assign_regions added

src/griml/metadata/__init__.py

@@ -1,5 +1,6 @@
 from griml.metadata.assign_certainty import *
 from griml.metadata.assign_id import *
 from griml.metadata.assign_names import *
+from griml.metadata.assign_regions import *
 from griml.metadata.assign_sources import *
 from griml.metadata.add_metadata import *

src/griml/metadata/add_metadata.py

@@ -1,30 +1,17 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 
-from griml.metadata import assign_id, assign_sources, assign_certainty, \
-    assign_names
 from griml.load import load
+from griml.metadata import assign_id, assign_sources, assign_certainty, \
+    assign_names, assign_regions
 import geopandas as gpd
 
-def add_metadata(iml_file, names_file, outfile=None):
-    '''Add metadata to collection of features
-    
-    Parameters
-    ----------
-    iml_file : str, geopandas.dataframe.DataFrame
-        Filepath or geopandas.dataframe.DataFrame object to assign metadata to
-
-    Returns
-    -------
-    iml: geopandas.dataframe.GeoDataFrame
-        Geodataframe with metadata
-    '''
-
-    # Load files
-    iml = load(iml_file)
-    names = load(names_file)
+def add_metadata(iml, names, regions, outfile=None):
+
+    iml = load(iml)
+    names = load(names)
+    regions = load(regions)
 
-    # Perform metadata steps
     print('Assigning ID...')
     iml = assign_id(iml)
 
@@ -35,24 +22,27 @@ def add_metadata(iml_file, names_file, outfile=None):
     n = ['S1','S2','ARCTICDEM']
     scores = [0.298, 0.398, 0.304]
     iml = assign_certainty(iml, n, scores)
+
+    print('Assigning regions...')
+    iml = assign_regions(iml, regions)
 
     print('Assigning placenames...')
     iml = assign_names(iml, names)
 
-    # Save to file if given
-    if outfile is not None:    
+    if outfile:    
         print('Saving file...')
         iml.to_file(outfile)
-        print('Saved to '+str(outfile))
-
-    return iml
+        print('Saved to '+str(outfile)+"_metadata.shp")
 
 
 if __name__ == "__main__": 
-    infile1 = '../test/test_merge_2.shp'            
-    infile2 = '../test/test_placenames.shp'              
-    add_metadata(infile1, infile2)
+    # indir = "/home/pho/python_workspace/GrIML/other/iml_2017/merged_vectors/griml_2017_inventory.shp"
+    indir = "/home/pho/python_workspace/GrIML/other/iml_2017/inspected_vectors/lakes_all-0000000000-0000037888_filtered.shp"
+
+
+    infile_names = '/home/pho/python_workspace/GrIML/other/datasets/placenames/oqaasileriffik_placenames.shp'
+    infile_regions = '/home/pho/python_workspace/GrIML/other/datasets/drainage_basins/greenland_basins_polarstereo.shp'
+
 
-    iml = gpd.read_file(infile1) 
-    names = gpd.read_file(infile2)
-    add_metadata(iml, names)
+    outfile = "/home/pho/python_workspace/GrIML/other/iml_2017/metadata_vectors/griml_2017_inventory_final.shp"
+    add_metadata(indir, infile_names, infile_regions, outfile)

src/griml/metadata/assign_names.py

@@ -1,7 +1,18 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 
-def assign_names(gdf, gdf_names, distance=500.0):
+import itertools
+from operator import itemgetter
+
+import geopandas as gpd
+import numpy as np
+import pandas as pd
+
+from scipy.spatial import cKDTree
+from shapely.geometry import Point, LineString, Polygon
+from griml.load import load
+
+def assign_names(gdf, gdf_names):
     '''Assign placenames to geodataframe geometries based on names in another 
     geodataframe point geometries
 
@@ -19,62 +30,41 @@ def assign_names(gdf, gdf_names, distance=500.0):
     gdf : pandas.GeoDataFrame
         Vectors with assigned IDs
     '''  
-    placenames = _compile_names(gdf_names)
+
+    # Load geodataframes
+    gdf1 = load(gdf)
+    gdf2 = load(gdf_names)
+
+    # Compile placenames into new dataframe
+    names = _compile_names(gdf2)
+    placenames = gpd.GeoDataFrame({"geometry": list(gdf2['geometry']),
+                                   "placename": names})
 
-    lakename=[]     
-    for i,v in gdf.iterrows():  
-
-        geom = v['geometry']
-
-        polynames=[] 
-        for pt in range(len(placenames)):
-            if geom.contains(gdf_names['geometry'][pt]) == True:
-                polynames.append(placenames[pt])  
-
-        if len(polynames)==0:
-            for pt in range(len(placenames)):  
-                if gdf_names['geometry'][pt].distance(geom) < distance: 
-                    polynames.append(placenames[pt]) 
-            lakename.append(polynames)
-
-        elif len(polynames)==1:  
-            lakename.append(polynames)    
-
-        else: 
-            out=[]          
-            for p in polynames:  
-                out.append(p) 
-            lakename.append(out) 
-
-    lakeid = gdf['unique_id'].tolist()      
-    lakename2 = []     
+    # Assign names based on proximity
+    a = _get_nearest_point(gdf1, placenames)
+    return a
+
+
+def _get_nearest_point(gdA, gdB, distance=500.0):
+    '''Return properties of nearest point in Y to geometry in X'''
+    nA = np.array(list(gdA.geometry.centroid.apply(lambda x: (x.x, x.y))))
+    nB = np.array(list(gdB.geometry.apply(lambda x: (x.x, x.y))))
+
+    btree = cKDTree(nB)
+    dist, idx = btree.query(nA, k=1)
+    gdB_nearest = gdB.iloc[idx].drop(columns="geometry").reset_index(drop=True)
+    gdf = pd.concat(
+        [
+            gdA.reset_index(drop=True),
+            gdB_nearest,
+            pd.Series(dist, name='dist')
+        ], 
+        axis=1)
 
-    for x in gdf.index:        
-        indx = _get_indices(lakeid, x)                                         
-        findname=[]
-        for l in indx:
-            if len(lakename[l])!=0: 
-                findname.append(lakename[l])
-
-        for i in range(len(indx)):  
-            if len(findname)==0:
-                lakename2.append('')
-
-            else:                                           
-                unique = set(findname[0])  
-                unique = list(unique)  
-
-                if len(unique)==1:
-                    lakename2.append(findname[0][0]) 
-
-                else:
-                    out2 = ', '
-                    out2 = out2.join(unique) 
-                    lakename2.append(out2) 
-    gdf['placename'] = lakename2
+    gdf.loc[gdf['dist']>=distance, 'placename'] = 'Unknown'
+    gdf = gdf.drop(columns=['dist'])    
     return gdf
 
-
 def _get_indices(mylist, value):
     '''Get indices for value in list'''
     return[i for i, x in enumerate(mylist) if x==value]
@@ -95,3 +85,14 @@ def _compile_names(gdf):
                 else:
                     placenames.append(None)
     return placenames 
+
+
+if __name__ == "__main__": 
+
+    # Define file attributes
+    infile = '/home/pho/Desktop/python_workspace/GrIML/src/griml/test/test_merge_1.shp'
+    names = '/home/pho/Desktop/python_workspace/GrIML/other/datasets/placenames/oqaasileriffik_placenames_polarstereo.shp'
+
+
+    # Perform conversion
+    a = assign_names(infile, names)