Legal Document RAG Pipeline

from mixpeek import Mixpeek

client = Mixpeek(api_key="YOUR_API_KEY")

namespace = client.namespaces.create(name="legal-docs")
collection = client.collections.create(
    namespace_id=namespace.id,
    name="contracts",
    extractors=["pdf-extraction", "text-embedding-v2"],
    chunk_strategy="paragraph"
)

# High BM25 weight for exact legal term matching
retriever = client.retrievers.create(
    namespace_id=namespace.id,
    name="legal-search",
    stages=[
        {"type": "hybrid_search", "vector_weight": 0.4, "bm25_weight": 0.6, "top_k": 50},
        {"type": "rerank", "model": "colbert-v2", "top_k": 10}
    ]
)

results = client.retrievers.execute(
    retriever_id=retriever.id,
    query="Section 14.2 indemnification obligations"
)