Regression results part2

[mortonjt]

No description provided.

[mortonjt]

@antgonza @josenavas @wasade review? Thanks!

[coveralls]

Coverage increased (+0.06%) to 99.468% when pulling c378361 on mortonjt:regression_results_part2 into 356896d on biocore:regression_results.

[coveralls]

Coverage increased (+0.06%) to 99.468% when pulling f3b7d94 on mortonjt:regression_results_part2 into 356896d on biocore:regression_results.

[coveralls]

Coverage increased (+0.06%) to 99.468% when pulling f3b7d94 on mortonjt:regression_results_part2 into 356896d on biocore:regression_results.

[antgonza]

check=True in the comment but False in the code, which one is it?

[mortonjt]

This is the reason why I added this comment.

Although this has been resolved in scikit-bio, it is only in the development version - which would mean that this would have to depend on the development version of scikit-bio.

I was thinking about raising an issue, and making the fix at the next release of scikit-bio. What do you think?

[antgonza]

That fix has been merged in skbio and you could use the dev version of skbio (in travis, setup.py, etc) until there is release, will this work?

[mortonjt]

Ok. Then this should be ok

On Jul 31, 2016 6:38 AM, "Antonio Gonzalez" notifications@github.com
wrote:

In gneiss/_summary.py
#28 (comment):

•    ValueError:
  

•        Cannot perform projection into Aitchison simplex
  

•        if `feature_names` is not specified.
  

•    """
  

•    self._check_projection(project)
  

•    coef = pd.DataFrame()
  

•    for r in self.results:
  

•        c = r.params
  

•        c.name = r.model.endog_names
  

•        coef = coef.append(c)
  

•    if project:
  

•        # `check=True` due to type issue resolved here
  

•        # https://github.com/biocore/scikit-bio/pull/1396
  

•        c = ilr_inv(coef.values.T, basis=self.basis, check=False).T
  

That fix has been merged in skbio and you could use the dev version of
skbio (in travis, setup.py, etc) until there is release, will this work?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/biocore/gneiss/pull/28/files/f3b7d9472d3eb4ebebe004b42141f66d20f9461a#r72904299,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AD_a3Tyk2nu8mJS-28A2xkXGhFb67ZI_ks5qbKU6gaJpZM4JXppU
.

[antgonza]

some comments

[coveralls]

Coverage increased (+0.05%) to 99.465% when pulling b3e11d6 on mortonjt:regression_results_part2 into 356896d on biocore:regression_results.

[coveralls]

Coverage increased (+0.05%) to 99.465% when pulling be4a07e on mortonjt:regression_results_part2 into 356896d on biocore:regression_results.