https://github.com/olahsymbo/news-categorizer

A Flask REST API for categorising news contents from Postgres DB fetched via SQLAlchemy ORM
https://github.com/olahsymbo/news-categorizer

beautifulsoup4 flask keras nltk postgres python3 sqlalchemy tensorflow

Last synced: 3 months ago
JSON representation

A Flask REST API for categorising news contents from Postgres DB fetched via SQLAlchemy ORM

Host: GitHub
URL: https://github.com/olahsymbo/news-categorizer
Owner: olahsymbo
Created: 2020-03-30T19:33:49.000Z (about 6 years ago)
Default Branch: master
Last Pushed: 2023-05-03T14:47:49.000Z (about 3 years ago)
Last Synced: 2025-03-26T16:17:20.450Z (over 1 year ago)
Topics: beautifulsoup4, flask, keras, nltk, postgres, python3, sqlalchemy, tensorflow
Language: Python
Homepage:
Size: 39.8 MB
Stars: 1
Watchers: 2
Forks: 0
Open Issues: 6
Metadata Files:
- Readme: README.md

README

## News Categorizer

This project is concerned with classifying news content into 5 different categories.

### Preprocessing

The raw texts of each news file is preprocessed by removing punctuations, performing
stemming, and lemmatization.

### Feature Extraction

TFIDF is used for extracting features from the proprocessed news data. We set the max length
of the features to 3000.

### Classification

Deep Neural Network (DNN) is used for performing classification.

### API

To convert the model into a REST web application, we used Flask and gunicorn as the web application server

## Getting Started

The following steps will guide you through setting up the project

First, clone the repo

git clone https://github.com/olahsymbo/NewsCategorizer.git

### Install and configure virtual environment

Goto project directory and setup virtual environment:

```
cd NewsCategorizer
pip3 install virtualenv
virtualenv news
source news/bin/activate
```

To ensure this app functions properly, install the dependencies in the requirements.txt Simply run:

`pip install -r requirements.txt`

### Create a Postgres DB

Install postgresql 12.2 (incase it's not installed already).

Create a new postgres database named `news` using:

`CREATEDB news;`

inside the db shell, create username and password for the database using:

`CREATE USER newsdev with encrypted password 'newsdev';`

### Run the NewsCategorizer

Launch the flask web server using:

` python3 api/news_api.py`

The base url is:

`http://127.0.0.1:6000/`