Multithreading i overlay, dissolve. Gridlooper

.github/workflows/tests.yml

@@ -49,7 +49,7 @@ jobs:
       - name: Run pytest with coverage
         if: ${{ (matrix.os != 'ubuntu-latest') }}
         run: |
-          poetry run pytest -k "not tests/test_raster.py" --verbose --durations=5
+          poetry run pytest -k "not raster and not cube" --verbose --durations=5
 
       - name: Run raster pytest with coverage
         if: ${{ (matrix.os == 'ubuntu-latest') }}

docs/reference/io/index.rst

@@ -8,4 +8,3 @@ Functions for reading and writing geodata.
 
    dapla
    read_parquet_url
-   write_municipality_data

docs/reference/raster/index.rst

@@ -7,4 +7,3 @@ Class for raster analysis from image files, arrays and GeoDataFrames.
    :maxdepth: 3
 
    raster
-   elevationraster

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "ssb-sgis"
-version = "0.3.11"
+version = "0.3.12"
 description = "GIS functions used at Statistics Norway."
 authors = ["Statistics Norway <ort@ssb.no>"]
 license = "MIT"
@@ -26,18 +26,19 @@ mapclassify = ">=2.5.0"
 matplotlib = ">=3.7.0"
 networkx = ">=3.0"
 numpy = ">=1.24.2"
-pandas = ">=1.5.3"
+pandas = "2.0.3"
 pyarrow = ">=11.0.0"
 requests = ">=2.28.2"
 scikit-learn = ">=1.2.1"
 shapely = ">=2.0.1"
 xyzservices = ">=2023.2.0"
 jenkspy = ">=0.3.2"
 ipython = ">=8.13.2"
-rtree = "^1.0.1"
-geocoder = "^1.38.1"
-rasterio = "^1.3.8"
+rtree = ">=1.0.1"
+geocoder = ">=1.38.1"
+rasterio = ">=1.3.8"
 pip = "23.2.1"
+dask = ">=2024.1.1"
 
 [tool.poetry.group.dev.dependencies]
 black = {extras = ["d", "jupyter"], version = ">=23.1.0"}
@@ -97,8 +98,8 @@ notebook_metadata_filter = "jupytext.text_representation,-jupytext.text_represen
 cell_metadata_filter = "-all"
 
 [tool.pytest.ini_options]
-  pythonpath = [
-   "src"
+pythonpath = [
+  "src"
 ]
 
 [build-system]

src/sgis/__init__.py

@@ -1,3 +1,11 @@
+config = {
+    "n_jobs": 1,
+}
+
+import sgis.raster.bands as bands
+import sgis.raster.indices as indices
+from sgis.raster.raster import Raster, get_shape_from_bounds, get_transform_from_bounds
+
 from .geopandas_tools.bounds import (
     Gridlooper,
     bounds_to_points,
@@ -8,7 +16,6 @@
     make_grid_from_bbox,
     make_ssb_grid,
     points_in_bounds,
-    to_bbox,
 )
 from .geopandas_tools.buffer_dissolve_explode import (
     buff,
@@ -22,12 +29,17 @@
 from .geopandas_tools.cleaning import (
     coverage_clean,
     remove_spikes,
+    snap_polygons,
+    snap_to_mask,
+    split_and_eliminate_by_longest,
+    split_by_neighbors,
     split_spiky_polygons,
 )
 from .geopandas_tools.conversion import (
     coordinate_array,
     from_4326,
     to_4326,
+    to_bbox,
     to_gdf,
     to_geoseries,
     to_shapely,
@@ -119,13 +131,16 @@
 )
 from .networkanalysis.traveling_salesman import traveling_salesman_problem
 from .parallel.parallel import Parallel
-from .raster.elevationraster import ElevationRaster
-from .raster.raster import Raster
-from .raster.sentinel import Sentinel2
+from .raster.cube import DataCube
+
+
+try:
+    import sgis.raster.torchgeo as torchgeo
+except ImportError:
+    pass
 
 
 try:
     from .io.dapla_functions import check_files, read_geopandas, write_geopandas
-    from .io.write_municipality_data import write_municipality_data
 except ImportError:
     pass

src/sgis/geopandas_tools/bounds.py

@@ -13,7 +13,7 @@
 from shapely.geometry import Polygon
 
 from ..parallel.parallel import Parallel
-from .conversion import to_gdf
+from .conversion import to_bbox, to_gdf
 from .general import clean_clip, is_bbox_like
 
 
@@ -93,17 +93,20 @@ class Gridlooper:
     gridsize: int
     mask: GeoDataFrame | GeoSeries | Geometry
     gridbuffer: int = 0
+    parallelizer: Parallel | None = None
+    concat: bool = False
     clip: bool = True
     keep_geom_type: bool = True
     verbose: bool = False
-    parallelizer: Parallel | None = None
 
     def __post_init__(self):
         if not isinstance(self.mask, GeoDataFrame):
             self.mask = to_gdf(self.mask)
 
     def run(self, func: Callable, *args, **kwargs):
-        intersects_mask = lambda df: df.index.isin(df.sjoin(self.mask).index)
+        def intersects_mask(df):
+            return df.index.isin(df.sjoin(self.mask).index)
+
         grid: GeoSeries = (
             make_grid(self.mask, gridsize=self.gridsize).loc[intersects_mask].geometry
         )
@@ -123,15 +126,19 @@ def run(self, func: Callable, *args, **kwargs):
             )
             results = self.parallelizer.map(func_with_clip, buffered_grid)
             if not self.gridbuffer or not self.clip:
-                return results
+                return (
+                    results
+                    if not self.concat
+                    else pd.concat(results, ignore_index=True)
+                )
             out = []
             for cell_res, unbuffered in zip(results, grid, strict=True):
                 out.append(
                     _clip_back_to_unbuffered_grid(
                         cell_res, unbuffered, self.keep_geom_type
                     )
                 )
-            return out
+            return out if not self.concat else pd.concat(out, ignore_index=True)
 
         results = []
         for i, (unbuffered, buffered) in enumerate(zip(grid, buffered_grid)):
@@ -159,7 +166,7 @@ def run(self, func: Callable, *args, **kwargs):
             if self.verbose:
                 print(f"Done with {i+1} of {n} grid cells", end="\r")
 
-        return results
+        return results if not self.concat else pd.concat(results, ignore_index=True)
 
 
 def gridloop(
@@ -639,64 +646,6 @@ def bounds_to_points(
     return gdf
 
 
-def to_bbox(
-    obj: GeoDataFrame | GeoSeries | Geometry | Collection | Mapping,
-) -> tuple[float, float, float, float]:
-    """Returns 4-length tuple of bounds if possible, else raises ValueError.
-
-    Args:
-        obj: Object to be converted to bounding box. Can be geopandas or shapely
-            objects, iterables of exactly four numbers or dictionary like/class
-            with a the keys/attributes "minx", "miny", "maxx", "maxy" or
-            "xmin", "ymin", "xmax", "ymax".
-    """
-    if isinstance(obj, (GeoDataFrame, GeoSeries)):
-        return tuple(obj.total_bounds)
-    if isinstance(obj, Geometry):
-        return tuple(obj.bounds)
-    if (
-        hasattr(obj, "__iter__")
-        and len(obj) == 4
-        and all(isinstance(x, numbers.Number) for x in obj)
-    ):
-        return tuple(obj)
-
-    if is_dict_like(obj) and all(x in obj for x in ["minx", "miny", "maxx", "maxy"]):
-        try:
-            minx = np.min(obj["minx"])
-            miny = np.min(obj["miny"])
-            maxx = np.max(obj["maxx"])
-            maxy = np.max(obj["maxy"])
-        except TypeError:
-            minx = np.min(obj.minx)
-            miny = np.min(obj.miny)
-            maxx = np.max(obj.maxx)
-            maxy = np.max(obj.maxy)
-        return minx, miny, maxx, maxy
-    if is_dict_like(obj) and all(x in obj for x in ["xmin", "ymin", "xmax", "ymax"]):
-        try:
-            xmin = np.min(obj["xmin"])
-            ymin = np.min(obj["ymin"])
-            xmax = np.max(obj["xmax"])
-            ymax = np.max(obj["ymax"])
-        except TypeError:
-            xmin = np.min(obj.xmin)
-            ymin = np.min(obj.ymin)
-            xmax = np.max(obj.xmax)
-            ymax = np.max(obj.ymax)
-        return xmin, ymin, xmax, ymax
-    if is_dict_like(obj) and hasattr(obj, "geometry"):
-        try:
-            return tuple(GeoSeries(obj["geometry"]).total_bounds)
-        except Exception:
-            return tuple(GeoSeries(obj.geometry).total_bounds)
-    try:
-        of_length = f" of length {len(obj)}"
-    except TypeError:
-        of_length = ""
-    raise TypeError(f"Cannot convert type {obj.__class__.__name__}{of_length} to bbox")
-
-
 def get_total_bounds(
     *geometries: GeoDataFrame | GeoSeries | Geometry,
 ) -> tuple[float, float, float, float]: