Label bot dataset

[xtuchyna]

Is your feature request related to a problem? Please describe.
Having a label bot inside our organization (see thoth-station/kebechet#586) requires training it on Issue / PR dataset consisting of stored labels.

Describe the solution you'd like

1. Original label bot https://github.com/machine-learning-apps/Issue-Label-Bot
2. Giorgos version https://github.com/GiorgosKarantonis/Github-Issues-Classifier

Acceptance criteria

• Look at the datasets used in 1. and 2.

• Consider what can be improved in means of simplicity or better accuracy

• Implement Store labels with timestamps #173 or consider it invalid if it does not imporve the quality of dataset

• Study used models above. Use the one from Giorgos if sufficient

• Integrate the bot as a new Kebechet Manager

Additional context
Add any other context or screenshots about the feature request here.

[sesheta]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

/lifecycle stale

[xtuchyna]

I don't think this should be part of mi, but rather a glyph submodule.
cc @goern @pacospace ?

[sesheta]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

/lifecycle rotten

[pacospace]

@goern someone from aiops team is also looking at this right?

[goern]

/remove-lifecycle rotten
/triage needs-information

@durandom could you point us at the right person?

[goern]

It might be good to move this issue to https://github.com/aicoe-aiops/github-labeler

[goern]

@xtuchyna ping

[xtuchyna]

@goern I think that this may be closed. Issue was referenced in new issue under github-labeler repo, so there's no need to keep it alive here. Closing