From d6aa0dfd1b5cc4e35bbfc2b9c316e82705afba2f Mon Sep 17 00:00:00 2001
From: Hasan Nasrallah <50121708+hasan-nn@users.noreply.github.com>
Date: Wed, 13 Dec 2023 16:54:57 +0200
Subject: [PATCH] add conversion.py #correction

script to convert instances mask to geodataframe
---
 data_processing/conversion.py | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100644 data_processing/conversion.py

diff --git a/data_processing/conversion.py b/data_processing/conversion.py
new file mode 100644
index 0000000..59c0339
--- /dev/null
+++ b/data_processing/conversion.py
@@ -0,0 +1,34 @@
+import rasterio as rio 
+import geopandas as gpd
+from rasterio.features import shapes
+
+def instance_mask_to_gdf(
+        instance_mask,
+        transform = None,
+        crs=None
+        ):
+    """
+    Input:
+        - instance_mask : np.array of shape (H,W), where each instance is labeled by a unique id/number
+        - transform : geospatial transform of the raster - default is None
+        - crs : crs of the raster - default is None
+    Output:
+        - GeoDataFrame of the shapes projected to the specified crs using the transform
+    """
+
+    #transform should be Identity if None is provided
+    transform = rio.transform.IDENTITY if transform is None else transform
+    
+    all_shapes = shapes(instance_mask,mask=None,transform=transform)
+    data = [
+            {'properties' : {'id' : v} , 'geometry' : s} for i,(s,v) in enumerate(all_shapes) if v!=0
+    ]
+    
+    if len(data) == 0:
+        ##return empty dataframe
+        return gpd.GeoDataFrame(columns=['id','geometry'], geometry='geometry',crs=crs)
+    
+    gdf = gpd.GeoDataFrame.from_features(data,crs=crs)
+    gdf = gdf.dissolve(by='id')
+    
+    return gdf
\ No newline at end of file