Word Predictor

🌐 Available Languages: Türkçe 🇹🇷 | English 🇬🇧

Welcome to the Word Predictor project...

Text-Based Word Predictor

This project is a simple language modeling application that learns word bigrams—and optionally trigrams—from a given text source. It predicts the most likely next word following a user-provided word or sentence. It serves as a starting point for basic language processing tasks such as autocomplete or text generation.

---

Features

• Text Cleaning: Input text is converted to lowercase and stripped of punctuation, which improves prediction accuracy.
• Bigram Generation: Builds a basic language model by counting the frequency of word pairs.
• Trigram Support (Optional): Supports trigram usage to make more contextual predictions by considering the two previous words. This feature can be enabled or disabled via config.json.
• Chained Prediction: After a user-provided word or sentence, it can predict a chain of words based on the configured predictionChainLength.
• Configuration Management: Application settings (input file path, prediction chain length, trigram usage) can be easily managed through the config.json file.
• Modular Structure: The code is divided into separate packages (config, textprocessor, languagemodel, utils) with distinct responsibilities, making it easier to read and maintain.
• Robust Error Handling: Unexpected conditions such as file read errors or configuration issues are reported to the user with clear messages.

---

Installation & Usage

Prerequisites

• Go must be installed (version 1.16 or later is recommended).

Steps

1. Clone or Download the Project: Download or copy the project files to your local machine.

2. Ensure Folder Structure: Make sure your project directory follows this structure:

   project_folder/
   ├── main.go
   ├── internal/
   │   ├── config/
   │   │   └── config.go
   │   ├── textprocessor/
   │   │   └── text_processing.go
   │   ├── languagemodel/
   │   │   └── language_model.go
   │   └── utils/
   │       └── utils.go
   ├── config.json
   └── input.json
   

3. Initialize the Go Module: Navigate to the root directory of your project (project_folder) in the terminal and initialize the Go module:

   go mod init bigram # 'bigram' is your module name.
                      # You can use a different name if preferred.

   This command will create a go.mod file in your project root.

4. Download (or Verify) Required Dependencies: While still in the project root (where go.mod is located), run:

   go mod tidy

   This command tidies up your go.mod file and ensures that all necessary packages are downloaded and updated.

5. Prepare the input.json File: The input.json file should contain the text to be used for model training. A sample input.json is included with the project. Its structure should look like this:

   {
     "text": "This is the entire text content that will be placed here. The model will learn from this text."
   }

6. Configure the config.json File: The config.json file controls the behavior of the application:

   {
     "inputFilePath": "input.json",        // Path to the input text file
     "predictionChainLength": 3,           // Number of words to predict
     "useTrigrams": true                   // Enable trigrams? (true/false)
   }

   You can adjust these values based on your needs.

7. Run the Application: From the project root (where the bigram module is defined), run the application:

   go run .

   The application will prompt you to enter a word or sentence:

   Enter a word or sentence:
   

   Enter your input and press Enter. The application will then display the predicted word sequence.

---

Usage Examples