You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When attempting to deploy a pipeline created with AzureOpenAIGenerator to Hayhooks with hayhooks deploy my_pipeline.yaml, I encountered the following issues:
Unexpected Parameters: Hayhooks did not seem to expect default_headers and generation_kwargs. I removed them from the YAML configuration, and it got better (into the following errors), but I’m unsure why these were problematic.
Pydantic type handling: I encountered type handling issues in Pydantic, specifically with types like Optional[Callable[[StreamingChunk], NoneType]].
Pipeline parameters mismatch: The deployed pipeline required additional parameters beyond those expected in pipe.run(). At this point the pipeline returned the desired results, but I had to pass empty parameters to make it work
500 Internal Server Error: After resolving issues 2 and 3, I encountered a new error but only when trying to access the /docs page. The logs show the following:
hayhooks-1 | INFO: 172.19.0.1:45200 - "GET /openapi.json HTTP/1.1" 500 Internal Server Error hayhooks-1 | ERROR: Exception in ASGI application
The error seems to originate from Pydantic: pydantic.errors.PydanticInvalidForJsonSchema: Cannot generate a JsonSchema for core_schema.CallableSchema
For reference, the pipeline consisted of a single PromptBuilder and a generator:
Hello @alex-stoica , @mpangrazzi worked on this issue and we tested and merged #41
In case it still doesn't work for you, feel free to reopen this issue.
When attempting to deploy a pipeline created with
AzureOpenAIGenerator
to Hayhooks withhayhooks deploy my_pipeline.yaml
, I encountered the following issues:Unexpected Parameters: Hayhooks did not seem to expect
default_headers
andgeneration_kwargs
. I removed them from the YAML configuration, and it got better (into the following errors), but I’m unsure why these were problematic.Pydantic type handling: I encountered type handling issues in Pydantic, specifically with types like
Optional[Callable[[StreamingChunk], NoneType]]
.Pipeline parameters mismatch: The deployed pipeline required additional parameters beyond those expected in
pipe.run()
. At this point the pipeline returned the desired results, but I had to pass empty parameters to make it work500 Internal Server Error: After resolving issues 2 and 3, I encountered a new error but only when trying to access the /docs page. The logs show the following:
The error seems to originate from Pydantic:
pydantic.errors.PydanticInvalidForJsonSchema: Cannot generate a JsonSchema for core_schema.CallableSchema
For reference, the pipeline consisted of a single
PromptBuilder
and a generator:I have fixes for issues 2 and 3, but they may not be useful as issue 4 persists.
Deployment: Docker with image deepset/hayhooks:main
Python Version: 3.12
The text was updated successfully, but these errors were encountered: