/multitemporal

Efficient, chainable time series processing of raster stacks.

Primary LanguagePythonGNU General Public License v2.0GPL-2.0

Multitemporal

(c) 2018 Applied Geosolutions, LLC

This library provides an efficient means of flexibly performing time series analysis on stacks of gridded data. There is a core python application that breaks the processing job into pieces and launches workers to perform the processing. Each worker has a configurable sequence of processing steps. All the inputs and each step are prescribed in a user-conigured JSON files.

Authors:

  • Bobby H. Braswell (rbraswell at ags.io)
  • Justin Fisk
  • Ian Cooke

Supported in part by NASA Interdisciplinary Science Grant (NASA-IDS) #NNX14AD31G -- Drought-induced vegetation change and fire in Amazonian forests: past, present, and future to University of New Hampshire (Michael Palace, PI)

Current supported modules:

Also see this directory

correlate.pyx
diff_ts.pyx
gapfill.pyx
interpolate.pyx
multiply.pyx
passthrough.pyx
phenology.pyx
recomposite.pyx
screen.pyx
simpletrend.pyx
summation.pyx
validmask.pyx

Dev Setup

Build a container, set an alias to let you run tests using your host machine's working copy, then run the test suite:

$ time docker build . -t mt --no-cache
$ alias rmt="docker run --rm -it -v ${HOME}/src/multitemporal/:/multitemporal"
$ time rmt mt python3 setup.py build_ext --inplace
$ rmt mt pytest -vv -s