Add pre-commit config file; run linting & formatting

.pre-commit-config.yaml

@@ -0,0 +1,24 @@
+default_language_version:
+  python: python3
+
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.5.0
+    hooks:
+      - id: check-added-large-files
+        args: ["--maxkb=1000"]
+
+  - repo: https://github.com/psf/black
+    rev: 24.2.0
+    hooks:
+      - id: black
+
+  - repo: https://github.com/astral-sh/ruff-pre-commit  # https://beta.ruff.rs/docs/usage/#github-action
+    rev: v0.3.0
+    hooks:
+      - id: ruff
+        args: [ --fix, --exit-non-zero-on-fix ]
+
+ci:  # https://pre-commit.ci/
+  autofix_prs: false
+  autoupdate_schedule: monthly

src/osmox/build.py

@@ -5,21 +5,15 @@
 import osmium
 import pandas as pd
 import shapely.wkb as wkblib
-from pyproj import Transformer, CRS
+from pyproj import CRS, Transformer
 from shapely.geometry import MultiPoint
-from shapely.ops import transform, nearest_points
+from shapely.ops import nearest_points, transform
 
 from osmox import helpers
 
-OSMTag = namedtuple('OSMtag', 'key value')
-OSMObject = namedtuple('OSMobject', 'idx, activity_tags, geom')
-AVAILABLE_FEATURES = [
-    "area",
-    "levels",
-    "floor_area",
-    "units",
-    "transit_distance"
-]
+OSMTag = namedtuple("OSMtag", "key value")
+OSMObject = namedtuple("OSMobject", "idx, activity_tags, geom")
+AVAILABLE_FEATURES = ["area", "levels", "floor_area", "units", "transit_distance"]
 
 
 class Object:
@@ -57,7 +51,7 @@ class Object:
         "university": 3,
         "college": 3,
         "sports_hall": 1,
-        "stadium": 1
+        "stadium": 1,
     }
 
     def __init__(self, idx, osm_tags, activity_tags, geom) -> None:
@@ -82,12 +76,12 @@ def area(self):
         return int(self.geom.area)
 
     def levels(self):
-        if 'building:levels' in self.osm_tags:
-            levels = self.osm_tags['building:levels']
+        if "building:levels" in self.osm_tags:
+            levels = self.osm_tags["building:levels"]
             if levels.isnumeric():
                 return float(levels)  # todo ensure integer
-        if 'height' in self.osm_tags:
-            height = helpers.height_to_m(self.osm_tags['height'])
+        if "height" in self.osm_tags:
+            height = helpers.height_to_m(self.osm_tags["height"])
             if height:
                 return float(height / 4)
         if self.osm_tags.get("building"):
@@ -100,8 +94,8 @@ def floor_area(self):
         return self.area() * self.levels()
 
     def units(self):
-        if 'building:flats' in self.osm_tags:
-            flats = self.osm_tags['building:flats']
+        if "building:flats" in self.osm_tags:
+            flats = self.osm_tags["building:flats"]
             if flats.isnumeric():
                 return float(flats)  # todo ensure integer
         return 1
@@ -169,8 +163,8 @@ def summary(self):
         """
         fixed = {
             "id": self.idx,
-            "activities": ','.join(self.activities),
-            "geometry": self.geom.centroid
+            "activities": ",".join(self.activities),
+            "geometry": self.geom.centroid,
         }
         return {**fixed, **self.features}
 
@@ -179,11 +173,7 @@ def single_activity_summaries(self):
         Yield (dict) summaries for each each activity of an object.
         """
         for act in self.activities:
-            fixed = {
-                "id": self.idx,
-                "activity": act,
-                "geometry": self.geom.centroid
-            }
+            fixed = {"id": self.idx, "activity": act, "geometry": self.geom.centroid}
             yield {**fixed, **self.features}
 
 
@@ -193,12 +183,7 @@ class ObjectHandler(osmium.SimpleHandler):
     logger = logging.getLogger(__name__)
 
     def __init__(
-        self,
-        config,
-        crs='epsg:27700',
-        from_crs='epsg:4326',
-        lazy=False,
-        level=logging.DEBUG
+        self, config, crs="epsg:27700", from_crs="epsg:4326", lazy=False, level=logging.DEBUG
     ):
 
         super().__init__()
@@ -216,12 +201,7 @@ def __init__(
         self.points = helpers.AutoTree()
         self.areas = helpers.AutoTree()
 
-        self.log = {
-            "existing": 0,
-            "points": 0,
-            "areas": 0,
-            "defaults": 0
-            }
+        self.log = {"existing": 0, "points": 0, "areas": 0, "defaults": 0}
 
     """
     On handler.apply_file() method; parse through all nodes and areas:
@@ -243,14 +223,19 @@ def get_filtered_tags(self, tags):
             found = []
             for osm_key, osm_val in tags.items():
                 if osm_key in self.activity_config:
-                    if osm_val in self.activity_config[osm_key] or self.activity_config[osm_key] == "*":
+                    if (
+                        osm_val in self.activity_config[osm_key]
+                        or self.activity_config[osm_key] == "*"
+                    ):
                         found.append(OSMTag(key=osm_key, value=osm_val))
             return found
 
     def add_object(self, idx, activity_tags, osm_tags, geom):
         if geom:
             geom = transform(self.transformer.transform, geom)
-            self.objects.auto_insert(Object(idx=idx, osm_tags=osm_tags, activity_tags=activity_tags, geom=geom))
+            self.objects.auto_insert(
+                Object(idx=idx, osm_tags=osm_tags, activity_tags=activity_tags, geom=geom)
+            )
 
     def add_point(self, idx, activity_tags, geom):
         if geom:
@@ -267,29 +252,33 @@ def fab_point(self, n):
             wkb = self.wkbfab.create_point(n)
             return wkblib.loads(wkb, hex=True)
         except RuntimeError:
-            self.logger.warning(f' RuntimeError encountered for point: {n}')
+            self.logger.warning(f" RuntimeError encountered for point: {n}")
             return None
 
     def fab_area(self, a):
         try:
             wkb = self.wkbfab.create_multipolygon(a)
             return wkblib.loads(wkb, hex=True)
         except RuntimeError:
-            self.logger.warning(f' RuntimeError encountered for polygon: {a}')
+            self.logger.warning(f" RuntimeError encountered for polygon: {a}")
             return None
 
     def node(self, n):
         activity_tags = self.get_filtered_tags(n.tags)
         # todo consider renaming activiity tags to filtered or selected tags
         if self.selects(n.tags):
-            self.add_object(idx=n.id, osm_tags=n.tags, activity_tags=activity_tags, geom=self.fab_point(n))
+            self.add_object(
+                idx=n.id, osm_tags=n.tags, activity_tags=activity_tags, geom=self.fab_point(n)
+            )
         elif activity_tags:
             self.add_point(idx=n.id, activity_tags=activity_tags, geom=self.fab_point(n))
 
     def area(self, a):
         activity_tags = self.get_filtered_tags(a.tags)
         if self.selects(a.tags):
-            self.add_object(idx=a.id, osm_tags=a.tags, activity_tags=activity_tags, geom=self.fab_area(a))
+            self.add_object(
+                idx=a.id, osm_tags=a.tags, activity_tags=activity_tags, geom=self.fab_area(a)
+            )
         elif activity_tags:
             self.add_area(idx=a.id, activity_tags=activity_tags, geom=self.fab_area(a))
 
@@ -307,7 +296,9 @@ def assign_tags_full(self):
         Assign unknown tags to buildings spatially.
         """
 
-        for obj in helpers.progressBar(self.objects, prefix='Progress:', suffix='Complete', length=50):
+        for obj in helpers.progressBar(
+            self.objects, prefix="Progress:", suffix="Complete", length=50
+        ):
 
             if obj.activity_tags:
                 # if an onject already has activity tags, continue
@@ -332,7 +323,9 @@ def assign_tags_full(self):
     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):
+        for obj in helpers.progressBar(
+            self.objects, prefix="Progress:", suffix="Complete", length=50
+        ):
 
             if obj.activity_tags:
                 # if an onject already has activity tags, continue
@@ -356,16 +349,18 @@ def assign_tags_lazy(self):
                     obj.apply_default_tag(a)
 
     def assign_activities(self):
-        for obj in helpers.progressBar(self.objects, prefix='Progress:', suffix='Complete', length=50):
+        for obj in helpers.progressBar(
+            self.objects, prefix="Progress:", suffix="Complete", length=50
+        ):
             obj.assign_activities(self.activity_config)
 
     def fill_missing_activities(
-            self,
-            area_tags=("landuse", "residential"),
-            required_acts="home",
-            new_tags=("building", "house"),
-            size=(10, 10),
-            spacing=(25, 25)
+        self,
+        area_tags=("landuse", "residential"),
+        required_acts="home",
+        new_tags=("building", "house"),
+        size=(10, 10),
+        spacing=(25, 25),
     ):
         """
         Fill "empty" areas with new objects. Empty areas are defined as areas with the select_tags but
@@ -389,7 +384,9 @@ def fill_missing_activities(
         new_osm_tags = [OSMTag(key=k, value=v) for k, v in area_tags]
         new_tags = [OSMTag(key=k, value=v) for k, v in new_tags]
 
-        for area in helpers.progressBar(self.areas, prefix='Progress:', suffix='Complete', length=50):
+        for area in helpers.progressBar(
+            self.areas, prefix="Progress:", suffix="Complete", length=50
+        ):
 
             if not helpers.tag_match(a=area_tags, b=area.activity_tags):
                 continue
@@ -402,7 +399,9 @@ def fill_missing_activities(
             # sample a grid
             points = helpers.area_grid(area=area.geom, spacing=spacing)
             for point in points:  # add objects built from grid
-                self.objects.auto_insert(helpers.fill_object(i, point, size, new_osm_tags, new_tags, required_acts))
+                self.objects.auto_insert(
+                    helpers.fill_object(i, point, size, new_osm_tags, new_tags, required_acts)
+                )
                 i += 1
 
         return empty_zones, i
@@ -420,15 +419,19 @@ def add_features(self):
         """
         ["units", "floors", "area", "floor_area"]
         """
-        for obj in helpers.progressBar(self.objects, prefix='Progress:', suffix='Complete', length=50):
+        for obj in helpers.progressBar(
+            self.objects, prefix="Progress:", suffix="Complete", length=50
+        ):
             obj.add_features(self.object_features)
 
     def assign_nearest_distance(self, target_act):
         """
         For each facility, calculate euclidean distance to targets of given activity type.
         """
         targets = self.extract_targets(target_act)
-        for obj in helpers.progressBar(self.objects, prefix='Progress:', suffix='Complete', length=50):
+        for obj in helpers.progressBar(
+            self.objects, prefix="Progress:", suffix="Complete", length=50
+        ):
             obj.get_closest_distance(targets, target_act)
 
     def extract_targets(self, target_act):
@@ -447,12 +450,10 @@ def geodataframe(self, single_use=False):
             df = pd.DataFrame(
                 (summary for o in self.objects for summary in o.single_activity_summaries())
             )
-            return gp.GeoDataFrame(df, geometry='geometry', crs=self.crs)
+            return gp.GeoDataFrame(df, geometry="geometry", crs=self.crs)
 
-        df = pd.DataFrame(
-            (o.summary() for o in self.objects)
-        )
-        return gp.GeoDataFrame(df, geometry='geometry', crs=self.crs)
+        df = pd.DataFrame((o.summary() for o in self.objects))
+        return gp.GeoDataFrame(df, geometry="geometry", crs=self.crs)
 
     # def extract(self):
     #     df = pd.DataFrame.from_records(

src/osmox/cli.py

@@ -3,9 +3,8 @@
 
 import click
 
-from osmox import config, build
-from osmox.helpers import PathPath
-from osmox.helpers import path_leaf
+from osmox import build, config
+from osmox.helpers import PathPath, path_leaf
 
 default_config_path = os.path.abspath(
     os.path.join(os.path.dirname(__file__), "../configs/config.json")
@@ -38,15 +37,28 @@ def validate(config_path):
 
 
 @cli.command()
-@click.argument('config_path', type=PathPath(exists=True), nargs=1, required=True)
-@click.argument('input_path', type=PathPath(exists=True), nargs=1, required=True)
-@click.argument('output_name', nargs=1, required=True)
-@click.option("-crs", "--crs", type=str, default="epsg:27700",
-              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")
-@click.option("-l", "--lazy", is_flag=True,
-              help="if filtered object already has a label, do not search for more (supresses multi-use)")
+@click.argument("config_path", type=PathPath(exists=True), nargs=1, required=True)
+@click.argument("input_path", type=PathPath(exists=True), nargs=1, required=True)
+@click.argument("output_name", nargs=1, required=True)
+@click.option(
+    "-crs",
+    "--crs",
+    type=str,
+    default="epsg:27700",
+    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",
+)
+@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_name, crs, single_use, lazy):
     logger.info(f" Loading config from {config_path}")
     cnfg = config.load(config_path)
@@ -58,13 +70,9 @@ def run(config_path, input_path, output_name, crs, single_use, lazy):
     if lazy:
         logger.info("Handler will be using lazy assignment, this may suppress some multi-use.")
 
-    handler = build.ObjectHandler(
-        config=cnfg,
-        crs=crs,
-        lazy=lazy
-        )
+    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')
+    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.")
     logger.info(f" Found {len(handler.areas)} areas with valid tags.")
@@ -84,15 +92,15 @@ def run(config_path, input_path, output_name, crs, single_use, lazy):
                 required_acts=group["required_acts"],
                 new_tags=group["new_tags"],
                 size=group["size"],
-                spacing=group["spacing"]
+                spacing=group["spacing"],
             )
             logger.info(f" Filled {zones} zones with {objects} objects.")
 
     if cnfg.get("object_features"):
         logger.info(f" Assigning object features: {cnfg['object_features']}.")
         handler.add_features()
 
-    if 'distance_to_nearest' in cnfg:
+    if "distance_to_nearest" in cnfg:
         for target_activity in cnfg["distance_to_nearest"]:
             logger.info(f" Assigning distances to nearest {target_activity}.")
             handler.assign_nearest_distance(target_activity)