search

Arize-ai / phoenix

AI Observability & Evaluation
https://docs.arize.com/phoenix
Other
3.02k stars 218 forks source link

[ENHANCEMENT] - Simplify Data Sending with Inline Call Back Support #3397

Open jlopatec opened 1 month ago

jlopatec commented 1 month ago

This notebook shows how to create a no boiler plate setup of Phoenix that supports inline (optional) call back. Also would suggest having a default callaback like one included here that teams can start with for retrieval.

Could we include cleaned up versions of:

It would hugely help with getting started with lower friction.

https://colab.research.google.com/drive/1_lGqhHz-V_SQ6Cs5iKLmkKYxJjrEySc1?authuser=2#scrollTo=PjtKA7ujZlWt

################################################# ########### Simple No Boiler Plate ############## #################################################

PhoenixEvalSpanProcessorExporter(project="llamaindex-llm-tracing", px_endpoint="http://localhost:6006/v1/traces", eval_callback=process_eval_hallu_qa) LlamaIndexInstrumentor().instrument()

https://colab.research.google.com/drive/1_lGqhHz-V_SQ6Cs5iKLmkKYxJjrEySc1?authuser=2#scrollTo=PjtKA7ujZlWt

mikeldking commented 2 weeks ago

@jlopatec notebook link is broken. Can you fix?