Some people like to kickstart their day as a researcher by reading the papers appearing on the Arxiv. Some wish they could listen to these abstracts while commuting or on their way to work. This projects aims at doing exactly this.
You can listen to the arxiv reader by following one of the podcast flux found here. Alternatively, you can find it on Google Podcasts, Spotify or your favourite podcast app.
This script is actually simple and simply brings together different already-existing tools. It fetches the paper (title, authors, abstract) from the ArXiV using the public API. A text-only description of the paper is then generated (title, author, abstract). The final step is to feed the text-only description to a text-to-speech synthesis (TTS) app. For now, the only supported TTS is Google's but it would be simple to use another one instead.
The script requires a working installation of Python 3 and or the Google cloud sdk.
To install the script, download it (using e.g. git clone) and in the downloaded folder, install the package with
pip install .That's it! Note that before using the package, you need to have set up a Google TTS account. Before running the script, do not forget to export your Google application credentials.
export GOOGLE_APPLICATION_CREDENTIALS=/path/to/your/credentials.jsonTo run the script, now simply do
arxiv-reader -d "two days ago"
arxiv-reader -d "01 June 2021"