Matching algo: Scoring based on history, match preferences, and industry. by Allimonae · Pull Request #135 · Patina-Network/patchats

Allimonae · 2026-10-09T23:26:37Z

Overview

This adds the scoring-based matching algorithm, which replaces the purely random pairing used by the matching CLI.

Greedy pairing

GreedyPairer scores every possible pair of members, then takes the highest-scoring pairs first and skips any pair that includes someone already taken.
Members are shuffled with a seeded Random before scoring, so ties break randomly but reproducibly. With no scorers, the result is identical to RandomPairer for the same seed.
If the member count is odd, one member is left unmatched.

Scorers, each added as a layer on top of the previous ones:

HistoryScorer applies a -100 penalty to any pair matched in a previous cycle, regardless of order. Repeats only happen when there is no alternative.
MatchPrefScorer gives +20 for compatible preferences (Mentor with Mentee, Peer with Peer, or either member having no preference) and -20 for conflicting ones. The history penalty outweighs the best preference bonus, so preferences never cause a repeat.
IndustryScorer gives +10 when both members share a specific industry. "Other", "No Preference" and differing industries are neutral, so industry only breaks ties between pairs the earlier layers score equally.

Preference parsing

Industry parses Member.industryPref, ignoring case and whitespace. Missing or unrecognized values count as "No Preference".
MatchPreference parses Member.matchPref. The stored value is the full dropdown text (e.g. "Mentor - I am looking for guidance ..."), so it is recognized by its leading label. Missing or unrecognized values count as "No Preference".

Matching CLI

MatchingCli now builds the scorer list for the selected layer and every layer before it, and runs GreedyPairer instead of RandomPairer.
The pairs table has new columns for each member's match preference, industry and pair score.
The totals section reports how many pairs have compatible, neutral and conflicting preferences, and how many share an industry.

Allimonae · 2026-10-09T23:26:57Z

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

Matching algo: Scoring based on history, match preferences, and industry. #135 👈 (View in Graphite)
Matching algo: Design doc and implementation of a random pairing algorithm. #134
Matching algo: Add mock members data and match history for matching algo #133
Matches: Implement CRUD operations and filtering #126 : 3 other dependent PRs (#127 , #128 , #132 )
main

This stack of pull requests is managed by Graphite. Learn more about stacking.

spiffyy99 · 2026-10-22T03:10:25Z

+ * The industry a member is in or wants to get into, from the sign-up dropdown stored in {@code Member.industryPref}.
+ * Comment - The current match preference dropdown includes description, the mock data does not
+ */
+public enum Industry {


also for the future - but i feel like it might be a good idea to eventually have a negative industry match - like someone wanting to talk to someone out of their industry too. we'd have to adjust the signup form as well, but just a thought.

spiffyy99 · 2026-10-22T03:12:25Z

+            return NO_PREFERENCE;
+        }
+        String trimmed = industryPref.trim();
+        for (Industry industry : values()) {


you can just do Industry.valueOf(industryPref) here

…try.

coderabbitai · 2026-10-26T19:03:12Z

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

Run configuration

Source mode: Auto
Repository: Unavailable
Branch: Unavailable
File: No configuration file from this repository applied
Commit: Unavailable
Layers: defaults
Review profile: CHILL
Plan: Advanced
Run ID: bd76056b-9fe5-4c7d-9194-a8bec394d435

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

🔍 Trigger review

Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

_{Comment @coderabbitai help to get the list of available commands.}

sonarqubecloud · 2026-10-26T19:04:10Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

graphite-app · 2026-10-26T19:08:08Z

+        return isCompatible(prefA, prefB) ? COMPATIBLE_BONUS : CONFLICT_PENALTY;
+    }
+
+    private static boolean isCompatible(MatchPreference prefA, MatchPreference prefB) {
+        return (prefA == MatchPreference.MENTOR && prefB == MatchPreference.MENTEE)
+                || (prefA == MatchPreference.MENTEE && prefB == MatchPreference.MENTOR)
+                || (prefA == MatchPreference.PEER && prefB == MatchPreference.PEER)
+                || (prefA == MatchPreference.NO_PREFERENCE || prefB == MatchPreference.NO_PREFERENCE);


The Javadoc says a member with no preference is neutral, but isCompatible returns true whenever either side is NO_PREFERENCE, so those pairs get COMPATIBLE_BONUS (+20) instead of 0. Missing or unrecognized matchPref values parse to NO_PREFERENCE, so they also get +20.

The greedy pairer then ranks "no preference" pairs level with true Mentor/Mentee and Peer/Peer pairs, and above pairs that merely lack a conflict. The CLI's "no preference" count is also always 0, because such pairs are counted as compatible.

Fix: return 0 when either side is NO_PREFERENCE, and keep +20 or -20 only for specific preferences.

if (prefA == NO_PREFERENCE || prefB == NO_PREFERENCE) return 0; return isCompatible(prefA, prefB) ? COMPATIBLE_BONUS : CONFLICT_PENALTY;

Suggested change

return isCompatible(prefA, prefB) ? COMPATIBLE_BONUS : CONFLICT_PENALTY;

}

private static boolean isCompatible(MatchPreference prefA, MatchPreference prefB) {

return (prefA == MatchPreference.MENTOR && prefB == MatchPreference.MENTEE)

|| (prefA == MatchPreference.MENTEE && prefB == MatchPreference.MENTOR)

|| (prefA == MatchPreference.PEER && prefB == MatchPreference.PEER)

|| (prefA == MatchPreference.NO_PREFERENCE || prefB == MatchPreference.NO_PREFERENCE);

if (prefA == MatchPreference.NO_PREFERENCE || prefB == MatchPreference.NO_PREFERENCE) {

return 0;

}

return isCompatible(prefA, prefB) ? COMPATIBLE_BONUS : CONFLICT_PENALTY;

}

private static boolean isCompatible(MatchPreference prefA, MatchPreference prefB) {

return (prefA == MatchPreference.MENTOR && prefB == MatchPreference.MENTEE)

|| (prefA == MatchPreference.MENTEE && prefB == MatchPreference.MENTOR)

|| (prefA == MatchPreference.PEER && prefB == MatchPreference.PEER);

Spotted by Graphite

Is this helpful? React 👍 or 👎 to let us know.

@Allimonae this is a valid comment - maybe we should assign 0 points to the score if its no preference instead?

spiffyy99 · 2026-10-27T18:29:02Z

+        }
+        String trimmed = matchPref.trim();
+        for (MatchPreference preference : values()) {
+            if (trimmed.regionMatches(true, 0, preference.label, 0, preference.label.length())) {


same as above for valueOf

spiffyy99 · 2026-10-27T18:31:47Z

+        return isCompatible(prefA, prefB) ? COMPATIBLE_BONUS : CONFLICT_PENALTY;
+    }
+
+    private static boolean isCompatible(MatchPreference prefA, MatchPreference prefB) {
+        return (prefA == MatchPreference.MENTOR && prefB == MatchPreference.MENTEE)
+                || (prefA == MatchPreference.MENTEE && prefB == MatchPreference.MENTOR)
+                || (prefA == MatchPreference.PEER && prefB == MatchPreference.PEER)
+                || (prefA == MatchPreference.NO_PREFERENCE || prefB == MatchPreference.NO_PREFERENCE);


@Allimonae this is a valid comment - maybe we should assign 0 points to the score if its no preference instead?

spiffyy99 · 2026-10-27T18:33:55Z

+    public int score(Member memberA, Member memberB) {
+        Industry industryA = Industry.fromIndustryPref(memberA.getIndustryPref());
+        Industry industryB = Industry.fromIndustryPref(memberB.getIndustryPref());
+        return industryA.isSpecific() && industryA == industryB ? SAME_INDUSTRY_BONUS : 0;


if were being biased towards same industry for those who want it, shouldnt we apply a small negative penalty if both industries are specific and different?

Allimonae mentioned this pull request Oct 9, 2026

Matching algo: Add mock members data and match history for matching algo #133

Open

Allimonae mentioned this pull request Oct 9, 2026

Matching algo: Design doc and implementation of a random pairing algorithm. #134

Open

Allimonae changed the title ~~Matching Algo: Scoring based on history, match preferences, and industry.~~ Matching algo: Scoring based on history, match preferences, and industry. Oct 9, 2026

Allimonae force-pushed the matching-algo-score branch from ca067f9 to abe319f Compare October 13, 2026 22:00

Allimonae force-pushed the match-random branch from ed55e0a to 744b77b Compare October 13, 2026 22:00

spiffyy99 reviewed Oct 22, 2026

View reviewed changes

Matching Algo: Scoring based on history, match preferences, and indus…

a4241b3

…try.

Allimonae force-pushed the match-random branch from 744b77b to cad8a11 Compare October 26, 2026 19:03

Allimonae force-pushed the matching-algo-score branch from abe319f to a4241b3 Compare October 26, 2026 19:03

Allimonae marked this pull request as ready for review October 26, 2026 19:06

graphite-app Bot reviewed Oct 26, 2026

View reviewed changes

spiffyy99 reviewed Oct 27, 2026

View reviewed changes

Navigation Menu

Repository navigation

Matching algo: Scoring based on history, match preferences, and industry. - #135

Matching algo: Scoring based on history, match preferences, and industry.#135
Allimonae wants to merge 1 commit into
match-randomfrom
matching-algo-score

Allimonae commented Oct 9, 2026 •

edited

Loading

Uh oh!

Allimonae commented Oct 9, 2026

Uh oh!

spiffyy99 Oct 22, 2026

Uh oh!

spiffyy99 Oct 22, 2026

Uh oh!

coderabbitai Bot commented Oct 26, 2026 •

edited

Loading

Review skipped

Uh oh!

sonarqubecloud Bot commented Oct 26, 2026

Uh oh!

graphite-app Bot Oct 26, 2026

Uh oh!

spiffyy99 Oct 27, 2026

Uh oh!

spiffyy99 Oct 27, 2026

Uh oh!

spiffyy99 Oct 27, 2026

Uh oh!

spiffyy99 Oct 27, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

Allimonae commented Oct 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Overview

Uh oh!

Allimonae commented Oct 9, 2026

Uh oh!

spiffyy99 Oct 22, 2026

Choose a reason for hiding this comment

Uh oh!

spiffyy99 Oct 22, 2026

Choose a reason for hiding this comment

Uh oh!

coderabbitai Bot commented Oct 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review skipped

Uh oh!

sonarqubecloud Bot commented Oct 26, 2026

Quality Gate passed

Uh oh!

graphite-app Bot Oct 26, 2026

Choose a reason for hiding this comment

Uh oh!

spiffyy99 Oct 27, 2026

Choose a reason for hiding this comment

Uh oh!

spiffyy99 Oct 27, 2026

Choose a reason for hiding this comment

Uh oh!

spiffyy99 Oct 27, 2026

Choose a reason for hiding this comment

Uh oh!

spiffyy99 Oct 27, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Allimonae commented Oct 9, 2026 •

edited

Loading

coderabbitai Bot commented Oct 26, 2026 •

edited

Loading