Provide defaults for shapes and scales

[janezd]

The Color widget allows the user to set the color that is used in subsequent visualizations. Add the functionality to set the variables' shape (for discrete variables) and scales (for numeric).

This will require some changes in Variable class: adding an attribute for storing shapes and scales, obviously.

Shapes should be, I suppose, handled in a similar way as colors: the related methods and attributes are monkey-patched into the base class. While monkey patching is not a common and very welcome practice in Python, it seems more appropriate than lettings the Variable (as a base class) know that it's being visualized (GUI is a level above the data storing/mining classes).

As for scales for numeric attributes, I would have them in the base class, so they can be set to defaults when the data is loaded. Defaults should be based on reasonably rounded min/max from the data.

• [ ] Add attributes for scale (range and whether to use log-scale)
• [ ] Inverting the color palette
• [ ] Add attributes for shapes
• [ ] Extend and rename the color widget
• [ ] Modify the visualizations to use the scale
• [ ] Modify some discretization methods (and any other applicable methods) to use the scale
• [ ] Widgets allow overriding the scale in plot options; there is also a button to copy the current scale from the widget