Development: Port new gender contracts

[ge94zec]

Checklist

General

• Language: I followed the guidelines for inclusive, diversity-sensitive, and appreciative language.
• I chose a title conforming to the naming conventions for pull requests.

Client

• Important: I implemented the changes with a very good performance, prevented too many (unnecessary) REST calls and made sure the UI is responsive, even with large data (e.g. using paging).
• I strictly followed the client coding and design guidelines.

Motivation and Context

This PR ports the Gender Decoder integration to the new BiasedIssue / JobAnalysisDTO contracts ahead of the dependent PRs being merged, so the required compatibility changes are already prepared and don’t block on merge order.
Fixes: #2637 resulting from #2630 and #2465 will collide head-on.

Description

The editor no longer subscribes to GenderBiasAnalysisService.getAnalysisForField(). It receives findings through a new biasedAnalysis input instead, so the removed endpoint is no longer referenced from the atom. toObservable, fieldIdChanges$ and the service injection are gone; genderBiasHighlights now derives from the input.

Contract migration

• GenderBiasAnalysisResponse / BiasedWordDTO → BiasedIssueDTO
• coding string values ('non-inclusive-coded', 'inclusive-coded') →
  computeCodingStatus() returning 'NON_INCLUSIVE' / 'INCLUSIVE' / 'NEUTRAL'
• type: 'non-inclusive' → BiasedIssueDTOTypeEnum.NON_INCLUSIVE
• ComplianceIssue → ComplianceIssueDTO in the assistant card and filter types

Prerequisites:

1. Log in to DocApply
2. Nav to Create position
3. Type a description containing non-inclusive wording, the pill count and the balance pointer update, and the wording is underlined in the editor
4. Select the Gender Decoder filter; only gender-bias highlights remain visible
5. Save and reload, the persisted description contains no highlight spans

Review Progress

Code Review

• Code Review 1

Manual Tests

• Test 1

Screenshots

Test Coverage

Client

Class/File	Line Coverage	Lines	Expects	Ratio
job-creation-form.component.ts	0.00%	1317	103	7.8
editor.component.ts	77.19%	382	29	7.6
ai-assistant-card.component.ts	90.76%	139	11	7.9
gender-bias-analysis.utils.ts	100.00%	19	2	10.5

Last updated: 2026-08-24 16:13:20 UTC

[codacy-production]

Not up to standards ⛔

🔴 Issues 1 medium · 3 minor

Alerts:
⚠ 4 issues (≤ 0 issues of at least minor severity)

Results:
4 new issues

Category	Results
ErrorProne	1 medium
CodeStyle	3 minor

View in Codacy

🟢 Metrics 22 complexity

Metric	Results
Complexity	22

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}

[github-actions]

📊 Client Test Coverage Too Low

🔍 View coverage locally:

pnpm run test:ci
open build/test-results/vitest/coverage/index.html

🌐 View coverage from GitHub:
Download the "coverage-report-client" artifact from this workflow run.

[github-actions]

🤖 No OpenAPI or client changes needed.

[github-actions]

🤖 No OpenAPI or client changes needed.

[az108]

This is in good shape — merges cleanly (git merge-tree origin/main HEAD exits 0, zero conflicts), which is worth calling out given the branch it replaces does not. One deletable file, inline.

Checks that passed

Check	Result
Merge into main	clean, 0 conflicts
Word-boundary fix carried over from #2630	isWordChar in the shared utils, applied at editor.component.ts:453
Empty-token guard	present on both loops (compliance and gender)
Boundary guard scoped correctly	gender loop only — right call, compliance snippets are multi-word phrases where it would be wrong
i18n parity	364/364 keys, genderDecoder complete in EN and DE
Test naming	all new it(...) start with should
New spec	gender-bias-analysis.utils.spec.ts covers trim/dedup/type-filter plus undefined and empty input

The isWordChar JSDoc is the good kind — it records why \p{L} rather than \w (umlauts, ß) and why the hyphen is excluded (mirrors deHyphenNonCodedWords server-side). That reasoning is not recoverable from the code alone, so thanks for writing it down.

One thing that improved in the port: isGenderAnalyzing is now bound as [isGenderAnalyzing]="isAnalyzing()", a plain signal, rather than the genderBiasService.isAnalyzing(...) service-method call the previous branch had in the template.

I deliberately did not flag the raw 'NON_INCLUSIVE' literal in getUniqueNonInclusiveWords — it matches computeCodingStatus in the same file, and type is a narrow 'NON_INCLUSIVE' | 'INCLUSIVE' union, so a typo would fail to compile.

Context

I have closed #2630 in favour of this PR. 11 of the 12 files here overlap with it, that branch could no longer merge after #2465 removed the API it was built on, and everything raised in review there was resolved before closing — the fixes carry forward in this branch.

[az108]

This whole file can go — it's a leftover from before 00000000000052_add_biased_issues_to_jobs.xml merged to main.

It is byte-identical to the merged 052, apart from the changeset id prefixes:

052a_add_biased_issues_to_jobs            ->  053a_add_biased_issues_to_jobs
052b_add_unique_constraint_biased_issues  ->  053b_add_unique_constraint_biased_issues
052c_rename_gender_bias_score_to_ai_score ->  053c_rename_gender_bias_score_to_ai_score

It is also not registered in master.xml — the diff doesn't touch that file at all — so it never runs.

Those two facts happen to cancel out into "harmless": as-is it is dead weight. But if someone later notices the missing include and "fixes" it, this would re-attempt createTable job_biased_issues and the gender_bias_score -> ai_score rename on a schema where 052 already did both. The preConditions onFail="MARK_RAN" guards would absorb it, so still no breakage — just a second changeset that does nothing. Cleanest is to delete the file rather than leave that trap.

job_biased_issues and ai_score are already on main via 052, so nothing here is needed.