Remove warnings supression as noted by Issue #229

[mmcky]

This PR removes the warning supression from arma.py and matrix_eqn.py.

TODO: The next step is to fix the underlying cause of the warning.

[mmcky]

The arma warnings are a result of:

w, spect = self.spectral_density(two_pi=False)

the spect array is complex. With the example on the quant-econ.net page the coeeficient on the j term are all zero. In general is it safe to use the real component of this array? If so we can add a.real on spec to send matplotlib an array of reals.

[mmcky]

Moving related comments here from Issue #229

Comment in arma.py

# == Ignore unnecessary warnings concerning casting complex variables back to
# floats == #

@jstac In general is it safe to use the real component of this array?

[mmcky]

Comment in matrix_eqn.py

# == Suppress warnings from checking conditioning of matrices == #

[mmcky]

@jstac would you mind reviewing this spectral density question when you have time? Its in relation to some warnings we are trying to remove in the quantecon.py package.

[oyamad]

What is the status of this PR?

[mmcky]

this PR has removed the instructions to ignore and suppress warnings. So the last step is to review arma.py and matrix_eqn.py to resolve the root cause of the issued warnings.

[oyamad]

Is it possible to merge this PR as is and open another issue for discussion on arma.py and matrix_eqn.py? I find it a bit annoying that warnings are all ignored in the whole session once quantecon is imported...

[mmcky]

Sure. I will have a quick review of this this afternoon to double check that all of the suppression statements have been removed and then I will setup a separate issue to fix arma.py and matrix_eqn.py.

[oyamad]

Thanks!

[coveralls]

Coverage increased (+0.7%) to 86.633% when pulling 0302277 on adjust-warnings into 712cae6 on master.

[mmcky]

This looks ready to merge (passes tests).

Note: this will cause warnings to appear when using the library.

#271
#272