Scikit-learn DecisionTreeClassifier Updater

This repository contains the necessary code and tools to upgrade DecisionTreeClassifier models from scikit-learn==1.3.x or lower to the newer versions of scikit-learn.

Features

• Model Converter: Extracts and upgrades components of a Scikit-learn model for compatibility with newer versions than 1.3.x.
• Model Upgrader: Reconstructs and updates the model using the modified components.
• Open-source and easy-to-use CLI commands.

Installation

Clone this repository:

git clone https://github.com/Infinitode/repo-name.git
cd repo-name

Usage

Python Script Files

You can modify the Python script files to convert and upgrade your models. These files are located in new_environment and old_environment respectively.

1. Set up old_environment

• Start with the old_environment.
• You'll need to install the older versions of both scikit-learn and numpy in this environment.
• Make sure to use the same Python version your model was saved in.

Note

It needs to be similar to your original environment in which you saved your model.

2. Modify the script under old_environment and run

Modify the script file, and run it, it will convert your model and store the files that we need for later.

3. Set up new_environment

• Create a new environment for new_environment.
• Install the latest/target versions of scikit-learn and numpy.
• Use the latest/target Python version.

4. Modify the script under new_environment and run

Modify the script file and run it. It will use the files created in step 2 to create a new model, that is compatible with your chosen version.

CLI

You can use the CLI tools to quickly convert your model files.

• model_converter.py - Initial conversion. Extracts and upgrades components of the model for compatibility with newer versions than 1.3.x. Works in original environment (older scikit-learn, etc.).
• model_upgrader.py - Final conversion. Reconstructs and updates the model using the modified components. Works in new environment (updated scikit-learn, etc.)

Model Converter

python model_converter.py \
    --model_path path/to/original_model.pkl \
    --shell_path path/to/classifier_shell.pkl \
    --tree_state_path path/to/tree_state.pkl \
    --upgraded_tree_state_path path/to/upgraded/tree_state.pkl \
    --model_txt_path path/to/model.txt

Explanation of arguments:

• --model_path: Path to your saved model.
• --shell_path: Output path for the classifier shell (e.g. shell.pkl).
• --tree_state_path: Output path for the tree state (e.g. tree_state.pkl).
• --upgraded_tree_state_path: Output path for the upgraded tree state (e.g. upgraded_tree_state.pkl).
• --model_txt_path: Output path for the model's inputs and outputs (e.g. model.txt).

Model Upgrader

python model_upgrader.py \
    --shell_path path/to/classifier_shell.pkl \
    --upgraded_tree_state_path path/to/upgraded/tree_state.pkl \
    --model_txt_path path/to/model.txt \
    --output_path path/to/updated_classifier.pkl

Explanation of arguments:

• --shell_path: Path to the classifier shell (e.g. shell.pkl).
• --upgraded_tree_state_path: Path to the upgraded tree state (e.g. upgraded_tree_state.pkl).
• --model_txt_path: Path to the model's inputs and outputs (e.g. model.txt).
• --output_path: Output path for the final model (e.g. upgraded_model.pkl).

Tip

You can use --help in the commandline for usage instructions.

python model_upgrader.py --help

License

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

Contributing

Contributions are welcome! Please open an issue or submit a pull request with your changes.

Acknowledgments

This project was inspired by the need to adapt Scikit-learn models to newer versions and ensure compatibility in various environments.