komanawa.easy_nc_accessor#

Support for easy access to NetCDF4 datasets produced by Komanawa Solutions Ltd.

Submodules#

Classes#

CompressedSpatialAccessor

This supports easy access to compressed spatial datasets. That is data which has n dimensions with only 1 dimension of space. For example, this could be a 2d Dataset of time, space. Where the "space" dimension has point values (e.g. unique x,y). This is opposed to an uncompressed spatial dataset which would have n dimensions with 2 dimensions of space (x, y)

TileIndexAccessor

Support to select geospatial tiles from a directory of netcdf files. The netcdf files must have the following attributes:

Package Contents#

class CompressedSpatialAccessor(datapath, active_index_name='active_index', grid_x_name='grid_x', grid_y_name='grid_y', loc_x_name='x', loc_y_name='y')#

Bases: _BaseAccessor

Inheritance diagram of komanawa.easy_nc_accessor.CompressedSpatialAccessor

This supports easy access to compressed spatial datasets. That is data which has n dimensions with only 1 dimension of space. For example, this could be a 2d Dataset of time, space. Where the “space” dimension has point values (e.g. unique x,y). This is opposed to an uncompressed spatial dataset which would have n dimensions with 2 dimensions of space (x, y)

Parameters:

  • datapath – path to the netcdf file

  • active_index_name – name of the active index variable in the netcdf file

  • grid_x_name – name of the grid x variable in the netcdf file

  • grid_y_name – name of the grid y variable in the netcdf file

  • loc_x_name – name of the x location variable in the netcdf file (to support compressed spatial dimensions)

  • loc_y_name – name of the y location variable in the netcdf file (to support compressed spatial dimensions)

check_raster_crs(raster_path)#

check the crs of a raster

Parameters:

raster_path – path to raster

Returns:

check_shape_crs(shapefile_path=None, shapefile=None)#

check the crs of a shapefile

Parameters:

  • shapefile_path – path to shapefile

  • shapefile – shapefile object (geodataframe/geoseries), note shapely geometries do not have a crs

Returns:

None

static close(fig=None)#

a shortcut to be able to close plots without importing matplotlib.pyplot

Parameters:

fig – see matplotlib.pyplot.close()

Returns:

get_2d_spatial_zero(dtype=float)#

get a 2d array of zeros with the same shape as the spatial 2d shape

Parameters:

dtype – data type of the array

Returns:

get_active_index()#

read the active index of the dataset, this is a boolean 2d array of the same shape as the spatial 2d shape.

True = has data, False = no data

Returns:

get_closest_loc_to_point(nztmx, nztmy, coords_out_domain='raise')#

get the closest spatial index(s) to a point in the spatial domain

Parameters:

  • nztmx – single or array of x coordinates

  • nztmy – single or array of y coordinates

  • coords_out_domain – [‘raise’, ‘coerce’ or ‘pass’]. What to do if the coordinates are outside the domain

  • ‘raise’: raise a ValueError

  • ‘coerce’: return -1 for out of domain coords (note that this may still be a valid index, so care must be taken)

  • ‘pass’: returns the closest index, but this may be WELL outside the domain

Returns:

index(s) of the closest point(s). an integer if single point, or an array of integers if multiple points

  • note that this is the index of the spatial points (which does not include inactivate cells)

get_xlim_ylim()#

read the dataset spatial limits

Returns:

x_min, x_max, y_min, y_max

plot_2d(array, vmin=None, vmax=None, title=None, ax=None, color_bar=True, base_map_path=None, cbar_lab=None, cbarlabelpad=15, contour=False, norm=None, contour_levels=None, cmap='plasma', label_contours=False, contour_label_format='%1.1f', figsize=(10, 10), **kwargs)#

Plot a 2D array with a color map and optional contours.

Parameters:

  • array – they array to plot (of i,j)

  • vmin – vmin to pass to the plot, None, number (as per matplotlib), str (e.g., 1th, 1st, 5th), the percentile to use

  • vmax – vmax to pass to the plot, None, number (as per matplotlib), str (e.g., 1th, 1st, 5th), the percentile to use

  • title – a title to include on the plot

  • ax – None or a matplotlib ax to plot on top of, not frequently used (execpt in subplots)

  • color_bar – Boolean if true plot a color bar scale

  • base_map_path – None or a path to a raster basemap to plot under the data. if the raster is RGB then it is transformed to grayscale.

  • cbar_lab – string to label the cbar

  • contour – boolean if true print black contours on the map

  • norm – None or matplotlib.colors.Normalize, if None then no normalisation is applied (other than vmin/vmax)

  • contour_levels – see levels in matplotlib.pyplot.contour, or float/int, If float/int then contour levels are calculated from array.min()//level x level to array.max()//level x level

  • cmap – color map to use

  • label_contours – boolean if true then print labels in line with the contours

  • contour_label_format – format for the contour label (see fmt in plt.clabel)

  • figsize – figsize (ignored if ax is not None)

  • kwargs – additional kwargs to plot to pass to pcolormesh

Returns:

fig, ax

static show()#

a shortcut to be able to show plots without importing matplotlib.pyplot

Returns:

spatial_1d_to_spatial_2d(array, missing_value=np.nan)#

convert a 1d (collapsed) spatial array to a 2d spatial array

Parameters:

  • array – 1d array to convert

  • missing_value – value to use for missing values (to support integer arrays)

Returns:

array (self.spatial_2d_shape)

spatial_2d_to_raster(path, array, dtype=np.float32, compression=True)#

saves a 2d array as a raster geotiff file

Parameters:

  • path – path to save the raster

  • array – array to save (must be 2d model array)

  • dtype – gdal data type to save as

  • compression – boolean if True use compression (LZW, options = ‘COMPRESS=LZW’, ‘PREDICTOR={p}’, ‘TILED=YES’) where p=2 for int and 3 for float

Returns:

spatial_2d_to_spatial_1d(array)#

convert a 2d spatial array to a 1d (collapsed) spatial array

Parameters:

array – 2d array to convert

Returns:

array (1d)

class TileIndexAccessor(data_dir, save_index_path)#

Bases: _common_functions

Inheritance diagram of komanawa.easy_nc_accessor.TileIndexAccessor

Support to select geospatial tiles from a directory of netcdf files. The netcdf files must have the following attributes:

  • xmin: the minimum x coordinate of the tile

  • ymin: the minimum y coordinate of the tile

  • xmax: the maximum x coordinate of the tile

  • ymax: the maximum y coordinate of the tile

The netcdf files can have the following attributes:

  • tile_number: the tile number

  • start_date: the start date of the tile (inclusive)

  • end_date: the end date of the tile (inclusive)

All NetCDF files within the directory and it’s children are searched for the above attributes.

The key mechanism is to create a tile index from the netcdf files in the directory and then return a dataframe of the tiles that fall within a bounding box, shapefile for all datasets. The end user can then do subsequent filtering based on the other parameters (e.g. tile number, start date, and end date)

Parameters:

  • data_dir – path to the directory containing the netcdf files

  • save_index_path – path to save the index file

check_raster_crs(raster_path)#

check the crs of a raster

Parameters:

raster_path – path to raster

Returns:

check_shape_crs(shapefile_path=None, shapefile=None)#

check the crs of a shapefile

Parameters:

  • shapefile_path – path to shapefile

  • shapefile – shapefile object (geodataframe/geoseries), note shapely geometries do not have a crs

Returns:

None

export_tiles_to_shapefile(outpath, tiles=None)#

Export the tile extents to a shapefile

Parameters:

  • outpath – path to save the shapefile

  • tiles – None (Export all) or tile numbers to export.

Returns:

None

get_index(recalc=False)#

get / make a tile index from the netcdf files in the data_dir

Returns:

dataframe of all tiles that fall within the shapefile. columns are:

  • ’tile_path’: path of the tile relative to the data_dir

  • ’tile_number’: the tile number. This can be missing without causing an exception

  • ’tile_xmin’: the minimum x coordinate of the tile

  • ’tile_ymin’: the minimum y coordinate of the tile

  • ’tile_xmax’: the maximum x coordinate of the tile

  • ’tile_ymax’: the maximum y coordinate of the tile

  • ’start_date’: the start date of the tile (inclusive) This can be missing without causing an exception

  • ’end_date’: the end date of the tile (inclusive) This can be missing without causing an exception

get_tiles_from_extent(xs, ys)#

get the tile paths from a bounding box

Parameters:

  • xs – the x coordinates of the bounding box / extent. the min/max of the passed x coordinates are used

  • ys – the y coordinates of the bounding box / extent. the min/max of the passed y coordinates are used

Returns:

dataframe of all tiles that fall within the shapefile. columns are:

  • ’tile_path’: pathlib.Path to the tile

  • ’tile_number’: the tile number. This can be missing without causing an exception

  • ’tile_xmin’: the minimum x coordinate of the tile

  • ’tile_ymin’: the minimum y coordinate of the tile

  • ’tile_xmax’: the maximum x coordinate of the tile

  • ’tile_ymax’: the maximum y coordinate of the tile

  • ’start_date’: the start date of the tile (inclusive) This can be missing without causing an exception

  • ’end_date’: the end date of the tile (inclusive) This can be missing without causing an exception

get_tiles_from_shapefile(shapefile_path, check_crs=True)#

get the tile paths from a shapefile

Parameters:

  • shapefile_path – path to the shapefile

  • check_crs – boolean if True check the crs of the shapefile

Returns:

dataframe of all tiles that fall within the shapefile. columns are:

  • ’tile_path’: pathlib.Path to the tile

  • ’tile_number’: the tile number. This can be missing without causing an exception

  • ’tile_xmin’: the minimum x coordinate of the tile

  • ’tile_ymin’: the minimum y coordinate of the tile

  • ’tile_xmax’: the maximum x coordinate of the tile

  • ’tile_ymax’: the maximum y coordinate of the tile

  • ’start_date’: the start date of the tile (inclusive) This can be missing without causing an exception

  • ’end_date’: the end date of the tile (inclusive) This can be missing without causing an exception

plot_tiles(tiles=None, basemap_path=None, ax=None, figsize=(10, 10), linewidth=4, linecolor='r', label_tiles=True)#

plot the tiles on an optional basemap

Parameters:

  • tiles – None (plot all) or tile numbers to export.

  • basemap_path – path to a basemap

  • ax – None or a matplotlib axis to plot on

  • figsize – if ax is None, the size of the figure to create

  • linewidth – line width of the tile edges

  • linecolor – color of the tile edges

  • label_tiles – if True, label the tiles with their tile number at the center of the tile

Returns:

fig, ax

On this page