Questions on Using Arize Phoenix with Python SDK and Security Concerns
Hello everyone, I was try arize pheonix using the python-sdk and had a few questions.
- 1.
Is there any way to specify response_format and other model parameters like reasoning_effort, temperature etc via the sdk
prompt_name = "planner_prompt"
prompt = Client().prompts.create(
name=prompt_name,
prompt_description="planner prompt",
version=PromptVersion(
[
{"role": "system", "content": planner_prompt},
{"role": "user", "content": user_prompt},
],
model_name="o3",
model_provider="OPENAI",
template_format="F_STRING",
),
)2. When using arize-pheonix via python -m phoenix.server.main serve, it says uvicorn running at Uvicorn running on http://0.0.0.0:6006 (Press CTRL+C to quit) . This introduces a security vulnerability. Binding to 0.0.0.0 makes the application accessible from all network interfaces, not just localhost. Can someone help me with how to change this to use my localhost instead? 3. Is it safe to put my openai-api-key in the browser in the pheonix ui? I am feeling uncomfortable doing that. Can the ui use it from my .env file or any other way? Any help is appreciated. Thanks.
