Metadata

# Generate metadata for all documents in a connector
results = client.classify.generate_batch(
    connector_name="my-s3-bucket",
    taxonomy_name="contract-analysis",
)

print(f"Processed {len(results.metadata)} documents")
for doc in results.metadata:
    print(f"\n{doc.file_name}:")
    print(f"  Type: {doc.tags.get('contract_type')}")
    print(f"  Value: ${doc.tags.get('total_value', 0):,.2f}")

# List metadata for documents
metadata = client.metadata.list(
    connector_name="my-s3-bucket",
    tag_names=["contract_type", "effective_date"],
)

for doc in metadata.documents:
    print(f"{doc.file_name}: {doc.tags}")

# Paginated listing for large datasets
page = client.metadata.list_paginated(
    connector_name="my-s3-bucket",
    page_size=100,
    page_number=1,
)
print(f"Page 1 of {page.total_pages}")

# Manually upsert metadata
client.metadata.upsert(
    file_name="contract.pdf",
    connector_name="my-s3-bucket",
    tags={
        "contract_type": "NDA",
        "reviewed": True,
        "reviewer": "John Doe",
    },
)

# Delete metadata for specific files
client.metadata.delete(
    connector_name="my-s3-bucket",
    file_names=["old-contract.pdf"],
)