refactor: handling of string fields

Author: rabii-chaarani

m2l/main/vectorLayerWrapper.py

@@ -23,22 +23,68 @@
 
 
 
+# def qgsLayerToGeoDataFrame(layer) -> gpd.GeoDataFrame:
+#     if layer is None:
+#         return None
+#     features = layer.getFeatures()
+#     fields = layer.fields()
+#     data = {'geometry': []}
+#     for f in fields:
+#         data[f.name()] = []
+#     for feature in features:
+#         geom = feature.geometry()
+#         if geom.isEmpty():
+#             continue
+#         data['geometry'].append(geom)
+#         for f in fields:
+#             data[f.name()].append(feature[f.name()])
+#     return gpd.GeoDataFrame(data, crs=layer.crs().authid())
+
 def qgsLayerToGeoDataFrame(layer) -> gpd.GeoDataFrame:
+    """
+    Convert a QgsVectorLayer to a GeoDataFrame with:
+      - Shapely geometries
+      - Pandas nullable string dtype for QGIS string fields
+    """
     if layer is None:
         return None
-    features = layer.getFeatures()
+
     fields = layer.fields()
-    data = {'geometry': []}
-    for f in fields:
-        data[f.name()] = []
-    for feature in features:
-        geom = feature.geometry()
+    string_field_names = {
+        f.name() for f in fields
+        if f.type() in (QVariant.String,)  # extend here if you use other text types
+    }
+
+    rows = []
+    for feat in layer.getFeatures():
+        geom = feat.geometry()
         if geom.isEmpty():
             continue
-        data['geometry'].append(geom)
+
+        row = {"geometry": wkb_loads(geom.asWkb())}
         for f in fields:
-            data[f.name()].append(feature[f.name()])
-    return gpd.GeoDataFrame(data, crs=layer.crs().authid())
+            val = feat[f.name()]
+            # Normalize None in string cols to pandas.NA so StringDtype works cleanly
+            if f.name() in string_field_names:
+                row[f.name()] = pd.NA if val is None else str(val)
+            else:
+                row[f.name()] = val
+        rows.append(row)
+
+    if not rows:
+        # Empty GeoDataFrame with correct schema & crs
+        gdf = gpd.GeoDataFrame(columns=["geometry"] + [f.name() for f in fields],
+                               geometry="geometry",
+                               crs=layer.crs().authid())
+    else:
+        gdf = gpd.GeoDataFrame(rows, geometry="geometry", crs=layer.crs().authid())
+
+    # Enforce pandas' nullable string dtype on QGIS string fields
+    for name in string_field_names:
+        if name in gdf.columns:
+            gdf[name] = gdf[name].astype("string")
+
+    return gdf
 
 
 def qgsLayerToDataFrame(layer, dtm) -> pd.DataFrame: