Automatic Image Tagging Pipeline

from mixpeek import Mixpeek

client = Mixpeek(api_key="YOUR_API_KEY", namespace="image-tags")

# 1. A bucket for the photos, and a collection whose Gemini description step
#    returns tags as structured JSON. There is no separate object detector or
#    color extractor; one structured answer per image carries all three.
bucket = client.buckets.create(
    bucket_name="product-photos",
    bucket_schema={"properties": {"photo": {"type": "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",
        "parameters": {
            "run_video_description": True,
            "description_prompt": "Tag this photo for a media library.",
            "response_shape": {
                "type": "object",
                "properties": {
                    "tags": {"type": "array", "items": {"type": "string"}},
                    "objects": {"type": "array", "items": {"type": "string"}},
                    "dominant_colors": {"type": "array", "items": {"type": "string"}},
                },
            },
        },
    },
)

# 2. Upload and process
client.buckets.upload(
    bucket["bucket_id"],
    blobs=[{"property": "photo", "type": "image", "data": "s3://your-bucket/photos/IMG_0001.jpg"}],
)
client.collections.trigger(collection["collection_id"])

# 3. Read the tags back
docs = client.documents.list(collection["collection_id"], page_size=50)
for doc in docs["results"]:
    print(doc["document_id"], doc.get("json_output"))