Sports Highlights Pipeline

import requests

from mixpeek import Mixpeek

client = Mixpeek(api_key="YOUR_API_KEY", namespace="sports")
API = "https://api.mixpeek.com/v1"
HEADERS = {"Authorization": "Bearer YOUR_API_KEY", "X-Namespace": "sports"}

# 1. Game footage, cut at scene changes, with commentary transcripts and embeddings
bucket = client.buckets.create(
    bucket_name="sports-footage",
    bucket_schema={"properties": {"broadcast": {"type": "video"}}},
)
collection = client.collections.create(
    collection_name="game-scenes",
    source={"type": "bucket", "bucket_ids": [bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "multimodal_extractor",
        "version": "v1",
        "parameters": {
            "split_method": "scene",
            "run_transcription": True,
            "run_transcription_embedding": True,
        },
    },
)
client.buckets.upload(
    bucket["bucket_id"],
    blobs=[{"property": "broadcast", "type": "video", "data": "s3://your-bucket/games/cl-final.mp4"}],
)
client.collections.trigger(collection["collection_id"])

# 2. Event exemplars (goal, save, foul, celebration) and the taxonomy over them
example_bucket = client.buckets.create(
    bucket_name="event-examples",
    bucket_schema={"properties": {"clip": {"type": "video"}}},
)
examples = client.collections.create(
    collection_name="event-examples",
    source={"type": "bucket", "bucket_ids": [example_bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "multimodal_extractor",
        "version": "v1",
    },
)
client.buckets.upload(
    example_bucket["bucket_id"],
    blobs=[{"property": "clip", "type": "video", "data": "s3://your-bucket/events/goal-01.mp4"}],
)
client.collections.trigger(examples["collection_id"])
matcher = client.retrievers.create(
    retriever_name="event-matcher",
    collection_identifiers=["event-examples"],
    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": 3,
                    },
                ],
                "final_top_k": 3,
            },
        },
    ],
)

# A flat taxonomy matches each document, through a retriever, against a
# collection of labeled examples. The SDK has no taxonomies resource, so this is REST.
taxonomy = requests.post(API + "/taxonomies", headers=HEADERS, json={
    "taxonomy_name": "soccer_events",
    "config": {
        "taxonomy_type": "flat",
        "retriever_id": matcher["retriever_id"],
        "input_mappings": [{"input_key": "query", "source_type": "payload", "path": "transcription"}],
        "source_collection": {"collection_id": examples["collection_id"]},
    },
}).json()

# 3. The highlights retriever: action and commentary search, event labels, top 20
highlights = client.retrievers.create(
    retriever_name="soccer-highlights",
    collection_identifiers=["game-scenes"],
    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://multimodal_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": "label", "stage_id": "taxonomy_enrich", "parameters": {"taxonomy_id": taxonomy["taxonomy_id"], "top_k": 1}},
        {"stage_name": "top", "stage_id": "limit", "parameters": {"limit": 20}},
    ],
)

results = client.retrievers.execute(highlights["retriever_id"], inputs={"query": "goal celebration crowd roars"})
for i, doc in enumerate(results["documents"], 1):
    print(i, doc["start_time"], doc["end_time"], doc["score"], doc.get("thumbnail_url"))