AI-Powered Catalog Search

from mixpeek import Mixpeek

client = Mixpeek(api_key="YOUR_API_KEY", namespace="catalog")

# 1. A bucket whose objects carry a product photo and its description, and one collection per extractor over it
bucket = client.buckets.create(
    bucket_name="products",
    bucket_schema={
        "properties": {
            "photo": {
                "type": "image",
            },
            "description": {
                "type": "text",
            },
        },
    },
)
image_collection = client.collections.create(
    collection_name="product_photos",
    source={"type": "bucket", "bucket_ids": [bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "multimodal_extractor",
        "version": "v1",
    },
)
text_collection = client.collections.create(
    collection_name="product_text",
    source={"type": "bucket", "bucket_ids": [bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "text_extractor",
        "version": "v1",
    },
)

# 2. Upload and process
client.buckets.upload(
    bucket["bucket_id"],
    blobs=[{"property": "photo", "type": "image", "data": "s3://your-bucket/products/linen-dress.jpg"}, {"property": "description", "type": "text", "data": "s3://your-bucket/products/linen-dress.txt"}],
)
client.collections.trigger(image_collection["collection_id"])
client.collections.trigger(text_collection["collection_id"])

# 3. The query searched against photos and descriptions, fused with RRF, then a rerank on description text
retriever = client.retrievers.create(
    retriever_name="catalog-search",
    collection_identifiers=["product_photos", "product_text"],
    input_schema={
        "query": {
            "type": "text",
            "required": True,
        },
    },
    stages=[
        {
            "stage_name": "search",
            "stage_id": "feature_search",
            "parameters": {
                "searches": [
                    {
                        "feature_uri": "mixpeek://multimodal_extractor@v1/vertex_multimodal_embedding",
                        "query": {
                            "input_mode": "text",
                            "value": "{{INPUT.query}}",
                        },
                        "top_k": 100,
                    },
                    {
                        "feature_uri": "mixpeek://text_extractor@v1/multilingual_e5_large_instruct_v1",
                        "query": {
                            "input_mode": "text",
                            "value": "{{INPUT.query}}",
                        },
                        "top_k": 100,
                    },
                ],
                "fusion": "rrf",
                "final_top_k": 100,
            },
        },
        {
            "stage_name": "rerank",
            "stage_id": "rerank",
            "parameters": {
                "inference_name": "BAAI__bge_reranker_v2_m3",
                "query": "{{INPUT.query}}",
                "document_field": "text",
                "top_k": 20,
            },
        },
    ],
)

# 4. Search
results = client.retrievers.execute(
    retriever["retriever_id"],
    inputs={
    "query": "lightweight summer dress under $50",
},
)

for doc in results["documents"]:
    print(doc["document_id"], doc["score"])