Multimodal Hybrid Search Pipeline

from mixpeek import Mixpeek

client = Mixpeek(api_key="YOUR_API_KEY")

# Create a multi-extractor collection
namespace = client.namespaces.create(name="multimodal-search")

# Configure hybrid retriever
retriever = client.retrievers.create(
    namespace_id=namespace.id,
    name="hybrid-multimodal",
    stages=[
        {
            "type": "hybrid_search",
            "vector_weight": 0.7,
            "bm25_weight": 0.3,
            "top_k": 100
        },
        {
            "type": "rerank",
            "model": "colbert-v2",
            "top_k": 20
        }
    ]
)

# Query across all modalities
results = client.retrievers.execute(
    retriever_id=retriever.id,
    query="quarterly revenue growth chart",
    filters={"modality": ["text", "image"]}
)