picoraster

Small Python library for processing large raster images.

Currently a work in progress.

Example usage

source = AWSLandsat8Source("LC08_L1TP_139045_20170304_20170316_01_T1", band="8")

# Lazily create a band and build a description of processing steps
band = Band(source) \
    .and_then(Resize(extents)) \
    .and_then(HistogramAdjust()) \
    .and_then(Reproject(crs))

# Forces computation
array = band.render_to_array()

band.render_to_file("output.tif")

Installation

Installing GDAL is the most challenging part. Installing directly from PyPI is historically unlikely to work.

First, install numpy: pip install numpy

Then, choose one of the following:

• install with a system package manager
  • Ubuntu: sudo apt install libgdal-dev
  • MacOS: brew install gdal
• install from conda-forge: conda install -c conda-forge gdal
• compile manually

Afterwards, the correct Python bindings can be installed with

pip install GDAL==$(gdal-config --version) --global-option=build_ext --global-option="-I/usr/include/gdal" 

Finally, pip install picoraster

Running tests

python -m src.tests