fix(translate): same-language short-circuit blocked multi-language translation

Root cause: _classify() immediately skipped (continue) when detected
language matched ANY target language, preventing cache lookup and LLM
translation for remaining targets. E.g., ru source + targets [ru, en]
→ only ru row created, en never processed (8978 ru vs 18 en in DB).

Fix:
1. _classify(): Only short-circuit when ALL targets match detected
   language. Partial match → mark _same_language but continue to
   cache check and LLM for non-matching targets.
2. _persist_pre(): Unify two code paths into single loop over all
   target languages. Same-language targets use source_text as-is;
   cached targets use cache value; fallback to approved_translation.
3. SIM102: Merge nested 'if cl: if all(...)' → 'if cl and all(...)'
   to keep cyclomatic complexity ≤ 10 (INV_7).

QA: All 69 translate tests pass. 7 scenario traces (A-G) verified.
No regressions in approved_translation, preview edits, or single-
target same-language flows.

backend/src/plugins/translate/_batch_proc.py

@@ -127,21 +127,27 @@ class BatchProcessingService:
         llm_rows, pre_rows = [], []
         tls_lower = [str(t).lower() for t in tls]
         for row in batch_rows:
-            # ★ NEW: Same-language pre-filter — detected language matches target
+            # ★ Same-language pre-filter: only short-circuit when ALL targets
+            #    match the detected language. If only SOME match, still process
+            #    other targets via cache or LLM.
             dl = row.get("_detected_lang")
             if dl and str(dl).lower() not in ("und", "") and str(dl).lower() in tls_lower:
+                non_matching = [t for t in tls if str(t).lower() != str(dl).lower()]
+                if not non_matching:
+                    # All targets are the same as detected language — no translation needed
+                    row["_same_language"] = True
+                    pre_rows.append(row)
+                    continue
+                # Partial match: mark same-language for later use, but don't skip
                 row["_same_language"] = True
-                pre_rows.append(row)
-                continue
 
             if row.get("approved_translation"):
                 pre_rows.append(row)
                 continue
             cl = row.get("_cached_lang_values")
-            if cl:
-                if all(lc in cl for lc in tls):
-                    pre_rows.append(row)
-                    continue
+            if cl and all(lc in cl for lc in tls):
+                pre_rows.append(row)
+                continue
             if preview_edits_cache:
                 sd = row.get("source_data") or {}
                 if sd:
@@ -159,34 +165,18 @@ class BatchProcessingService:
     def _persist_pre(self, pre_rows, bid, run_id, tls):
         count = 0
         for row in pre_rows:
-            # ★ Same-language row: source language = target → no translation needed
-            if row.get("_same_language"):
-                dl = str(row["_detected_lang"])
-                source_text = row.get("source_text", "")
-                rec = TranslationRecord(
-                    id=str(uuid.uuid4()), batch_id=bid, run_id=run_id,
-                    source_sql=source_text, target_sql=source_text,
-                    source_object_type="table_row", source_object_id=row.get("row_index"),
-                    source_object_name=row.get("source_object_name", ""),
-                    source_data=row.get("source_data"), source_hash=row.get("_source_hash"),
-                    status="SUCCESS",
-                )
-                self.db.add(rec)
-                # One TranslationLanguage for the detected (same) language
-                self.db.add(TranslationLanguage(
-                    id=str(uuid.uuid4()), record_id=rec.id, language_code=dl,
-                    source_language_detected=dl, translated_value=source_text,
-                    final_value=source_text, status="translated", needs_review=False,
-                ))
-                count += 1
-                continue
-
-            # Existing: cached / approved rows
             cl = row.get("_cached_lang_values")
             detected_lang = row.get("_detected_lang", "und") or "und"
+            source_text = row.get("source_text", "")
+            is_same = row.get("_same_language")
+
+            # target_sql: prefer approved_translation, then first cached value, then source
+            primary_cached = next(iter(cl.values()), "") if cl else ""
+            target_sql = row.get("approved_translation") or primary_cached or source_text
+
             rec = TranslationRecord(
                 id=str(uuid.uuid4()), batch_id=bid, run_id=run_id,
-                source_sql=row.get("source_text", ""), target_sql=row.get("approved_translation", ""),
+                source_sql=source_text, target_sql=target_sql,
                 source_object_type="table_row", source_object_id=row.get("row_index"),
                 source_object_name=row.get("source_object_name", ""),
                 source_data=row.get("source_data"), source_hash=row.get("_source_hash"),
@@ -194,7 +184,13 @@ class BatchProcessingService:
             )
             self.db.add(rec)
             for lc in tls:
-                fv = cl[lc] if (cl and lc in cl) else row.get("approved_translation", "")
+                if is_same and str(lc).lower() == str(detected_lang).lower():
+                    # Same language: use source text as-is (no translation needed)
+                    fv = source_text
+                elif cl and lc in cl:
+                    fv = cl[lc]
+                else:
+                    fv = row.get("approved_translation", "")
                 self.db.add(TranslationLanguage(
                     id=str(uuid.uuid4()), record_id=rec.id, language_code=lc,
                     source_language_detected=detected_lang, translated_value=fv,