Add support for KNN vector similarity search

[Pwuts]

Added components

• KNNExpression, to be passed to RedisModel.find(). Can be used in conjunction with normal operators, e.g.:

  v_query = openai.Embedding.create(input=query).data[0]["embedding"]
  v_query = numpy.array(v_query, numpy.float32)   # may be redundant
  
  MyRedisModel.find(
      MyRedisModel.prop == "value",
      knn=KNNExpression(
          k=5,
          vector_field=MyRedisModel.__fields__["embedding"],
          reference_vector=v_query.tobytes(),
      )
  )

• VectorFieldOptions, a simple dataclass to encapsulate the settings for a vector field; passed to Field(vector_options=.

Changed components

• JsonModel and HashModel, to accommodate vector fields
  • RedisModel.from_redis(), to add a field containing the vector score when applicable
• FindQuery, to incorporate KNN vector search support

[Pwuts]

@dvora-h @chayim 0:)

[Pwuts]

@chayim where should I add instructions for using this feature?

[chayim]

IMHO we should start a separate vectorsimilarity doc... @elena-kolevska given what you're up to - any preference?

[elena-kolevska]

Looping in @bsbodden, who's starting to work on a VSS quickstart for redis.io.

[Pwuts]

bump

[chayim]

Type hint asks mostly

[codecov-commenter]

Codecov Report

Patch coverage: 70.10% and project coverage change: -0.82 ⚠️

Comparison is base (70f6401) 78.94% compared to head (583ae4a) 78.13%.

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #513      +/-   ##
==========================================
- Coverage   78.94%   78.13%   -0.82%     
==========================================
  Files          14       14              
  Lines        1178     1267      +89     
==========================================
+ Hits          930      990      +60     
- Misses        248      277      +29     

Flag	Coverage Δ
unit	78.13% <70.10%> (-0.82%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
aredis_om/__init__.py	100.00% <ø> (ø)
aredis_om/model/__init__.py	100.00% <ø> (ø)
aredis_om/model/model.py	85.30% <70.10%> (-1.79%)	⬇️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[XChikuX]

Okay, it's been more than a month. And this hasn't been merged even though all that's being corrected are type hints.
@chayim, I'm aware you are a very busy individual. Please spare 5 minutes of your busy schedule for approving this!

[chayim]

Hi folks. I just merged in #534 so that we can have a CI that is properly reflective and runs this, and tests. That was the first step.

Logic wise, this looks proper, and as if it fits with the rest of this library. Merging in the master changes, and giving it another look.

I'm pro merging this - and issuing a release.