Now catching 'LinAlgErrors' in '.fit' method

[FelSiq]

Resolves #92, partially.

Changes

• Catching an unified set of exceptions in pre-computations, extraction and post-computations. In particular, now pre-computations will not crash when a np.linalg.LinAlgError is raised.

Still up to discussion

• How to handle non-categorical y values? (opened the issue Support for regression tasks or issue an error while fitting real-valued 'y' #98 to address this in the future)
• Is it worth link the user custom arguments with the CCA max_iter argument?

[codecov]

Codecov Report

Merging #94 into master will increase coverage by 0.07%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #94      +/-   ##
==========================================
+ Coverage   98.85%   98.93%   +0.07%     
==========================================
  Files          15       15              
  Lines        2187     2528     +341     
==========================================
+ Hits         2162     2501     +339     
- Misses         25       27       +2     

Impacted Files	Coverage Δ
pymfe/complexity.py	99.61% <ø> (-0.39%)	⬇️
pymfe/info_theory.py	99.01% <ø> (ø)
pymfe/_internal.py	97.40% <100.00%> (+0.03%)	⬆️
pymfe/_utils.py	100.00% <100.00%> (ø)
pymfe/mfe.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 80a0fc8...69b729f. Read the comment docs.

[ealcobaca]

Thanks, Felipe, we can find other errors that can be caught to increase the '_EXCEPTIONS' list later :)