Manual Coding Helper GUI for Text Classification

This project is a simple Python GUI tool designed to assist with manual classification of segmented text data stored in Excel or CSV files. The motivation behind this project was to provide a user-friendly alternative for creating training data for machine learning tasks, avoiding the complexities of Doccano (https://github.com/doccano/doccano) or tedious manual spreadsheet work.

Features

• User-Friendly Interface: A straightforward GUI designed to streamline the process of annotating and classifying text data.
• Customizable: Easily adaptable to different datasets by configuring JSON files with specific codes and descriptions. Select your columns to display.
• Navigation Controls: Navigate through rows of data, update annotations, and manage classification status with buttons and keyboard shortcuts.
• Font Size Adjustment: Ability to customize the font size for better readability.
• Dark Mode Support: Toggle between light and dark modes for optimal viewing experience.

Known Bugs

• Correct toggle is bugged and does not save

Planned Features

• Setting persistence between sessions
• Merging the json files into one
• Submenu to edit categories, category values and category descriptions within the GUI
• Multi-Label classification support

Not Planned Features

• Span annotations. This is only for table-based manual classification

Getting Started

Prerequisites

• Python 3.x installed on your system.
• Required Python packages can be installed using pip: pip install pandas ttkbootstrap

Installation and Usage

Clone the repository:

git clone https://github.com/your-username/manual-coding-helper.git

cd manual-coding-helper

Run the application:

python manual_coding_gui.py

Select your dataset file (in Excel or CSV format) when prompted. Use the GUI to navigate through your data and annotate classifications. Your process is automatically saved. For best performance, use csv.

Configuration

JSON Files

Two JSON files (descriptive_sentences.json and category_mapping.json) are used for configuring codes, descriptions, and category mappings. Update these files according to your dataset's classification needs.

Display Customization

Modify the display_row function in the code to specify the column you want to display within the GUI interface. (Replace 'sentence_text')

Contributing

Contributions to enhance this tool are welcome! Feel free to open issues, submit pull requests, or suggest improvements.

License

tbd