cv2 transform support 2 modes; (#885)

* cv2 transform support 2 modes;

* update docs

Author: zsdonghao

tensorlayer/prepro.py

@@ -551,7 +551,7 @@ def affine_transform(x, transform_matrix, channel_index=2, fill_mode='nearest',
 apply_transform = affine_transform
 
 
-def affine_transform_cv2(x, transform_matrix, flags=None, borderMode=None):
+def affine_transform_cv2(x, transform_matrix, flags=None, border_mode='constant'):
     """Return transformed images by given an affine matrix in OpenCV format (x is width). (Powered by OpenCV2, faster than ``tl.prepro.affine_transform``)
 
     Parameters
@@ -560,21 +560,28 @@ def affine_transform_cv2(x, transform_matrix, flags=None, borderMode=None):
         An image with dimension of [row, col, channel] (default).
     transform_matrix : numpy.array
         A transform matrix, OpenCV format.
+    border_mode : str
+        - `constant`, pad the image with a constant value (i.e. black or 0)
+        - `replicate`, the row or column at the very edge of the original is replicated to the extra border.
 
     Examples
     --------
     >>> M_shear = tl.prepro.affine_shear_matrix(intensity=0.2, is_random=False)
     >>> M_zoom = tl.prepro.affine_zoom_matrix(zoom_range=0.8)
     >>> M_combined = M_shear.dot(M_zoom)
-    >>> result = affine_transform_cv2(image, M_combined)
+    >>> result = tl.prepro.affine_transform_cv2(image, M_combined)
     """
     rows, cols = x.shape[0], x.shape[1]
     if flags is None:
         flags = cv2.INTER_AREA
-    if borderMode is None:
-        borderMode = cv2.BORDER_CONSTANT
+    if border_mode is 'constant':
+        border_mode = cv2.BORDER_CONSTANT
+    elif border_mode is 'replicate':
+        border_mode = cv2.BORDER_REPLICATE
+    else:
+        raise Exception("unsupport border_mode, check cv.BORDER_ for more details.")
     return cv2.warpAffine(x, transform_matrix[0:2,:], \
-            (cols,rows), flags=flags, borderMode=borderMode)
+            (cols,rows), flags=flags, borderMode=border_mode)
 
 
 def affine_transform_keypoints(coords_list, transform_matrix):