fix: consistent value-to-legend-item matching and small fixes [PR1] [DHIS2-18242]

[BRaimbault]

Parent

• DHIS2-18242

Overview

Foundational refactor and small bug fixes extracted from the parent epic, landing first as a prerequisite for the classification and legend work to follow. No user-visible feature change (except the clamp fix, see below); mostly internal plumbing.

Changes

Thread valueFormat through value-to-legend-item matching

getLegendItems and its helpers (getEqualIntervals, getQuantiles) now return { items, valueFormat } instead of a bare array. getAutomaticLegendItems propagates the valueFormat, and both callers (thematicLoader.js and styleByDataItem.js) pass it to getLegendItemForValue. Incoming values are rounded to the same precision as bin boundaries before comparison, eliminating a class of off-by-epsilon mismatches at bin edges.

Files: src/util/classify.js, src/util/legend.js, src/loaders/thematicLoader.js, src/util/styleByDataItem.js

Named args for getAutomaticLegendItems

Switched from positional to named parameters; removes the eslint-disable max-params override.

Unified clamp behavior between thematic and event layers

Event style-by-data-item with automatic classification didn't clamp out-of-range values while thematic layer did — a pre-existing asymmetry in master. Both callers now use clamp: method !== CLASSIFICATION_PREDEFINED.

Consequence: events with values slightly outside the auto-classified range (typically a rounding-gap artefact) are now pulled into the nearest bin instead of being dropped to the "Other" colour. The "Other" bucket still catches values that genuinely have no data.

Files: src/loaders/thematicLoader.js, src/util/styleByDataItem.js

Fix min/max calculation in thematicLoader.js for legend sets

When a predefined legendSet was used, minValue/maxValue were read from legend.items[0] and legend.items[legend.items.length - 1]. If a noData item was present at either end, its sentinel value was picked
up and corrupted the bubble-radius scaleSqrt domain. Fixed by filtering noData items first. The .at(-1) rewrite is cosmetic.

Files: src/loaders/thematicLoader.js

Robust legend-item sorting

sortLegendItems rewritten:

• no longer mutates input ([...items].sort(...))
• normalises both { from, to } and { startValue, endValue } shapes on both sides of the comparator (previously only a was inspected)
• handles items with no range shape (sorted to the end deterministically)
• tiebreaker on equal start values

Files: src/util/legend.js

Filter noData items by flag in styleByDataItem.js

Replaced config.legend.items.slice(0, -1) with legend.items.filter((item) => !item.noData). The "Other" legend item
now carries noData: true so the filter captures it. Mirrors the pattern already used in thematicLoader.js; forward-compatible with additional noData items that later PRs may introduce.

Files: src/util/styleByDataItem.js

Local variable renames and dead-guard removal in classify.js

Pure readability cleanup in getQuantiles / getEqualIntervals: bins → items, binSize → classSize, binCount → valuesCount, binLastValPos → lastValuePosition. Dropped the unreachable binCount === 0 ? 0 : binCount guard (always equivalent to binCount because the enclosing if (values.length > 0) is skipped when it would matter). Added ?? {} fallback in getLegendItems for unknown methods.

Files: src/util/classify.js

Drive-by modernisations

• String.prototype.replace(/…/g, …) → replaceAll(…, …) in 9 files
• isNaN(x) → Number.isNaN(x) in PeriodSelect.jsx and RadiusSelect.jsx

No behaviour change. Addresses SonarQube findings.

Tests update

• src/util/__tests__/classify.spec.js
• src/util/__tests__/legend.spec.js

Quality checklist

Add N/A to items that are not applicable.

• Jest tests added/updated
• Docs added N/A
• d2-ci dependencies replaced N/A
• Include plugin in testing N/A
• Tester approved N/A

[dhis2-bot]

🚀 Deployed on https://pr-3645.maps.netlify.dhis2.org

[BRaimbault]

Ready for review.

[sonarqubecloud]

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