You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 2912415
Browse filesBrowse the repository at this point in the historyBrowse files
Copy file name to clipboardExpand all lines: doc/architecture.md
+21-1Lines changed: 21 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -70,6 +70,20 @@ flowchart LR
70
70
71
71
Candidate issues are enriched where possible with repository metadata, recent discussion, assignment state, linked pull requests, and Hacktoberfest signals. Repository health uses recent pushes, open issue and pull-request activity, stars, forks, and issue-tracker availability to distinguish active, moderate, and stale projects. The health score contributes a small optional boost to issue quality ranking and is shown directly on issue cards. The application ranks the enriched results for contributor relevance and returns paginated results to the browser.
72
72
73
+
Maintainer responsiveness remains separate from repository health. For at most 12
74
+
candidate repositories per search, a cached six-hour GitHub GraphQL request samples
75
+
up to 20 recently updated issues and pull requests from the previous 90 days. The
76
+
summary considers the first response from an owner, member, or collaborator (while
77
+
excluding comments from the issue author), unanswered contributor-friendly issues,
78
+
recent closures, and the external pull-request merge ratio. A responsive status
79
+
requires at least half of sampled external pull requests to be merged when any are
80
+
present; multiple external pull requests with a merge ratio below 25% indicate a
81
+
slow repository. Fewer than four contribution
82
+
samples, missing authentication, repositories outside the bounded batch, and partial
83
+
GitHub failures produce `Unknown`. Responsive repositories receive a five-point
84
+
quality boost and variable repositories receive two points; repository health is
85
+
calculated and displayed independently.
86
+
73
87
Search responses report repository-metadata, discussion-analysis, and linked-PR
74
88
availability separately. Optional enrichment failures do not discard usable issue
75
89
results; the interface identifies partial data and score explanations omit signals
@@ -85,6 +99,7 @@ Search filters currently include:
85
99
- Any experience, first contribution, beginner, or intermediate
86
100
- Documentation, tests, bug fix, or feature contribution type
87
101
- All scopes or small scope
102
+
- Any, responsive, variable, slow, or unknown maintainer responsiveness
88
103
89
104
Experience, contribution type, and scope classifications use explicit issue
90
105
labels and structured issue-template fields. Results show the signals that
@@ -185,7 +200,12 @@ observation is explicitly presented as a baseline.
185
200
186
201
Digest issue links open GitHub directly. Saved-search links include the existing
187
202
filter query parameters; the issue finder reads those parameters and runs the
188
-
linked search on load.
203
+
linked search on load. Weekly recommendations honor the saved responsiveness
204
+
filter and show each repository's status and sample size. Repository-alert
205
+
emails also show the cached responsiveness status, sample period, and
206
+
contributing signals beside each repository. Weekly recommendations display the
207
+
same quality score used to rank results in the portal. Repository alerts remain
208
+
newest-issue notifications and do not assign portal search-quality scores.
189
209
190
210
Authenticated users can also request their own digest immediately from the
191
211
saved-search card. The manual route uses the same delivery pipeline and six-day
The first migration creates Better Auth's user, session, account, and verification tables. The second creates user-owned saved searches. Migration files intentionally contain structure only—never credentials or production data.
54
55
@@ -70,6 +71,8 @@ deduplicated opportunity record for each issue a signed-in user saves or opens;
70
71
it does not store the user's GitHub contribution history.
71
72
The ninth adds experience, contribution-type, and scope preferences to cloud
72
73
saved searches. Existing records default to unrestricted filters.
74
+
The tenth adds the optional maintainer-responsiveness preference to cloud saved
75
+
searches. Existing records default to unrestricted responsiveness.
0 commit comments