scaling/scaling/scaling.py

import pint


class DimensionConverter(object):
    def __init__(self):
        # Initialise unit definitions
        self.ureg = pint.UnitRegistry()

    def _convert(self, x, length_scale_factor, input_unit, target_unit):

        # Calculate unit conversion factor
        input_unit = self.ureg(input_unit)
        target_unit = self.ureg(target_unit)
        unit_conversion_factor = input_unit.to(target_unit).magnitude

        # Calculate Froude scaling factor
        froude_scale_factor = length_scale_factor**(
            input_unit.dimensionality['[length]'] * self.LENGTH_EXPONENT +
            input_unit.dimensionality['[time]'] * self.TIME_EXPONENT +
            input_unit.dimensionality['[mass]'] * self.MASS_EXPONENT)

        # Scale values
        x_scaled = x * froude_scale_factor

        # Convert to output units
        x_scaled *= unit_conversion_factor

        return x_scaled

    def proto_to_model(self,
                       x_proto,
                       length_scale,
                       input_unit,
                       target_unit,
                       index_input_unit=None,
                       index_target_unit=None):
        """Convert prototype value(s) to model value(s) in specified units.

        Args:
            x_proto:            prototype values (array_like, or pandas dataframe)
            length_scale:       ratio between proto and model dimensions (float)
            input_unit:         unit of input (string)
            target_unit:        unit of output (string)
            index_input_unit:   unit of input index (dataframe only)
            index_target_unit:  unit of output index (dataframe only)

        Returns:
            input values in model scale
        """

        length_scale_factor = 1 / length_scale

        # Convert values
        x_model = self._convert(x_proto, length_scale_factor, input_unit,
                                target_unit)

        # Convert index (dataframe or series only)
        if (index_input_unit is not None) and (index_target_unit is not None):
            if type(x_model).__name__ in ['DataFrame', 'Series']:
                x_model.index = self._convert(
                    x_model.index, length_scale_factor, index_input_unit,
                    index_target_unit)
            else:
                raise ValueError("'index_input_unit' and 'index_target_unit' "
                                 "can only be used when input is dataframe")

        return x_model

    def model_to_proto(self,
                       x_model,
                       length_scale,
                       input_unit,
                       target_unit,
                       index_input_unit=None,
                       index_target_unit=None):
        """Convert model value(s) to prototype value(s) in specified units.

        Args:
            x_model:            model values (array_like, or pandas dataframe)
            length_scale:       ratio between proto and model dimensions (float)
            input_unit:         unit of input (string)
            target_unit:        unit of output (string)
            index_input_unit:   unit of input index (dataframe only)
            index_target_unit:  unit of output index (dataframe only)

        Returns:
            input values in prototype scale
        """

        length_scale_factor = length_scale

        # Convert values
        x_proto = self._convert(x_model, length_scale_factor, input_unit,
                                target_unit)

        # Convert index (dataframe or series only)
        if (index_input_unit is not None) and (index_target_unit is not None):
            if type(x_proto).__name__ in ['DataFrame', 'Series']:
                x_proto.index = self._convert(
                    x_proto.index, length_scale_factor, index_input_unit,
                    index_target_unit)
            else:
                raise ValueError("'index_input_unit' and 'index_target_unit' "
                                 "can only be used when input is dataframe")

        return x_proto

    def dimensions(self, unit):
        """Get unit dimensions.

        Args:
            unit:  unit name or symbol (string)

        Returns:
            string containing unit dimensions
        """

        # Get dimensions of unit
        dims = self.ureg(unit).dimensionality

        dim_data = {'L': '[length]', 'M': '[mass]', 'T': '[time]'}

        s = ''
        for symbol, key in dim_data.items():
            exponent = dims[key]
            if exponent != 0:
                s += '{}^{:g} '.format(symbol, exponent)

        return s.strip()

    def scaling_exponent(self, unit):
        """Convert prototype value(s) to model value(s) in specified units.

        Args:
            unit:  unit of quantity to be scaled (string)

        Returns:
            scaling factor
        """
        # Calculate Froude scaling factor
        scaling_exponent = (
            self.ureg(unit).dimensionality['[length]'] * self.LENGTH_EXPONENT +
            self.ureg(unit).dimensionality['[time]'] * self.TIME_EXPONENT +
            self.ureg(unit).dimensionality['[mass]'] * self.MASS_EXPONENT)

        return scaling_exponent


class FroudeConverter(DimensionConverter):
    def __init__(self):
        super(FroudeConverter, self).__init__()

        # Define Froude scaling relationships
        self.LENGTH_EXPONENT = 1
        self.TIME_EXPONENT = 1 / 2
        self.MASS_EXPONENT = 3


class ReynoldsConverter(DimensionConverter):
    def __init__(self):
        super(ReynoldsConverter, self).__init__()

        # Define Reynolds scaling relationships
        self.LENGTH_EXPONENT = 1
        self.TIME_EXPONENT = 2
        self.MASS_EXPONENT = 3