Replacing matplotlib.mlab.griddata by scipy.interpolate.griddata

[hbueno]

As discussed in issue #146, I'd some problems when I tried to run PrismMesh.carvetopo because of matplotlib.mlab.griddata. So, @leouieda suggested to replace the matplotlib.mlab.griddata functions by scipy.interpolate.griddata.

@leouieda I don't know what's the best interpolation method to set as default. With scipy.interpolate.griddata our options are linear, nearest and cubic. Here I'm using cubic.

Checklist:

• Make tests for new code
• Create/update docstrings
• Code follows PEP8 style conventions
• Code and docs have been spellchecked
• Changelog entry
• Include new dependencies in docs, requirements.txt, README
• Documentation builds properly
• All tests pass
• Can be merged

[leouieda]

Thanks for this Henrique!

@leouieda I don't know what's the best interpolation method to set as default. With scipy.interpolate.griddata our options are linear, nearest and cubic. Here I'm using cubic.

Cubic is fine.

Please replace the mlab call in fatiando.gridder and any other places you can find. I think that is the only one though. To make sure, check that all tests and recipes run on you machine.

[hbueno]

@leouieda, please is there any special application for the nearest neighbor method 'nn' of mlab?
If not, I think we can remove it from the gridder.py#L248 and use only the nearest of scipy.interpolate.griddata.

[eusoubrasileiro]

Good bye 'nearest'. We say farewell
On Nov 24, 2014 8:20 PM, "Henrique Bueno dos Santos" <
notifications@github.com> wrote:

@leouieda https://github.com/leouieda, please is there any special
application for the nearest neighbor method 'nn' of mlab?
If not, I think we can remove it from the gridder.py#L248
https://github.com/fatiando/fatiando/blob/master/fatiando/gridder.py#L248
and use only the nearest of scipy.interpolate.griddata.

—
Reply to this email directly or view it on GitHub
#148 (comment).

[eusoubrasileiro]

Ops farewell to the wrong defunct.
Ops two defunct with same name and just different nicknames.

Cheers fatiandovillage
On Nov 25, 2014 8:51 AM, "André" eusoubrasileiro@gmail.com wrote:

Good bye 'nearest'. We say farewell
On Nov 24, 2014 8:20 PM, "Henrique Bueno dos Santos" <
notifications@github.com> wrote:

@leouieda https://github.com/leouieda, please is there any special
application for the nearest neighbor method 'nn' of mlab?
If not, I think we can remove it from the gridder.py#L248
https://github.com/fatiando/fatiando/blob/master/fatiando/gridder.py#L248
and use only the nearest of scipy.interpolate.griddata.

—
Reply to this email directly or view it on GitHub
#148 (comment).

[leouieda]

@hbueno you can tick the "Update docstrings" task (you've already update the interp docstring).

You should leave the changelog for last. I've update the develop docs putting it last in the checklist.
The reason for that is that the changelog almost always causes merge conflicts (like the one you're having now). It happens because master updated the changelog at the same time you did. To avoid that, merge the master branch into yours before editing the changelog. For now, you'll have to update your master and merge it into your branch:

git checkout master
git pull https://github.com/fatiando/fatiando.git master
git checkout YOUR_BRANCH
git merge master

When you run this now you'll have "merge conflicts". You'll need to resolve them before we can merge the PR. See https://help.github.com/articles/resolving-a-merge-conflict-from-the-command-line/ for a guide on how to do that.

[leouieda]

@hbueno great job! After this one little fix I think it's ready to merge.

[leouieda]

@hbueno looks good to merge. Do you have anything else to add?

[hbueno]

No. Thanks...