Braille Autocorrect and Suggestion System

A sophisticated autocorrect and suggestion system for Braille input using the QWERTY keyboard format. This project aims to improve the typing experience for visually impaired individuals by providing real-time suggestions, auto-correction, and visual feedback.

Features

• Efficient Word Lookup: Uses a Trie data structure for fast dictionary searches
• Similarity Matching: Implements Levenshtein distance algorithm optimized for Braille cells
• Learning Mechanism: Improves suggestions based on user's previous inputs
• Real-time Suggestions: Optimized for speed to support large dictionaries
• Auto-replace Options: Shows top 3 suggestions for quick selection

Installation

1. Clone the repository

git clone https://github.com/akash-kumar-dev/braille-autocorrect.git
cd braille-autocorrect

2. Install dependencies

npm install

3. Run the development server

npm run start

Technical Approach

Data Structures

1. Trie: For efficient dictionary storage and prefix matching
2. Levenshtein Distance: For finding similar words, optimized for Braille cells

Algorithms

1. Word Suggestion:

   • First attempts exact matching using the Trie
   • If no exact match, finds similar words using Levenshtein distance
   • Ranks suggestions by similarity and user frequency
   • Returns the top N suggestions

2. Learning Mechanism:

   • Tracks word usage frequency
   • Prioritizes frequently used words in ambiguous cases
   • Adapts to user's typing patterns over time

Usage

QWERTY Braille Typing Format

The system uses the following key mapping for Braille dots:

• D → Dot 1 (top left)
• W → Dot 2 (middle left)
• Q → Dot 3 (bottom left)
• K → Dot 4 (top right)
• O → Dot 5 (middle right)
• P → Dot 6 (bottom right)

1. Type using the D, W, Q, K, O, P keys to input Braille dots
2. Press Space to complete a cell
3. See the Braille visualization and English translation update in real-time
4. Press Tab to cycle through auto-replace options
5. Press Enter to select the highlighted suggestion

How It Works

1. Input Parsing:

   • QWERTY keys (D, W, Q, K, O, P) are mapped to Braille dots (1-6)
   • Multiple keys pressed simultaneously form a single Braille cell
   • Cells are separated by spaces

2. Suggestion Generation:

   • First attempts exact matching using the Trie
   • If no exact match, finds similar words using Levenshtein distance
   • Ranks suggestions by similarity and user frequency
   • Returns the top N suggestions

3. Learning Mechanism:

   • Tracks word usage frequency
   • Prioritizes frequently used words in ambiguous cases
   • Adapts to user's typing patterns over time