it in llm_monitors

v0.1.0a17

Fixes openai issues

v0.1.0a16

Rename goals to tests

v0.1.0a15

Fixes issues with tz-aware vs. tz-naive timestamps