adds lazy mode

Author: fredshone

README.md

@@ -55,10 +55,12 @@ pytest
 
 ## Quick Start
 
-Extract `home`, `work`, `education`, `shop` and various other activity locations ("facilities") for the Isle of Man, using the following command (the path is given from OSMOX project root):
+Extract `home`, `work`, `education`, `shop` and various other activity locations ("facilities") for the Isle of Man, using the following steps (paths is given from OSMOX project root):
+
+First download `isle-of-man-latest.osm.pbf` from [geobabrik](https://download.geofabrik.de/europe/isle-of-man.html) and place in an `example` directory.
 
 ```{sh}
-osmox run configs/example.json example_data/isle-of-man.osm example -crs "epsg:27700"
+osmox run configs/example.json example/isle-of-man.osm example -crs "epsg:27700"
 ```
 
 After about 30 seconds, you should find the outputs in geojson format in the specified `example` directory. The geojson file contains locations for the extracted facilities, and each facility includes a number of features with coordinates given in WGS-84 (EPSG:4326) coordinate reference system (CRS), so that they can be quickly inspected via [kepler](https://kepler.gl) or equivalent.
@@ -91,8 +93,6 @@ After about 30 seconds, you should find the outputs in geojson format in the spe
         ...
 ```
 
-
-
 ![isle of man floor areas](./readme_fixtures/floor-areas.png)
 *^ Isle of Man facility `floor_area` feature. Approximated based on polygon areas and floor labels or sensible defaults.*
 
@@ -106,9 +106,12 @@ After about 30 seconds, you should find the outputs in geojson format in the spe
 ```{sh}
 osmox run --help
 
+Usage: osmox run [OPTIONS] CONFIG_PATH INPUT_PATH OUTPUT_PATH
+
 Options:
   -crs, --crs TEXT  crs string eg (default): 'epsg:27700' (UK grid)
   -s, --single_use  split multi-activity facilities into multiple single-activity facilities
+  -l, --lazy        if filtered object already has a label, do not search for more (supresses multi-use but is faster)
   --help            Show this message and exit.
 ```
 

osmox/build.py

@@ -117,22 +117,29 @@ def __str__(self):
             """
 
     def add_tags(self, osm_objects):
+        print(f"adding tag to {self.idx}")
+        print(f"existing = {self.activity_tags}")
         for o in osm_objects:
             if o.activity_tags:
                 self.activity_tags.extend(o.activity_tags)
+        print(f"updated = {self.activity_tags}")
 
     def apply_default_tag(self, tag):
         self.activity_tags = [OSMTag(tag[0], tag[1])]
 
     def assign_points(self, points):
+        print("assigning points")
         snaps = [c for c in points.intersection(self.geom.bounds)]
+        print(f"snaps = {snaps}")
         if snaps:
             self.add_tags(snaps)
             return True
 
     def assign_areas(self, areas):
+        print("assigning areas")
         snaps = [c for c in areas.intersection(self.geom.bounds)]
         snaps = [c for c in snaps if c.geom.contains(self.geom.centroid)]
+        print(f"snaps = {snaps}")
         if snaps:
             self.add_tags(snaps)
             return True
@@ -192,12 +199,20 @@ class ObjectHandler(osmium.SimpleHandler):
     wkbfab = osmium.geom.WKBFactory()
     logger = logging.getLogger(__name__)
 
-    def __init__(self, config, crs='epsg:27700', from_crs='epsg:4326', level=logging.DEBUG):
+    def __init__(
+        self,
+        config,
+        crs='epsg:27700',
+        from_crs='epsg:4326',
+        lazy=False,
+        level=logging.DEBUG
+        ):
 
         super().__init__()
         logging.basicConfig(level=level)
         self.cnfg = config
         self.crs = crs
+        self.lazy = lazy
         self.filter = self.cnfg["filter"]
         self.object_features = self.cnfg["object_features"]
         self.default_tags = self.cnfg["default_tags"]
@@ -285,11 +300,50 @@ def area(self, a):
         elif activity_tags:
             self.add_area(idx=a.id, activity_tags=activity_tags, geom=self.fab_area(a))
 
-    """
-    Assign unknown tags to buildings spatially.
-    """
 
     def assign_tags(self):
+        """
+        Assign unknown tags to buildings spatially.
+        """
+        if not self.lazy:
+            self.assign_tags_full()
+        else:
+            self.assign_tags_lazy()
+
+    def assign_tags_full(self):
+        """
+        Assign unknown tags to buildings spatially.
+        """
+
+        for obj in helpers.progressBar(self.objects, prefix='Progress:', suffix='Complete', length=50):
+            print(obj.idx, obj.activity_tags)
+
+            if obj.activity_tags:
+                print("existing:", obj.idx, obj.activity_tags)
+                # if an onject already has activity tags, continue
+                self.log["existing"] += 1
+
+            if obj.assign_points(self.points):
+                print("pre point assignment:", obj.idx, obj.activity_tags)
+                # else try to assign activity tags based on contained point objects
+                self.log["points"] += 1
+                continue
+
+            if obj.assign_areas(self.areas):
+                print("pre area assignment:", obj.idx, obj.activity_tags)
+                # else try to assign activity tags based on containing area objects
+                self.log["areas"] += 1
+                continue
+
+            if self.default_tags and not obj.activity_tags:
+                print("defaults need to be assigned...", obj.idx, obj.activity_tags)
+                # otherwise apply defaults if set
+                self.log["defaults"] += 1
+                for a in self.default_tags:
+                    obj.apply_default_tag(a)
+
+    def assign_tags_lazy(self):
+        """Assign tags if filtered object does not already have useful tags."""
 
         for obj in helpers.progressBar(self.objects, prefix='Progress:', suffix='Complete', length=50):
 
@@ -401,11 +455,19 @@ def extract_targets(self, target_act):
         return MultiPoint(targets)
 
     def geodataframe(self, single_use=False):
+        print("+++++++++++++++++++")
+        # print(self.objects)
+        print(single_use)
+
         if single_use:
+            print("SINGLE USE")
 
             df = pd.DataFrame(
                 (summary for o in self.objects for summary in o.single_activity_summaries())
             )
+            print("====")
+            print(df)
+            print("====")
             return gp.GeoDataFrame(df, geometry='geometry', crs=self.crs)
 
         df = pd.DataFrame(

osmox/main.py

@@ -1,3 +1,4 @@
+from linecache import lazycache
 import logging
 import os
 
@@ -45,17 +46,31 @@ def validate(config_path):
               help="crs string eg (default): 'epsg:27700' (UK grid)")
 @click.option("-s", "--single_use", is_flag=True,
               help="split multi-activity facilities into multiple single-activity facilities")
-def run(config_path, input_path, output_path, crs, single_use):
+@click.option("-l", "--lazy", is_flag=True,
+              help="if filtered object already has a label, do not search for more (supresses multi-use)")
+def run(config_path, input_path, output_path, crs, single_use, lazy):
     logger.info(f" Loading config from {config_path}")
     cnfg = config.load(config_path)
     config.validate_activity_config(cnfg)
 
     if not os.path.exists(output_path):
         logger.info(f'Creating output directory: {output_path}')
         os.mkdir(output_path)
-
-    handler = build.ObjectHandler(cnfg, crs)
-    logger.info(f" Filtering all objects found in {input_path}.")
+    
+    logger.info(f"Creating handler with crs: {crs}.")
+    print(crs)
+    print(lazy)
+    if single_use:
+        logger.info(f"Handler is single-use, different activities will get unique locations. even if these are duplicated.")
+    if lazy:
+        logger.info(f"Handler will be using lazy assignment, this may suppress some multi-use.")
+
+    handler = build.ObjectHandler(
+        config=cnfg,
+        crs=crs,
+        lazy=lazy
+        )
+    logger.info(f" Filtering all objects found in {input_path}. This may take a long while.")
     handler.apply_file(str(input_path), locations=True, idx='flex_mem')
     logger.info(f" Found {len(handler.objects)} buildings.")
     logger.info(f" Found {len(handler.points)} nodes with valid tags.")

tests/fixtures/test_config_leisure.json

@@ -3,7 +3,7 @@
         "leisure": ["pitch", "park", "playground"]
     },
 
-    "object_features": ["units", "transit_distance", "levels", "area", "floor_area"],
+    "object_features": ["units", "levels", "area", "floor_area"],
 
     "default_tags": [["building", "residential"]],
 

tests/test_activityhandler.py

@@ -126,7 +126,7 @@ def test_leisure_config():
 
 @pytest.fixture()
 def leisureHandler(test_leisure_config):
-    return build.ObjectHandler(test_leisure_config, crs='epsg:4326')
+    return build.ObjectHandler(test_leisure_config, crs='epsg:4326', lazy=False)
 
 
 def test_leisure_handler(leisureHandler):