Integration of bot data

[bockthom]

As we aim at integrating bot data to be able to filter bots, @hechtlC and I discussed on how we can achieve this. We agreed on the following adaption to our toolchain:

From an external tool, we get a prediction on which GitHub user should be treated as bot in a certain project. The output of this tool contains a row for each user, where the first column is the GitHub username and the last column is the predicted user type (Bot, Human, Unknown).

As we just have a user name, but neither name nor e-mail address, we cannot just pass the username to the id service. However, as all the users should already appear in the GitHub issue data, we just need to get a mapping from username to name and e-mail address from the GitHub issue processing. So, the bot processing will look like as follows:

(0) Within GitHub issue processing: Dump final user table as a csv file, to be able to map username to name and e-mail.

Bot processing:
(1) Read bot data (created from external tool).
(2) Read user table (dumped from GitHub issue processing).
(3) Get name and e-mail from user table for each username appearing in the bot data.
(4) Dump final bots.list file, consisting of three columns: "name", "e-mail", "user type"

(5) Within author post-processing: Update name and e-mail of bot data also during post-processing.

This implementation is will be the baseline for se-sic/coronet#202.

---

As we dump the user table after the GitHub issue processing, this might also be useful when dealing with issue #24 (at some point in some)