From d39f47a8bcc4b29f02d0a0b6e72cba821669776f Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Mon, 21 Sep 2020 23:44:53 +0530
Subject: [PATCH 01/11] Introduce 'balance' arg in flow_from_directory

---
 keras_preprocessing/image/image_data_generator.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/keras_preprocessing/image/image_data_generator.py b/keras_preprocessing/image/image_data_generator.py
index 77d88147..f5f1e9c9 100644
--- a/keras_preprocessing/image/image_data_generator.py
+++ b/keras_preprocessing/image/image_data_generator.py
@@ -441,6 +441,7 @@ def flow(self,
 
     def flow_from_directory(self,
                             directory,
+                            balance = False,
                             target_size=(256, 256),
                             color_mode='rgb',
                             classes=None,
@@ -465,6 +466,7 @@ def flow_from_directory(self,
                 See [this script](
                 https://gist.github.com/fchollet/0830affa1f7f19fd47b06d4cf89ed44d)
                 for more details.
+            balance: Boolean, handles data imbalance if True
             target_size: Tuple of integers `(height, width)`,
                 default: `(256, 256)`.
                 The dimensions to which all images found will be resized.

From e594e3cbf55fe30bea6f03a6cb44f7f78d41f8da Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Tue, 22 Sep 2020 00:20:29 +0530
Subject: [PATCH 02/11] Introduce 'balance' arg in DirectoryIterator init

---
 keras_preprocessing/image/directory_iterator.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/keras_preprocessing/image/directory_iterator.py b/keras_preprocessing/image/directory_iterator.py
index 3a829b4f..36c3eb95 100644
--- a/keras_preprocessing/image/directory_iterator.py
+++ b/keras_preprocessing/image/directory_iterator.py
@@ -25,6 +25,7 @@ class DirectoryIterator(BatchFromFilesMixin, Iterator):
             via the `classes` argument.
         image_data_generator: Instance of `ImageDataGenerator`
             to use for random transformations and normalization.
+        balance: Boolean, will handle data imbalance if set to True.
         target_size: tuple of integers, dimensions to resize input images to.
         color_mode: One of `"rgb"`, `"rgba"`, `"grayscale"`.
             Color mode to read images.
@@ -76,6 +77,7 @@ def __new__(cls, *args, **kwargs):
     def __init__(self,
                  directory,
                  image_data_generator,
+                 balance = False,
                  target_size=(256, 256),
                  color_mode='rgb',
                  classes=None,

From 8b9110fee1e1c5264a8ce17002ee6bf09cdb425d Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Tue, 22 Sep 2020 00:25:19 +0530
Subject: [PATCH 03/11] Add '_balance_config' attr to DirectoryIterator

_balance_config: dict, supposed to store relevant key-value pairs
important for further steps to handle data imbalance

_balance_config is not generated for validation subset as it does not
need resampling (oversampling/handing data imbalance)

Yet to implement _make_balance_config
---
 keras_preprocessing/image/directory_iterator.py | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/keras_preprocessing/image/directory_iterator.py b/keras_preprocessing/image/directory_iterator.py
index 36c3eb95..aa05dd92 100644
--- a/keras_preprocessing/image/directory_iterator.py
+++ b/keras_preprocessing/image/directory_iterator.py
@@ -11,7 +11,8 @@
 import numpy as np
 
 from .iterator import BatchFromFilesMixin, Iterator
-from .utils import _list_valid_filenames_in_directory
+from .utils import (_list_valid_filenames_in_directory,
+                    _make_balance_config)
 
 
 class DirectoryIterator(BatchFromFilesMixin, Iterator):
@@ -103,6 +104,13 @@ def __init__(self,
                                                             subset,
                                                             interpolation)
         self.directory = directory
+
+        if balance and subset != 'validation':
+            self._balance_config = _make_balance_config(directory, 
+                image_data_generator._validation_split)
+        else:
+            self._balance_config = None
+
         self.classes = classes
         if class_mode not in self.allowed_class_modes:
             raise ValueError('Invalid class_mode: {}; expected one of: {}'

From 0be974b0a40357ee35e674e21b3fbbc8277a6f8e Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Tue, 22 Sep 2020 00:47:19 +0530
Subject: [PATCH 04/11] Implement '_make_balance_config'

Scans the directory and generates a dict object which stores the
configurations that will be used for handling data imbalance.

Yet to implement '_generate_class_count'
---
 keras_preprocessing/image/utils.py | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index bc3e6886..c9fb63e2 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -228,6 +228,29 @@ def _list_valid_filenames_in_directory(directory, white_list_formats, split,
     return classes, filenames
 
 
+def _make_balance_config(directory, validation_split):
+    """Scans the directory to make a config dictionary to handle data imbalance.
+    # Arguments
+        directory: string, absolute path to the directory
+        validation_split: float, validation split
+            Default: None
+    # Returns
+        balance_config: dictionary, specs needed to handle data imbalance
+            'majority': integer, number of samples in the majority class
+    """
+    class_count = _generate_class_count(directory)
+
+    # Get the sample count of the majority class
+    majority_class_count = class_count[max(class_count, key = class_count.get)]
+    if validation_split:
+        majority_class_count = int(majority_class_count*(1 - validation_split)) + 1        
+
+    balance_config = {
+        'majority': majority_class_count
+    }
+
+    return balance_config
+
 def array_to_img(x, data_format='channels_last', scale=True, dtype='float32'):
     """Converts a 3D Numpy array to a PIL Image instance.
 

From 12d844d0442f4a79912d7797538098da80e15232 Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Tue, 22 Sep 2020 00:52:40 +0530
Subject: [PATCH 05/11] Implement '_generate_class_count'

Scans the directory and generates a dict object which maintains a
class count (no. of image samples in each class/category)
---
 keras_preprocessing/image/utils.py | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index c9fb63e2..b5871b18 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -227,9 +227,28 @@ def _list_valid_filenames_in_directory(directory, white_list_formats, split,
 
     return classes, filenames
 
+def _generate_class_count(directory):
+    """Maintain sample count of each class in the directory.
+    
+    # Arguments
+        directory: string, absolute path to the directory
+    # Returns
+        class_count: dictionary, sample count for each class
+    """
+    
+    class_count = {}
+
+    if directory[-1] != '/':
+        directory += '/'
+
+    for category in os.listdir(directory):
+      class_count[category] = len(os.listdir(directory + category))
+
+    return class_count
 
 def _make_balance_config(directory, validation_split):
     """Scans the directory to make a config dictionary to handle data imbalance.
+    
     # Arguments
         directory: string, absolute path to the directory
         validation_split: float, validation split

From 7cb2c103de381235a05bd0348f8f196c29e067a2 Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Fri, 2 Oct 2020 18:49:26 +0530
Subject: [PATCH 06/11] Refactor '_generate_class_count'

---
 keras_preprocessing/image/utils.py | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index b5871b18..aa777a21 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -238,11 +238,9 @@ def _generate_class_count(directory):
     
     class_count = {}
 
-    if directory[-1] != '/':
-        directory += '/'
-
     for category in os.listdir(directory):
-      class_count[category] = len(os.listdir(directory + category))
+      category_directory = os.path.join(directory, category)
+      class_count[category] = len(os.listdir(category_directory))
 
     return class_count
 

From a81511fd988a0672e39fc30539b61cf3464647a6 Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Fri, 2 Oct 2020 19:17:50 +0530
Subject: [PATCH 07/11] Add oversampling in list valid files in directory

Yet to implement '_settle_debt'
---
 keras_preprocessing/image/directory_iterator.py |  3 ++-
 keras_preprocessing/image/utils.py              | 13 ++++++++++++-
 2 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/keras_preprocessing/image/directory_iterator.py b/keras_preprocessing/image/directory_iterator.py
index aa05dd92..8da1dc05 100644
--- a/keras_preprocessing/image/directory_iterator.py
+++ b/keras_preprocessing/image/directory_iterator.py
@@ -139,7 +139,8 @@ def __init__(self,
             results.append(
                 pool.apply_async(_list_valid_filenames_in_directory,
                                  (dirpath, self.white_list_formats, self.split,
-                                  self.class_indices, follow_links)))
+                                  self.class_indices, follow_links,
+                                  self._balance_config)))
         classes_list = []
         for res in results:
             classes, filenames = res.get()
diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index aa777a21..8e9d1038 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -183,7 +183,8 @@ def _recursive_list(subpath):
 
 
 def _list_valid_filenames_in_directory(directory, white_list_formats, split,
-                                       class_indices, follow_links):
+                                       class_indices, follow_links, 
+                                       balance_config = None):
     """Lists paths of files in `subdir` with extensions in `white_list_formats`.
 
     # Arguments
@@ -198,6 +199,7 @@ def _list_valid_filenames_in_directory(directory, white_list_formats, split,
             of images in each directory.
         class_indices: dictionary mapping a class name to its index.
         follow_links: boolean, follow symbolic links to subdirectories.
+        balance_config: dict, stores configurations for handling data imbalance.
 
     # Returns
          classes: a list of class indices
@@ -225,6 +227,15 @@ def _list_valid_filenames_in_directory(directory, white_list_formats, split,
             dirname, os.path.relpath(absolute_path, directory))
         filenames.append(relative_path)
 
+    if balance_config:
+        filenames_copy = filenames.copy()
+
+        debt = balance_config['majority'] - len(filenames_copy)
+
+        for filename in _settle_debt(filenames_copy, debt):
+            classes.append(class_indices[dirname])
+            filenames.append(filename)
+
     return classes, filenames
 
 def _generate_class_count(directory):

From d9d2993cb6c291cc6dc43473c5a3ae09dedaf298 Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Fri, 2 Oct 2020 20:30:47 +0530
Subject: [PATCH 08/11] Implement '_settle_debt'

---
 keras_preprocessing/image/utils.py | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index 8e9d1038..307ac4f0 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -7,6 +7,7 @@
 import io
 import os
 import warnings
+import random
 
 import numpy as np
 
@@ -181,6 +182,19 @@ def _recursive_list(subpath):
             if fname.lower().endswith(white_list_formats):
                 yield root, fname
 
+def _settle_debt(list_valid_files, debt):
+    """Iterates over list_valid_files and resamples to settle debt.
+
+    # Arguments:
+        list_valid_files: List of strings, list that contains valid filenames
+        debt: Integer, required number of samples to be resampled from
+            valid_file_names
+
+    # Yields:
+        randomly chosen filename from list_valid_files 
+    """
+    for i in range(debt):
+        yield random.choice(list_valid_files)
 
 def _list_valid_filenames_in_directory(directory, white_list_formats, split,
                                        class_indices, follow_links, 

From 71cabcf53690792a8bc762944bf96729e9d9165d Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Fri, 2 Oct 2020 21:32:06 +0530
Subject: [PATCH 09/11] Set balance arg in flow from directory

---
 keras_preprocessing/image/image_data_generator.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/keras_preprocessing/image/image_data_generator.py b/keras_preprocessing/image/image_data_generator.py
index f5f1e9c9..20e17fb0 100644
--- a/keras_preprocessing/image/image_data_generator.py
+++ b/keras_preprocessing/image/image_data_generator.py
@@ -533,6 +533,7 @@ class subdirectories (default: False).
         return DirectoryIterator(
             directory,
             self,
+            balance=balance,
             target_size=target_size,
             color_mode=color_mode,
             classes=classes,

From 6e4e9d0dd375d9591f51334d29420657941ed454 Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Fri, 9 Oct 2020 16:59:48 +0530
Subject: [PATCH 10/11] Make changes based on autopep8

---
 .../image/directory_iterator.py               |  6 ++---
 .../image/image_data_generator.py             |  2 +-
 keras_preprocessing/image/utils.py            | 23 +++++++++++--------
 3 files changed, 18 insertions(+), 13 deletions(-)

diff --git a/keras_preprocessing/image/directory_iterator.py b/keras_preprocessing/image/directory_iterator.py
index 8da1dc05..ecb8c654 100644
--- a/keras_preprocessing/image/directory_iterator.py
+++ b/keras_preprocessing/image/directory_iterator.py
@@ -78,7 +78,7 @@ def __new__(cls, *args, **kwargs):
     def __init__(self,
                  directory,
                  image_data_generator,
-                 balance = False,
+                 balance=False,
                  target_size=(256, 256),
                  color_mode='rgb',
                  classes=None,
@@ -106,8 +106,8 @@ def __init__(self,
         self.directory = directory
 
         if balance and subset != 'validation':
-            self._balance_config = _make_balance_config(directory, 
-                image_data_generator._validation_split)
+            self._balance_config = _make_balance_config(directory,
+                                                        image_data_generator._validation_split)
         else:
             self._balance_config = None
 
diff --git a/keras_preprocessing/image/image_data_generator.py b/keras_preprocessing/image/image_data_generator.py
index 20e17fb0..a978166e 100644
--- a/keras_preprocessing/image/image_data_generator.py
+++ b/keras_preprocessing/image/image_data_generator.py
@@ -441,7 +441,7 @@ def flow(self,
 
     def flow_from_directory(self,
                             directory,
-                            balance = False,
+                            balance=False,
                             target_size=(256, 256),
                             color_mode='rgb',
                             classes=None,
diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index 307ac4f0..da2d3b18 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -182,6 +182,7 @@ def _recursive_list(subpath):
             if fname.lower().endswith(white_list_formats):
                 yield root, fname
 
+
 def _settle_debt(list_valid_files, debt):
     """Iterates over list_valid_files and resamples to settle debt.
 
@@ -196,9 +197,10 @@ def _settle_debt(list_valid_files, debt):
     for i in range(debt):
         yield random.choice(list_valid_files)
 
+
 def _list_valid_filenames_in_directory(directory, white_list_formats, split,
-                                       class_indices, follow_links, 
-                                       balance_config = None):
+                                       class_indices, follow_links,
+                                       balance_config=None):
     """Lists paths of files in `subdir` with extensions in `white_list_formats`.
 
     # Arguments
@@ -252,26 +254,28 @@ def _list_valid_filenames_in_directory(directory, white_list_formats, split,
 
     return classes, filenames
 
+
 def _generate_class_count(directory):
     """Maintain sample count of each class in the directory.
-    
+
     # Arguments
         directory: string, absolute path to the directory
     # Returns
         class_count: dictionary, sample count for each class
     """
-    
+
     class_count = {}
 
     for category in os.listdir(directory):
-      category_directory = os.path.join(directory, category)
-      class_count[category] = len(os.listdir(category_directory))
+        category_directory = os.path.join(directory, category)
+        class_count[category] = len(os.listdir(category_directory))
 
     return class_count
 
+
 def _make_balance_config(directory, validation_split):
     """Scans the directory to make a config dictionary to handle data imbalance.
-    
+
     # Arguments
         directory: string, absolute path to the directory
         validation_split: float, validation split
@@ -283,9 +287,9 @@ def _make_balance_config(directory, validation_split):
     class_count = _generate_class_count(directory)
 
     # Get the sample count of the majority class
-    majority_class_count = class_count[max(class_count, key = class_count.get)]
+    majority_class_count = class_count[max(class_count, key=class_count.get)]
     if validation_split:
-        majority_class_count = int(majority_class_count*(1 - validation_split)) + 1        
+        majority_class_count = int(majority_class_count*(1 - validation_split)) + 1
 
     balance_config = {
         'majority': majority_class_count
@@ -293,6 +297,7 @@ def _make_balance_config(directory, validation_split):
 
     return balance_config
 
+
 def array_to_img(x, data_format='channels_last', scale=True, dtype='float32'):
     """Converts a 3D Numpy array to a PIL Image instance.
 

From 80ed78660140fea91da730f00dab04a082dfd43c Mon Sep 17 00:00:00 2001
From: Divyanshu Gupta <divyanshu30gupta@gmail.com>
Date: Fri, 9 Oct 2020 17:15:11 +0530
Subject: [PATCH 11/11] Make changes acc to pep8 conventions

---
 keras_preprocessing/image/directory_iterator.py | 3 ++-
 keras_preprocessing/image/utils.py              | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/keras_preprocessing/image/directory_iterator.py b/keras_preprocessing/image/directory_iterator.py
index ecb8c654..11fef919 100644
--- a/keras_preprocessing/image/directory_iterator.py
+++ b/keras_preprocessing/image/directory_iterator.py
@@ -106,8 +106,9 @@ def __init__(self,
         self.directory = directory
 
         if balance and subset != 'validation':
+            validation_split = image_data_generator._validation_split
             self._balance_config = _make_balance_config(directory,
-                                                        image_data_generator._validation_split)
+                                                        validation_split)
         else:
             self._balance_config = None
 
diff --git a/keras_preprocessing/image/utils.py b/keras_preprocessing/image/utils.py
index da2d3b18..30f67426 100644
--- a/keras_preprocessing/image/utils.py
+++ b/keras_preprocessing/image/utils.py
@@ -192,7 +192,7 @@ def _settle_debt(list_valid_files, debt):
             valid_file_names
 
     # Yields:
-        randomly chosen filename from list_valid_files 
+        randomly chosen filename from list_valid_files
     """
     for i in range(debt):
         yield random.choice(list_valid_files)