Hoops Finance's CTO Tim Baker details their indexer upgrades on Stellar for Protocol 23 support, resumable ingestion, and historical contract specs. They store raw transaction data before parsing, resume from exact pages to avoid re-ingestion, and maintain historical WASM bytecode to correctly parse upgraded contracts. This ensures precise, durable data across Postgres, MongoDB, and Neo4j for market analytics and dashboards.
Tim Baker, CTO at Hoops Finance, shares a devlog on making their Stellar indexer bulletproof amid Protocol 23 and Soroban v4 changes. Key improvements include storing raw transaction XDRs in MongoDB before parsing for deterministic replays, implementing resumable ingestion via stored paging URLs from StellarExpert for validation without re-fetching, and creating a historical ContractCode store keyed by WASM hash to resolve function names and args from the exact bytecode active at transaction time. Their architecture triangulates data from captive core, local Horizon/RPC, Stellar ETL's Galaxie dataset, and third-party sources like StellarExpert. Postgres handles ledger state and market data, MongoDB provides fast API caching, and Neo4j powers relationship graphs for path capacity and risk analysis. These changes deliver stable semantics, real-time processing at ledger close, and richer context for swaps, liquidity, and off-ramp paths to stables like USDC.
