Add Support for Storing Repo Labels #1683
Conversation
Adds two new tables: `repo_labels` and `labels` to store label data Created a new service class to store and associate labels for repos if the same `label` exists on multiple `repos`, it does not make duplicate `label` records. Add simple rake task to iterate through all `repos` `in_batches` and fetch one by one
| return unless github_bub_response.success? | ||
|
|
||
| remote_labels.each do |label_hash| | ||
| label_name = label_hash['name'].downcase |
There was a problem hiding this comment.
The issue stated that we wanted to only store hacktoberfest related labels. I'm happy to add a next unless line here if that's still the case but this is pretty simple as is.
| desc 'fetch and assign labels for repos' | ||
| task fetch_labels_and_assign: :environment do | ||
| Repo.find_each(batch_size: 100) do |repo| | ||
| RepoLabelAssigner.new(repo: repo).create_and_associate_labels! |
There was a problem hiding this comment.
I considered creating a new background job here and enqueueing it for 10 repos at a time but worried about rate-limiting and not knowing the concurrency of background processing in production. I also wasn't sure what the total count of repos in production was.
There was a problem hiding this comment.
Repo count is on the main page it's ~6,000. The way we do API requests is that we rotate through everyone's tokens on the service for each request.
CodeTriage/config/initializers/git_hub_bub.rb
Lines 48 to 54 in 31bd0e7
It's a good idea not to make a bunch of extra requests, but we're not that constrained. If this is N API requests where N is the number of repos in the system, that's absolutely no problem.
|
|
||
| require 'test_helper' | ||
|
|
||
| class RepoLabelsAssignerTest < ActiveSupport::TestCase |
There was a problem hiding this comment.
Outside of my comfort zone here (I'm used to RSpec) so if this needs to change please let me know!
|
@schneems thanks for accepting! It seems like Hacktoberfest wants you to add
|
|
Updated |
@schneems thanks for the opportunity to contribute!
Description
Adds two new tables:
repo_labelsandlabelsto store label dataCreated a new service class to store and associate labels for repos
if the same
labelexists on multiplerepos, it does not makeduplicate
labelrecords.Add simple rake task to iterate through all
reposin_batchesand fetch one by oneRelated Issue
Partially resolved #1680
Motivation and Context
People should be able to target Hacktoberfest
How Has This Been Tested?
Some simple integration testing of the newly created PORO
Types of changes
Checklist: