samsja
samsja
Could you try to run the Executor code locally directly on your machine ? ```python from docarray import Document, DocumentArray, Executor executor = Executor.from_hub(uri='jinahub://CLIPTextEncoder', install_requirements=True) executor.encode(DocumentArray([Document(text="your text")])) ``` this will...
@hsalehgh which indexer are you using ?
If you are storing a lot of data SimpleIndexer will suffer from performance issue, as it's name suggest this indexer is for prototyping but not for real application. I would...
> Accoridng to diagram, in `D` I would call `respond` and not `send` I see something like this `jina_respond_request_bytes` instead of `jina_send_request_bytes`
> my suggestion regarding naming, is to introduce a breaking change in `jina_request_size_bytes` and have more clear naming: use `received`/`sent`, `request`/`response`, just bytes instead of size_bytes + we add runtime...
> it may help future readers to add comments above each defined metric with some description @alaeddine-13 you mean in the documentation ? How would you see it ?
Lets keep this open
any news on this issue ? Big fan of this project and ruff is great
Yes indeed `json_encoders` is a way to do it. Unfortunately it will be called not directly on the BaseModel but at its child levels. In the code it first goes...
> let > I think you could use `def dict` in Pydantic. FastAPI uses `jsonable_encoder` to create a dict from a Pydantic model then serializes the result with `json`, `orjson`,...