Installing optional dependencies#
Depending on the installation method (Conda, pip) and the IDAES version, not all of IDAES’s dependencies will be installed by default.
Installing these optional dependencies might require additional steps, described below, in addition to the default installation steps described elsewhere in the documentation.
For pip installations#
Important
Users who installed IDAES by running pip install idaes-pse
should follow these steps.
When installing IDAES using pip, optional dependencies can be installed by specifying one or more “extras_require
targets”:
# install the `ui` target
pip install "idaes-pse[ui]"
# install the `ui` and `omlt` targets
pip install "idaes-pse[ui,omlt]"
Important
The pip install
argument should be wrapped in double quotes ("
) when it contains square brackets. Otherwise, it might cause an error to be reported by the shell/command interface used to invoke the command.
Available optional dependencies targets#
As of IDAES 2.6, the following extras_require
targets are available:
ui
: for the IDAES User Interface (UI) (UI documentation)grid
: for the IDAES Grid integrationomlt
: for the OMLT integrationcoolprop
: for theidaes.models.properties.modular_properties.coolprop
property package
Specifying optional dependencies when installing a prerelease version of IDAES#
pip install "idaes-pse[ui,omlt] @ git+https://github.com/IDAES/idaes-pse@main"
For Conda installations#
Important
These steps apply to users who installed IDAES by running conda install <...>
. They do not apply if IDAES was installed using pip.
As of IDAES 2.1, the IDAES Conda package available already includes several dependencies that are optional for pip installations.
This difference is due to a variety of reasons, including the fact that not all optional dependencies are available as Conda packages (e.g. OMLT); or that Conda does not have an equivalent of extras_require
targets to specify optional dependencies at install-time.
The consequence of this is that some optional dependencies might be unavailable if IDAES was installed using Conda.
Therefore, we encourage users who encounter issues after a Conda installation of IDAES to try installing IDAES again using pip instead and see if the issues are resolved.