Getting Started

We start from setting up a Python environment for modelx, and learn about core modelx objects before sample exercises.

Setting up Python and modelx

Setting up a Python environment for modelx is pretty straight-forward. Download the WinPython distribution customized for modelx from here.

This tutorial assumes you’re using the latest modelx and spyde-modelx, so if you have downloaded the distribution before, make sure you update modelx and spyder-modelx to the latest versions by following the instruction here, or download the latest distribution from the site above.

Unzip the downloaded zip file. You can unzip it from the Windows default context menu by right-clicking on the file and select Extract All…, or instead you can use your favorite third party tool for unzipping.


Windows context menu for extracting zip file

In this tutorial, we use Spyder as a graphical user interface to modelx. The downloaded WinPython comes with Spyder and Spyder plugin for modelx pre-installed and pre-configured, so no need to install them separately.

Within the unzipped folder, find Spyder.exe and start Spyder by double-clicking it.


Spyder.exe in unzipped folder

The Spyder window shows up. You may have modelx widgets showing already upon startup, but if you don’t, bring them up by going to View menu in the menu bar and select Pane and then select three modelx items (MxExplorer, MxDataView, MxAnalyzer) at the bottom.


modelx Widgets in View menu

Now you should be able to see the 3 widgets. You can move them around and change their locations by unlocking panes. To unlock panes, go to View menu and uncheck Lock panes and toolbars item.


Lock and Unlock Panes

The last widget to prepare for modelx is MxConsole, an IPython console that communicates with the modelx widgets. modelx works fine in Spyder’s default IPython consoles, but the default consoles do not communicate with the modelx widgets, so you want to use MxConsoles instead. You should have IPython console pane and a tab named Console 1/A. Right-click on the tab, and from the context menu, select New MxConsole.


Console Tab Context Menu

A new tab named MxConsole 2/A is created, and after a few seconds, an IPython session starts in the MxConsole and waits for your input.



Overview of core modelx objects

modelx is designed to let the users build models consisting of a few types of objects. Model, Space and Cells are the most important types of objects. Before getting started with the first example, you want to have an idea on what these types of objects are.

Model, Space and Cells are to modelx what workbook, worksheet and cells are to a spreadsheet program respectively, although there are differences. The diagram below illustrates containment relationships between those objects.

blockdiag Model1 Space1 Space2 Cells1 Space3

Model, Space and Cells

Models are the top level objects that contain all the other types of modelx objects. Models can be saved to files and loaded back again.

Directly under Models, there are Spaces. Spaces serve as containers, separating contents in Models into components. Spaces contain Cells objects and other Spaces, allowing tree structures of objects to form within Models.

Spaces also serve as the namespaces for the formulas associated to the Spaces themselves or to the Cells contained in them.

Cells are objects that can have formulas and hold values, just like spreadsheet cells can have formulas and values. Cells values are either calculated by their formulas or assigned as input by the user. We will learn how to define Cells formulas through the examples soon.