Contribute
=======

Report your past augmentation practice
---------------------------

Fill in this  `short survey <https://forms.office.com/pages/responsepage.aspx?id=Wht7-jR7h0OUrtLBeN7O4YAi1YW5UytAn2uCO4MRqSBUODhKUjdTUVFJM0tTN0NIOU5NSE01V1dUQS4u&fswReload=1&fswNavStart=1625240091445>`_ 


Propose your customize functions
---------------------------



``beacon_aug`` will automatically mode your operator to augmentation operator


* Name: Class name has to be the same with python file name

  * if classname exist in json file in `standard <https://github.com/adobe-research/beacon-aug/tree/main/beacon_aug/generator/standard>`_ :

    * append a ``custom`` library function to the existing beacon_aug class

  * else:

    * create a beacon_aug class with the same name that only support the ``custom`` library

See example in `KeepSizeCrop.py <https://github.com/adobe-research/beacon-aug/tree/main/beacon_aug/generator/custom/KeepSizeCrop.py>`_

Test the operator function itself:    

.. code-block:: python

   from beacon_aug.generator.custom.KeepSizeCrop import KeepSizeCrop
   op = KeepSizeCrop( scale=(0.08, 1.0))
   img_transformed = op(img )

Test the augmentation wrapper:   

.. code-block:: python

   import beacon_aug as BA
   aug = BA.KeepSizeCrop(p=1, height=64,width=64, library="custom")
   image_auged = aug(image=image)["image"]



Contribute to the repo
---------------------------
Create your own branch ``git branch -d {Your_LDAP}/{branch_name}``