Course automation: Automatic duplicate proposal detection

[johennin]

Automatic duplicate proposal detection

Members

Lee Badal (badal@kth.se)

GitHub: @LeeBadal

Johan Henning (johennin@kth.se)

GitHub @johennin

Proposal

We plan to create a github action that validates that no previous/similar proposals have been made before.

To verify that the PR is a proposal we intend to test the folder structure or validate that a proposal label is assigned (if #949 is implemented)
We will then run a semantic analysis comparator to previous PR(Title and description) proposals and notify the user/TA's by writing a comment if they are above a certain threshold of similarity or if manual revision by a TA is required. (In this case the URLs to the similar PR will be provided)

Proposed tools

• Github Actions
• NLTK

[cesarsotovalero]

@LeeBadal: is this your KTH account? If not, please change.

[LeeBadal]

@cesarsotovalero if you are wondering if this account is linked to my kth mail address, then yes.

[cesarsotovalero]

@cesarsotovalero if you are wondering if this account is linked to my kth mail address, then yes.

No, I mean that you should use badal instead.

[LeeBadal]

@cesarsotovalero if you are wondering if this account is linked to my kth mail address, then yes.

No, I mean that you should use badal instead.

Sorry I'm not quite following, that is not possible because that username is already taken. Please clarify why/where this account is an issue and perhaps I will understand what you want me to change.

[cesarsotovalero]

Group Rules

• When you send a pull request for registration, please follow the name convention of using email addresses of two members to create the folder: email-email.

Your email is badal@kth.se, so the name of the folder in your pull request should be contributions/course-automation/badal-johennin/README.md

Is this clear to you now?

[johennin]

Hi, changes made, sorry about that!
@cesarsotovalero

[cesarsotovalero]

LGTM now, I'll merge.