Fix on eigenvalue signs in quadric surfaces

[SteSeg]

I believe there was a wrong sign definition in the eigenvalue signs of the coefficient matrix of quadric surfaces (gqs.py):

  eigenvalues, eigenvectors = np.linalg.eig(Aa)
  signs = np.array([0, 0, 0])
  for i in range(0, 3):
      if eigenvalues[i] > -1 * gq_tol:
          signs[i] = 1  # should be -1 
      else:
          signs[i] = -1  # should be 1

Now the conversion of fng_str and fng_w is not giving error anymore:

[pshriwise]

Thanks for this fix @SteSeg! Is it possible to identify one of the surfaces in the model for which this fix is successful so we can use it to design a test?

Also, it appears that the first of the two commits here are mainly style updates. I'd love to go through and update style at some point, but for now can you remove those (for now) to keep the changes relevant to the purpose of the PR?

[SteSeg]

My bad, I realized it was the previous merged PR to solve the problems I was facing with quadrics. I'll close this PR.