Spyder plugin#

Spyder is a popular open-source Python IDE, and it’s bundled in with major Python distributions such as Anaconda and WinPython by default. Spyder allows plugins to be installed to add extra features to itself.

Spyder plugin for modelx enriches user interface to modelx in Spyder. The plugin adds custom IPython consoles and GUI widgets for using modelx in Spyder.

Th plugin consists of 4 components, MxConsole, MxExplorer, MxDataViewer and MxAnalyzer.

MxConsoles appear as tabs in Spyder’s default IPython console widget, and host custom IPython shells. The custom shells function exactly as the default IPython shells, except that the custom shells are connected to the plugin widgets. You should use the custom shells instead of Spyder’s default shells in order for the other plugin widgets to interface with the user’s Python sessions. The plugin widgets do not interface with Python running in default IPython consoles.

MxExplorer is the main plugin widgets, and it contains an object tree pane showing the object tree of a selected Model. From the context menu on the object tree, you can perform various operations, such as creating and deleting modelx objects, defining names bound to objects in the selected Model, etc. On the right-hand side of MxExplorer, there are a property pane and formula pane for showing detail information.

MxDataView lets you see values of vector and tabular data in modelx objects, such as list, set, tuple, dict, numpy array, in addition to pandas DataFrame, Series, and Index in a tabular format. It also shows the values and types of scalar objects, such as int and str.

MxAnalyzer enables you to visually trace calculation dependency. For a specified combination of a Cell and arguments called node, MxAnalyzer shows a tree of preceding nodes or descending nodes.

The plugin widgets are “dockable” as Spyder’s default widgets, meaning you can detach those widgets from the Spyder’s main window to have their own separate windows, and “dock” them back in the main window at different positions to rearrange the widgets positions in the main window as you like.

MxExplorer and MxConsole#

If the plugin is installed successfully, then you should see the plugin components showing or hiding in the Spyder window, except for MxConsole.

If you don’t find any plugin widgets, go to View->Panes menu, and check MxExplorer, MxDataViewer and MxAnalyzer. If you don’t find such menu items then check the plugin supports your Spyder version and update your Spyder if it’s not supported. If you still can’t find plugin widgets, post an issue on modelx github site.

_images/SpyderMainMenuForModelx.png

MxExplorer looks like below just after startup.

_images/NoMxConsole.png

Right-click ont the IPython console tab in the lower right pane, then click Open a MxConsole menu.

_images/IPythonConsoleMenu.png

Once the MxConsole is ready, then the plugin widgets get populated with clean panes.

_images/BlankMxExplorer.png

The MxConsole works exactly the same as a regular IPython console, except that the console is connected to the plugin widgets and they can be used to interact with modelx components in the Python session in the console. To test the behaviour, create a new model and space in the MxConsole like this:

>>> import modelx as mx

>>> model, space = mx.new_model(), mx.new_space()

>>> cells = space.new_cells()

The modelx explorer shows the component tree of the created space.

_images/MxExplorerTreeSample.png

Select Cells2 in the tree by double-clicking on it. You will see the pane on the right-hand side of MxExplorer gets populated with properties of Cells2 and its Formula appears in the lower half of the pane.

_images/MxExplorerTreeSample2.png

MxExplorer can also list the formulas of all Cells in a selected Space. Select a Space in the tree, right-click to bring up the context menu, and the click Show Formulas. The list of the formulas appears to the right of the model tree in MxExplorer.

_images/CodeListExample.png

MxDataViewer#

MxDataViewer lets you see values of vector and tabular data in modelx objects, such as list, set, tuple, dict, numpy array, in addition to pandas DataFrame, Series, and Index in a tabular format. It also shows the values and types of scalar objects, such as int and str.

_images/MxDataViewerDataFrameExample.png

MxDataViewer showing a pandas DataFrame#

_images/MxDataViewerListInDict1.png

MxDataViewer showing a dict and its nested lists#

If MxDataViewer widget is not shown, Go to View->Panes menu as you did with MxExploer, and check MxDataViewer.

_images/SpyderMainMenuForModelx.png

To specify the object to examine, click the object in the object tree in MxExplorer, right click to open the context menu, and select either Select in DataView or Select in New DataView. You can alternatively click one of the tool bar buttons located above the tree. Then enter arguments in the box in MxDataView if the object has parameters, and click the Update button.

_images/MenuInMxExplorerForMxDataView1.png

Items in toolbar and context menu of MxExplorer for selecting objects in MxDataView#

MxAnalyzer#

MxAnalyzer enables you to visually trace calculation dependency.

If MxDataView widget is not shown, Go to View->Panes menu as you did with MxExploer, and check MxAnalyzer.

_images/MxAnalyzerMenu.png

MxAnalyzer has two tabs, one for tracing the precedents, nodes that the selected node depends on, and the other for tracing the dependents, nodes that depends on the selected nodes.

Select Precedents tab and enter an expression that returns a Cells object in the text box titled Expression, and arguments to the Cell in the Args box next to it. The Python expression is evaluated in the global namespace of the Python session in the active MxConsole. Then MxAnalyzer shows a tree of Cell nodes directly or indirectly used in calculating the value of the specified node.

Alternatively, you can set the object to analyze from MxExplorer. By clicking the Analyze Selected item in the MxExplorer context menu, the selected object is set in MxAnalyzer (currently, only Cells can be selected).

_images/MxExplorerAnalyzeMenu.png

Analyze Selected context menu item on MxExplorer#

Each tab in MxAnalyzer has two split panes in it. The upper pane is for the dependency tree, and the lower pane shows the formula of the selected object in the tree.

_images/EnhancedMxAnalyzer.png

Dependency tree and Formula pane in MxAnalyzer#

The Value column in the dependency tree shows the type of each object if it’s not a scalar value. By double-clicking on the Value column or right-clicking and selecting Show Value from the context menu, the value of the selected element is shown in a pop-up window. The pop-up window shows values of list, set, tuple, dict, numpy array, in addition to pandas DataFrame, Series, and Index in a tabular format.

_images/MxAnalyzerShowValueMenu.png

Show Value context menu item on MxAnalyzer#

_images/MxAnalyzerPopUpSeries.png

Pop-up window showing the values of a Series#