Full Metal Weaviate

High level wrapper for Weaviate: iterate faster while reducing query and load boilerplate.

QUERY DATA WITH A SIMPLE SYNTAX

Query deeply nested data with . dot notation.

collection.metal_query('hasChildren.hasProperty.name=name_value')

attribute: name=value_name
logical: name=value_name&age=value_age|desc=desc_value
refs: hasChildren.name=name_value
refs deeply nested: hasChildren.hasProperty.name=name_value

LOAD TWO-WAY REFS WITH A SINGLE LINER

Using <> operator:

collection.metal_load([<uuid_source>,"<>yourRef",<uuid_target>])

loads uuid_source->yourRef->uuid_target and uuid_target->oppositeOfYourRef->uuid_source

EXAMPLE

It turn this search query:

response = collection.query.fetch_objects(
    filters=Filter.by_ref(link_on="hasChildren").by_ref(link_on="hasProperty").by_property('name').equal(name_value),
    return_references=[
        QueryReference(
            link_on="hasChildren",
            return_references=QueryReference(link_on="hasProperty", 
            return_properties=["name"])
        ),
    ]
)

into this one-liner:

response = collection.metal_query('hasChildren.hasProperty.name=name_value', 'hasChildren.hasProperty:name')

Installation

pip install git+https://github.com/thedeepengine/full-metal-weaviate.git

Documentation

https://thedeepengine.github.io/full-metal-weaviate/docs/quickstart/

Limitation/next steps/things that needs to change

Does not handle tenants
Currently this library is monkey patching Weaviate, should be handled with proper class inheritance or something
Not handling near_vector queries and other hybrid search, it might be nice to handle this as well

Disclaimer

Full Metal Weaviate is currently in alpha and under active development. It is not recommended for production use as it may contain bugs, incomplete features, and can undergo significant changes. Use at your own risk.

thedeepengine / full-metal-weaviate

readme