Code Summary: Create a Search Query Using Vector Search

Embedding Model: The embedding model used in the examples below is the voyage-3.5-lite. The voyage-3.5-lite model from Voyage AI is a state-of-the-art embedding model designed for efficient and high-quality text retrieval. It supports multiple embedding dimensions—2048, 1024, 512, and 256—and offers various quantization options, including int8 and binary. Voyage-3.5-lite is suitable for a wide range of domains, such as technical documentation, code, law, finance, web reviews, long documents, and conversations.

To get started with Voyage AI, check out the Voyage AI documentation.

Create a Vector Search Query

The following code uses the $vectorSearch aggregation stage to perform a vector search.

pipeline = [
    {
        "$vectorSearch": {
            "index": "vectorPlotIndex",
            "path": "plot_embedding",
            "queryVector": embedding,
            "numCandidates": 100,
            "limit": 10
        }
    },
    {
        "$project": {
            "title": 1,
            "plot": 1,
            "score": {"$meta": "vectorSearchScore"}
        }
    }
]

x = collection.aggregate(pipeline)

Create a Vector Search Query with a Filter

The following code uses $vectorSearch with a filter on the year field to pre-filter the data before performing a vector search.

pipeline = [
    {
        "$vectorSearch": {
            "index": "vectorPlotIndex",
            "path": "plot_embedding",
            "queryVector": embedding,
            "numCandidates": 100,
            "filter": {"year": {"$gt": 2010}},
            "limit": 10
        }
    },
    {
        "$project": {
            "title": 1,
            "plot": 1,
            "year": 1,
            "score": {"$meta": "vectorSearchScore"}
        }
    }
]

x = collection.aggregate(pipeline)