Enhance Progression Path Comparison and Slot Evaluation Features
Deploy Development / deploy (push) Successful in 43s
Test Suite / pytest-backend (push) Successful in 44s
Test Suite / lint-backend (push) Successful in 0s
Test Suite / build-frontend (push) Successful in 14s
Test Suite / k6 /health Baseline (push) Successful in 33s
Test Suite / playwright-tests (push) Successful in 1m12s

- Introduced new fields in `ProgressionPathSuggestRequest` for baseline evaluation and incremental scoring, improving the assessment of proposed paths.
- Implemented `_apply_slot_diff_to_steps` and `_score_incremental_slot_diffs` functions to manage slot differences and evaluate their impact on quality scores.
- Updated `ProgressionGraphEditor` to streamline the match comparison flow, integrating new evaluation parameters and improving user notifications.
- Enhanced `ProgressionOptimizeCompareModal` to better display proposed path suggestions, including pro/con evaluations and quality delta metrics.
- Refactored utility functions for clearer handling of slot differences and improved overall data management in the progression graph editor.
This commit is contained in:
Lars committed 2026-06-13 10:11:10 +02:00
1 parent 19bbcdaf50
commit 85fccdd093
5 files changed
+1098 -282

No files matched your search

@@ -28,11 +28,13 @@ import {
applyEvaluateResponseToDraft,
applyGapOfferToDraft,
applySelectedCompareSteps,
applySelectedSlotSuggestions,
applyResolvedStructuredToDraft,
buildPlanningArtifactFromDraft,
buildProgressionComparePayload,
collectGapOffersFromApiResponse,
compareSlotDiffs,
compareDiffsForDialog,
dedupeGapOffersBySlot,
draftHasLibrarySlotAssignments,
draftRetrievalBoostExerciseIds,
@@ -47,7 +49,7 @@ import {
patchSlotInDraft,
pathQaQualityPercent,
planningCatalogContextToApi,
recommendedCompareDiffs,
rejectedCompareDiffs,
removeSlotFromDraft,
saveProgressionGraphDraft,
setCatalogSelectItems,
@@ -491,24 +493,20 @@ export default function ProgressionGraphEditor({ graphId, embedded = false, onSa
}
}
const fetchFullMatch = async (synced) =>
api.suggestProgressionPath({
const runMatchCompareFlow = async (synced, { source = 'match' } = {}) => {
setMatchNotice('Pfad bewerten und je Slot passende Verbesserungen prüfen…')
const reviewRes = await api.suggestProgressionPath({
...buildMatchRequestBase(synced),
preserve_slot_assignments: false,
unified_slot_review: true,
baseline_evaluate_steps: slotsToEvaluateSteps(synced),
include_llm_intent: false,
include_llm_path_qa: false,
auto_rematch_after_qa: false,
})
setPathQa(reviewRes?.path_qa || null)
const runMatchCompareFlow = async (synced, { source = 'match' } = {}) => {
setMatchNotice('Schritt 1/2: Aktuellen Pfad bewerten…')
const baselineRes = await fetchPathEvaluate(synced)
setPathQa(baselineRes?.path_qa || null)
setMatchNotice('Schritt 2/2: Match für alle Slots (Bibliothek + Lücken)…')
const matchRes = await fetchFullMatch(synced)
const compareRes = buildProgressionComparePayload(baselineRes, matchRes)
setGapFillOffers(mergeGapOffersForDraft(synced, baselineRes, matchRes))
const compareRes = buildProgressionComparePayload(null, reviewRes)
setGapFillOffers(mergeGapOffersForDraft(synced, reviewRes, reviewRes))
presentMatchCompare(compareRes, { source })
return compareRes
}
@@ -522,31 +520,20 @@ export default function ProgressionGraphEditor({ graphId, embedded = false, onSa
setCompareOpen(true)
const baselineQa = res?.baseline_path_qa || null
const proposedQa = res?.proposed_path_qa || res?.path_qa || null
const diffCount =
res?.slot_diff_count_recommended
?? recommendedCompareDiffs(res).length
?? res?.slot_diff_count
?? compareSlotDiffs(res, { actionableOnly: true }).length
const replaceCount = (res?.slot_diffs || []).filter(
(d) => d?.diff_kind === 'replace',
).length
const diffCount = res?.slot_diff_count ?? compareDiffsForDialog(res).length
const rejectedCount = res?.slot_diff_count_rejected ?? rejectedCompareDiffs(res).length
const bPct = pathQaQualityPercent(baselineQa)
const pPct = pathQaQualityPercent(proposedQa)
let notice =
diffCount > 0
? `Match: ${diffCount} Lückenfüllung(en) im Dialog — nur diese sind vorausgewählt.`
: 'Match: Keine Bibliotheks-Lückenfüllungen — Dialog zur Kontrolle geöffnet.'
if (replaceCount > 0) {
notice += ` ${replaceCount} optionale Ersetzung(en) bestehender Slots — standardmäßig abgewählt.`
? `Match: ${diffCount} Verbesserung(en) — je Slot gegen deinen Pfad (${bPct != null ? `${bPct} %` : 'QS'}) geprüft.`
: 'Match: Keine messbare Verbesserung gegenüber deinem Pfad.'
if (rejectedCount > 0) {
notice += ` ${rejectedCount} Vorschlag/Vorschläge verworfen (Verschlechterung oder neutral).`
}
const gapCount = collectGapOffersFromApiResponse(res).length
if (gapCount > 0) {
notice += ` ${gapCount} KI-Angebot(e) für leere Slots im Panel „Graph-Bewertung“.`
}
if (bPct != null && pPct != null && pPct !== bPct) {
notice += ` Pfad-QS Vorschlag fair bewertet: ${bPct} % → ${pPct} %.`
}
setMatchNotice(notice)
}
@@ -603,11 +590,13 @@ export default function ProgressionGraphEditor({ graphId, embedded = false, onSa
setCompareApplying(true)
try {
const synced = syncProgressionRoadmapFromSlots(draft)
const nextDraft = applySelectedCompareSteps(
synced,
comparePayload.proposed_steps || comparePayload.steps,
selectedMajorIndices,
)
const nextDraft = comparePayload?.unified_slot_review
? applySelectedSlotSuggestions(synced, comparePayload, selectedMajorIndices)
: applySelectedCompareSteps(
synced,
comparePayload.proposed_steps || comparePayload.steps,
selectedMajorIndices,
)
const syncedNext = syncProgressionRoadmapFromSlots(nextDraft)
const evalRes = await fetchPathEvaluate(syncedNext)
const { draft: evaluated, remainingOffers } = applyEvaluateResult(syncedNext, evalRes)
@@ -1,14 +1,13 @@
/**
* Gegenüberstellung: bestehender Pfad vs. optimierter Match-Vorschlag.
* Gegenüberstellung: Verbesserungsvorschläge mit Slot-Bewertung (Pro/Contra).
*/
import React, { useMemo, useState } from 'react'
import {
compareDiffsForDialog,
defaultSelectedCompareDiffs,
gapOnlyCompareDiffs,
optionalReplaceCompareDiffs,
pathQaQualityPercent,
recommendedCompareDiffs,
qualityDeltaPercent,
rejectedCompareDiffs,
} from '../utils/progressionGraphDraft'
function qaLabel(pathQa) {
@@ -18,25 +17,44 @@ function qaLabel(pathQa) {
return ok ? 'OK' : 'Hinweise'
}
function DiffRow({ diff, checked, onToggle, applying, tone = 'neutral' }) {
function deltaLabel(diff) {
const pct = qualityDeltaPercent(diff)
if (pct == null) return null
if (pct > 0) return `+${pct} % Pfad-QS`
if (pct === 0) return '±0 % Pfad-QS'
return `${pct} % Pfad-QS`
}
function ProContraList({ title, items, tone = 'neutral' }) {
if (!items?.length) return null
const color =
tone === 'pro' ? 'var(--accent-dark)' : tone === 'contra' ? 'var(--danger)' : 'var(--text2)'
return (
<div style={{ marginTop: '6px' }}>
<div style={{ fontSize: '10px', fontWeight: 600, color: 'var(--text3)' }}>{title}</div>
<ul style={{ margin: '4px 0 0', paddingLeft: '16px', color, fontSize: '11px' }}>
{items.map((text, i) => (
<li key={`${title}-${i}`}>{text}</li>
))}
</ul>
</div>
)
}
function DiffRow({ diff, checked, onToggle, applying }) {
const midx = Number(diff.roadmap_major_step_index)
const border =
tone === 'warn'
? '1px solid color-mix(in srgb, var(--danger) 35%, var(--border))'
: '1px solid var(--border)'
const bg = checked
? tone === 'warn'
? 'color-mix(in srgb, var(--danger) 6%, var(--surface2))'
: 'var(--surface2)'
: 'var(--surface)'
const delta = deltaLabel(diff)
const pc = diff.pro_contra || {}
const isAi = diff.suggestion_type === 'ai_gap' || diff.proposed_is_ai_proposal
const isFill = diff.baseline_exercise_id == null && !isAi
return (
<li
style={{
padding: '10px 12px',
borderRadius: '8px',
border,
background: bg,
border: '1px solid var(--border)',
background: checked ? 'var(--surface2)' : 'var(--surface)',
fontSize: '12px',
}}
>
@@ -49,28 +67,60 @@ function DiffRow({ diff, checked, onToggle, applying, tone = 'neutral' }) {
style={{ marginTop: '3px' }}
/>
<span style={{ flex: 1 }}>
<strong>Slot {midx + 1}</strong>
{tone === 'warn' ? (
<span style={{ marginLeft: '6px', fontSize: '10px', color: 'var(--danger)' }}>
Ersetzt deine Zuordnung
</span>
) : null}
<div style={{ display: 'flex', flexWrap: 'wrap', gap: '6px', alignItems: 'center' }}>
<strong>Slot {midx + 1}</strong>
{isFill ? (
<span style={{ fontSize: '10px', color: 'var(--accent-dark)' }}>Lücke füllen</span>
) : isAi ? (
<span style={{ fontSize: '10px', color: 'var(--accent-dark)' }}>KI-Alternative</span>
) : diff.off_topic ? (
<span style={{ fontSize: '10px', color: 'var(--danger)' }}>Passt nicht — Ersatz</span>
) : (
<span style={{ fontSize: '10px', color: 'var(--text2)' }}>Bessere Übung</span>
)}
{delta ? (
<span
style={{
fontSize: '10px',
fontWeight: 600,
color: qualityDeltaPercent(diff) > 0 ? 'var(--accent-dark)' : 'var(--text2)',
}}
>
{delta}
</span>
) : null}
</div>
<div
style={{
display: 'grid',
gridTemplateColumns: '1fr 1fr',
gap: '8px',
marginTop: '6px',
gap: '10px',
marginTop: '8px',
}}
>
<span style={{ color: 'var(--text2)' }}>
Bisher: {diff.baseline_title || '— leer —'}
{diff.baseline_exercise_id != null ? ` (#${diff.baseline_exercise_id})` : ''}
</span>
<span style={{ color: tone === 'warn' ? 'var(--danger)' : 'var(--accent-dark)' }}>
Neu: {diff.proposed_title || '— leer —'}
{diff.proposed_exercise_id != null ? ` (#${diff.proposed_exercise_id})` : ''}
</span>
<div>
<div style={{ fontSize: '10px', fontWeight: 600, color: 'var(--text3)', marginBottom: '4px' }}>
Aktuell
</div>
<div style={{ color: 'var(--text2)' }}>
{diff.baseline_title || '— leer —'}
{diff.baseline_exercise_id != null ? ` (#${diff.baseline_exercise_id})` : ''}
</div>
<ProContraList title="Pro" items={pc.current_pro} tone="pro" />
<ProContraList title="Contra" items={pc.current_contra} tone="contra" />
</div>
<div>
<div style={{ fontSize: '10px', fontWeight: 600, color: 'var(--text3)', marginBottom: '4px' }}>
Vorschlag
</div>
<div style={{ color: 'var(--accent-dark)' }}>
{diff.proposed_title || '—'}
{diff.proposed_exercise_id != null ? ` (#${diff.proposed_exercise_id})` : isAi ? ' (KI)' : ''}
</div>
<ProContraList title="Pro" items={pc.proposed_pro} tone="pro" />
<ProContraList title="Contra" items={pc.proposed_contra} tone="contra" />
</div>
</div>
</span>
</label>
@@ -86,16 +136,8 @@ export default function ProgressionOptimizeCompareModal({
onApplySelected,
applying = false,
}) {
const recommended = useMemo(
() => recommendedCompareDiffs(comparison),
[comparison],
)
const optionalReplace = useMemo(
() => optionalReplaceCompareDiffs(comparison),
[comparison],
)
const gapOnly = useMemo(() => gapOnlyCompareDiffs(comparison), [comparison])
const dialogDiffs = useMemo(() => compareDiffsForDialog(comparison), [comparison])
const rejected = useMemo(() => rejectedCompareDiffs(comparison), [comparison])
const defaultSelected = useMemo(
() => defaultSelectedCompareDiffs(comparison),
[comparison],
@@ -111,18 +153,8 @@ export default function ProgressionOptimizeCompareModal({
if (!open || !comparison) return null
const baselineQa = comparison.baseline_path_qa
const pipelineQa = comparison.proposed_path_qa_pipeline
const baselinePct = pathQaQualityPercent(baselineQa)
const pipelinePct = pathQaQualityPercent(pipelineQa)
const rematchRounds = pipelineQa?.rematch_rounds
const rematchCount = Array.isArray(pipelineQa?.rematch_log) ? pipelineQa.rematch_log.length : 0
const refineCount = Array.isArray(pipelineQa?.refine_log) ? pipelineQa.refine_log.length : 0
const hintCount = Number(pipelineQa?.optimization_hint_count || 0)
const tierCount = Array.isArray(pipelineQa?.qa_tiers) ? pipelineQa.qa_tiers.length : 0
const selectedReplaceCount = optionalReplace.filter((d) =>
selected.has(Number(d.roadmap_major_step_index)),
).length
const rejectedCount = rejected.length
const toggle = (midx) => {
setSelected((prev) => {
@@ -133,20 +165,12 @@ export default function ProgressionOptimizeCompareModal({
})
}
const toggleGroup = (diffs, on) => {
setSelected((prev) => {
const next = new Set(prev)
for (const d of diffs) {
const midx = Number(d.roadmap_major_step_index)
if (on) next.add(midx)
else next.delete(midx)
}
return next
})
const toggleAll = (on) => {
setSelected(on ? new Set(dialogDiffs.map((d) => Number(d.roadmap_major_step_index))) : new Set())
}
const title =
mode === 'match' ? 'Übungs-Match — Vorschläge prüfen' : 'Optimierung vergleichen'
mode === 'match' ? 'Übungs-Match — Verbesserungen' : 'Optimierung vergleichen'
return (
<div
@@ -161,201 +185,90 @@ export default function ProgressionOptimizeCompareModal({
>
<div
className="card modal-content"
style={{ maxWidth: '720px', width: '100%', maxHeight: '90vh', overflow: 'auto' }}
style={{ maxWidth: '760px', width: '100%', maxHeight: '90vh', overflow: 'auto' }}
onClick={(e) => e.stopPropagation()}
>
<h3 id="optimize-compare-title" style={{ marginTop: 0 }}>
{title}
</h3>
<p style={{ fontSize: '12px', color: 'var(--text3)', marginTop: 0, lineHeight: 1.45 }}>
Übernimm nur, was deinen Pfad verbessert. Leere Slots mit Bibliotheks-Treffer sind
vorausgewählt; Ersetzungen bestehender Übungen sind optional und oft schlechter.
KI-Entwürfe ohne Bibliotheks-ID gehören ins Panel „KI-Angebote“, nicht hierher.
Bewertung und Vorschläge in einem Durchlauf: je Slot wird geprüft, ob eine passendere
Übung (Bibliothek oder KI) den Pfad verbessert. Nur messbare Verbesserungen erscheinen
hier — mit Pro- und Contra-Punkten auf Slot-Ebene.
</p>
<div
style={{
marginBottom: '12px',
padding: '10px 12px',
borderRadius: '8px',
border: '1px solid color-mix(in srgb, var(--accent) 30%, var(--border))',
background: 'color-mix(in srgb, var(--accent) 6%, var(--surface2))',
border: '1px solid var(--border)',
background: 'var(--surface2)',
fontSize: '12px',
lineHeight: 1.45,
}}
>
<strong>Warum nicht einfach alles übernehmen?</strong>
<p style={{ margin: '6px 0 0', color: 'var(--text2)' }}>
Der Match-Lauf optimiert den <em>gesamten</em> Pfad neu (inkl. Rematch). Das kann
bereits gute Slots verschlechtern. Die Prozentzahl rechts bezieht sich auf diesen
Ganzpfad — nicht darauf, dass deine Auswahl besser ist. Nimm deshalb standardmäßig
nur Lückenfüllungen an.
</p>
</div>
<div
style={{
display: 'grid',
gridTemplateColumns: '1fr 1fr',
gap: '12px',
marginBottom: '14px',
}}
>
<div
style={{
padding: '10px 12px',
borderRadius: '8px',
border: '1px solid var(--border)',
background: 'var(--surface2)',
fontSize: '12px',
}}
>
<strong>Dein Pfad (bewertet)</strong>
<div style={{ marginTop: '6px' }}>{qaLabel(baselineQa)}</div>
{baselineQa?.topic_coverage ? (
<p style={{ margin: '6px 0 0', color: 'var(--text2)' }}>{baselineQa.topic_coverage}</p>
) : null}
</div>
<div
style={{
padding: '10px 12px',
borderRadius: '8px',
border: '1px solid color-mix(in srgb, var(--text3) 35%, var(--border))',
background: 'var(--surface2)',
fontSize: '12px',
}}
>
<strong>Match-Ganzpfad (nur Info)</strong>
<div style={{ marginTop: '6px', color: 'var(--text2)' }}>
{pipelineQa ? qaLabel(pipelineQa) : '—'}
</div>
<p style={{ margin: '6px 0 0', fontSize: '11px', color: 'var(--text3)' }}>
Rematch-Prozess — kein Versprechen für deine Checkbox-Auswahl.
{pipelinePct != null && baselinePct != null && pipelinePct < baselinePct
? ` (${pipelinePct} % < ${baselinePct} % bei voller Übernahme).`
: ''}
</p>
</div>
<strong>Dein Pfad</strong>
<div style={{ marginTop: '6px' }}>{qaLabel(baselineQa)}</div>
{baselineQa?.topic_coverage ? (
<p style={{ margin: '6px 0 0', color: 'var(--text2)' }}>{baselineQa.topic_coverage}</p>
) : null}
</div>
{tierCount > 0 || rematchCount > 0 || refineCount > 0 || hintCount > 0 ? (
{rejectedCount > 0 ? (
<p style={{ fontSize: '11px', color: 'var(--text2)', margin: '0 0 14px', lineHeight: 1.45 }}>
Rematch-Protokoll
{tierCount > 0 ? ` · ${tierCount} QS-Stufen` : ''}
{rematchCount > 0
? ` · ${rematchRounds != null ? `${rematchRounds} Runde(n)` : ''}: ${rematchCount} Anpassung(en)`
: ''}
{refineCount > 0 ? ` · ${refineCount} Stufen-Spec verfeinert` : ''}
{hintCount > 0 ? ` · ${hintCount} Handlungshinweis(e)` : ''}
</p>
) : null}
{gapOnly.length > 0 ? (
<p style={{ fontSize: '11px', color: 'var(--text2)', margin: '0 0 14px', lineHeight: 1.45 }}>
Slot{gapOnly.length > 1 ? 's' : ''}{' '}
{gapOnly.map((d) => Number(d.roadmap_major_step_index) + 1).join(', ')}: kein
Bibliotheks-Treffer — bitte „KI-Angebote“ im Panel nutzen (eigenständig pro Slot).
{rejectedCount} Alternative(n) verworfen — kein QS-Gewinn gegenüber deinem Pfad
{baselinePct != null ? ` (${baselinePct} %)` : ''}.
</p>
) : null}
{dialogDiffs.length === 0 ? (
<p style={{ fontSize: '12px', color: 'var(--text2)' }}>
Keine übernehmbaren Bibliotheks-Änderungen. Leere Slots ggf. über KI-Angebote im Panel
befüllen — nichts am Pfad ändern ist oft die richtige Wahl.
Keine Verbesserung gefunden — dein Pfad ist für alle Slots bereits optimal bewertet
oder es fehlen passende Bibliotheks-Treffer (KI-Angebote im Bewertungs-Panel).
</p>
) : (
<>
{recommended.length > 0 ? (
<>
<h4 style={{ margin: '0 0 8px', fontSize: '0.9rem' }}>
Lücken füllen (empfohlen)
</h4>
<div style={{ display: 'flex', gap: '8px', marginBottom: '8px', flexWrap: 'wrap' }}>
<button
type="button"
className="btn btn-secondary"
style={{ fontSize: '11px' }}
onClick={() => toggleGroup(recommended, true)}
>
Alle Lücken wählen
</button>
<button
type="button"
className="btn btn-secondary"
style={{ fontSize: '11px' }}
onClick={() => toggleGroup(recommended, false)}
>
Keine
</button>
</div>
<ul
style={{
listStyle: 'none',
padding: 0,
margin: '0 0 16px',
display: 'flex',
flexDirection: 'column',
gap: '8px',
}}
>
{recommended.map((diff) => (
<DiffRow
key={`fill-${diff.roadmap_major_step_index}`}
diff={diff}
checked={selected.has(Number(diff.roadmap_major_step_index))}
onToggle={toggle}
applying={applying}
/>
))}
</ul>
</>
) : null}
{optionalReplace.length > 0 ? (
<>
<h4 style={{ margin: '0 0 8px', fontSize: '0.9rem', color: 'var(--danger)' }}>
Bestehende Slots ersetzen (optional — oft Verschlechterung)
</h4>
<p style={{ fontSize: '11px', color: 'var(--text2)', margin: '0 0 8px' }}>
Standard: abgewählt. Nur aktivieren, wenn du die konkrete Übung bewusst tauschen
willst.
</p>
<ul
style={{
listStyle: 'none',
padding: 0,
margin: 0,
display: 'flex',
flexDirection: 'column',
gap: '8px',
}}
>
{optionalReplace.map((diff) => (
<DiffRow
key={`replace-${diff.roadmap_major_step_index}`}
diff={diff}
checked={selected.has(Number(diff.roadmap_major_step_index))}
onToggle={toggle}
applying={applying}
tone="warn"
/>
))}
</ul>
</>
) : null}
<div style={{ display: 'flex', gap: '8px', marginBottom: '8px', flexWrap: 'wrap' }}>
<button
type="button"
className="btn btn-secondary"
style={{ fontSize: '11px' }}
onClick={() => toggleAll(true)}
>
Alle wählen
</button>
<button
type="button"
className="btn btn-secondary"
style={{ fontSize: '11px' }}
onClick={() => toggleAll(false)}
>
Keine
</button>
</div>
<ul
style={{
listStyle: 'none',
padding: 0,
margin: 0,
display: 'flex',
flexDirection: 'column',
gap: '8px',
}}
>
{dialogDiffs.map((diff) => (
<DiffRow
key={`improve-${diff.roadmap_major_step_index}-${diff.suggestion_type || 'lib'}`}
diff={diff}
checked={selected.has(Number(diff.roadmap_major_step_index))}
onToggle={toggle}
applying={applying}
/>
))}
</ul>
</>
)}
{selectedReplaceCount > 0 ? (
<p
className="form-error"
style={{ marginTop: '12px', fontSize: '11px' }}
>
{selectedReplaceCount} Ersetzung(en) gewählt — kann Pfad-QS senken. Lückenfüllungen
sind unkritischer.
</p>
) : null}
<div
style={{
display: 'flex',