v6: Aggregate NearVector and OverAll

[bevzzz]

Merge #346 first to reduce diff

This PR adds the Aggregate namespace to the collection handle with 2 aggregation types: over all and near vector. I decided to add both at once to explore the differences in the API (one has a search filter, the other one doesn't) and see where their APIs have touching points.

As usual, a large part of the diff is dedicated to tests. Since NearVector and OverAll share the same underlying implementation (see aggregate.aggregate function), the tests focus on the OverAll query and only verify that NearVector parameter is set if Aggregate.NearVector() is called.

The syntax is something like this:

songs := c.Collections.Use("Songs")
overAll := aggregate.OverAll{
  TotalCount: true,
  Boolean: aggregate.Boolean{{ Property: "isSingle", PercentageTrue: true }},
}
songs.Aggregate.OverAll(ctx, overAll)

songs.Aggregate.OverAll.GroupBy(ctx, overAll, aggregate.GroupBy{
  Property: "album",
  Limit: 5,
})

// And for NearVector (Hybrid, BM25, etc)
songs.Aggregate.NearVector(ctx, aggregate.NearVector{
  Boolean: aggregate.Boolean{{ Property: "isSingle", PercentageTrue: true }},
  Query: query.NearVector{
    Target: types.Vector{Single: []float32{...}},
    Similarity: query.Distance(0.45),
  },
})

[orca-security-eu]

Orca Security Scan Summary

Status	Check	Issues by priority
Passed	Secrets	0   0   0   0	View in Orca