The classes here provide support for using custom classes with matplotlib, eg those that do not expose the array interface but know how to converter themselves to arrays. It also supoprts classes with units and units conversion. Use cases include converters for custom objects, eg a list of datetime objects, as well as for objects that are unit aware. We don’t assume any particular units implementation, rather a units implementation must provide a ConversionInterface, and the register with the Registry converter dictionary. For example, here is a complete implementation which support plotting with native datetime objects
import matplotlib.units as units import matplotlib.dates as dates import matplotlib.ticker as ticker import datetime
class DateConverter(units.ConversionInterface):
@staticmethod def convert(value, unit, axis):
‘convert value to a scalar or array’ return dates.date2num(value)@staticmethod def axisinfo(unit, axis):
‘return major and minor tick locators and formatters’ if unit!=’date’: return None majloc = dates.AutoDateLocator() majfmt = dates.AutoDateFormatter(majloc) return AxisInfo(majloc=majloc,
majfmt=majfmt, label=’date’)@staticmethod def default_units(x, axis):
‘return the default unit for x or None’ return ‘date’# finally we register our object type with a converter units.registry[datetime.date] = DateConverter()
information to support default axis labeling and tick labeling, and default limits
majloc and minloc: TickLocators for the major and minor ticks majfmt and minfmt: TickFormatters for the major and minor ticks label: the default axis label default_limits: the default min, max of the axis if no data is present If any of the above are None, the axis will simply use the default
The minimal interface for a converter to take custom instances (or sequences) and convert them to values mpl can use