Newvelles

Simple tool to get summarized news from RSS feeds.

Requirements

• Python 3.12 or higher
• Docker (for containerized deployment)

Installation

Local Development

First you need to create a log folder where the latest news will be stored, together with some metadata.

mkdir /var/data/newvelles

Then, create a virtual environment where you can install the newvelles python command:

# Create virtual environment
python3.12 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install the package
pip install -e .

Docker Deployment

Build the Docker image:

docker build -t newvelles:latest .

Then, run the script to pull the news and generate the visualization data.

newvelles --rss_file data/rss_source_short.txt 

If you want to keep the news updated every N minutes (configured in the ./config/newvelles.ini) file, you can run

newvelles --rss_file data/rss_source_short.txt --daemon

Finally, start a server in the root path of this project so the web page can be accessed within your localhost.

python -m http.server 

And then, open the web version of newvelles in your browser,

http://localhost:8000/

Note that you need to run the newvelles command before launching the webapp so the latest_news.json file is created, which will be needed to visualize the output.

Changelog

v0.0.1 (2021-05-09)

• Initial release of the command line version of newvelles with the UI tool.

v0.0.2 (2021-05-16)

• Upgraded output format for visualization.
• Daemon support to keep news updated every N minutes.
• Configuration support for debugging and daemon.

v0.0.3 (2021-05-21)

• Added link, date, and full title to each visualized news item.

v0.0.5 (2021-06-06)

• Upload data to S3 so it can be visualized by newvelles_web.
• Only consider news published no more than 2 weeks ago.

v0.0.6 (2021-07-19)

• Added new modeling alternative using universal-sentence-encoder-lite to be used for AWS Lambda.

v0.0.7 (2021-09-12)

• Added Spacy NLP layer (using nouns and verbs) to generate a better summary header for a group of news.

v1.1.0 (2025-01-16)

• BREAKING CHANGE: Upgraded from Python 3.8 to Python 3.12
• Updated all dependencies to latest compatible versions
• Improved Docker build process using Amazon Linux 2023
• Removed TensorFlow dependencies for lighter container image
• Updated Lambda runtime to Python 3.12