rast_converter

Conversion of rasters into tables. This module provides a command line tool and a function that both ingest a raster file and produce 3 column parquet table with coordinates and band value.

See the data_ingestion notebook for templates of this.

scalenav.rast_converter.check_crs(source, in_crs)

Check that the source CRS is provided.

Parameters:
  • source (DatasetReader) – Source raster file

  • in_crs (str) – Alternative CRS to use

Returns:

Returns the source csrs if it is provided, or an alternative.

Return type:

string

scalenav.rast_converter.check_nodata(source)

Checking the nodata field in a raster, assume it’s 0 if not present

Parameters:

source (DatasetReader) – A raster data source

Returns:

Based on whether a nodata values can be read from the source, return this value or 0 if not

Return type:

numeric

scalenav.rast_converter.check_out_crs(val)

Check the requested out crs is valid

Parameters:

val (str) – A string representing a CRS, for example ‘epsg:4326’

Returns:

A CRS object

Return type:

rasterio.CRS

scalenav.rast_converter.check_path(in_path)
What are we checking ?

UNDER DEV - input contains one of the desired file resolutions. - if folder file, extract all raster format files from it. - if specific file, then just use that. - some robustness to user input should be embedded here. For example when providing folder path: ‘/the/folder/with/rast/’ and /the/folder/with/rast’ as two potential accepted values. - also relative and absolute paths.

Parameters:

in_path (str) – Path to a raster file or folder with rasters.

Returns:

A string path or a list of paths depending whether the input is a file path or folder respectively

Return type:

str

scalenav.rast_converter.infer_dtype(source)

Simple function to avoid failures when data type is not recongnised, default assign float32()

Parameters:

source (DatasetReader) – A raster data source

Returns:

A type for the data schema specification, default is float32()

Return type:

np.dtype

scalenav.rast_converter.rast_convert_core(src, transform, win=None, band=1)

The core of the rast conversion can be put here in order to centralise the most efficient workflow that can be then used in the CL tool and function.

Parameters:
  • src (DatasetReader) – Data read from a filestream

  • transform (a rasterio transform)

  • win (a rasterio window, optional) – default None

  • band (int, optional) – by default 1

Returns:

A data frame of the transformed raster window.

Return type:

DataFrame | pyarrow.table