erlexec
erlexec copied to clipboard
docarray
Represent, send, store and search multimodal data
`root_id` is like `parent_id`, but instead of pointing one level above, it points all the way to the top (root). https://github.com/jina-ai/docarray/pull/460
This **proof of concept** shows how we can make our dataclass types more flexible, for example by allowing the user to set the desired Image size right there. This POC...
In many places (such as, e.g., [here](https://github.com/jina-ai/docarray/commit/b905405bf2d74daab9c95f68ffceb5ca82523e74#diff-15a3e68c0ad54692e1edab9a84feee923a43d3fb8de748f91f46bf64914d5d6bR141-R143)), we use string keys to access information in a Document's `_metadata`. This is not very clean and prone to bugs/typos/etc. We should collect...
# Description Initializing a Document with a dataclass and other kwargs ignores the dataclass, and only the top level kwargs get set. # How to reproduce ```python from docarray import...
Maybe it could be interesting to have an `async` version of `match` and `find` specially for `Storage Backends` which are not CPU-bound. This would help turn the `Executor` endpoints into...
Hi, I have a question about the `load_uri_xxxx` methods does it support customization? supposed I have my images stored in to service where need to retrieve with authentication(compute some signature)
ElasticSearch by default only allows to retrieve 10k docs at once. It could be interesting to see how to use the scroll API to retrieve beyond this. Related to #430
closes: https://github.com/jina-ai/docarray/issues/376
Goals: ClickHouse allows quick insertion and searches over a large quantity of data. We use Clickhouse internally for search over a large corpus of embeddings. As Docarray provide a great...
