Simple tool to get summarized news from RSS feeds.
First you need to create a log folder where the latest news will be stored, together with some metadata.
bash mkdir /var/data/newvelles
Then, create a virtualenv where you can install the newvelles
python command,
python setup.py install
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.
- Initial release of the command line version of newvelles with the UI tool.
- Upgraded output format for visualization.
- Daemon support to keep news updated every N minutes.
- Configuration support for debugging and daemon.
- Added link, date, and full title to each visualized news item.
- Upload data to S3 so it can be visualized by
newvelles_web
. - Only consider news published no more than 2 weeks ago.
- Added new modeling alternative using universal-sentence-encoder-lite to be used for AWS Lambda.
- Added Spacy NLP layer (using nouns and verbs) to generate a better summary header for a group of news.