vector2dggs 0.6.3__tar.gz → 0.9.0__tar.gz

{vector2dggs-0.6.3 → vector2dggs-0.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vector2dggs
-Version: 0.6.3
+Version: 0.9.0
 Summary: CLI DGGS indexer for vector geospatial data
 Home-page: https://github.com/manaakiwhenua/vector2dggs
 License: LGPL-3.0-or-later
@@ -29,8 +29,10 @@ Requires-Dist: pillow (>=11.2.1,<12.0.0)
 Requires-Dist: psycopg2 (>=2.9.9,<3.0.0)
 Requires-Dist: pyarrow (>=20.0,<21.0)
 Requires-Dist: pyproj (>=3.7,<4.0)
-Requires-Dist: rhealpixdggs (>=0.5.5,<0.6.0)
-Requires-Dist: rhppandas (>=0.1.2,<0.2.0)
+Requires-Dist: python-geohash (>=0.8.5,<0.9.0)
+Requires-Dist: rhppandas (>=0.2.0,<0.3.0)
+Requires-Dist: rusty-polygon-geohasher (>=0.2.3,<0.3.0)
+Requires-Dist: s2geometry (>=0.9.0,<0.10.0)
 Requires-Dist: shapely (>=2.1,<3.0)
 Requires-Dist: sqlalchemy (>=2.0.32,<3.0.0)
 Requires-Dist: tqdm (>=4.67,<5.0)
@@ -47,8 +49,13 @@ This is the vector equivalent of [raster2dggs](https://github.com/manaakiwhenua/
 
 Currently this tool supports the following DGGSs:
 
-- H3 (polygons, linestrings)
-- rHEALPix (polygons)
+- [H3](https://h3geo.org/)
+- [rHEALPix](https://datastore.landcareresearch.co.nz/dataset/rhealpix-discrete-global-grid-system)
+- [S2](https://s2geometry.io/)
+
+... and the following geocode systems:
+
+- [Geohash](https://en.wikipedia.org/wiki/Geohash) (points, polygons)
 
 Contributions (espeically for other DGGSs), suggestions, bug reports and strongly worded letters are all welcome.
 
@@ -63,7 +70,8 @@ pip install vector2dggs
 ## Usage
 
 ```bash
-vector2dggs --help                                                                                                                                                                                                           [11:22:14]
+vector2dggs --help
+
 Usage: vector2dggs [OPTIONS] COMMAND [ARGS]...
 
 Options:
@@ -71,8 +79,10 @@ Options:
   --help     Show this message and exit.
 
 Commands:
-  h3   Ingest a vector dataset and index it to the H3 DGGS.
-  rhp  Ingest a vector dataset and index it to the rHEALPix DGGS.
+  geohash  Ingest a vector dataset and index it using the Geohash geocode...
+  h3       Ingest a vector dataset and index it to the H3 DGGS.
+  rhp      Ingest a vector dataset and index it to the rHEALPix DGGS.
+  s2       Ingest a vector dataset and index it to the S2 DGGS.
 ```
 
 ```bash
@@ -119,8 +129,8 @@ Options:
                                   [default: 5000; required]
   -t, --threads INTEGER           Amount of threads used for operation
                                   [default: 7]
-  -tbl, --table TEXT              Name of the table to read when using a
-                                  spatial database connection as input
+  -lyr, --layer TEXT              Name of the layer or table to read when using a
+                                  an input that supports layers or tables
   -g, --geom_col TEXT             Column name to use when using a spatial
                                   database connection as input  [default:
                                   geom]
@@ -137,9 +147,9 @@ Options:
 
 Output is in the Apache Parquet format, a directory with one file per partition.
 
-For a quick view of your output, you can read Apache Parquet with pandas, and then use h3-pandas and geopandas to convert this into a GeoPackage or GeoParquet for visualisation in a desktop GIS, such as QGIS. The Apache Parquet output is indexed by an ID column (which you can specify), so it should be ready for two intended use-cases:
+For a quick view of your output, you can read Apache Parquet with pandas, and then use tools like h3-pandas and geopandas to convert this into a GeoPackage or GeoParquet for visualisation in a desktop GIS, such as QGIS. The Apache Parquet output is indexed by an ID column (which you can specify), so it should be ready for two intended use-cases:
 - Joining attribute data from the original feature-level data onto computer DGGS cells.
-- Joining other data to this output on the H3 cell ID. (The output has a column like `h3_\d{2}`, e.g. `h3_09` or `h3_12` according to the target resolution.)
+- Joining other data to this output on the DGGS cell ID. (The output has a column like `{dggs}_\d`, e.g. `h3_09` or `h3_12` according to the target resolution, zero-padded to account for the maximum resolution of the DGGS)
 
 Geoparquet output (hexagon boundaries):
 
@@ -166,6 +176,34 @@ h3_12
 >>> g.to_parquet('./output-data/parcels.12.geo.parquet')
 ```
 
+An example for S2 output (using `s2sphere`):
+
+
+```python
+import pandas as pd
+import geopandas as gpd
+import s2sphere
+from shapely.geometry import Polygon
+
+RES = 18
+df = pd.read_parquet(f'~/output-data/ponds-with-holes.s2.{RES}.pq')
+df = df.reset_index()
+
+def s2id_to_polygon(s2_id_hex):
+    cell_id = s2sphere.CellId.from_token(s2_id_hex)
+    cell = s2sphere.Cell(cell_id)
+    vertices = []
+    for i in range(4):
+        vertex = cell.get_vertex(i)
+        lat_lng = s2sphere.LatLng.from_point(vertex)
+        vertices.append((lat_lng.lng().degrees, lat_lng.lat().degrees))  # (lon, lat)
+    return Polygon(vertices)
+
+df['geometry'] = df[f's2_{RES}'].apply(s2id_to_polygon)
+df = gpd.GeoDataFrame(df, geometry='geometry', crs='EPSG:4326')  # WGS84
+df.to_parquet(f'sample-{RES}.parquet')
+```
+
 ### For development
 
 In brief, to get started:
@@ -175,6 +213,7 @@ In brief, to get started:
     - If you're on Windows, `pip install gdal` may be necessary before running the subsequent commands.
     - On Linux, install GDAL 3.8+ according to your platform-specific instructions, including development headers, i.e. `libgdal-dev`.
 - Create the virtual environment with `poetry init`. This will install necessary dependencies.
+  - If the installation of `s2geometry` fails, you may require SWIG to build it. (A command like `conda install swig` or `sudo dnf install swig` depending on your platform).
 - Subsequently, the virtual environment can be re-activated with `poetry shell`.
 
 If you run `poetry install`, the CLI tool will be aliased so you can simply use `vector2dggs` rather than `poetry run vector2dggs`, which is the alternative if you do not `poetry install`.
@@ -199,7 +238,7 @@ vector2dggs h3 -v DEBUG -id title_no -r 12 -o ~/Downloads/nz-property-titles.gpk
 With a PostgreSQL/PostGIS connection:
 
 ```bash
-vector2dggs h3 -v DEBUG -id ogc_fid -r 9 -p 5 -t 4 --overwrite -tbl topo50_lake postgresql://user:password@host:port/db ./topo50_lake.parquet
+vector2dggs h3 -v DEBUG -id ogc_fid -r 9 -p 5 -t 4 --overwrite -lyr topo50_lake postgresql://user:password@host:port/db ./topo50_lake.parquet
 ```
 
 ## Citation
@@ -209,14 +248,14 @@ vector2dggs h3 -v DEBUG -id ogc_fid -r 9 -p 5 -t 4 --overwrite -tbl topo50_lake
   title={{vector2dggs}},
   author={Ardo, James and Law, Richard},
   url={https://github.com/manaakiwhenua/vector2dggs},
-  version={0.6.3},
+  version={0.9.0},
   date={2023-04-20}
 }
 ```
 
 APA/Harvard
 
-> Ardo, J., & Law, R. (2023). vector2dggs (0.6.3) [Computer software]. https://github.com/manaakiwhenua/vector2dggs
+> Ardo, J., & Law, R. (2023). vector2dggs (0.9.0) [Computer software]. https://github.com/manaakiwhenua/vector2dggs
 
 [![manaakiwhenua-standards](https://github.com/manaakiwhenua/vector2dggs/workflows/manaakiwhenua-standards/badge.svg)](https://github.com/manaakiwhenua/manaakiwhenua-standards)
 

{vector2dggs-0.6.3 → vector2dggs-0.9.0}/README.md

@@ -8,8 +8,13 @@ This is the vector equivalent of [raster2dggs](https://github.com/manaakiwhenua/
 
 Currently this tool supports the following DGGSs:
 
-- H3 (polygons, linestrings)
-- rHEALPix (polygons)
+- [H3](https://h3geo.org/)
+- [rHEALPix](https://datastore.landcareresearch.co.nz/dataset/rhealpix-discrete-global-grid-system)
+- [S2](https://s2geometry.io/)
+
+... and the following geocode systems:
+
+- [Geohash](https://en.wikipedia.org/wiki/Geohash) (points, polygons)
 
 Contributions (espeically for other DGGSs), suggestions, bug reports and strongly worded letters are all welcome.
 
@@ -24,7 +29,8 @@ pip install vector2dggs
 ## Usage
 
 ```bash
-vector2dggs --help                                                                                                                                                                                                           [11:22:14]
+vector2dggs --help
+
 Usage: vector2dggs [OPTIONS] COMMAND [ARGS]...
 
 Options:
@@ -32,8 +38,10 @@ Options:
   --help     Show this message and exit.
 
 Commands:
-  h3   Ingest a vector dataset and index it to the H3 DGGS.
-  rhp  Ingest a vector dataset and index it to the rHEALPix DGGS.
+  geohash  Ingest a vector dataset and index it using the Geohash geocode...
+  h3       Ingest a vector dataset and index it to the H3 DGGS.
+  rhp      Ingest a vector dataset and index it to the rHEALPix DGGS.
+  s2       Ingest a vector dataset and index it to the S2 DGGS.
 ```
 
 ```bash
@@ -80,8 +88,8 @@ Options:
                                   [default: 5000; required]
   -t, --threads INTEGER           Amount of threads used for operation
                                   [default: 7]
-  -tbl, --table TEXT              Name of the table to read when using a
-                                  spatial database connection as input
+  -lyr, --layer TEXT              Name of the layer or table to read when using a
+                                  an input that supports layers or tables
   -g, --geom_col TEXT             Column name to use when using a spatial
                                   database connection as input  [default:
                                   geom]
@@ -98,9 +106,9 @@ Options:
 
 Output is in the Apache Parquet format, a directory with one file per partition.
 
-For a quick view of your output, you can read Apache Parquet with pandas, and then use h3-pandas and geopandas to convert this into a GeoPackage or GeoParquet for visualisation in a desktop GIS, such as QGIS. The Apache Parquet output is indexed by an ID column (which you can specify), so it should be ready for two intended use-cases:
+For a quick view of your output, you can read Apache Parquet with pandas, and then use tools like h3-pandas and geopandas to convert this into a GeoPackage or GeoParquet for visualisation in a desktop GIS, such as QGIS. The Apache Parquet output is indexed by an ID column (which you can specify), so it should be ready for two intended use-cases:
 - Joining attribute data from the original feature-level data onto computer DGGS cells.
-- Joining other data to this output on the H3 cell ID. (The output has a column like `h3_\d{2}`, e.g. `h3_09` or `h3_12` according to the target resolution.)
+- Joining other data to this output on the DGGS cell ID. (The output has a column like `{dggs}_\d`, e.g. `h3_09` or `h3_12` according to the target resolution, zero-padded to account for the maximum resolution of the DGGS)
 
 Geoparquet output (hexagon boundaries):
 
@@ -127,6 +135,34 @@ h3_12
 >>> g.to_parquet('./output-data/parcels.12.geo.parquet')
 ```
 
+An example for S2 output (using `s2sphere`):
+
+
+```python
+import pandas as pd
+import geopandas as gpd
+import s2sphere
+from shapely.geometry import Polygon
+
+RES = 18
+df = pd.read_parquet(f'~/output-data/ponds-with-holes.s2.{RES}.pq')
+df = df.reset_index()
+
+def s2id_to_polygon(s2_id_hex):
+    cell_id = s2sphere.CellId.from_token(s2_id_hex)
+    cell = s2sphere.Cell(cell_id)
+    vertices = []
+    for i in range(4):
+        vertex = cell.get_vertex(i)
+        lat_lng = s2sphere.LatLng.from_point(vertex)
+        vertices.append((lat_lng.lng().degrees, lat_lng.lat().degrees))  # (lon, lat)
+    return Polygon(vertices)
+
+df['geometry'] = df[f's2_{RES}'].apply(s2id_to_polygon)
+df = gpd.GeoDataFrame(df, geometry='geometry', crs='EPSG:4326')  # WGS84
+df.to_parquet(f'sample-{RES}.parquet')
+```
+
 ### For development
 
 In brief, to get started:
@@ -136,6 +172,7 @@ In brief, to get started:
     - If you're on Windows, `pip install gdal` may be necessary before running the subsequent commands.
     - On Linux, install GDAL 3.8+ according to your platform-specific instructions, including development headers, i.e. `libgdal-dev`.
 - Create the virtual environment with `poetry init`. This will install necessary dependencies.
+  - If the installation of `s2geometry` fails, you may require SWIG to build it. (A command like `conda install swig` or `sudo dnf install swig` depending on your platform).
 - Subsequently, the virtual environment can be re-activated with `poetry shell`.
 
 If you run `poetry install`, the CLI tool will be aliased so you can simply use `vector2dggs` rather than `poetry run vector2dggs`, which is the alternative if you do not `poetry install`.
@@ -160,7 +197,7 @@ vector2dggs h3 -v DEBUG -id title_no -r 12 -o ~/Downloads/nz-property-titles.gpk
 With a PostgreSQL/PostGIS connection:
 
 ```bash
-vector2dggs h3 -v DEBUG -id ogc_fid -r 9 -p 5 -t 4 --overwrite -tbl topo50_lake postgresql://user:password@host:port/db ./topo50_lake.parquet
+vector2dggs h3 -v DEBUG -id ogc_fid -r 9 -p 5 -t 4 --overwrite -lyr topo50_lake postgresql://user:password@host:port/db ./topo50_lake.parquet
 ```
 
 ## Citation
@@ -170,13 +207,13 @@ vector2dggs h3 -v DEBUG -id ogc_fid -r 9 -p 5 -t 4 --overwrite -tbl topo50_lake
   title={{vector2dggs}},
   author={Ardo, James and Law, Richard},
   url={https://github.com/manaakiwhenua/vector2dggs},
-  version={0.6.3},
+  version={0.9.0},
   date={2023-04-20}
 }
 ```
 
 APA/Harvard
 
-> Ardo, J., & Law, R. (2023). vector2dggs (0.6.3) [Computer software]. https://github.com/manaakiwhenua/vector2dggs
+> Ardo, J., & Law, R. (2023). vector2dggs (0.9.0) [Computer software]. https://github.com/manaakiwhenua/vector2dggs
 
 [![manaakiwhenua-standards](https://github.com/manaakiwhenua/vector2dggs/workflows/manaakiwhenua-standards/badge.svg)](https://github.com/manaakiwhenua/manaakiwhenua-standards)

{vector2dggs-0.6.3 → vector2dggs-0.9.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "vector2dggs"
-version = "0.6.3"
+version = "0.9.0"
 description = "CLI DGGS indexer for vector geospatial data"
 authors = ["James Ardo <ardoj@landcareresearch.co.nz>"]
 maintainers = ["Richard Law <lawr@landcareresearch.co.nz>"]
@@ -30,9 +30,11 @@ sqlalchemy = "^2.0.32"
 psycopg2 = "^2.9.9"
 shapely = "^2.1"
 numpy = "^2"
-rhppandas = "^0.1.2"
-rhealpixdggs = "^0.5.5"
+rhppandas = "^0.2.0"
 pillow = "^11.2.1"
+s2geometry = "^0.9.0"
+rusty-polygon-geohasher = "^0.2.3"
+python-geohash = "^0.8.5"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.2.2"

vector2dggs-0.9.0/vector2dggs/__init__.py

@@ -0,0 +1 @@
+__version__: str = "0.9.0"

{vector2dggs-0.6.3 → vector2dggs-0.9.0}/vector2dggs/cli.py

@@ -3,6 +3,8 @@ import click
 from vector2dggs import __version__
 from vector2dggs.h3 import h3
 from vector2dggs.rHP import rhp
+from vector2dggs.s2 import s2
+from vector2dggs.geohash import geohash
 
 #   If the program does terminal interaction, make it output a short
 # notice like this when it starts in an interactive mode:
@@ -21,6 +23,8 @@ def cli():
 
 cli.add_command(h3)
 cli.add_command(rhp)
+cli.add_command(s2)
+cli.add_command(geohash)
 
 
 def main():

{vector2dggs-0.6.3 → vector2dggs-0.9.0}/vector2dggs/common.py

@@ -104,13 +104,15 @@ def drop_condition(
     _diff = _before - _after
     if _diff:
         log_method = (
-            LOGGER.info if (_diff / float(_before)) < warning_threshold else LOGGER.warn
+            LOGGER.info
+            if (_diff / float(_before)) < warning_threshold
+            else LOGGER.warning
         )
         log_method(f"Dropped {_diff} rows ({_diff/float(_before)*100:.2f}%)")
     return df
 
 
-def get_parent_res(dggs: str, parent_res: Union[None, int], resolution: int):
+def get_parent_res(dggs: str, parent_res: Union[None, str], resolution: int) -> int:
     """
     Uses a parent resolution,
     OR,
@@ -118,22 +120,17 @@ def get_parent_res(dggs: str, parent_res: Union[None, int], resolution: int):
 
     Used for intermediate re-partioning.
     """
-    if dggs == "h3":
-        return (
-            parent_res
-            if parent_res is not None
-            else max(const.MIN_H3, (resolution - const.DEFAULT_PARENT_OFFSET))
-        )
-    elif dggs == "rhp":
-        return (
-            parent_res
-            if parent_res is not None
-            else max(const.MIN_RHP, (resolution - const.DEFAULT_PARENT_OFFSET))
-        )
-    else:
+    if not dggs in const.DEFAULT_DGGS_PARENT_RES.keys():
         raise RuntimeError(
-            "Unknown dggs {dggs}) -  must be one of [ 'h3', 'rhp' ]".format(dggs=dggs)
+            "Unknown dggs {dggs}) -  must be one of [ {options} ]".format(
+                dggs=dggs, options=", ".join(const.DEFAULT_DGGS_PARENT_RES.keys())
+            )
         )
+    return (
+        int(parent_res)
+        if parent_res is not None
+        else const.DEFAULT_DGGS_PARENT_RES[dggs](resolution)
+    )
 
 
 def parent_partitioning(
@@ -141,10 +138,9 @@ def parent_partitioning(
     input_dir: Path,
     output_dir: Path,
     resolution: int,
-    parent_res: Union[None, int],
+    parent_res: int,
     **kwargs,
 ) -> None:
-    parent_res: int = get_parent_res(dggs, parent_res, resolution)
     partition_col = f"{dggs}_{parent_res:02}"
 
     with TqdmCallback(desc="Repartitioning"):
@@ -174,30 +170,29 @@ def polyfill(
     pq_in: Path,
     spatial_sort_col: str,
     resolution: int,
-    parent_res: Union[None, int],
+    parent_res: int,
     output_directory: str,
 ) -> None:
     """
     Reads a geoparquet, performs polyfilling (for Polygon),
-    linetracing (for LineString), and writes out to parquet.
+    linetracing (for LineString), or indexing (for Point),
+    and writes out to parquet.
     """
-    df = gpd.read_parquet(pq_in).reset_index().drop(columns=[spatial_sort_col])
+    df = gpd.read_parquet(pq_in).reset_index()
+    if spatial_sort_col != "none":
+        df = df.drop(columns=[spatial_sort_col])
     if len(df.index) == 0:
-        # Input is empty, nothing to polyfill
+        # Input is empty, nothing to convert
         return None
 
-    # DGGS specific polyfill
+    # DGGS specific conversion
     df = dggsfunc(df, resolution)
 
     if len(df.index) == 0:
-        # Polyfill resulted in empty output (e.g. large cell, small feature)
+        # Conversion resulted in empty output (e.g. large cell, small feature)
         return None
 
     df.index.rename(f"{dggs}_{resolution:02}", inplace=True)
-    parent_res: int = get_parent_res(dggs, parent_res, resolution)
-    # print(parent_res)
-    # print(df.index)
-    # print(df.columns)
 
     # Secondary (parent) index, used later for partitioning
     df = secondary_index_func(df, parent_res)
@@ -228,22 +223,23 @@ def index(
     id_field: str = None,
     cut_crs: pyproj.CRS = None,
     con: SQLConnectionType = None,
-    table: str = None,
+    layer: str = None,
     geom_col: str = "geom",
     overwrite: bool = False,
 ) -> Path:
     """
-    Performs multi-threaded polyfilling on (multi)polygons.
+    Performs multi-threaded DGGS indexing on geometries (including multipart and collections).
     """
+    parent_res = get_parent_res(dggs, parent_res, resolution)
 
-    if table and con:
+    if layer and con:
         # Database connection
         if keep_attributes:
-            q = sqlalchemy.text(f"SELECT * FROM {table}")
+            q = sqlalchemy.text(f"SELECT * FROM {layer}")
         elif id_field and not keep_attributes:
-            q = sqlalchemy.text(f"SELECT {id_field}, {geom_col} FROM {table}")
+            q = sqlalchemy.text(f"SELECT {id_field}, {geom_col} FROM {layer}")
         else:
-            q = sqlalchemy.text(f"SELECT {geom_col} FROM {table}")
+            q = sqlalchemy.text(f"SELECT {geom_col} FROM {layer}")
         df = gpd.read_postgis(q, con.connect(), geom_col=geom_col).rename_geometry(
             "geometry"
         )
@@ -291,7 +287,8 @@ def index(
             "index": lambda frame: frame[
                 (frame.geometry.geom_type != "Polygon")
                 & (frame.geometry.geom_type != "LineString")
-            ],  # NB currently points and other types are lost; in principle, these could be indexed
+                & (frame.geometry.geom_type != "Point")
+            ],
             "message": "Considering unsupported geometries",
         },
     ]
@@ -300,11 +297,12 @@ def index(
 
     ddf = dgpd.from_geopandas(df, chunksize=max(1, chunksize), sort=True)
 
-    LOGGER.debug("Spatially sorting and partitioning (%s)", spatial_sorting)
-    ddf = ddf.spatial_shuffle(by=spatial_sorting)
+    if spatial_sorting != "none":
+        LOGGER.debug("Spatially sorting and partitioning (%s)", spatial_sorting)
+        ddf = ddf.spatial_shuffle(by=spatial_sorting)
     spatial_sort_col = (
         spatial_sorting
-        if spatial_sorting == "geohash"
+        if (spatial_sorting == "geohash" or spatial_sorting == "none")
         else f"{spatial_sorting}_distance"
     )
 
@@ -314,9 +312,9 @@ def index(
 
         filepaths = list(map(lambda f: f.absolute(), Path(tmpdir).glob("*")))
 
-        # Multithreaded polyfilling
+        # Multithreaded DGGS indexing
         LOGGER.debug(
-            "Indexing on spatial partitions by polyfill with resolution: %d",
+            "DGGS indexing by spatial partitions with resolution: %d",
             resolution,
         )
         with tempfile.TemporaryDirectory(suffix=".parquet") as tmpdir2:
@@ -344,7 +342,7 @@ def index(
 
             parent_partitioning(
                 dggs,
-                tmpdir2,
+                Path(tmpdir2),
                 output_directory,
                 resolution,
                 parent_res,

vector2dggs-0.9.0/vector2dggs/constants.py

@@ -0,0 +1,75 @@
+import multiprocessing
+import warnings
+import tempfile
+
+
+MIN_H3, MAX_H3 = 0, 15
+MIN_RHP, MAX_RHP = 0, 15
+MIN_S2, MAX_S2 = 0, 30
+MIN_GEOHASH, MAX_GEOHASH = 1, 12
+
+DEFAULTS = {
+    "id": None,
+    "k": False,
+    "ch": 50,
+    "s": "none",
+    "crs": None,
+    "c": 5000,
+    "t": (multiprocessing.cpu_count() - 1),
+    "lyr": None,
+    "g": "geom",
+    "tempdir": tempfile.tempdir,
+}
+
+SPATIAL_SORTING_METHODS = ["hilbert", "morton", "geohash", "none"]
+
+DEFAULT_DGGS_PARENT_RES = {
+    "h3": lambda resolution: max(MIN_H3, (resolution - DEFAULT_PARENT_OFFSET)),
+    "rhp": lambda resolution: max(MIN_RHP, (resolution - DEFAULT_PARENT_OFFSET)),
+    "geohash": lambda resolution: max(
+        MIN_GEOHASH, (resolution - DEFAULT_PARENT_OFFSET)
+    ),
+    "s2": lambda resolution: max(MIN_S2, (resolution - DEFAULT_PARENT_OFFSET)),
+}
+
+DEFAULT_PARENT_OFFSET = 6
+
+# http://s2geometry.io/resources/s2cell_statistics.html
+S2_CELLS_MAX_AREA_M2_BY_LEVEL = {
+    0: 85011012.19 * 1e6,
+    1: 21252753.05 * 1e6,
+    2: 6026521.16 * 1e6,
+    3: 1646455.50 * 1e6,
+    4: 413918.15 * 1e6,
+    5: 104297.91 * 1e6,
+    6: 26113.30 * 1e6,
+    7: 6529.09 * 1e6,
+    8: 1632.45 * 1e6,
+    9: 408.12 * 1e6,
+    10: 102.03 * 1e6,
+    11: 25.51 * 1e6,
+    12: 6.38 * 1e6,
+    13: 1.59 * 1e6,
+    14: 0.40 * 1e6,
+    15: 99638.93,
+    16: 24909.73,
+    17: 6227.43,
+    18: 1556.86,
+    19: 389.22,
+    20: 97.30,
+    21: 24.33,
+    22: 6.08,
+    23: 1.52,
+    24: 0.38,
+    25: 950.23 * 1e-4,
+    26: 237.56 * 1e-4,
+    27: 59.39 * 1e-4,
+    28: 14.85 * 1e-4,
+    29: 3.71 * 1e-4,
+    30: 0.93 * 1e-4,
+}
+
+
+warnings.filterwarnings(
+    "ignore"
+)  # This is to filter out the polyfill warnings when rows failed to get indexed at a resolution, can be commented out to find missing rows