geographer.cutters package

Submodules

geographer.cutters.cut_every_raster_to_grid module

Cut rasters in the source dataset to a grid of rasters.

Dataset cutter that cuts rasters in the source dataset to a grid of rasters.

geographer.cutters.cut_every_raster_to_grid.get_cutter_every_raster_to_grid(source_data_dir, target_data_dir, name, new_raster_size=512, raster_filter_predicate=None)[source]

Return dataset cutter that cuts every raster to a grid.

Return dataset cutter that cuts every raster in the source dataset to a grid of rasters.

Return type:

DSCutterIterOverRasters

Parameters:

  • source_data_dir (Path) – source data dir

  • target_data_dir (Path) – target data dir

  • name (str) – name of cutter, used when saving the cutter

  • new_raster_size (int | Tuple[int, int]) – size of new rasters. Defaults to 512.

  • raster_filter_predicate (RasterFilterPredicate | None) – raster filter predicate to select rasters. Defaults to None (i.e. cut all rasters that have not been previously cut).

Returns:

dataset cutter

Return type:

DSCutterIterOverRasters

geographer.cutters.cut_iter_over_rasters module

Dataset cutter that iterates over rasters.

Implements a general-purpose higher order function to create or update datasets of GeoTiffs from existing ones by iterating over rasters.

class geographer.cutters.cut_iter_over_rasters.DSCutterIterOverRasters(**data)[source]

Bases: DSCreatorFromSourceWithBands

Dataset cutter that iterates over rasters.

Implements a general-purpose higher order function to create or update datasets of GeoTiffs from existing ones by iterating over rasters.

Parameters:
Return type:

None

field bands: Optional[dict[str, Optional[list[int]]]] = None

keys: raster directory names, values: list of band indices to keep, starting with 1

Validated by:

  • validate_connectors

field cut_rasters: list[str] [Optional]

Names of cut rasters in source_data_dir. Usually not to be set by hand!

Validated by:

  • validate_connectors

field label_maker: Optional[LabelMaker] = None

Optional label maker. If given, will be used to recompute labels when necessary. Defaults to None

Validated by:

  • validate_connectors

field name: str [Required]

Name of dataset creator. Used as part of filename when saving.

Validated by:

  • validate_connectors

field raster_cutter: SingleRasterCutter [Required]
Validated by:

  • validate_connectors

field raster_filter_predicate: RasterFilterPredicate = AlwaysTrue()
Validated by:

  • validate_connectors

field source_data_dir: Path [Required]
Validated by:

  • validate_connectors

  • validate_source_data_dir

field target_data_dir: Path [Required]
Validated by:

  • validate_connectors

create(*args, **kwargs)

Create a new dataset by cutting the source dataset.

Return type:

Connector

create_or_update()[source]

Create or update target dataset.

Return type:

Connector

Returns:

connector of target dataset

cut()[source]

Cut a dataset.

Alternate name for the create method. See create_or_update for docstring.

Return type:

Connector

classmethod from_json_file(json_file_path, constructor_symbol_table=None)

Load and return saved BaseModel.

Return type:

Any

Parameters:

  • json_file_path (Path | str)

  • constructor_symbol_table (dict[str, Any] | None)

save()

Save to update folder in source_data_dir.

update(*args, **kwargs)

Update the target dataset from the source dataset.

Return type:

Connector

property source_connector

Connector in source_data_dir.

property target_connector

Connector in target_data_dir.

geographer.cutters.cut_iter_over_vectors module

Dataset cutter that iterates over vector features.

Implements a general-purpose higher order function to create or update datasets of GeoTiffs from existing ones by iterating over vector features.

class geographer.cutters.cut_iter_over_vectors.DSCutterIterOverVectors(**data)[source]

Bases: DSCreatorFromSourceWithBands

Dataset cutter that iterates over vector features.

Implements a general-purpose higher order function to create or update datasets of GeoTiffs from existing ones by iterating over vector features: Adds all features in the source dataset to the target dataset and iterate over all features in the target dataset. For each vector feature if the vector_filter_predicate is met uses the raster_selector to select a subset of the rasters in the source dataset for which no rasters for this vector feature have previously been cut from. Each of the rasters is then cut using the raster_cutter, and the new rasters are added to the target dataset/connector.

Parameters:
Return type:

None

field bands: Optional[dict[str, Optional[list[int]]]] = None

keys: raster directory names, values: list of band indices to keep, starting with 1

Validated by:

  • validate_connectors

field cut_rasters: dict[str, list[str]] [Optional]

Normally, should not be set by hand! Dict with vector features as keys and lists of rasters cut for each vector feature as values

Validated by:

  • validate_connectors

field label_maker: Optional[LabelMaker] = None

Optional label maker. If given, will be used to recompute labels when necessary. Defaults to None

Validated by:

  • validate_connectors

field name: str [Required]

Name of dataset creator. Used as part of filename when saving.

Validated by:

  • validate_connectors

field raster_cutter: SingleRasterCutter [Required]
Validated by:

  • validate_connectors

field raster_selector: RasterSelector [Required]

Selects rasters from source to cut for a given vector feature

Validated by:

  • validate_connectors

field source_data_dir: Path [Required]
Validated by:

  • validate_connectors

  • validate_source_data_dir

field target_data_dir: Path [Required]
Validated by:

  • validate_connectors

field vector_filter_predicate: VectorFilterPredicate [Optional]

Filters vector features to be cut

Validated by:

  • validate_connectors

create(*args, **kwargs)

Create a new dataset by cutting the source dataset.

Return type:

Connector

create_or_update()[source]

Create or update target dataset.

Return type:

Connector

Returns:

connector of target dataset

cut()[source]

Cut a dataset.

Alternate name for the create method. See create_or_update for docstring.

classmethod from_json_file(json_file_path, constructor_symbol_table=None)

Load and return saved BaseModel.

Return type:

Any

Parameters:

  • json_file_path (Path | str)

  • constructor_symbol_table (dict[str, Any] | None)

save()

Save to update folder in source_data_dir.

update(*args, **kwargs)

Update the target dataset from the source dataset.

Return type:

Connector

property source_connector

Connector in source_data_dir.

property target_connector

Connector in target_data_dir.

geographer.cutters.cut_iter_over_vectors.remove_duplicates(from_list)[source]

Remove duplicates from list.

Return type:

list

Parameters:

from_list (list)

geographer.cutters.cut_rasters_around_every_vector module

Dataset cutter that cuts out rasters around vector features.

geographer.cutters.cut_rasters_around_every_vector.get_cutter_rasters_around_every_vector(source_data_dir, target_data_dir, name, mode='random', new_raster_size=512, min_new_raster_size=None, scaling_factor=None, target_raster_count=1, bands=None, random_seed=10)[source]

Return dataset cutter that creates cutouts around vector features.

In “random” or “centered” all cutouts will be of the same size. In “variable” mode the size of a cutout is the maximum of scaling_factor*size_of_bounding_box or min_new_raster_size

Return type:

DSCutterIterOverVectors

Parameters:

  • source_data_dir (Path | str) – source data dir

  • target_data_dir (Path | str) – target data dir

  • name (str) – name

  • new_raster_size (int | Tuple[int, int] | None) – new raster size for “random” and “centered” modes. Defaults to 512.

  • min_new_raster_size (int | Tuple[int, int] | None) – lower bound on raster size for “variable” mode. Defaults to None.

  • scaling_factor (float | None) – scaling factor for “variable” mode. Defaults to None.

  • target_raster_count (int) – targeted number of rasters per vector feature to create. Defaults to 1.

  • mode (Literal['random', 'centered', 'variable']) – On. Defaults to “random”.

  • bands (dict | None) – bands dict. Defaults to None.

  • random_seed (int) – random seed. Defaults to 10.

Returns:

dataset cutter

Return type:

DSCutterIterOverVectors

geographer.cutters.raster_filter_predicates module

ABC for predicates for filtering rasters.

Used in cutting functions.

class geographer.cutters.raster_filter_predicates.AlwaysTrue(**data)[source]

Bases: RasterFilterPredicate

Default raster filter predicate that always returns True.

Used when filtering is not desired.

Return type:

None

save(json_path)

Save the predicate.

Return type:

None

Parameters:

json_path (Path)

class geographer.cutters.raster_filter_predicates.RasterFilterPredicate(**data)[source]

Bases: ABC, Callable, BaseModel

ABC for predicates used to filter rasters in cutting functions.

Subclasses should implement a __call__method that has the arguments and behavior given below.

Return type:

None

save(json_path)[source]

Save the predicate.

Return type:

None

Parameters:

json_path (Path)

class geographer.cutters.raster_filter_predicates.RasterFilterRowCondition(row_series_predicate)[source]

Bases: RasterFilterPredicate

Simple RasterFilter based on row condition.

Applies a given predicate to the row in source_connector.rasters corresponding to the raster name in question.

Parameters:

row_series_predicate (RowSeriesPredicate)

Return type:

None

field row_series_predicate: RowSeriesPredicate [Required]
save(json_path)

Save the predicate.

Return type:

None

Parameters:

json_path (Path)

class geographer.cutters.raster_filter_predicates.RastersNotPreviouslyCutOnly(**data)[source]

Bases: RasterFilterPredicate

Select rasters not previously cut.

Return type:

None

save(json_path)

Save the predicate.

Return type:

None

Parameters:

json_path (Path)

class geographer.cutters.raster_filter_predicates.RowSeriesPredicate(**data)[source]

Bases: ABC, BaseModel

Row series predicate.

Apply to series, i.e. single row.

Return type:

None

geographer.cutters.raster_filter_predicates.wrap_function_as_RowSeriesPredicate(fun)[source]

Wrap a function as a RowSeriesPredicate.

Return type:

RowSeriesPredicate

Parameters:

fun (Callable[[GeoSeries | Series], bool]) – Function to wrap.

Returns:

RowSeriesPredicate.

geographer.cutters.raster_selectors module

Callable classes for selecting a sublist from a list of rasters.

Used by cutting functions.

class geographer.cutters.raster_selectors.RandomRasterSelector(**data)[source]

Bases: RasterSelector

RasterSelector that randomly selects randomly from a list of rasters.

Parameters:

target_raster_count (int)

Return type:

None

field target_raster_count: int = 1
class geographer.cutters.raster_selectors.RasterSelector(**data)[source]

Bases: Callable, BaseModel

ABC for selecting from a list of rasters.

Subclasses are used by DSCutterIterOverVectors.

Subclasses should implement a __call__method that has the arguments and behavior given below.

Return type:

None

geographer.cutters.single_raster_cutter_around_vector module

SingleRasterCutter that creates a cutout around a vector feature.

class geographer.cutters.single_raster_cutter_around_vector.SingleRasterCutterAroundVector(mode, new_raster_size=None, scaling_factor=1.2, min_new_raster_size=None, random_seed=42, **kwargs)[source]

Bases: SingleRasterCutter

SingleRasterCutter that creates a cutout around a vector feature.

SingleRasterCutter that cuts a small raster (or several contiguous such rasters if the vector feature does not fit into a single one) around each vector feature in the raster accepted by the feature filter predicate.

Parameters:

  • mode (Literal['random', 'centered', 'variable'])

  • new_raster_size (int | Tuple[int, int] | None)

  • scaling_factor (float | None)

  • min_new_raster_size (int | Tuple[int, int] | None)

  • random_seed (int)

Return type:

None

field min_new_raster_size: Optional[RasterSize] = None
field mode: Literal['random', 'centered', 'variable'] [Required]
field new_raster_size: Optional[RasterSize] = None
field random_seed: int = 42
field scaling_factor: Optional[float] = 1.2

geographer.cutters.single_raster_cutter_base module

Abstract base class for single raster cutters.

class geographer.cutters.single_raster_cutter_base.SingleRasterCutter(**data)[source]

Bases: ABC, BaseModel, RasterBandsGetterMixIn

Base class for SingleRasterCUtter.

Return type:

None

geographer.cutters.single_raster_cutter_bbox module

SingleRasterCutter that extracts pre defined bboxes from a raster.

class geographer.cutters.single_raster_cutter_bbox.SingleRasterCutterFromBBoxes(**data)[source]

Bases: SingleRasterCutter

SingleRasterCutter that extracts pre defined bboxes from a raster.

The new size of the rasters must be specified as it is used to ensure a standardised output.

Parameters:

  • new_raster_size (int | Tuple[int, int])

  • bbox_geojson_path (Path)

Return type:

None

field bbox_geojson_path: Path [Required]
Validated by:

  • path_points_to_geojson

field new_raster_size: RasterSize [Required]
Validated by:

  • new_raster_size_side_lengths_must_be_positive

  • new_raster_size_type_correctness

property new_raster_size_cols: int

Return number of columns of new raster size.

property new_raster_size_rows: int

Return number of rows of new raster size.

geographer.cutters.single_raster_cutter_grid module

SingleRasterCutter that cuts a raster to a grid of rasters.

class geographer.cutters.single_raster_cutter_grid.SingleRasterCutterToGrid(**data)[source]

Bases: SingleRasterCutter

SingleRasterCutter that cuts a raster into a grid of rasters.

Parameters:

new_raster_size (int | Tuple[int, int])

Return type:

None

field new_raster_size: RasterSize [Required]
Validated by:

  • new_raster_size_side_lengths_must_be_positive

  • new_raster_size_type_correctness

property new_raster_size_cols: int

Return number of columns of new raster size.

property new_raster_size_rows: int

Return number of rows of new raster size.

geographer.cutters.type_aliases module

Type aliases.

geographer.cutters.vector_filter_predicates module

Predicates for filtering vector features.

class geographer.cutters.vector_filter_predicates.AlwaysTrue(**data)[source]

Bases: VectorFilterPredicate

Simple vector feature filter predicate that always returns True.

Return type:

None

class geographer.cutters.vector_filter_predicates.FilterVectorByRowCondition(row_series_predicate, mode)[source]

Bases: VectorFilterPredicate

Simple GeomFilterPredicate that uses a predicate on rows.

Applies a predicate to the row in the source or target vectors corresponding to the vector feature name in question.

Parameters:

  • row_series_predicate (collections.abc.Callable[[GeoSeries | Series], bool])

  • mode (Literal['source', 'target'])

Return type:

None

class geographer.cutters.vector_filter_predicates.IsVectorMissingRasters(**data)[source]

Bases: VectorFilterPredicate

VectorFilterPredicate that uses raster counts as criterion.

Checks whether a vector feature has fewer rasters than a specified target raster count.

Parameters:

target_raster_count (int)

Return type:

None

field target_raster_count: int = 1
class geographer.cutters.vector_filter_predicates.OnlyThisVector(this_vector_name)[source]

Bases: VectorFilterPredicate

Filter out all vector features except a given one.

Simple vector feature filter initialized with a vector feature this_vector_name.

Returns True if and only if the vector feature under consideration is equal to this_vector_name.

Parameters:

this_vector_name (str | int)

Return type:

None

class geographer.cutters.vector_filter_predicates.VectorFilterPredicate(**data)[source]

Bases: BaseModel, Callable

ABC for predicates for filtering vector features.

To be used in cutting functions.

Subclasses should implement a __call__method that has the arguments and behavior given below.

Return type:

None

Module contents