dircraft 0.1.0

A tool to generate project directory structures from specifications.

dircraft

dircraft is a versatile Python package that automatically generates a project folder structure with files based on an input specification. It supports input in various formats – a YAML file, a plain text tree structure, or even direct terminal input – to create the corresponding directories and files.

Features

• Multi-format Input:
  • YAML Files: Define your folder structure with nested dictionaries.
  • Tree/Text Files: Provide a tree-like text representation of the desired structure.
  • Direct Terminal Input: Paste a file structure directly via command-line parameters.
• Customizable File Contents:
  Include file content in the input definition. If no content is provided, empty files are created.
• Recursive Directory Creation:
  Automatically creates nested directories and files.
• Command Line Interface (CLI):
  Easy-to-use CLI using Click.
• Extendable and Modular:
  Designed to be extended with additional input formats or templating engines.

Installation

Clone the repository and install it using pip:

git clone https://github.com/yourusername/dircraft.git
cd dircraft
pip install -e .

Alternatively, you can install it directly from PyPI (if published):

pip install dircraft

Usage

dircraftoldGen provides a CLI tool called dircraft that you can use to generate your project structure.

Using a YAML File

Create a YAML file (e.g., structure.yaml) that defines your file structure. For example:

modal_decomposition:
  data:
    train: {}
    val: {}
    test: {}
  models:
    vgg16_modified.py: "# TODO: VGG16 model implementation"
    pnorm_layer.py: "# TODO: Custom normalization layer"
  scripts:
    train.py: "# TODO: Training script"
    evaluate.py: "# TODO: Evaluation script"
    infer.py: "# TODO: Inference script"
  utils:
    data_loader.py: "# TODO: Data loader implementation"
    visualize.py: "# TODO: Visualization functions"
    losses.py: "# TODO: Custom loss functions"
  config:
    config.yaml: |
      learning_rate: 1e-4
      batch_size: 64
  main.py: "# Main entry point"
  README.md: "# Project documentation"
  requirements.txt: "torch\nnumpy\n..."

Then run:

dircraft structure.yaml /path/to/target/directory

Using a Tree/Text File

If you have a tree-like structure defined in a text file (e.g., structure.txt), the file might look like:

modal_decomposition/
├── data/
│   ├── train/
│   ├── val/
│   └── test/
├── models/
│   ├── vgg16_modified.py # TODO: VGG16 model implementation
│   └── pnorm_layer.py    # TODO: Custom normalization layer
├── scripts/
│   ├── train.py    # TODO: Training script
│   ├── evaluate.py # TODO: Evaluation script
│   └── infer.py    # TODO: Inference script
├── utils/
│   ├── data_loader.py   # TODO: Data loader implementation
│   ├── visualize.py     # TODO: Visualization functions
│   └── losses.py        # TODO: Custom loss functions
├── config/
│   └── config.yaml      # learning_rate: 1e-4, batch_size: 64
├── main.py     # Main entry point
├── README.md   # Project documentation
└── requirements.txt  # torch, numpy, ...

dircraft can be extended to parse such tree-like formats. (For now, the basic version uses YAML but check out our roadmap for future support.)

Direct Terminal Input

You can also pass a file structure definition directly as a command-line argument. For example:

dircraft --structure "modal_decomposition/main.py:# Main entry point; modal_decomposition/README.md:# Documentation" /path/to/target/directory

In this mode, separate multiple file definitions with a semicolon (;) and use a colon (:) to separate the file path from its content.

Note: The direct terminal input is ideal for small structures and quick tests.

Roadmap

• Enhanced Input Parsers:
  Support for more input formats (e.g., JSON, indented text trees).
• Templating Support:
  Integration with Jinja2 for templated file contents.
• GUI Frontend:
  A simple web interface to define and preview the project structure before generation.

License

This project is licensed under the MIT License – see the LICENSE file for details.

Contact

For questions or suggestions, please open an issue.