This is a scientific vibe estimator based on ADS abstracts!

Project for Code/Astro 2022 by Mireya Arora, Steph Merritt, and Luna Zagorac.

This code requires that you provide your own ADS API key. This is easy to generate: go to your account settings and choose "API Token" on the right-hand menu, then click the "Generate a new key" button.

All the dependencies are specified in requirements.txt, so cloning the git directory onto your computer and running

pip install scivibes -r requirements.txt

within the directory should take care of all dependencies.

Please see demo Jupyter notebook to test your vibes!

You can also run the software directly from the terminal. You need either the author's name or ORCID ID (the example includes both, please only use one!).

python scivibes_.py --ads_config_token [Your ADS Configuration Token] --ORCID [Your ORCID id here] --author_name [Your author name here] --filename [optional name for file]

The vibe files are saved in the format "filename_Vibe.png" and "filename_AntiVibe.png". The vibestogram of total vibes is saved as "vibestogram.jpg".

The files in the subreddits/ folder were created by the authors of the SocialSent project. The process between producing these domain-specific lexicons is described in their paper, Inducing Domain-Specific Sentiment Lexicons from Unlabeled Corpora.

The code used to create the Vibe and Anti_Vibe images is based on the WatercolorClouds code by Eric Davidson.