Installing InVEST¶

Note

The natcap.invest python package is currently only supported in Python 2.7. Other versions of python may be supported at a later date.

Warning

Python 2.7.11 or later is required to be able to use the InVEST Recreation model on Windows.

Binary Dependencies¶

InVEST itself depends only on python packages, but many of these package dependencies depend on low-level libraries or have complex build processes. In recent history, some of these packages (notably, numpy and scipy) have started to release precompiled binary packages of their own, removing the need to install these packages through a system package manager. Others, however, remain easiest to install through a package manager.

Linux¶

Linux users have it easy, as almost every package required to use natcap.invest is available in the package repositories. The provided commands will install only the libararies and binaries that are needed, allowing pip to install the rest.

Ubuntu & Debian¶

Attention

The package versions in the debian:stable repositories often lag far behind the latest releases. It may be necessary to install a later version of a libarary from a different package repository, or else build the library from source.



Mac OS X¶

The easiest way to install binary packages on Mac OS X is through a package manager such as Homebrew:

$brew install gdal spatialindex pyqt matplotlib  The GDAL, PyQt and matplotlib packages include their respective python packages. The others will allow their corresponding python packages to be compiled against these binaries via pip. Windows¶ While many packages are available for Windows on the Python Package Index, some may need to be fetched from a different source. Many are available from Christogh Gohlke’s unofficial build page: http://www.lfd.uci.edu/~gohlke/pythonlibs/ PyQt4 installers can also be downloaded from the Riverbank Computing website. Python Dependencies¶ Dependencies for natcap.invest are listed in requirements.txt: # # Any lines with "# pip-only" at the end will be processed by # scripts/convert-requirements-to-conda-yml.py as though it can only be found # on pip. GDAL>=2.0,<3.0 Pyro4==4.41 # pip-only pandas>=0.22.0 numpy>=1.11.0 Rtree>=0.8.2 scipy>=0.16.1 Shapely six pygeoprocessing>=1.2.3 taskgraph[niced_processes]>=0.6.1 chardet>=3.0.4 matplotlib  Optional Qt User Interface¶ InVEST’s user interface is built with PyQt. Because of the hefty binary requirement of Qt and the relative difficulty of installing PyQt, these dependencies will not be installed with the standard pip install natcap.invest. Several of these dependencies are available as extras, however, and can be installed via pip: $ pip install natcap.invest[ui]


These extras do not include a distribution of PyQt, so you will need to install PyQt in an appropriate way on your system. PyQt4 is not currently available from the Python Package Index, but other sources and package managers allow for straightforward installation on Windows, Mac OS X, and Linux.

The InVEST user interface uses a wrapper layer to support both PyQt4 and PyQt5, one of which must be installed on your system for the UI to be able to run. If both are installed, PyQt5 is preferred, but you can force the UI to use PyQt4 by defining an environment variable before launching the UI:

$QT_API=pyqt4 invest pollination  Installing from Source¶ Note Windows users will find best compilation results by using the MSVC compiler, which can be downloaded from the Microsoft website. See the python wiki page on compilation under Windows for more information. Assuming you have a C/C++ compiler installed and configured for your system, and dependencies installed, the easiest way to install InVEST as a python package is: $ pip install natcap.invest


If you are working within virtual environments, there is a documented issue with namespaces in setuptools that may cause problems when importing packages within the natcap namespace. The current workaround is to use these extra pip flags: