Add OCR Kombi Pipeline: modular 11-step architecture with multi-page support

Phase 1 of the clean architecture refactor: Replaces the 751-line ocr-overlay
monolith with a modular pipeline. Each step gets its own component file.

Frontend: /ai/ocr-kombi route with 11 steps (Upload, Orientation, PageSplit,
Deskew, Dewarp, ContentCrop, OCR, Structure, GridBuild, GridReview, GroundTruth).
Session list supports document grouping for multi-page uploads.

Backend: New ocr_kombi/ module with multi-page PDF upload (splits PDF into N
sessions with shared document_group_id). DB migration adds document_group_id
and page_number columns.

Old /ai/ocr-overlay remains fully functional for A/B testing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

admin-lehrer/app/(admin)/ai/ocr-kombi/page.tsx

@@ -0,0 +1,174 @@
+'use client'
+
+import { Suspense } from 'react'
+import { PagePurpose } from '@/components/common/PagePurpose'
+import { BoxSessionTabs } from '@/components/ocr-pipeline/BoxSessionTabs'
+import { KombiStepper } from '@/components/ocr-kombi/KombiStepper'
+import { SessionList } from '@/components/ocr-kombi/SessionList'
+import { SessionHeader } from '@/components/ocr-kombi/SessionHeader'
+import { StepUpload } from '@/components/ocr-kombi/StepUpload'
+import { StepOrientation } from '@/components/ocr-kombi/StepOrientation'
+import { StepPageSplit } from '@/components/ocr-kombi/StepPageSplit'
+import { StepDeskew } from '@/components/ocr-kombi/StepDeskew'
+import { StepDewarp } from '@/components/ocr-kombi/StepDewarp'
+import { StepContentCrop } from '@/components/ocr-kombi/StepContentCrop'
+import { StepOcr } from '@/components/ocr-kombi/StepOcr'
+import { StepStructure } from '@/components/ocr-kombi/StepStructure'
+import { StepGridBuild } from '@/components/ocr-kombi/StepGridBuild'
+import { StepGridReview } from '@/components/ocr-kombi/StepGridReview'
+import { StepGroundTruth } from '@/components/ocr-kombi/StepGroundTruth'
+import { useKombiPipeline } from './useKombiPipeline'
+
+function OcrKombiContent() {
+  const {
+    currentStep,
+    sessionId,
+    sessionName,
+    loadingSessions,
+    activeCategory,
+    isGroundTruth,
+    subSessions,
+    parentSessionId,
+    steps,
+    gridSaveRef,
+    groupedSessions,
+    loadSessions,
+    openSession,
+    handleStepClick,
+    handleNext,
+    handleNewSession,
+    deleteSession,
+    renameSession,
+    updateCategory,
+    handleOrientationComplete,
+    handleSessionChange,
+    setSessionId,
+    setSubSessions,
+    setParentSessionId,
+    setIsGroundTruth,
+  } = useKombiPipeline()
+
+  const renderStep = () => {
+    switch (currentStep) {
+      case 0:
+        return (
+          <StepUpload
+            onUploaded={(sid) => {
+              setSessionId(sid)
+              loadSessions()
+              openSession(sid)
+              handleNext()
+            }}
+          />
+        )
+      case 1:
+        return (
+          <StepOrientation
+            sessionId={sessionId}
+            onNext={handleOrientationComplete}
+            onSessionList={() => { loadSessions(); handleNewSession() }}
+          />
+        )
+      case 2:
+        return (
+          <StepPageSplit
+            sessionId={sessionId}
+            onNext={handleNext}
+            onSubSessionsCreated={(subs) => {
+              setSubSessions(subs)
+              if (sessionId) setParentSessionId(sessionId)
+            }}
+          />
+        )
+      case 3:
+        return <StepDeskew sessionId={sessionId} onNext={handleNext} />
+      case 4:
+        return <StepDewarp sessionId={sessionId} onNext={handleNext} />
+      case 5:
+        return <StepContentCrop sessionId={sessionId} onNext={handleNext} />
+      case 6:
+        return <StepOcr sessionId={sessionId} onNext={handleNext} />
+      case 7:
+        return <StepStructure sessionId={sessionId} onNext={handleNext} />
+      case 8:
+        return <StepGridBuild sessionId={sessionId} onNext={handleNext} />
+      case 9:
+        return <StepGridReview sessionId={sessionId} onNext={handleNext} saveRef={gridSaveRef} />
+      case 10:
+        return (
+          <StepGroundTruth
+            sessionId={sessionId}
+            isGroundTruth={isGroundTruth}
+            onMarked={() => setIsGroundTruth(true)}
+            gridSaveRef={gridSaveRef}
+          />
+        )
+      default:
+        return null
+    }
+  }
+
+  return (
+    <div className="space-y-6">
+      <PagePurpose
+        title="OCR Kombi Pipeline"
+        purpose="Modulare 11-Schritt-Pipeline: Upload, Vorverarbeitung, Dual-Engine-OCR (PP-OCRv5 + Tesseract), Strukturerkennung, Grid-Aufbau und Review. Multi-Page-Dokument-Unterstuetzung."
+        audience={['Entwickler']}
+        architecture={{
+          services: ['klausur-service (FastAPI)', 'OpenCV', 'Tesseract', 'PaddleOCR'],
+          databases: ['PostgreSQL Sessions'],
+        }}
+        relatedPages={[
+          { name: 'OCR Overlay (Legacy)', href: '/ai/ocr-overlay', description: 'Alter 3-Modi-Monolith' },
+          { name: 'OCR Regression', href: '/ai/ocr-regression', description: 'Regressionstests' },
+        ]}
+        defaultCollapsed
+      />
+
+      <SessionList
+        items={groupedSessions()}
+        loading={loadingSessions}
+        activeSessionId={sessionId}
+        onOpenSession={(sid) => openSession(sid)}
+        onNewSession={handleNewSession}
+        onDeleteSession={deleteSession}
+        onRenameSession={renameSession}
+        onUpdateCategory={updateCategory}
+      />
+
+      {sessionId && sessionName && (
+        <SessionHeader
+          sessionName={sessionName}
+          activeCategory={activeCategory}
+          isGroundTruth={isGroundTruth}
+          onUpdateCategory={(cat) => updateCategory(sessionId, cat)}
+        />
+      )}
+
+      <KombiStepper
+        steps={steps}
+        currentStep={currentStep}
+        onStepClick={handleStepClick}
+      />
+
+      {subSessions.length > 0 && parentSessionId && sessionId && (
+        <BoxSessionTabs
+          parentSessionId={parentSessionId}
+          subSessions={subSessions}
+          activeSessionId={sessionId}
+          onSessionChange={handleSessionChange}
+        />
+      )}
+
+      <div className="min-h-[400px]">{renderStep()}</div>
+    </div>
+  )
+}
+
+export default function OcrKombiPage() {
+  return (
+    <Suspense fallback={<div className="p-4 text-sm text-gray-400">Lade...</div>}>
+      <OcrKombiContent />
+    </Suspense>
+  )
+}