Simple but powerful batch data generator for deterministic or randomized datasets.

• Lean Python DSL for rapid data synthesis.
• Batch-friendly CLI with stdin streaming and file-based workflows.
• Extensible via inline Python blocks and a rich prebuilt toolbox (math, random, rint, rstr, rfloat).

pip install mkdata

# Assuming you have UV installed, run:
uv tool install mkdata

1. Clone the repository: git clone https://github.com/RayZh-hs/mkdata.git && cd mkdata.
2. Build a wheel: python -m build.
3. Install the freshest artifact: pip install "$(ls dist/mkdata-*-py3-none-any.whl | sort | tail -n 1)".

Run a generator script (conventionally *.gen):

mkdata path/to/script.gen

Reading from stdin instead of a file:

mkdata -

You can find examples to run in the examples/ directory.

Save this file to hello.gen

@run {
  n: rint(1, 100) \n        # Random integer between 1 and 100, inclusive
  @loop n {                 # Repeat n times
    rint(1, 20)             # By default entries concat with space
  }
  \n                        # End with newline
}

mkdata hello.gen

You should see the output in stdout, eg.

5
17 3 19 8 4

A full set of examples is available in the examples/ directory.

(%)(variable:) expression (\suffix) (# comment)

• % hides output while still evaluating the expression.
• variable stores the evaluated result in the runtime environment.
• expression must be valid Python.
• \ suffix customizes the string appended to the output (\n for newline, omit for empty).

Use @ to introduce a directive. They can be nested and combined freely.

You must use braces { ... } to denote the scope of a directive. Scopes should follow the Google convention for braces (opening brace on the same line, closing brace on its own line, unless an opening brace follows immediately).

1. Create a virtual environment and activate it.
2. Install locally in editable mode: pip install -e ..
3. Install tooling: pip install pytest.
4. Run the test suite: pytest.

Formatters, linters, and additional tooling are welcome. See CONTRIBUTING.md for coordination tips.

• Contribution guide: CONTRIBUTING.md
• Release highlights: CHANGELOG.md
• License: LICENSE

This project uses the following open source packages:

• pytest for end-to-end testing.

The project's logo is designed by apien from Flaticon.