tse_tick: A Python Library for Parsing and Querying Nikkei NEEDS Tick Data from the Tokyo Stock Exchange

Battle Count: 50

Highly relevant as foundational infrastructure for TSE market microstructure research. Provides the tick-level trade-and-quote data (10-level order book, execution records) essential for studying liquidity, spread dynamics, order flow, and intraday patterns on the Tokyo Stock Exchange. The event-window extraction and order-book feature computation (spread, depth, flow imbalance, volatility) directly support quantitative strategy development. However, it is a data-access tool rather than a trading model or strategy itself. The 410× query speedup and memory-bounded ingestion make previously impractical large-scale TSE tick analyses feasible on commodity hardware.

End-user complexity is low: pip install tse-tick and use two calls (create_df/read_ticks or extract_to_store/query_ticks). The library handles four data types, two schema eras, coded fields, and Japanese layouts automatically. However, understanding the full architecture (morsel-driven streaming, part-pruning, RAM-aware parallelism, Hive partitioning, coverage markers) requires significant systems knowledge. The CLI provides scriptable access without Python. Optional dependencies (DuckDB for query path) add minimal complexity. The translation layer bridges familiar API names from yfinance/Polygon/ccxt.

The library is open-source under MIT license, available on PyPI (pip install tse-tick), with full source on GitHub. The benchmark suite ships with the library (benchmarks/ directory, run_all.py) and is reproducible by any NEEDS licensee. The test suite (637 tests) is synthetic-first, requiring no licensed data. Usage examples in the paper are locked by tests (tests/test_paper_examples.py). CI runs Python 3.9, 3.11, and 3.13. All benchmark configurations, hardware specs, and software versions are fully documented. Pre-release alpha and beta validation passes are reported in Appendix A.