CLI Reference

All functionality is available through the single tdw command. Run tdw --help or tdw <command> --help for the authoritative, installed help text; this page is an overview.

tdw analyze

Reflect a database schema and report what the tool infers about it.

tdw analyze <connection_string> [--format table|json|summary]
                                [--output FILE] [--metadata-only]

  • connection_string is any SQLAlchemy URL (sqlite:///shop.db, postgresql://user:pass@host/db).

  • --format selects the report style. table is the rich default; json is machine-readable; summary is a short digest.

  • --output writes the report to a file.

  • --metadata-only reflects structure without issuing any SQL that reads row data. See Privacy and Data Access.

The report covers tables and inferred entity types, detected relationships, candidate business rules, and columns flagged as potential PII.

tdw deploy

Run the full pipeline and write a generator set.

tdw deploy <connection_string> --output DIR [--seed N]
                               [--metadata-only] [--target-time N]

  • --output is the directory to write the generated tree into.

  • --seed makes the run reproducible to the byte. See Design Principles.

  • --metadata-only applies the same no-row-data guarantee as analyze.

  • --target-time sets the soft time budget reported in the phase table.

tdw demo

Build a small SQLite e-commerce database and run analyze plus deploy against it. Needs no database of your own.

tdw demo [--output DIR] [--seed N]

tdw validate

Check that a generated or hand-edited generator meets the required interface and safety rules. Validation is AST-level (syntax, unsafe imports, required methods), not string matching.

tdw validate <file_or_dir> [--check-all] [--strict]

tdw team

Configure a shared development workspace for contributors working on a generated set.

tdw team [--project DIR] [--member NAME --skill LEVEL] [--git-hooks]

tdw --version

Print the installed version.