normalized windows

app.py

@@ -274,7 +274,7 @@ async def upload_batch_dataset(payload: BatchUploadPayload):
                 # 새 데이터 정규화
                 normalized = [normalize_windows(r) for r in records]
                 new_df = pd.DataFrame(normalized)
-                new_dataset = Dataset.from_pandas(new_df)
+                new_dataset = df_to_dataset(new_df)
 
                 if user_id in existing:
                     # 기존 데이터 정규화 및 병합
@@ -325,3 +325,36 @@ async def upload_batch_dataset(payload: BatchUploadPayload):
         print(f"❌ 배치 푸시 실패: {e}")
         raise HTTPException(status_code=500, detail=f"배치 푸시 실패: {str(e)}")
 
+def normalize_windows(record):
+    w = record.get("windows")
+    result = []
+
+    if isinstance(w, list):
+        for item in w:
+            if isinstance(item, dict):
+                result.extend([str(v) for v in item.values() if v is not None])
+            elif item is not None:
+                result.append(str(item))
+    elif isinstance(w, dict):
+        result = [str(v) for v in w.values() if v is not None]
+
+    record["windows"] = [x for x in result if x not in ("", "None")]
+    return record
+
+def df_to_dataset(df):
+    import pyarrow as pa
+    schema = pa.schema([
+        ("session_id", pa.string()),
+        ("measure_date", pa.string()),
+        ("rms", pa.float64()),
+        ("freq", pa.float64()),
+        ("fatigue", pa.float64()),
+        ("mode", pa.string()),
+        ("window_count", pa.int64()),
+        ("windows", pa.list_(pa.string())),
+        ("measurement_count", pa.int64()),
+        ("batch_date", pa.string()),
+        ("batch_size", pa.int64()),
+        ("timestamp", pa.string())
+    ])
+    return Dataset.from_pandas(df, schema=schema)