Module enthought.​contexts.​adapter.​unit_conversion_adapter

Experimental adapter that, given a context, adapts it to a particular set of units.

For example, say you have a context with logs that have unknown units, and you want to ensure logs that you use are of a specific set. Below, we have a log 'depth' in meters that we want to ensure is in 'feet'. The adapter enforces this.

>>> from enthought.units.length import meters, feet
>>> from enthought.numerical_modeling.units.api import UnitArray
>>> from enthought.contexts.api import AdaptedDataContext, DataContext
>>> from enthought.contexts.api import UnitConversionAdapter
>>> old_log = UnitArray((1,2,3),units=meters)
>>> context=AdaptedDataContext(context=DataContext())
>>> context['depth'] = old_log
>>> getitem_units = {'depth':feet}
>>> adapter = UnitConversionAdapter(getitem_units=getitem_units)
>>> context.push_adapter(adapter)
>>> new_log = context['depth']
>>> print new_log.units
0.3048*m
>>> print new_log
[ 3.2808399   6.56167979  9.84251969]

Note that currently , get and set both convert to the same unit system. This is quite possibly not what we want to happen. You want gets in one units system and sets in another.

Also Note that by using this adapter approach with multiple adapter layers, we potentially have a serious performance impact. Imagine having multiple unit conversion adapters chained together. Each of these would convert the data to new units individually resulting in a large number of new log creations and calculation steps. It would be much better to "reduce" these conversions by multiplying them together and then apply the conversion once to the entire array. The desire to do this may lead to an alternative architecture where we have a list of "converters" or adapters that are attached to a GeoContext that are used when dealing with array conversions. This would allow some external object to examine all of them at once and them make intelligent decisions about how to optimize their application.