Document Classification Pipeline

import requests

from mixpeek import Mixpeek

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

# 1. Exemplars: sample invoices, contracts, reports, forms and letters
example_bucket = client.buckets.create(
    bucket_name="document-examples",
    bucket_schema={"properties": {"document": {"type": "pdf"}}},
)
examples = client.collections.create(
    collection_name="document-examples",
    source={"type": "bucket", "bucket_ids": [example_bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "document_graph_extractor",
        "version": "v1",
    },
)
client.buckets.upload(
    example_bucket["bucket_id"],
    blobs=[{"property": "document", "type": "pdf", "data": "s3://your-bucket/examples/invoice-01.pdf"}],
)
client.collections.trigger(examples["collection_id"])

# 2. The retriever the taxonomy uses to find the closest exemplar block
matcher = client.retrievers.create(
    retriever_name="document-type-matcher",
    collection_identifiers=["document-examples"],
    input_schema={"query": {"type": "text", "required": True}},
    stages=[
        {
            "stage_name": "search",
            "stage_id": "feature_search",
            "parameters": {
                "searches": [
                    {
                        "feature_uri": "mixpeek://document_graph_extractor@v1/intfloat__multilingual_e5_large_instruct",
                        "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": "document_types",
    "config": {
        "taxonomy_type": "flat",
        "retriever_id": matcher["retriever_id"],
        "input_mappings": [{"input_key": "query", "source_type": "payload", "path": "text_raw"}],
        "source_collection": {"collection_id": examples["collection_id"]},
    },
}).json()

# 3. Incoming documents are labeled as they are written
bucket = client.buckets.create(
    bucket_name="incoming",
    bucket_schema={"properties": {"document": {"type": "pdf"}}},
)
incoming = client.collections.create(
    collection_name="incoming-documents",
    source={"type": "bucket", "bucket_ids": [bucket["bucket_id"]]},
    feature_extractor={
        "feature_extractor_name": "document_graph_extractor",
        "version": "v1",
    },
    taxonomy_applications=[{"taxonomy_id": taxonomy["taxonomy_id"], "execution_mode": "materialize"}],
)
client.buckets.upload(
    bucket["bucket_id"],
    blobs=[{"property": "document", "type": "pdf", "data": "s3://your-bucket/incoming/scan-2291.pdf"}],
)
client.collections.trigger(incoming["collection_id"])

# 4. An LLM check for blocks whose label needs a second look
review = client.retrievers.create(
    retriever_name="classification-review",
    collection_identifiers=["incoming-documents"],
    input_schema={"query": {"type": "text", "required": True}},
    stages=[
        {
            "stage_name": "search",
            "stage_id": "feature_search",
            "parameters": {
                "searches": [
                    {
                        "feature_uri": "mixpeek://document_graph_extractor@v1/intfloat__multilingual_e5_large_instruct",
                        "query": {"input_mode": "text", "value": "{{INPUT.query}}"},
                        "top_k": 50,
                    },
                ],
                "final_top_k": 50,
            },
        },
        {"stage_name": "validate", "stage_id": "llm_filter", "parameters": {"criteria": "Keep blocks that read like an invoice: line items, totals and a due date.", "provider": "google", "model_name": "gemini-2.5-flash-lite"}},
    ],
)

results = client.retrievers.execute(review["retriever_id"], inputs={"query": "amount due and payment terms"})
print(len(results["documents"]), "invoice-like blocks")