Sentiment-Analysis

Sentiment analysis of Amazon phone review :The analysis was done through distilbert-base-uncased model from hugging face. The program was then integrated with fastAPI and Flask and was developed into a web application were a user can first sign up and then sign in, then upload the csv file which contains the review. The sentiment analysis was done as a background task and the result was stored in PostgreSQL database in json format. The result was then displayed as a table by interacting with the database using SQLAlchemy. JWT tokens were used for authentication and authorization of users. The API was designed to handle HTTP exceptions. Pydantic models were used to validate requests and responses and python logging to log errors and information. Typing was implemented for all functions. Pytest was used to test the functions used in the application. The application was finally deployed in docker. A docker compose was used, with separate docker containers for PostgreSQL and FastAPI.