Tests failing (v0.5) on OSX

[TimKeeping]

OS: Mac Sierra 10.12 on airbook early 2014
Fatiando version: 0.5
Python distribution: Anaconda
Python version: 2.7
Script: pytest
Error: 1 failed, 214 passed, 8 warnings in 52.74 seconds

Description:
I created a new environment in Anaconda, opened a terminal through the environment and cut and paste from the website and ran the following commands:

conda install fatiando --channel conda-forge
conda install fatiando mayavi --channel conda-forge
conda install pytest
python -c "import fatiando; fatiando.test()"

I then deleted all environments except root, creted new environments and installed all but Mayavi and still have the same result.

However, running the pytest in the root environment does not result in errors.

The full result is shown below.

  assert_array_almost_equal(eql[0].predicted(), data, decimal=1)

E AssertionError:

E Arrays are not almost equal to 1 decimals

E

E (mismatch 0.5%)

E x: array([-0.1, -0.5, -0.5, ..., -0.3, -0.3, -0.3])

E y: array([-0.3, -0.4, -0.4, ..., -0.3, -0.3, -0.3])

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/gravmag/tests/test_eqlayer.py:33: AssertionError

=============================== warnings summary ===============================

::fatiando.mesher.PointGrid.split

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/gridder.py:319: RuntimeWarning: invalid value encountered in double_scalars

dy = (y2 - y1)/(ny - 1)

::fatiando.geothermal.climsig.SingleChange

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/inversion/optimization.py:282: RuntimeWarning: Exited because couldn't take a step without increasing the objective function. Might not have achieved convergence. Try inscreasing the max number of step attempts allowed.

RuntimeWarning)

::fatiando.inversion

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/inversion/optimization.py:177: RuntimeWarning: invalid value encountered in double_scalars

if newmisfit > misfit or abs((newmisfit - misfit) / misfit) < tol:

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/inversion/optimization.py:185: RuntimeWarning: Exited because maximum iterations reached. Might not have achieved convergence. Try inscreasing the maximum number of iterations allowed.

RuntimeWarning)

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/inversion/init.py:8: RuntimeWarning: overflow encountered in exp

predicted (or modeled) data and (optionally) the Jacobian (or sensitivity)

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/inversion/init.py:8: RuntimeWarning: overflow encountered in multiply

predicted (or modeled) data and (optionally) the Jacobian (or sensitivity)

::test_impulse_response

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/seismic/conv.py:214: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future

ricker = np.zeros(nw)

::test_rc_shorter_than_wavelet

//anaconda/envs/fatiando/lib/python2.7/site-packages/fatiando/seismic/conv.py:71: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future

synth_l[:, j] = np.convolve(rc[:, j], w, mode='full')[aux:-aux]

-- Docs: http://doc.pytest.org/en/latest/warnings.html

============== 1 failed, 214 passed, 8 warnings in 52.74 seconds ===============

(//anaconda/envs/fatiando)

[leouieda]

@TimKeeping thanks for creating the issue!

This might have been something that we've already fixed but hasn't been released yet.
Could you try installing the latest development version and running the tests? Instructions are here: http://www.fatiando.org/dev/install.html#installing-the-latest-development-version

Replace the conda install fatiando --channel conda-forge with:

pip install https://github.com/fatiando/fatiando/archive/master.zip

We've been overdue for the 0.6 release but I can't find the time to fix some of the final issues. I might just release what we have right now and leave some things for a later release.