tkfmRecruitment

This repository contains a Telegram bot for automatically querying recruitment tag combinations from the TenkafuMA! toolbox for the TenkafuMA! mobile game.

Table of Contents

• Demo
• Installation
  • Python Environment
  • Docker Container
• Usage
• Configuration
  • User-defined OCR Word Replacement
• Additional Information
  • Enable GPU
  • Cannot Read Text from the Screenshot
  • Multi-language Support
• Credit
• License

Demo

Installation

Python Environment

1. Clone the Repository

   git clone https://github.com/gordon0907/tkfmRecruitment.git
   cd tkfmRecruitment

2. Install Required Modules

   pip install -r requirements.txt

3. Define Telegram Bot Token

   If you don't already have a Telegram bot token, you can obtain one from BotFather.

   • Unix

     # Replace `your-telegram-bot-token` with your actual Telegram bot token
     export TELEGRAM_BOT_TOKEN="your-telegram-bot-token"

   • Windows (PowerShell)

     # Replace `your-telegram-bot-token` with your actual Telegram bot token
     $env:TELEGRAM_BOT_TOKEN = "your-telegram-bot-token"

4. Run the Script

   python main.py

   Note: You may need to install additional dependencies as per the module's instructions.

Docker Container

1. Clone the Repository

   git clone https://github.com/gordon0907/tkfmRecruitment.git
   cd tkfmRecruitment

2. Build the Docker Image

   sudo docker build --tag tkfm-recruitment .

3. Run the Docker Image in a Container

   If you don't already have a Telegram bot token, you can obtain one from BotFather.

   # Replace `your-telegram-bot-token` with your actual Telegram bot token
   # Replace `/path/to/local/data` with the path to your local data directory
   sudo docker run --name tkfm-recruitment-container --env TELEGRAM_BOT_TOKEN="your-telegram-bot-token" --volume /path/to/local/data:/app/data tkfm-recruitment

4. Start the Existing Container

   After the initial run, you can start the existing container without recreating it:

   sudo docker start --attach tkfm-recruitment-container

5. Save the Docker Image

   If you need to save the Docker image for future use:

   sudo docker save --output tkfm-recruitment.tar tkfm-recruitment

Usage

1. Start the Bot: Use the /start command in the Telegram chat to receive a welcome message.
2. Send a Screenshot: Send a screenshot of the Recruitment page in TKFM to the bot via Telegram as a photo (not as a file). The bot will reply with the query result in PDF format in the chat.

Configuration

User-defined OCR Word Replacement

To handle OCR misreading similar words, you can add word mappings for replacement in ./data/word_mappings.yaml. This file is generated after the first run and supports hot-editing, meaning you can edit word_mappings.yaml anytime without restarting the bot.

Additional Information

Enable GPU

By default, the OCR process does not utilize the GPU. To enable CUDA or MPS, you must install torch and torchvision before installing easyocr. For the installation command, please refer to the official PyTorch page.

For Docker users, you may need to modify the Dockerfile accordingly.

Refer to EasyOCR for more details on enabling GPU for the OCR process.

Cannot Read Text from the Screenshot

If the OCR fails to read text from the image, it is likely due to low image quality. In this case, try cropping the tags section and sending it to the bot. Ensure the word 招募條件 is included in the image, as it is used as a reference point during the tags filtering process.

For iOS users, do not use the Drag-and-Drop screenshot function, as it significantly reduces image quality.

Multi-language Support

Currently, there are no plans to support other languages for the game interface. However, you can manually change the language settings by editing image_ocr.py and tkfmtools.py. You can also modify main.py to support other languages for the Telegram bot.

Credit

TenkafuMA! toolbox

License

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

© 2024 gordon0907