Recent GitClear Improvements
Get your own free automated changelog at the link
Add redundant call site to ensure that a commit is designated as "present in main branch" if we find its commit site has kin in default branch
Deployed November 3
Worked October 31 in gitclear/gitclear
100 Δ
Add "Test Delta Cohorts" and "Copy/Paste Cohorts" to lineup of AI Cohort stats; updates to chart color palette;
* Fix color palette retrieval relying on calling segment when it was not being passed to AreaChart (but could be derived from chartDataPath)
Deployed November 3
Worked October 30 in gitclear/gitclear
100 Δ
Transform gross "Commit count per AI Cohort" to "Commit Count per Count in AI Cohort" ✨✨✨
Deployed October 30
Worked October 30 in gitclear/gitclear
100 Δ
Rework repo_import_redirect_path such that, if no evidence exists that user has completed GK redirect, that will take precedence over other types of redirects (besides usage profile)
Awaiting deploy
Worked October 27 in gitclear/kraken
100 Δ
Extract GK redirect to its own concern, so it can be invoked via import or user usage profile
Awaiting deploy
Worked October 27 in gitclear/kraken
100 Δ
Passing test for fetching a missing Github remote ID via GraphQL
* Move CaptureCommitterDetail out of committer callback and into post-work method called by ProcessCommit / ClosedPR. Potentially should also be called after creating PR review, not done yet since that'd imply additional architecture needed for the GraphQL input during what's already an unscheduled task
Deployed October 23
Worked October 23 in gitclear/gitclear
100 Δ
Add default_branch_processed_back_through as another column returned by import_progress to help deduce how far back the import has processed
Deployed October 24
Worked October 23 in gitclear/gitclear
100 Δ
Assorted updates to ensure that commits participating in a closed PR will check whether they are in default branch + test
Deployed October 23
Worked October 22 in gitclear/gitclear
100 Δ
Add provider_email to data returned in developer_list. 0
Add provider_email to data returned in developer_list. Fall back to looking up developer by email when the `git_provider_id` given does not match any known developer from the resource in question.
Deployed October 23
Worked October 21 in gitclear/gitclear
100 Δ
Add stat for teammates' review minutes, to measure extent to which AI increases review burden
* LlmTeamCohortStat builds* Add teammates_review_minutes committer_stat segment* Move validation for start_date & date_type_em out of StatCacheConcerns and into standalone concern since AI stat models don't use digests
Deployed October 23
Worked October 20 in gitclear/gitclear
100 Δ
Add "remove_repo_paths" to complement "repo_paths" of team API
Deployed October 10
Worked on from October 9 through October 15 in gitclear/gitclear
100 Δ
Passing test that we can process Copilot stats for Bitbucket/Gitlab/Azure organizations
Pull request
closed
Deployed October 10
Worked October 8 in gitclear/gitclear
100 Δ
Sync upstream
Passing test that we can process Copilot stats for Bitbucket/Gitlab/Azure organizations
Pull request
closed
Deployed October 10
Worked October 8 in gitclear/kraken
100 Δ
Refactor ImportRepo job
* Ensure that job gets a final attempt to import if it can find a valid git connection to use* Employ exponential backoff with some jitter, for cases where 100+ repos might have been simultaneously queued, invoking quota issues* Ensure that, for cases where the repo import has met its import limit, that we add some message in customer_error_detail to explain more
Pull request
closed
Deployed October 8
Worked October 7 in gitclear/gitclear
100 Δ
Ensure that "List import API call" provides detail on nature of error, when available
Pull request
closed
Deployed October 8
Worked October 7 in gitclear/gitclear
100 Δ
Sync upstream
Refactor ImportRepo job
Pull request
closed
Deployed October 10
Worked October 7 in gitclear/kraken
100 Δ
Revise AI stats to automatically aggregate by whichever method will produce most interesting results, in a model that should probably soon be extended to at least ResourceTeamStat
* Rework ai_impact_charts_test.rb to reflect variation in optimal aggregate type+resource returned* Alphabetize class & instance methods of StatCacheConcerns no I'm not ocd thx for asking
Pull request
closed
Deployed October 3
Worked October 2 in gitclear/gitclear
100 Δ
Sync upstream
Revise AI stats to automatically aggregate by whichever method will produce most interesting results, in a model that should probably soon be extended to at least ResourceTeamStat
Pull request
closed
Deployed October 10
Worked October 2 in gitclear/kraken
100 Δ
Add 2025 developer productivity research to the Research page
Pull request
closed
Deployed October 1
Worked October 1 in gitclear/gitclear
100 Δ
Sync upstream
Add 2025 developer productivity research to the Research page
Pull request
closed
Deployed October 10
Worked October 1 in gitclear/kraken
100 Δ