mamiksik/parrot-intellij

AI Commit Messages Autocomplete for InteliJ

Parrot - AI Autocomplete for Commit Messages

➡️ Get it on Jetbrains Marketplace

Welcome to Parrot, the AI autocomplete plugin for commit messages! With Parrot, you can save time and improve the quality of your commit messages by using machine learning to generate suggestions for what to include in your message.

⚠️ Don't forget to run the inference server

For the plugin to work you have host completion server somewhere. To do that, you can install the Parrot Inference Server and run it in background.

🔌 Plugins for IDEs

• Parrot for IntellJ
• Parrot for VSCode

🦾 Trained Models

• Parrot commit generation Model
• Parrot autocomplete model

📚 Others

• Parrot Dataset
• Parrot Inference Server

Usage

To use Parrot, simply start typing your commit message as usual. When you want to get a suggestion for the next word, use the autocomplete shortcut (typically Ctrl + Space or Cmd + Space). Parrot will then generate a list of suggestions based on the words you have already typed. You can select the suggestion you want to use and it will be automatically inserted into your message.

parrot-intellij's Issues

Error while sending request to Parrot API

I've configured and run the parrot-server on Windows and installed the parrot-intellij plugin in PHP Storm.
I've updated the Parrot Configurations in PHP Storm settings (see screenshot below). But it gives me the error Error while sending request to Parrot API.

.

error: No mask_token ( ) found on the input

When I try to use the autocomplete with the Hugging Face Hosted inference API, I get:
[400] {"error":"No mask_token ( ) found on the input","warnings":["There was an inference error: No mask_token ( ) found on the input"]}

I have the Inference api URL set to: https://api-inference.huggingface.co/models/mamiksik/CodeBERTa-commit-message-autocomplete
and I have "Fill Token api token" set to my Hugging Face access token.