codemodder-java

This is the Java version of the Codemodder Framework, which builds on traditional codemod frameworks by providing codemods with additional context and services. Codemodder plugins inject codemods with the context and services they need to perform complex transforms.

Pluggability and the complex transforms they enable distinguish codemodder codemods from traditional codemods.

For directions on how to write your own codemods, visit the Codemodder documentation site.

Codemodder is sponsored by pixee.ai.

Developing

Follow these instructions if you intend to modify and build this project from source.

First Time Set Up

1. Install JDK 17 for building this project. We recommend Eclipse Adoptium

2. Install Semgrep CLI. See here for instructions. It can usually be done via pip:

   pip install semgrep

If your python library paths contain your home directory as a root folder (i.e. due to the use of the $HOME environment variable), you may need to manually set up your PYTHONPATH for tests:

PYTHONPATH=$HOME/<subpath-to-python-libs-folder> ./gradlew check

You can check your python paths with:

python -m site

Building

./gradlew assemble

Running Tests

./gradlew check

You need to be able to run Docker for some tests. Make sure you have it installed and have the necessary permissions to run with your user.

OpenAI Integration Tests

Some tests integrate with the OpenAI platform. To run these tests, configure an OpenAI key in the Gradle property codemodderOpenAIKey. When no API key is configured, the tests will be skipped.

You can add the Gradle property to your ~/.gradle/gradle.properties file:

codemodderOpenAIKey=your-openai-key

Run the Core Codemods

You can download and run the latest release from this repository in order to run the core codemods as a CLI:

$ unzip codemodder-java-codemods-<VERSION>.zip
$ sh core-codemods-<VERSION>/bin/core-codemods

In order to run them during active development from Gradle, you can use the run task:

./gradlew :core-codemods:run --args='--output /path/to/my.codetf --dry-run /path/to/project/'