How to get values given by `mpmath`?

I have no experience in using mpmath package. I just recently used it for some small tests and had the following returns:

from mpmath import *
loggamma(3+2j)
mpc(real='-0.031639059373961193', imag='2.022193197501327')
atan2(2,3)
mpf('0.5880026035475675')

I would just like to know how to retrieve the real values indicated by mpmath (-0.031..., 2.022..., 0.588...) in order to manipulate them for further calculations.

Edit - as an example, using numpy, I immediately get a real number that I can immediately use:

import numpy as np
np.arctan2(2,3)
0.5880026035475675

So, what is the best way to have the same thing with mpmath?

Solution

First, do you realize that numpy functions produce np.float64 values, not 'plain' python floats?

In [43]: np.atan2(2,3)
Out[43]: np.float64(0.5880026035475675)

The latest numpy version makes this explicit. On earlier ones you can verify this with type(...).

With mpmath functions, the return value is mpf/c object, which has all the needed extended precision attributes.

In [44]: x=mp.atan2(2,3)
In [46]: x
Out[46]: mpf('0.5880026035475675')

It's easy to get the equivalent python float:

In [47]: float(x)
Out[47]: 0.5880026035475675

To display the mpf value, use:

In [48]: mp.nstr(x)
Out[48]: '0.588003'

or with more precision values:

In [49]: mp.nstr(x,n=30)
Out[49]: '0.588002603547567503916582154488'

Or as direct print (rather than string):

In [50]: mp.nprint(x,n=25)
0.5880026035475675039165822

Similarly for the complex value:

In [52]: y = mp.loggamma(3+2j)    
In [53]: y
Out[53]: mpc(real='-0.031639059373961193', imag='2.022193197501327')

In [54]: complex(y)
Out[54]: (-0.03163905937396119+2.022193197501327j)

In [55]: mp.nprint(y, n=30)
(-0.0316390593739611925006016690531 + 2.02219319750132697777189605404j)

To get the two parts separately:

In [56]: y.real, y.imag
Out[56]: (mpf('-0.031639059373961193'), mpf('2.022193197501327'))

While nprint is nice for displaying values, you usually don't need to convert things to base python. You can do a lot math directly with the mp objects:

In [57]: x*y
Out[57]: mpc(real='-0.018603849285685252', imag='1.1890548650069606')