Merge pull request #22 from TaskarCenterAtUW/dev

Merge from Dev to Stage

Author: susrisha

.gitignore

@@ -1,5 +1,6 @@
 # Byte-compiled / optimized / DLL files
 __pycache__/
+cache/
 *.py[cod]
 *$py.class
 
@@ -160,4 +161,4 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 #.idea/
-.ve
+.ve

README.md

@@ -1,2 +1,22 @@
 # TDEI-python-osw-quality-metric
 Quality metric calculator for OSW record
+
+
+# Incoming message
+
+```json
+{
+    "datasetId":"",
+    "intersection_file":""
+}
+
+```
+
+# Outgoing message
+```json
+{
+    "datasetId":"",
+    "metrics_file":""
+}
+
+```

requirements.txt

@@ -1,4 +1,14 @@
 fastapi~=0.112.1
 python_ms_core~=0.0.22
 pydantic_settings~=2.4.0
-uvicorn~=0.30.6
+uvicorn~=0.30.6
+networkx==3.2.1
+geopandas==0.12.2
+osmnx==1.6.0
+dask==2024.5.2
+dask-geopandas==0.3.1
+geonetworkx==0.5.3
+shapely==2.0.1
+numpy==1.26.4
+pandas==1.3.4
+fiona==1.9.6

src/assets/messages/incoming.json

@@ -3,7 +3,8 @@
     "messageId": "message-id-from-msg",
     "data": {
       "jobId": "0b41ebc5-350c-42d3-90af-3af4ad3628fb",
-      "data_file": "https://tdeisamplestorage.blob.core.windows.net/osw/test_upload/Archive.zip",
-      "algorithms": "fixed"
+      "data_file": "https://tdeisamplestorage.blob.core.windows.net/osw/test/wenatchee.zip",
+      "algorithm": "fixed",
+      "sub_regions_file":""
     }
   }

src/calculators/__init__.py

@@ -1,2 +1,3 @@
 from .qm_calculator import QMCalculator
-from .qm_fixed_calculator import QMFixedCalculator, QMRandomCalculator
+from .qm_fixed_calculator import QMFixedCalculator
+from .qm_xn_lib_calculator import QMXNLibCalculator

src/calculators/qm_calculator.py

@@ -1,9 +1,15 @@
 from abc import ABC, abstractmethod
+from typing import NamedTuple
+
+class QualityMetricResult(NamedTuple):
+     success: bool
+     message: str
+     output_file: str
 
 class QMCalculator(ABC):
     @abstractmethod
-    def calculate_quality_metric(self, feature:dict) -> float:
+    def calculate_quality_metric(self) -> QualityMetricResult:
         pass
     @abstractmethod
-    def qm_metric_tag(self) -> str:
-        pass
+    def algorithm_name(self) -> str:
+        pass

src/calculators/qm_fixed_calculator.py

@@ -1,23 +1,41 @@
-from .qm_calculator import QMCalculator
+from src.calculators.qm_calculator import QMCalculator, QualityMetricResult
 import random
-
+import geopandas as gpd
+import sys
 
 class QMFixedCalculator(QMCalculator):
-    def __init__(self):
+    '''
+    Dummy quality metric calculator that assigns a random score to each edge in the input file
+    '''
+
+    def __init__(self, edges_file_path:str, output_file_path:str, polygon_file_path:str=None):
+        self.edges_file_path = edges_file_path
+        self.output_file_path = output_file_path
+        self.polygon_file_path = polygon_file_path
         pass
-    
-    def calculate_quality_metric(self, feature: dict) -> float:
-        return 10.5
-    def qm_metric_tag(self) -> str:
-        return "ext:qm:fixed"
 
+    def calculate_quality_metric(self):
+        gdf = gpd.read_file(self.edges_file_path)
+        gdf['fixed_score'] = random.randint(0, 100)
+        gdf.to_file(self.output_file_path)
+        return QualityMetricResult(success=True, message="QMFixedCalculator", output_file=self.output_file_path)
 
-class QMRandomCalculator(QMCalculator):
-    def __init__(self):
-        pass
-    
-    def calculate_quality_metric(self, feature: dict) -> float:
-        return random.uniform(30, 100)
+    def algorithm_name(self):
+        return "QMFixedCalculator"
+
+
+if __name__ == '__main__':
+    osw_edge_file_path = sys.argv[1]  # First argument: OSW edge file path
+    qm_file_path = sys.argv[2]        # Second argument: Quality metric output file path
 
-    def qm_metric_tag(self) -> str:
-        return "ext:qm:random"
+    # Check if the optional third argument (xn_polygon_path) is provided
+    if len(sys.argv) > 3:
+        xn_polygon_path = sys.argv[3]  # Third argument: Intersection polygon file path (optional)
+        qm_calculator = QMFixedCalculator(osw_edge_file_path, qm_file_path, xn_polygon_path)
+        print(qm_calculator.calculate_quality_metric())
+         
+    else:
+        # If the third argument is not provided, call without xn_polygon_path
+        qm_calculator = QMFixedCalculator(osw_edge_file_path, qm_file_path)
+        print(qm_calculator.calculate_quality_metric())
+         

src/calculators/qm_xn_lib_calculator.py

@@ -0,0 +1,179 @@
+from src.calculators.qm_calculator import QMCalculator, QualityMetricResult
+import geopandas as gpd
+import sys
+import warnings
+import networkx as nx
+import traceback
+import geonetworkx as gnx
+import osmnx as ox
+import dask_geopandas 
+from shapely import Point, LineString, MultiLineString, Polygon, MultiPolygon
+from shapely.ops import voronoi_diagram
+import itertools
+import numpy as np
+import pandas as pd
+
+
+class QMXNLibCalculator(QMCalculator):
+    def __init__(self, edges_file_path:str, output_file_path:str, polygon_file_path:str=None):
+        """
+        Initializes the QMXNLibCalculator class.
+
+        Args:
+            edges_file_path (str): Path to the file containing the OSW edge data.
+            output_file_path (str): Path to where the output quality metric file will be saved.
+            polygon_file_path (str, optional): Path to the intersection polygon file. If not provided, will use the polygon computed from the convex hull of OSW edge data. Defaults to None.
+        """
+        self.edges_file_path = edges_file_path
+        self.output_file_path = output_file_path
+        self.polygon_file_path = polygon_file_path
+        warnings.filterwarnings("ignore")
+        self.default_projection = 'epsg:26910'
+        self.output_projection = 'epsg:4326'
+        self.precision = 1e-5
+
+    def add_edges_from_linestring(self, graph, linestring, edge_attrs):
+        points = list(linestring.coords)
+        for start, end in zip(points[:-1], points[1:]):
+            graph.add_edge(start, end, **edge_attrs)
+    
+    def graph_from_gdf(self, gdf):
+        G = nx.Graph()
+        for index, row in gdf.iterrows():
+            geom = row.geometry
+            if isinstance(geom, LineString):
+                self.add_edges_from_linestring(G, geom, row.to_dict())
+            elif isinstance(geom, MultiLineString):
+                for linestring in geom.geoms:
+                    self.add_edges_from_linestring(G, linestring, row.to_dict())
+        return G
+    
+    def group_G_pts(self, G, poly):
+        P = poly
+        node_pts = [Point(x) for x in G.nodes()]
+        boundary = list(P.boundary.coords)
+        segments = [LineString([boundary[i], boundary[i + 1]]) for i in range(len(boundary) - 1)]
+        segment_point_map = {index: [] for index in range(len(segments))}
+        for point in node_pts:
+            for idx, segment in enumerate(segments):
+                if segment.distance(point) < self.precision:
+                    segment_point_map[idx].append((point.x, point.y))
+                    break
+        return segment_point_map
+    
+    def edges_are_connected(self, G, e1_pts, e2_pts):
+        for pt1 in e1_pts:
+            for pt2 in e2_pts:
+                if nx.has_path(G, pt1, pt2):
+                    return True
+        return False
+
+    def algorithm_name(self):
+        return "QMXNLibCalculator"
+
+    def tile_tra_score(self, G, polygon):
+        # assign each point to a polygon line
+        pts_line_map = self.group_G_pts(G, polygon)
+        boundary_nodes = [item for sublist in pts_line_map.values() for item in sublist]
+
+        # find all pair of edges
+        edge_pairs = list(itertools.combinations_with_replacement(pts_line_map.keys(), 2))
+
+        n_total = len(edge_pairs)
+        n_connected = 0
+        connected_pairs = list()
+        for pair in edge_pairs:
+            is_connected = self.edges_are_connected(G, pts_line_map[pair[0]], pts_line_map[pair[1]])
+            if is_connected:
+                n_connected += 1
+                connected_pairs.append(pair)
+        return n_total, n_connected, connected_pairs
+
+    def get_stats(self, polygon, G, gdf):
+        stats = {}
+        undirected_g = nx.Graph(G)
+
+        try:
+            n_total, n_connected, connected_pairs = self.tile_tra_score(G, polygon)
+            stats['tra_score'] = n_connected / n_total
+        except Exception as e:
+            print(f"Unexpected {e}, {type(e)} with polygon {polygon} when getting number of connected edge pairs")
+            #traceback.print_exc()
+            stats["tra_score"] = -1
+        return stats
+    
+    def get_measures_from_polygon(self, polygon, gdf):
+        if isinstance(polygon, MultiPolygon) and len(polygon.geoms)==1:
+            polygon = polygon.geoms[0]
+        # crop gdf to the polygon
+        cropped_gdf = gpd.clip(gdf, polygon)
+        
+        G = self.graph_from_gdf(cropped_gdf)
+        stats = self.get_stats(polygon, G, cropped_gdf)
+        return stats
+    
+    def qm_func(self, feature, gdf):
+        poly = feature.geometry
+        if (poly.geom_type == 'Polygon' or poly.geom_type == 'MultiPolygon'):
+            measures = self.get_measures_from_polygon(poly, gdf)
+            feature.loc['tra_score'] = measures['tra_score']
+            return feature
+        else:
+            return feature
+    
+    def create_voronoi_diagram(self, G_roads_simplified, bounds):
+        # first thin the nodes
+        gdf_roads_simplified = gnx.graph_edges_to_gdf(G_roads_simplified)
+        voronoi = voronoi_diagram(gdf_roads_simplified.boundary.unary_union, envelope=bounds)
+        voronoi_gdf = gpd.GeoDataFrame({'geometry':voronoi.geoms})
+        voronoi_gdf = voronoi_gdf.set_crs(gdf_roads_simplified.crs)
+        voronoi_gdf_clipped = gpd.clip(voronoi_gdf, bounds)
+        voronoi_gdf_clipped = voronoi_gdf_clipped.to_crs(self.default_projection)
+
+        return voronoi_gdf_clipped
+    
+    def calculate_quality_metric(self):
+        try:
+            gdf = gpd.read_file(self.edges_file_path)
+
+            if self.polygon_file_path:
+                 tile_gdf = gpd.read_file(self.polygon_file_path)
+            else:
+                unified_geom = gdf.unary_union
+                bounding_polygon = unified_geom.convex_hull
+                g_roads_simplified = ox.graph.graph_from_polygon(bounding_polygon, network_type='drive', simplify=True, retain_all=True)
+                tile_gdf = self.create_voronoi_diagram(g_roads_simplified, bounding_polygon)
+            
+            gdf = gdf.to_crs(self.default_projection)
+            tile_gdf = tile_gdf.to_crs(self.default_projection)
+            tile_gdf = tile_gdf[['geometry']]
+
+            df_dask = dask_geopandas.from_geopandas(tile_gdf, npartitions=64)
+
+            output = df_dask.apply(self.qm_func,axis=1, meta=[
+                ('geometry', 'geometry'),
+                ('tra_score', 'object')
+            ], gdf=gdf).compute(scheduler='multiprocessing')
+            output = output.to_crs(self.output_projection) # The output should be in WGS84 (epsg:4326)
+            output.to_file(self.output_file_path, driver='GeoJSON')
+            return QualityMetricResult(success=True, message='QMXNLibCalculator', output_file=self.output_file_path)
+
+        except Exception as e:
+            print(f"Error {e} occurred when calculating quality metric for data {self.edges_file_path}")
+            return QualityMetricResult(success=False, message=f'Error: {e}', output_file="")
+
+
+if __name__ == '__main__':
+    osw_edge_file_path = sys.argv[1]  # First argument: OSW edge file path
+    qm_file_path = sys.argv[2]        # Second argument: Quality metric output file path
+
+    # Check if the optional third argument (xn_polygon_path) is provided
+    if len(sys.argv) > 3:
+        xn_polygon_path = sys.argv[3]  # Third argument: Intersection polygon file path (optional)
+        qm_calculator = QMXNLibCalculator(osw_edge_file_path, qm_file_path, xn_polygon_path)
+        print(qm_calculator.calculate_quality_metric())
+         
+    else:
+        # If the third argument is not provided, call without xn_polygon_path
+        qm_calculator = QMXNLibCalculator(osw_edge_file_path, qm_file_path)
+        print(qm_calculator.calculate_quality_metric())