Chess_db: A Framework for Working with Large Chess Game Datasets

The paper presents Chess_db, a suite of logic-programming tools for manipulating large chess-game collections in memory and through backend databases. It supports creating databases from PGN files and investigates open-source key-value stores for position tables. The intended queries include a player’s historical games and continuations from a given position, along with outcomes for White and Black. The abstract emphasizes near-instant access over substantially large game collections, but does not provide concrete benchmark datasets, latency measurements, database configurations, or comparative results.