GeoscienceAustralia / wagl

Python library for standardising satellite imagery into an Analysis Ready Data (ARD) form
Apache License 2.0
30 stars 7 forks source link

wagl


wagl is a Python package for producing standarised imagery in the form of:

The luigi task workflow for producing NBAR for a Landsat 5TM scene is given below.

Supported Satellites and Sensors


Requirements


Installation


wagl Package

The wagl pacakage can be installed via:

$ python setup.py install --prefix=<prefix>

Additional HDF5 compression filters (optional)

Additional compression filters can be used via HDF5's dynamically loaded filters. Essentially the filter needs to be compiled against the HDF5 library, and installed into HDF5's plugin path, or a path of your choosing, and set the HDF5_PLUGIN_PATH environment variable. The filters are then automatically accessible by HDF5 via the integer code assigned to the filter.

Mafisc compression filter

Mafisc combines both a bitshuffling filter and lzma compression filter in order to get the best compression possible at the cost of lower compression speeds. To install the mafisc compression filter, follow these instructions.

Bitshuffle

The bitshuffle filter can be installed from source, or conda via the supplied conda recipe. It utilises a bitshuffling filter on top of either a lz4 or lzf compression filter.

Basic command line useage


Using the local scheduler:

$ luigi --module wagl.multifile_workflow ARD --workflow NBAR --level1-list scenes.txt --outdir /some/path --local-scheduler --workers 4

Using the central scheduler:

$ luigid --background --pidfile <PATH_TO_PIDFILE> --logdir <PATH_TO_LOGDIR> --state-path <PATH_TO_STATEFILE>

$ luigi --module wagl.multifile_workflow ARD --level1-list scenes.txt --workflow STANDARD --outdir /some/path --workers 4

$ luigi --module wagl.multifile_workflow ARD --level1-list scenes.txt --workflow NBAR --outdir /some/path --workers 4

$ luigi --module wagl.multifile_workflow ARD --level1-list scenes.txt --workflow SBT --outdir /some/path --workers 4