Sentiment Analyser Application

A Flask Application that performs the sentiment prediction of the given review.

The application can be used via API as well as Web Application

Files and Folder Information

• app.py is the main application file

• sentiment_analyzer.py is the python script that runs the sentiment analysis model.

• sentiment-analyzer-training-notebook.ipynb is the training notebook

• model folder contains the trained model and tokenizer

  Model's Accuracy = 88.807 %
  
  Model's Loss = 0.274
  

• templates folder contains the html files

How to set up the application ?

• Create virtual environment

  python3 -m venv env
  

• Activate the virtual environment

  source env/bin/activate
  

• Install the required libraries

  pip install -r requirements.txt
  

How to run the application on localhost ?

• Run the flask application on localhost

  python app.py
  
  # Go to http://localhost:5000/  and perform analysis
  

How to run the application on localhost ?

• Install the Web Server Gateway Interface (WSGI) Gunicorn

   pip install gunicorn
  

• Create access.log and error.log file

• Run the flask application using Gunicorn server

  gunicorn -b 0.0.0.0:5000 app:app --access-logfile access.log --error-logfile error.log --max-requests 10 --timeout 120 --daemon
  

---

RUN IT AS AN API SERVICE

• Open API Running tool like Postman.

• Run the following API

  URL = localhost:5000/api/detect 
  
  Method = POST
  
  Body ( form-data )
  
       review = "<review text>"
  

• See the below screenshot

---

RUN IT AS A WEB APPLICATION

Application Screenshots

INPUT SCREEN

RESULT SCREENS