feat: add text_search() and hybrid_text_vector_search() to CoordinodeClient

[polaz]

Feature

Add full-text search capability to the Python SDK.

API design

coordinode core client:

# Full-text BM25 search
results = client.text_search(label="Article", query="machine learning", limit=10)
# → List[TextResult]

# Hybrid BM25 + cosine vector search (RRF fusion)
results = client.hybrid_text_vector_search(
    label="Article",
    text_query="machine learning",
    vector=[0.1, 0.2, ...],
    limit=10,
)
# → List[HybridResult]

Prerequisites:

• A full-text index must exist on the target label before searching.
  Create one with client.create_text_index(name, label, properties) or
  a CREATE TEXT INDEX Cypher statement.

langchain-coordinode adapter (CoordinodeGraph):

• keyword_search() — to be added in a follow-up

llama-index-coordinode adapter (CoordinodePropertyGraphStore):

• Evaluate whether to expose as a custom method or via structured_query()

Acceptance criteria

• client.text_search() returns results for text queries
• client.hybrid_text_vector_search() combines full-text + vector scores
• client.create_text_index() / drop_text_index() manage indexes
• LangChain CoordinodeGraph.keyword_search() implemented
• Integration tests with real CoordiNode instance
• Documented in API reference