RPG patches joined with municipalities and exploitants

PATBuilder.py

@@ -12,9 +12,9 @@ def load_MAJIC(majic_dir):
         patches.append(gpd.GeoDataFrame.from_file(f))
     return patches
 
-def build_initial_PAT(rpg_filename, pat_municipalities_dir,
-    pat_cultural_classes_filename, adminexpress_filename,
-    bdv_filename, majic_dir):
+def build_initial_PAT(rpg_parcelles_filename, rpg_ilots_filename, pat_cultural_classes_filename, majic_dir,
+    pat_municipalities_dir, bdv_filename, bdv_fixes_filename,
+    adminexpress_com_filename, adminexpress_epci_filename):
     '''
     This method build the initial patches set by consolidating the data sources
     used by the project:
@@ -27,8 +27,32 @@ def build_initial_PAT(rpg_filename, pat_municipalities_dir,
 
      Returns an object
     '''
-    # TODO
-    pass
+    municipalities = build_PAT_municipalities(pat_municipalities_dir, bdv_filename, bdv_fixes_filename,
+        adminexpress_com_filename, adminexpress_epci_filename)
+    municipalities = gpd.GeoDataFrame(municipalities[['INSEE_COM','CODE_EPCI','geometry','POPULATION','BV2012']])
+    rpg = gpd.GeoDataFrame.from_file(rpg_parcelles_filename, encoding='utf-8')[['ID_PARCEL','CODE_CULTU','geometry']]
+    # we will use the centroid for doing the sjoin with municipalities
+    rpg['patches'] = rpg.geometry
+    rpg.geometry = rpg.geometry.centroid
+    patches = gpd.sjoin(rpg, municipalities, op='intersects') # 'op' is useless because sjoin between points and polygons
+    # sjoin with rpg_ilots for retrieving id of "exploitant"
+    rpg_ilots = gpd.GeoDataFrame.from_file(rpg_ilots_filename, encoding='utf-8')[['id_ilot','id_expl','geometry']]
+    # restoring geometry instead of centroid
+    patches.geometry = patches['patches']
+    del patches['patches']
+    del patches['index_right']
+    patches_expl = gpd.sjoin(patches, rpg_ilots, how='left', op='intersects')
+    del patches_expl['index_right']
+    orphan_patches = patches_expl[patches_expl['id_ilot'].isnull()]
+    print('{} patches without "id_expl" ({:.4f} % of total area)'.format(
+        len(orphan_patches),
+        orphan_patches.geometry.area.sum() / patches_expl.geometry.area.sum()
+        ))
+    patches_expl = patches_expl.fillna(-1)
+    patches_expl['id_ilot'] = patches_expl['id_ilot'].astype(int)
+    patches_expl['id_expl'] = patches_expl['id_expl'].astype(int)
+    # TODO join with MAJIC
+    return [patches,rpg_ilots,patches_expl]
 
 def build_PAT_municipalities(pat_municipalities_dir, bdv_filename, bdv_fixes_filename,
         adminexpress_com_filename, adminexpress_epci_filename):
@@ -68,7 +92,11 @@ if __name__ == '__main__':
     resources={}
     for k,v in yaml.load(open('resources/INDEX.yml','r')).items():
         resources[v['variable']] = 'resources/'+v['file'] if 'file' in v else k
-    municipalities = build_PAT_municipalities(resources['pat_municipalities_dir'],
+    patches,rpg_ilots,patches_expl = build_initial_PAT(resources['rpg_parcelles_filename'], resources['rpg_ilots_filename'],
+        resources['pat_cultural_classes_filename'],
+        resources['majic_filename'],
+        resources['pat_municipalities_dir'],
         resources['bdv_filename'], resources['bdv_fixes'], resources['adminexpress_com_filename'],
         resources['adminexpress_epci_filename'])
+    patches_expl.to_file('output/patches', encoding='utf-8')
     gpd.GeoDataFrame(municipalities).to_file('output/municipalities', encoding='utf-8')

resources/INDEX.yml

@@ -29,8 +29,15 @@ ADMIN-EXPRESS-COG_1-1__SHP__FRA_2018-04-03-EPCI:
   url: https://wxs-telechargement.ign.fr/x02uy2aiwjo9bm8ce5plwqmr/telechargement/prepackage/ADMINEXPRESS-COG-PACK_2018-05-04$ADMIN-EXPRESS-COG_1-1__SHP__FRA_2018-04-03/file/ADMIN-EXPRESS-COG_1-1__SHP__FRA_2018-04-03.7z
   download-filename: ADMIN-EXPRESS-COG_1-1__SHP__FRA_2018-04-03.7z
 
+RPG_2-0__SHP_LAMB93_R84-2016_2016-01-01:
+  file: dl/RPG_2-0__SHP_LAMB93_R84-2016_2016-01-01/RPG/1_DONNEES_LIVRAISON_2016/RPG_2-0_SHP_LAMB93_R84-2016/PARCELLES_GRAPHIQUES.shp
+  variable: rpg_parcelles_filename
+  webpage: https://www.data.gouv.fr/fr/datasets/registre-parcellaire-graphique-rpg-contours-des-parcelles-et-ilots-culturaux-et-leur-groupe-de-cultures-majoritaire/
+  url: https://wxs-telechargement.ign.fr/0zf5kvnyfgyss0dk5dvvq9n7/telechargement/prepackage/RPG_REGION_PACK_DIFF_2016$RPG_2-0__SHP_LAMB93_R84-2016_2016-01-01/file/RPG_2-0__SHP_LAMB93_R84-2016_2016-01-01.7z
+  download-filename: RPG_2-0__SHP_LAMB93_R84-2016_2016-01-01.7z
+
 l_ilot_anonyme_groupe_dominant_s_r84_2016.shp:
-  variable: rpg_filename
+  variable: rpg_ilots_filename
   file: private/0f2f3338-d845-45a9-a124-384032d70219_1549881773_7572/l_ilot_anonyme_groupe_dominant_s_r84_2016.shp
   description: RPG with exploitants IDs
   webpage: https://catalogue.datara.gouv.fr/geosource/panierDownloadFrontalParametrage/0f2f3338-d845-45a9-a124-384032d70219