Configuration¶

DO NOT EDIT OUTSIDE MARKERS

The pipeline is configured through a config.yaml file that combines settings for every stage. Values shown below are the defaults unless marked as required.

Schema¶

Key	Type	Default	Required
`base_folder`	string	`"output"`	yes
`download.site_code`	string	—	yes
`download.year_month`	string (YYYYMM)	—	yes
`download.flight_lines`	list[string]	—	yes
`download.product_code`	string	`"DP1.30006.001"`	no
`convert.export_ancillary`	bool	`true`	no
`convert.export_brdf_config`	bool	`true`	no
`topo_brdf.num_cpus`	int	`8`	no
`topo_brdf.file_type`	string	`"envi"`	no
`topo_brdf.corrections`	list[string]	`["topo","brdf"]`	no
`topo_brdf.bad_bands`	list[int]	`[]`	no
`topo_brdf.anc_files`	map[string,str]	—	conditional†
`topo_brdf.export.output_dir`	string	`"./"`	no
`topo_brdf.export.suffix`	string	`"_corrected_envi"`	no
`topo_brdf.export.image`	bool	`true`	no
`topo_brdf.export.masks`	bool	`true`	no
`topo_brdf.export.coeffs`	bool	`true`	no
`resample.method`	string	`"convolution"`	no
`resample.sensors`	list[string]	`["Landsat_8"]`	no
`mask.polygon_layer`	string	—	no
`mask.raster_crs_override`	string\|int	—	no
`mask.polygons_crs_override`	string\|int	—	no
`mask.plot_output`	bool	`false`	no
`sort.remote_prefix`	string	`""`	no
`sort.sync_files`	bool	`true`	no
`postprocess.reflectance_offset`	int	`0`	no

† required when topo_brdf.file_type is "envi".

Example¶

base_folder: output

download:
  site_code: NIWO
  year_month: "202008"
  flight_lines: ["FL1", "FL2"]
  product_code: DP1.30006.001

convert:
  export_ancillary: true
  export_brdf_config: true

topo_brdf:
  num_cpus: 8
  file_type: envi
  corrections: ["topo", "brdf"]
  bad_bands: []
  anc_files: {}
  export:
    output_dir: ./corrected
    suffix: _corrected_envi
    image: true
    masks: true
    coeffs: true

resample:
  method: convolution
  sensors: ["Landsat_8"]

mask:
  polygon_layer: polygons.geojson
  plot_output: false

sort:
  remote_prefix: ""
  sync_files: true

postprocess:
  reflectance_offset: 0

CLI overrides¶

The cscal-pipeline entry point automatically runs the download stage before spinning up per-flightline workers. Use --max-workers to opt into parallel processing once the .h5 files are present. Command-line options override the corresponding entries in config.yaml:

bin/jefe.py BASE_FOLDER SITE YEAR_MONTH FL1,FL2 sets base_folder, download.site_code, download.year_month and download.flight_lines.
--polygon_layer_path → mask.polygon_layer
--reflectance-offset → postprocess.reflectance_offset
--remote-prefix → sort.remote_prefix
--no-sync sets sort.sync_files to false
--max-workers sets the ThreadPool concurrency for go_forth_and_multiply()