About

This project is for redmine issues auto reply by using a local knowleage database.

How it works

graph TD
    A[Start] --> B[Fetch unanswered questions from Redmine]
    B --> C[Send questions to large language model]
    C --> D[Retrieve answers from local knowledge base]
    D --> E[Post answers to Redmine]
    E --> F[Add answered questions and answers to local knowledge base]
    F --> G[Wait for next interval]
    G --> B

Loading

Components Redmine Integration

1. Redmine Integration

   • Periodically fetch unanswered questions from Redmine.
   • Post answers back to Redmine.

2. Large Language Model (LLM) Integration

   • Send unanswered questions to the LLM.
   • Retrieve answers from the LLM based on the local knowledge base.

3. Local Knowledge Base Management

   • Query the local knowledge base for answers.
   • Update the local knowledge base with new answered questions and their answers.

4. Scheduler

   • Manage the periodic fetching and processing of unanswered questions.

Detailed Steps

1. Fetch Unanswered Questions from Redmine

   • Connect to the Redmine API.
   • Retrieve a list of questions that have not been answered.

2. Send Questions to Large Language Model

   • Format the questions appropriately.
   • Send the questions to the large language model (LLM).

3. Retrieve Answers from Local Knowledge Base

   • Query the local knowledge base using the LLM.
   • Fetch relevant information to answer the questions.

4. Post Answers to Redmine

   • Format the answers.
   • Post the answers back to Redmine via the Redmine API.

5. Add Answered Questions and Answers to Local Knowledge Base

   • Update the local knowledge base with the newly answered questions and their answers.

6. Wait for Next Interval

   • Wait for a predefined interval before repeating the process.

Technologies Used

• Redmine API: For fetching and posting questions.
• Langchain: For integrating with the large language model.
• Local Knowledge Base: For storing and retrieving knowledge.
• Python: For scripting and automation.
• Scheduler (e.g., Cron, APScheduler): For periodic execution.

Usage

To set up this project, follow these steps:

1. Clone the repository:

   git clone https://github.com/Xianleewu/redminegpt
   cd redminegpt

2. Install the required dependencies:

   python3 -m pip install -r requirements.txt

3. Configure the Redmine and LLM API credentials in the config.py file.

   redmine server
   
   REDMINE_SERVER = {
    "host": 'http://127.0.0.1:3000',
    "user": 'gptbot',
    "password": 'password',
   }
   
   # QAnything remote server
   
   QANYTHING_REMOTE_SERVER = {
    "app_key": '',
    "app_secret": '',
    "kb_id": '',
   }

4. Run the bot:

   python3 main.py

Depoly with docker

TODO

TODO

• Connect with Redmine via REST API

• Auto reply new issues

• QAnythingRemote support

• Auto gen local database acording to redmine closed issues

• QAnythingLocal support

• LangchainChatChat support

• Dify support

Contributing

If you would like to contribute to this project, please submit a pull request or open an issue on GitHub.