A ~2000 Elo Python chess engine (github.com)

Moonfish is a didactic Python chess engine (Python 3.10) that packages modern chess‑programming techniques into readable, easy‑to‑inspect code. Available via pip (pip install moonfish) and runnable as a UCI engine, web API, or Lichess bot, it achieves roughly ~2000 Elo on Lichess (beats level 5, loses to level 6) and ships with test suites including the Bratko‑Kopec tactical set. Core search and evaluation features include negamax alpha‑beta pruning, quiescence search, null‑move pruning, transposition tables, MVV‑LVA move ordering, PeSTO piece‑square (Rofchade) tapered evaluation, Syzygy tablebase support, opening book integration, and a RESTful API for remote queries. It also exposes runtime flags (search depth, quiescence depth, null‑move settings, algorithm choice) for easy experimentation. Significance: Moonfish lowers the bar for anyone wanting to study or prototype advanced search and parallelization ideas without diving into C++. Notably it implements Lazy SMP (shared‑memory parallel search) and layer‑based parallelization to illustrate multi‑core scaling strategies, making it a practical teaching and research platform for exploring search algorithms, evaluation tweaks, neural‑net integrations, time‑constrained search, and benchmarking. Its clear code, CI/test focus, and contributor roadmap make it a useful sandbox for students, researchers, and hobbyists to iterate on modern engine concepts and port ideas to higher‑performance implementations.