backend/src/sandbox/get_streets.py

@@ -7,6 +7,7 @@ from sklearn.decomposition import PCA
 import matplotlib.pyplot as plt
 from pydantic import BaseModel
 from OSMPythonTools.overpass import Overpass, overpassQueryBuilder
+from OSMPythonTools.cachingStrategy import CachingStrategy, JSON
 from math import sin, cos, sqrt, atan2, radians
 
 
@@ -177,7 +178,7 @@ def fit_lines(points, labels):
         else :
             loc = ShoppingLocation(
                 type='street',
-                centroid=tuple(centroid),
+                centroid=tuple((centroid[1], centroid[0])),
                 importance = len(cluster_points),
                 start=start_point,
                 end=end_point
@@ -199,24 +200,13 @@ def create_landmark(shopping_location: ShoppingLocation):
 
     # Define the bounding box for a given radius around the coordinates
     lat, lon = shopping_location.centroid
-    bbox = ("around:500", str(lat), str(lon))
+    bbox = ("around:1000", str(lat), str(lon))
     
     overpass = Overpass()
+    # CachingStrategy.use(JSON, cacheDir=OSM_CACHE_DIR)
 
     # Query neighborhoods and shopping malls
-    query = overpassQueryBuilder(
+    selectors = ['"place"~"^(suburb|neighborhood|city_block)$"', '"shop"="mall"']
-        bbox = bbox,
-        elementType = ['node'],
-        selector = ['"place"="suburb"'],
-        includeCenter = True,
-        out = 'body'
-    )
-
-    try:
-        result = overpass.query(query)
-        print(f'query OK with {len(result.elements())} elements')
-    except Exception as e:
-        raise Exception("query unsuccessful")
 
     min_dist = float('inf')
     new_name = 'Shopping Area'
@@ -224,15 +214,39 @@ def create_landmark(shopping_location: ShoppingLocation):
     osm_id = 0
     osm_type = 'node'
 
+    for sel in selectors : 
+        query = overpassQueryBuilder(
+            bbox = bbox,
+            elementType = ['node', 'way', 'relation'],
+            selector = sel,
+            includeCenter = True,
+            out = 'body'
+        )
+
+        try:
+            result = overpass.query(query)
+            # print(f'query OK with {len(result.elements())} elements')
+        except Exception as e:
+            raise Exception("query unsuccessful")
+
         for elem in result.elements():
+
+            location = (elem.lat(), elem.lon())
+
+            if location[0] is None : 
                 location = (elem.centerLat(), elem.centerLon())
-        print(f"Distance : {get_distance(shopping_location.centroid, location)}")
+                if location[0] is None : 
-        if get_distance(shopping_location.centroid, location) < min_dist :
+                    continue
+
+            # print(f"Distance : {get_distance(shopping_location.centroid, location)}")
+            d = get_distance(shopping_location.centroid, location)
+            if  d < min_dist :
+                min_dist = d
                 new_name = elem.tag('name')
                 osm_type = elem.type()              # Add type: 'way' or 'relation'
                 osm_id = elem.id()                  # Add OSM id 
 
-            print("closer thing found")
+                # print("closer thing found")
 
                 # add english name if it exists
                 try :
@@ -253,7 +267,9 @@ def create_landmark(shopping_location: ShoppingLocation):
 
 
 # Extract points
-points = extract_points('lyon_data.json')
+points = extract_points('newyork_data.json')
+
+# print(len(points))
 
 ######## Create a figure with 1 row and 3 columns for side-by-side plots
 fig, axes = plt.subplots(1, 3, figsize=(15, 5))
@@ -280,6 +296,7 @@ axes[1].set_title('DBSCAN Clusters')
 axes[1].scatter(clustered_points[:, 0], clustered_points[:, 1], c=clustered_labels, cmap='rainbow', s=20)
 axes[1].scatter(noise_points[:, 0], noise_points[:, 1], c='blue', s=7, label='Noise')
 
+# Keep the 5 biggest clusters
 clustered_points, clustered_labels = filter_clusters(clustered_points, clustered_labels)
 
 # Fit lines
@@ -289,26 +306,26 @@ corners, locations = fit_lines(clustered_points, clustered_labels)
 
 ######## Plot clustered points in normal size and noise points separately
 axes[2].scatter(clustered_points[:, 0], clustered_points[:, 1], c=clustered_labels, cmap='rainbow', s=30)
-
+axes[2].set_title('PCA Fitted Lines on Clusters')
 
 # Create a list of Landmarks for the shopping things
 shopping_landmarks = []
 for loc in locations :
+    axes[2].scatter(loc.centroid[0], loc.centroid[1], color='lime', marker='x', s=200, linewidth=3)
     landmark = create_landmark(loc)
     shopping_landmarks.append(landmark)
-    print(f"{landmark.name} is a shopping area with a score of {landmark.attractiveness}")
+    
     
 
 ####### Plot the detected lines in the final plot #######
-for loc in locations:
+# for loc in locations:
-    if loc.type == 'street' :
+#     if loc.type == 'street' :
-        line_x = loc.start
+#         line_x = loc.start
-        line_y = loc.end
+#         line_y = loc.end
-        axes[2].plot(line_x, line_y, color='lime', linewidth=3)
+#         axes[2].plot(line_x, line_y, color='lime', linewidth=3)
-    else :
+#     else :
-        axes[2].scatter(loc.centroid[0], loc.centroid[1], color='None', edgecolors='lime', s=200, linewidth=3)
+
 
-axes[2].set_title('PCA Fitted Lines on Clusters')
 
 axes[0].set_xlim(xmin-0.01, xmax+0.01)
 axes[0].set_ylim(ymin-0.01, ymax+0.01)
@@ -319,5 +336,11 @@ axes[1].set_ylim(ymin-0.01, ymax+0.01)
 axes[2].set_xlim(xmin-0.01, xmax+0.01)
 axes[2].set_ylim(ymin-0.01, ymax+0.01)
 
-# plt.tight_layout()
+
-# plt.show()
+print("\n\n\n")
+for landmark in shopping_landmarks :
+    print(f"{landmark.name} is a shopping area with a score of {landmark.attractiveness}")
+
+
+plt.tight_layout()
+plt.show()