This bot has been developed as an automatic processor of "call for papers" (CFP) emails received in the academic circles.
It was developed with the necessities of the Adaptive Technologies Lab in mind by one of its (former) students (me), not as a final product for widespread use. However, anyone is free to try it at his/hers own risk and, even better, collaborate for improvements on the existing code.
So far, it is still licensed under MIT License. Maybe, someday, it will be even more open.
We make use of Datejs library for parsing of the several human ways of writing dates. The library is licensed under the MIT License.
This repository is originally "the house" of the old version written in Python. If you are used with GitHub and git, feel free to navigate and learn from it. Nonetheless...
The current (new) version of this bot is a port of the old code from Python to
Javascript—or, more correctly,
Google Apps Script.
Why? Very simple: because the bot is totally based in Google products and so we
could use the Google Apps Script to trigger our bot automatically, moving to
Google the responsibility of maintaining our cron jobs. Simple like this.
I know it is not a perfect bot, but it does an amazing job in a very simple way.
Could I have used machine learning? Yes.
Could I have used neural network? Yes.
Could I have used Bayesian networks? Yes.
Could I have used mechanical turks? Yes, I could.
But, believe me, it was a night (or two) project to solve a problem we had and was making us crazy: full inbox of CFP emails.
I am very interested in using more intelligent approaches to solve this problem, but the probability it will happen tends to zero.
Anyhow, your collaboration, fork or whatever is welcome! 😄