fixing lat lon logic

services/geometry_detector.py

@@ -132,6 +132,44 @@ def is_geom_empty(g):
 
 
 
+import math
+
+def normalize_dynamic(val, is_lat=False):
+    if pd.isna(val):
+        return None
+    try:
+        v = float(val)
+    except:
+        return None
+
+    av = abs(v)
+    if av == 0:
+        return v
+
+    if (-180 <= v <= 180 and not is_lat) or (-90 <= v <= 90 and is_lat):
+        return v
+    
+    for factor in [1, 10, 100, 1e3, 1e4, 1e5, 1e6, 1e7, 1e8, 1e9]:
+        nv = v / factor
+        if (not is_lat and -180 <= nv <= 180) or (is_lat and -90 <= nv <= 90):
+            return nv
+
+    return None
+
+
+
+def normalize_lat(val):
+    if pd.isna(val): 
+        return None
+    v = float(val)
+    av = abs(v)
+    if av > 1e9:   # contoh: -8167413802 (10 digit)
+        return v / 1e9
+    elif av > 1e8: # fallback jika ada variasi
+        return v / 1e8
+    else:
+        return v
+
 
 # ============================================================
 # FUNGSI UTAMA GEOMETRY DETECTION (LAT/LON / PATH)
@@ -149,25 +187,18 @@ def detect_and_build_geometry(df: pd.DataFrame, master_polygons: gpd.GeoDataFram
             print(f"[INFO] Detected existing geometry in GeoDataFrame ({geom_count} features, {geom_type}).")
             return df
 
-    lat_col = next(
-        (c for c in df.columns if re.search(r'\b(lat|latitude|y[_\s]*coord|y$)\b', c.lower())), None
-    )
-    lon_col = next(
-        (c for c in df.columns if re.search(r'\b(lon|long|longitude|x[_\s]*coord|x$)\b', c.lower())), None
-    )
+    lat_col = next((c for c in df.columns if re.search(r'\b(lat|latitude|y[_\s]*coord|y$)\b', c.lower())), None)
+    lon_col = next((c for c in df.columns if re.search(r'\b(lon|long|longitude|x[_\s]*coord|x$)\b', c.lower())), None)
 
     if lat_col and lon_col:
         df[lat_col] = pd.to_numeric(df[lat_col], errors='coerce')
         df[lon_col] = pd.to_numeric(df[lon_col], errors='coerce')
 
-        lon_median = df[lon_col].abs().median()
-        lat_median = df[lat_col].abs().median()
-
-        if lon_median > 1000 or lat_median > 1000:
-            df[lon_col] = df[lon_col] / 1e7
-            df[lat_col] = df[lat_col] / 1e7
+        df[lon_col] = df[lon_col].apply(lambda x: normalize_dynamic(x, is_lat=False))
+        df[lat_col] = df[lat_col].apply(normalize_lat)
 
         gdf = gpd.GeoDataFrame(df, geometry=gpd.points_from_xy(df[lon_col], df[lat_col]), crs="EPSG:4326")
+        print("[INFO] Geometry dibangun dari kolom lat/lon.")
         return gdf
 
     coord_col = next(

services/reader_csv.py

@@ -47,19 +47,24 @@ def read_csv(path: str):
             delimiter = detect_delimiter(path)
             print(f"[INFO] Detected header line: {header_line + 1}, delimiter: '{delimiter}'")
 
-            df = pd.read_csv(path, header=header_line, sep=delimiter, encoding='utf-8', low_memory=False)
+            df = pd.read_csv(path, header=header_line, sep=delimiter, encoding='utf-8', low_memory=False, thousands=',')
 
         elif ext in ['.xlsx', '.xls']:
             # === Baca file Excel ===
             print(f"[INFO] Membaca file Excel: {os.path.basename(path)}")
-            df = pd.read_excel(path, header=0)  # default header baris pertama
+            pre_df = pd.read_excel(path, header=0, dtype=str)  # baca semua sebagai string
+            df = pre_df.copy()
+            for col in df.columns:
+                if df[col].str.replace(',', '', regex=False).str.match(r'^-?\d+(\.\d+)?$').any():
+                    df[col] = df[col].str.replace(',', '', regex=False)
+                    df[col] = pd.to_numeric(df[col], errors='ignore')
 
         else:
             raise ValueError("Format file tidak dikenali (hanya .csv, .txt, .xlsx, .xls)")
 
     except Exception as e:
         print(f"[WARN] Gagal membaca file ({e}), fallback ke default")
-        df = pd.read_csv(path, encoding='utf-8', low_memory=False)
+        df = pd.read_csv(path, encoding='utf-8', low_memory=False, thousands=',')
 
     # Bersihkan kolom dan baris kosong
     df = df.loc[:, ~df.columns.astype(str).str.contains('^Unnamed')]