Python: Wrapper + Samples#5177
Python: Wrapper + Samples#5177TaoChenOSU wants to merge 10 commits intofeature/python-foundry-hosted-agent-vnextfrom
Conversation
eavanvalkenburg
left a comment
eavanvalkenburg
left a comment
There was a problem hiding this comment.
left a bunch of comments, and also really wondering if there is way to make the samples not need 6 files?
python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py
Outdated
Show resolved
Hide resolved
python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py
Outdated
Show resolved
Hide resolved
python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py
Outdated
Show resolved
Hide resolved
python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py
Outdated
Show resolved
Hide resolved
python/samples/05-end-to-end/hosted_agents/foundry_hosted_responses/01_basic/.dockerignore
Show resolved
Hide resolved
python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py
Outdated
Show resolved
Hide resolved
python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py
Show resolved
Hide resolved
python/samples/05-end-to-end/hosted_agents/foundry_hosted_responses/02_local_tools/main.py
Show resolved
Hide resolved
| prefix: The prefix to prepend (e.g. "foundry-hosting-responses"). | ||
| """ | ||
| if prefix and prefix not in _user_agent_prefixes: | ||
| _user_agent_prefixes.append(prefix) |
There was a problem hiding this comment.
does this mean we are now keeping basically a global list of prefixes added? I don't really like that idea, can't we solve this within the foundry hosting package?
| yield update.text | ||
|
|
||
| return StreamingResponse( | ||
| stream_response(), |
There was a problem hiding this comment.
I think we should be able to add a transform_hook on the stream object that is returned from run(stream=True) because a ResponseStream should behave like a AsyncGenerator
| if self._stream: | ||
|
|
||
| async def stream_response() -> AsyncGenerator[str]: | ||
| async for update in self._agent.run(user_message, session=session, stream=True): |
There was a problem hiding this comment.
same thing here as with the responses, we need ways to pass options along
| history = await context.get_history() | ||
| messages = [*_to_messages(history), input_items] | ||
|
|
||
| chat_options = _to_chat_options(request) |
There was a problem hiding this comment.
this is a good start, I still think we need more configurability here, but that might also be something we need from the hosted team
3 participants
Motivation and Context
Initial PR to add Python Foundry hosted agent wrapper and samples.
Description
Contribution Checklist