Sentiment Analysis - Web service

A simple web service for sentiment analysis built with FastAPI. It currently uses a basic keyword-based approach to determine sentiment scores.

Local Development

Prerequisites

• uv installed.

Run locally

1. Install dependencies:

uv sync

2. Run the main sentiment analysis service (backend):

uv run uvicorn main:app --reload --port 8000

The server will be available at http://localhost:8000. You can access the Swagger UI at http://localhost:8000/docs.

3. Run the demo frontend app:

uv run uvicorn app:app --reload --port 8001

The frontend UI will be available at http://localhost:8001.

Sentiment Analysis Versions

This project implements different versions of sentiment analysis:

• v1: Basic keyword-based approach (hardcoded).
• v2: Uses the AFINN lexicon for sentiment scoring.
• v3 (spaCy): An advanced NLP approach using spaCy was implemented but is currently disabled to keep the Docker image size small.

Enabling spaCy (v3)

If you want to use the spaCy implementation locally (v3 endpoint), you need to reinstall the dependencies:

uv add spacy spacytextblob https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.8.0/en_core_web_sm-3.8.0-py3-none-any.whl

Then uncomment the relevant lines in main.py.

Docker

Build the image

docker build -t sentiment-analysis .

Run the container

docker run -p 8000:8000 sentiment-analysis

TASKs

• git repo
• uv
• sentiment analysis
  • hardcoded s.a.
  • Training a machine learning classifier would also be possible.
  • The Python library spaCy would also be possible.
  • LLM API.
• fastapi
• swagger
• docker
• test