[yt-dev] Issue #1259: magnetic field units for athena frontend (yt_analysis/yt)

[Chang-Goo Kim]

New issue 1259: magnetic field units for athena frontend
https://bitbucket.org/yt_analysis/yt/issues/1259/magnetic-field-units-for-athena-frontend

Chang-Goo Kim:

I set unit conversion factors with 'unit_override' keyword. These are the units I'm using

>>> print ya.unit_base

{'length_unit': (1.0, 'pc'),
 'mass_unit': (2.38858753789e-24, 'g/cm**3*pc**3'),
 'time_unit': (1.0, 's*pc/km')}

This should set the magnetic field unit as follows:

>>> mu=yt.YTQuantity(ya.unit_base['mass_unit'][0],ya.unit_base['mass_unit'][1])
>>> lu=yt.YTQuantity(ya.unit_base['length_unit'][0],ya.unit_base['length_unit'][1])
>>> tu=yt.YTQuantity(ya.unit_base['time_unit'][0],ya.unit_base['time_unit'][1])
>>> mag_unit=(np.sqrt(4*np.pi*mu/lu)/tu)
>>> print mag_unit.convert_to_units('gauss')

5.4786746797e-07 gauss

But, what I found in the unit_registry is a bit strange number.

>>> print ds.unit_registry['code_magnetic']

(3.5449077018110318,
 sqrt((mass))/(sqrt((length))*(time)),
 0.0,
 '\\rm{code\\ magnetic}')

It seems that "magnetic_unit" was set correctly in yt/frontends/athena/data_structures.py:

        self.magnetic_unit = np.sqrt(4*np.pi * self.mass_unit /
                                     (self.time_unit**2 * self.length_unit))
        self.magnetic_unit.convert_to_units("gauss")