/nlp-sentiment-analysis-api

This API uses spaCy's pre-trained models for Named Entity Recognition (NER) and Sentiment Analysis. NER is powered by the spaCy en_core_web_trf transformer model, while sentiment analysis uses a custom-trained transformer model. Built with Flask, the API offers simple endpoints for processing JSON requests and responses.

Primary LanguagePythonMIT LicenseMIT

NLP API (Sentiment Analysis & NER)

Welcome to the NLP API! This simple API is built with Flask and spaCy, and it helps you analyze the sentiment of a text and identify named entities (NER). It’s perfect for understanding the vibe of a text or extracting key details like names, places, and more.

Setup

  1. Clone the repository:

    git clone https://github.com/asma-sassi/nlp-sentiment-analysis-api.git
cd nlp-sentiment-analysis-api

  2. Create and activate the virtual environment:

    For MacOS/Linux:

    python3 -m venv venv
source venv/bin/activate

    For Windows:

    python -m venv venv
venv\Scripts\activate

  3. Install the dependencies:

    pip install -r requirements.txt

  4. Run the Flask API:

    python app.py

Endpoints

POST /analyze_sentiment

Request body:

Send a JSON object with the text you want to analyze.

{
  "text": "I am feeling good today!"
}

Response:

The API will return the sentiment (positive/negative/neutral) and entities (like people, places, or dates) found in the text.

{
  "sentiment": "positive",
  "entities": [["good", "ADJ"]],
  "text": "I am feeling good today!"
}

Technologies Used

  • Flask: A simple framework for building APIs.
  • spaCy: A library for natural language processing (NLP).
  • Transformers (Hugging Face): Used for advanced NLP tasks like sentiment analysis and named entity recognition (NER).