Create Sloth bot role model

[ftdgoodluck]

Create a role model for the bot.

Tasks

Preview Give feedback

1. bot message - basic include #92
   bot bot-message +2
   
2. bot message - scoring devbot#97
   bot bot-message +2
   
3. bot message - pause/unpause devbot#98
   bot bot-message +2
   
4. bot message - error handling devbot#99
   bot bot-message +2
   
5. bot message - merge & finalization #100
   bot bot-message +2
   
6. bot message - 3rd party repo #111
   bot bot-message +2
   

[ftdgoodluck]

List of supported commands:

For contributors:

• @sloth include - includes particular PR in Sloth race

For maintainers:

• @sloth score [1-10] - rates the specific PR that participates in the Sloth race
• @sloth exclude - excludes particular PR from the Sloth race
• @sloth pause - pauses bot for a particular repository

Workflow:

1. Contributor creates a PR with a mention of the Sloth bot using “@sloth include” command
2. The Sloth bot responds with a like emoji over the description and writes a comment“This pull request was submitted for the Sloth race [active link]”.
3. Contributor sends PR for a review
4. Sloth bot writes a comment "This pull request is a part of Sloth race[active link]. Please use the "@sloth reject" command to exclude this pull request from the Sloth race, "@sloth score [1-10]" to rate it, or "@sloth pause" to pause all Sloth race notifications across the project”
5. If the maintainer writes a comment with “@sloth pause” command, the PR is automatically excluded from Sloth race and all bot actions inside the repo are paused
6. If the maintainer writes a comment with “@sloth exclude” command, the PR is excluded from the Sloth race
7. If the maintainer writes a comment with @sloth score X” command, the bot responds with a like emoji over the message and writes a comment “Thanks for submitting your score for the Sloth race [active link]. This score would be recorded once the pull request is merged
8. If the maintainer writes a comment with “@sloth score” command but with an incorrect token that we can’t parse, the bot writes a comment “Please use the score command with the following syntax: @sloth score + a number between 1 and 10 after a space”

Some thoughts:

• any type of scoring with multiple parameters would result in a longer explanation message, more mistakes from reviewers and thus more useless messages from the bot asking to write a correct response. Are we OK with that, or it would make more sense to have only one parameter? @frol @botezatunear
• this flow does not touch some edge cases (multiple reviewers, unpausing bot, etc.)
• Once we agree on the flow I'll create a diagram for visual explanation

[botezatunear]

I agree that one score will be easier for everyone, I've written thoughts in the specific task:

Let's discuss tomorrow and finalise

[ftdgoodluck]

blocked by #46 - once the PRD is finalized this task would be updated

[ftdgoodluck]

@frol @botezatunear @akorchyn waiting for your comments, gentlemen