This is a set of shell and Python scripts that provides AI-driven formative feedback on Git repositories containing Java projects. The scripts combine LLM prompts with data extracted from the Git repository to provide feedback on the following areas:
- Repository contents
- Project architecture
- Source code style
- Project build specification
- Configuration management
An example of the generated report can be seen here.
The feedback.sh script can be executed to output Markdown or HTML regarding
a specified local Git repository.
It can also be provided with a list of repository URLs and email addresses.
In this case it will assess each repository and send email to the
specified email address.
To run the script you must set the environment variable OPENAI_API_KEY
to an OpenAI API key that provides access to the GPT models.
Furthermore, to send emails you must set the environment variables
SMTP_SERVER, SMTP_USERNAME, and SMTP_PASSWORD to the values
of an SMTP server and credentials that allow you to send out email.
export SMTP_SERVER=smtp.example.com
export SMTP_USERNAME=myname
export SMTP_PASSWORD=mypassword
export OPENAI_API_KEY=sk-.......
# Create a Markdown report from a local repo
feedback.sh -d repo-dir >comments.md
# Send out reports
feedback.sh -l repo-list -f mary@example.com -F 'Mary Zhu' -s FeedbackContributions to this project via GitHub pull requests are welcomed, provided they are general enough to be used in diverse contexts. For example these could improve the output format, change prompting to provide better feedback, add new feedback types, add support for another programming language or framework, or work with a different AI engine. Contributions should try to be self-contained with minimal external dependencies in order to simplify the project's long-term maintenance.