Text Emotion Prediction API

This project is a machine learning application that predicts emotions from text input. It consists of a trained model that can classify text into different emotion categories and a FastAPI-based REST API for serving predictions.

Project Overview

The system can identify 8 different emotions from text:

• Joy
• Sadness
• Fear
• Anger
• Surprise
• Neutral
• Disgust
• Shame

Project Structure

├── api/
│   └── main.py          # FastAPI application
├── data/
│   └── emotion_dataset_raw.csv  # Dataset used for training
├── model/
│   └── Text_emotions_model.pkl  # Trained model file
├── notebook/
│   └── Model_training.ipynb     # Jupyter notebook with model training process
├── python-version               # Python version specification
├── readme.md                    # This file
└── requirements.txt             # Project dependencies

Dataset

The model was trained on a dataset containing text samples labeled with emotions. The dataset includes 34,792 entries across 8 emotion categories with the following distribution:

• Joy: 11,045 samples
• Sadness: 6,722 samples
• Fear: 5,410 samples
• Anger: 4,297 samples
• Surprise: 4,062 samples
• Neutral: 2,254 samples
• Disgust: 856 samples
• Shame: 146 samples

Model Training

The model training process is documented in the Jupyter notebook notebook/Model_training.ipynb. The notebook includes:

• Data loading and exploration
• Data preprocessing
• Model training
• Model evaluation
• Model serialization

API

The API is built using FastAPI and provides endpoints for emotion prediction.

Endpoints

• GET /: Health check endpoint
• POST /predict: Endpoint for emotion prediction

Request Format for Prediction

{
  "text": "Your text here"
}

Response Format

{
  "Emotion": "The predicted emotion"
}

Live API

The API is deployed on Render and can be accessed at:

https://textemotions-8p94.onrender.com/

You can use this URL in place of localhost in the usage examples below.

Setup and Installation

1. Clone the repository
2. Install dependencies:

   pip install -r requirements.txt

3. Run the API:

   cd api
   python main.py

   The API will be available at http://localhost:8000

Dependencies

• seaborn
• neattext
• joblib
• pandas
• scikit-learn
• numpy
• fastapi
• uvicorn
• pydantic

Usage Examples

Using curl

CURL -X POST "http://localhost:8000/predict" -H "Content-Type: application/json" -d '{"text":"I am so happy today!"}'

Using Python requests

import requests

response = requests.post(
    "http://localhost:8000/predict",
    json={"text": "I am so happy today!"}
)

print(response.json())

Social Handles

• GitHub
• LinkedIn
• Twitter
• Email