Clustering & Theme Discovery

import time
import requests

from mixpeek import Mixpeek

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

# 1. A collection that embeds every image
bucket = client.buckets.create(
    bucket_name="content",
    bucket_schema={"properties": {"image": {"type": "image"}}},
)
collection = client.collections.create(
    collection_name="content",
    source={"type": "bucket", "bucket_ids": [bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "multimodal_extractor",
        "version": "v1",
    },
)
client.buckets.upload(
    bucket["bucket_id"],
    blobs=[{"property": "image", "type": "image", "data": "s3://your-bucket/content/post-0001.jpg"}],
)
client.collections.trigger(collection["collection_id"])

# 2. HDBSCAN finds dense groups without a preset count, and an LLM names each one.
#    The SDK has no clusters resource, so this part is REST.
cluster = requests.post(API + "/clusters", headers=HEADERS, json={
    "cluster_name": "content-themes",
    "collection_ids": [collection["collection_id"]],
    "cluster_type": "vector",
    "vector_config": {
        "feature_uris": ["mixpeek://multimodal_extractor@v1/vertex_multimodal_embedding"],
        "clustering_method": "hdbscan",
        "algorithm_params": {"min_cluster_size": 15},
    },
    "llm_labeling": {"enabled": True, "provider": "openai", "model_name": "gpt-4o-mini-2024-07-18"},
}).json()
task = requests.post(API + "/clusters/" + cluster["cluster_id"] + "/execute", headers=HEADERS, json={}).json()
while client.tasks.get(task["task_id"])["status"] not in ("COMPLETED", "COMPLETED_WITH_ERRORS", "FAILED"):
    time.sleep(15)

# 3. Each group carries its label, size and keywords
groups = requests.get(API + "/clusters/" + cluster["cluster_id"] + "/groups", headers=HEADERS).json()
for group in groups["groups"]:
    print(group["label"], group["member_count"], ", ".join(group.get("keywords") or []))