feat(graph): add Neo4j-based GraphRAG query engine

[bernardev254]

Summary

This PR introduces a query engine built on top of Neo4j and Cypher, enabling AI agents within the Flare AI Kit to retrieve structured blockchain transaction context for use in reasoning and decision-making processes. The engine supports contextual augmentation by exposing a set of parameterized queries for transaction histories, account relationships, and contract interactions.

Key Features

• Implemented a Cypher-based query engine (graph/engine.py) to support complex, structured queries over Flare transaction data.
• Developed a Neo4j ingestion layer (graph/indexers/neo4j_indexer.py) that stores account, contract, and transaction relationships in the graph DB.
• Exposed SDK methods for querying transaction context (e.g., transactions by account, contracts interacting with X, recent txs).
  closes [GraphRAG] Implement Query Engine and AI Agent connectors for Contextual Graph Queries #17

[dineshpinto]

Hi @bernardev254 can you add an example of using this and fix the lockfile conflict