PapersBot

Geomstats PapersBot is an academic Twitter bot: it reads RSS feeds from journals and preprint archives, selects papers based on keyword matching, and tweets them.

Geomstats PapersBot was inspired by (and borrows some code from) feedr and @MOF_papers from @fxcoudert.

---

Requirements

Python 3 and the following modules: beautifulsoup4, feedparser, tweepy.

If you have Python 3 installed, you can install those modules with pip install bs4 feedparser tweepy.

Setup

In order to run PapersBot, you need to do the following:

• Create a file credentials.yml which will contain your Twitter app credentials, with four lines:

CONSUMER_KEY: "x1F3s..."
CONSUMER_SECRET: "3VNg..."
ACCESS_KEY: "7109..."
ACCESS_SECRET: "AdnA..."

If you do not know how to get your Twitter credentials, follow steps #1 and #2 in this tutorial to register your app with Twitter and get credentials.

• Adjust the file feeds.txt which contains the list of RSS feeds you want to crawl. Lines starting with # are ignored.
• Inside the code, adjust the regular expression that selects the papers of interest
• Some extra parameters can be tweaked in configuration file config.yml.

How to run

PapersBot tracks in a file named posted.dat (which it will create) the papers that have already been tweeted. The first time you run it, if there is no posted.dat from a prior run, PapersBot can thus post a lot of papers. If you want to avoid this, especially on the first run or if it hasn't been run for a long time, use the --do-not-tweet option.

papersbot.py --do-not-tweet will list the papers it would tweet, without actually tweeting. But papers will still be recorded as tweeted in the posted.dat file.

Other features

• Running papersbot.py --top-tweets will give you a list of the 5 top tweets, from the bot's 200 latest tweets. It sorts tweets by adding number of retweets and likes.