fix(ocr-pipeline): always run fresh word detection, skip stale cache

Word-lookup is now ~0.03s (vs seconds with per-cell Tesseract), so
always re-run detection when entering Step 5 instead of showing
potentially stale cached word_result from the session DB.

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

admin-lehrer/components/ocr-pipeline/StepWordRecognition.tsx

@@ -72,24 +72,9 @@ export function StepWordRecognition({ sessionId, onNext, goToStep }: StepWordRec
 
   useEffect(() => {
     if (!sessionId) return
-
-    const fetchSession = async () => {
-      try {
-        const res = await fetch(`${KLAUSUR_API}/api/v1/ocr-pipeline/sessions/${sessionId}`)
-        if (res.ok) {
-          const info = await res.json()
-          if (info.word_result) {
-            applyGridResult(info.word_result)
-            return
-          }
-        }
-      } catch (e) {
-        console.error('Failed to fetch session info:', e)
-      }
-      runAutoDetection()
-    }
-
-    fetchSession()
+    // Always run fresh detection — word-lookup is fast (~0.03s)
+    // and avoids stale cached results from previous pipeline versions.
+    runAutoDetection()
   // eslint-disable-next-line react-hooks/exhaustive-deps
   }, [sessionId])