filesql — Cursor Rules

• Input file formats: CSV, TSV, LTSV, Parquet, Excel + compressed versions (.gz, .bz2, .xz, .zst) • Input methods: File path, directory, io.Reader, embed.FS • Streaming: Chunk-based processing for large files • Type inference: Automatic column data type detection (INTEGER/REAL/TEXT) • Auto-save: Automatic saving at Commit or db.Close timing

filesql is a library that implements the common functionality of the following commands: "Load files such as CSV into SQLite3 and execute SQL queries against CSV, etc." • sqly: easily execute SQL against CSV/TSV/LTSV/JSON and Microsoft Excel™ with shell. • sqluv: simple terminal UI for RDBMS & CSV/TSV/LTSV at local/https/s3 filesql reads the contents of specified input files and returns sql.DB. Developers can execute SQL queries in the same way as the standard sql.DB they normally use, making it easy to learn.

• make test: Run tests and measure coverage (generates cover.out file, viewable in browser with cover.html) • make lint: Code inspection with golangci-lint (.golangci.yml configuration) • make clean: Delete generated files • make tools: Install dependency tools (golangci-lint, octocov)

• Test-Driven Development: We adopt the test-driven development promoted by t-wada (Takuto Wada). Always write test code and be mindful of the test pyramid. • Working code: Ensure that make test and make lint succeed after completing work. • Sponsor acquisition: Since development incurs financial costs, we seek sponsors via https://github.com/sponsors/nao1215. Include sponsor links in README and documentation. • Contributor acquisition: Create developer documentation so anyone can participate in development and recruit contributors. • Comments in English: Write code comments in English to accept international contributors. • User-friendly documentation comments: Write detailed explanations and example code for public functions so users can understand usage at a glance.