Remove problematic characters from feature dirs

[xylar]

This merge removes characters that are sometimes problematic from the directory names of features.

This merge also fixes some incorrect docstrings, adding a missing component and correcting the use of tags for reading features.

[pep8speaks]

Hello @xylar! Thanks for opening this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file geometric_features/geometric_features.py:

Line 83:81: E501 line too long (100 > 80 characters)
Line 176:81: E501 line too long (100 > 80 characters)
Line 226:81: E501 line too long (100 > 80 characters)
Line 301:81: E501 line too long (96 > 80 characters)

[xylar]

@proteanplanet, this should make it safe to use the feature names you originally had, in case you want to do that.

[xylar]

Testing

I merged all the features and split them again, proving that none of the existing names is affected by removing these potential problem characters. Here was my script:

#!/usr/bin/env python
from geometric_features import GeometricFeatures, FeatureCollection


gf = GeometricFeatures(cacheLocation='./geometric_data')

fc = FeatureCollection()
for component in ['bedmachine', 'bedmap2', 'iceshelves', 'landice',
                  'natural_earth', 'ocean']:
    for objectType in ['point', 'transect', 'region']:
        if objectType in gf.allFeaturesAndTags[component]:
            print(component, objectType)
            fc.merge(gf.read(componentName=component, objectType=objectType))


gf = GeometricFeatures(cacheLocation='./new_geometric_data')
gf.split(fc)