Currently, the Reference documentation is not correctly extracted.
See for instance http://eartrack.readthedocs.io/en/latest/source/references/binarisation.html

When everyone is OK, we can plan a final release:

• Check the documentation
• Add Zenodo to create a doi from the release:
  • see https://guides.github.com/activities/citable-code/
• Create a release : https://help.github.com/articles/creating-releases
  • Update metainfo.ini (if needed) and conda/meta.yaml
  • Tag v1.0.0
• Generate the conda packages on Linux, Windows and Mac OS X
• Add the doi to the doc

@artzet-s
Build the documentation
Upload the doc on readthedoc

Add a Citation section in the README and the Doc.

Add a conda package.
Upload the conda package on openalea channel

On the reference documentation there are missing docstrings. Either exclude these functions or add adoc string (the latter action is recommended ;) )

In binarisation (http://eartrack.readthedocs.io/en/latest/source/references/binarisation.html):

• meanshift_hsv
• color_tree
• color_threshold_tree_cabin_1
  NB: you can alias the last function with a better name

In eartrack, everything is documented. Congrat!!!
However, you can improve and homogeneize the docstring:

• This function analyses -> Analyse (Remove This function when possible)
• Recall for doc string: A first line as a summary. Then blank line. Then detail description of what the function do.
  Modify get_view_angles

There is currently only a large test.
There is a need to add small unit test to test each function independantly.
This can be linked to issue #4

The Notebook ear_tracking_tutorial failed to load on my computer with the install following the documentation and also on nbviewer.

Import function from a readable string describing the decision tree (R output).

Use AppVeyor and Travis to generate conda packages and upload them on anaconda

I think that OptParse is more enhanced than argparse.
What I propose is to replace (in main.py) argparse by OptParse
(see vclimb.py)
Then, I recommend to add a function main.py which will be called at command execution.

@brichet: What do you think about it?

Make a notebook example with more details on method @brichet