Merge pull request #4 from Atashnezhad/Atashnezhad-patch-1

Update bit_vision.py

.gitignore

@@ -30,3 +30,5 @@ __pycache__/
 /dataset_augmented/
 /dataset_train_test_val/
 /deep_model/
+/tests/augmented_dataset/
+/tests/dataset/

README.md

@@ -83,10 +83,10 @@ from neural_network_model.bit_vision import BitVision
 if __name__ == "__main__":
  # download the images
  obj = Preprocessing(dataset_address=Path(__file__).parent / "dataset")
- obj.download_images(limit=25)
+ obj.download_images(limit=10)
  print(obj.image_dict)
  obj.augment_data(
- number_of_images_tobe_gen=100,
+ number_of_images_tobe_gen=10,
  augment_data_address=Path(__file__).parent / "augmented_dataset"
  )
  obj.train_test_split(
@@ -99,29 +99,33 @@ if __name__ == "__main__":
  print(obj.data_details)
  obj.plot_image_category()
  obj.compile_model()
-
+ #
  model_name = "model_epoch_{epoch:02d}_loss_{loss:.2f}_acc_{accuracy:.2f}_val_acc_{val_accuracy:.2f}_.h5"
  obj.train_model(
+ epochs=8,
  model_save_address=Path(__file__).parent / "deep_model",
- model_name=model_name,
- epochs=40,
+ model_name=model_name
  )
  obj.plot_history(fig_folder_address=Path(__file__).parent / "figures")
 
+ best_model = obj.return_best_model_name(directory="deep_model")
+
  obj.predict(
  fig_save_address=Path(__file__).parent / "figures",
- model_path=Path(__file__).parent / "deep_model" / model_name,
+ model_path=Path(__file__).parent / "deep_model" / best_model,
  test_folder_address=Path(__file__).parent / "dataset_train_test_val" / "test"
  )
 
  # find list of images in the Path(__file__).parent / "dataset_train_test_val" / "test" / "pdc_bit"
  directory_path = Path(__file__).parent / "dataset_train_test_val" / "test" / "pdc_bit"
  list_of_images = [str(x) for x in directory_path.glob("*.jpeg")]
+
  obj.grad_cam_viz(
- model_path=Path(__file__).parent / "deep_model" / model_name,
+ model_path=Path(__file__).parent / "deep_model" / best_model,
  fig_to_save_address=Path(__file__).parent / "figures",
  img_to_be_applied_path=Path(__file__).parent / "dataset_train_test_val" / "test" / "pdc_bit" / list_of_images[0],
  output_gradcam_fig_name="gradcam.png"
  )
 
 
+

neural_network_model/bit_vision.py

@@ -236,7 +236,7 @@ def _calculate_number_from_dict(self, my_dict: dict) -> int:
  return total
 
  def train_model(
- self, model_save_address: str = SETTING.MODEL_SETTING.MODEL_PATH, **kwargs
+ self, model_save_address: Path = SETTING.MODEL_SETTING.MODEL_PATH, **kwargs
  ) -> None:
  """
  This function is used to train the model.
@@ -281,10 +281,10 @@ def train_model(
  callbacks=[self._check_points(model_save_address, model_name)],
  )
 
- self.model.save(
- model_save_address / model_name
- or SETTING.MODEL_SETTING.MODEL_PATH / SETTING.MODEL_SETTING.MODEL_NAME
- )
+ # self.model.save(
+ #  model_save_address / model_name
+ #  or SETTING.MODEL_SETTING.MODEL_PATH / SETTING.MODEL_SETTING.MODEL_NAME
+ # )
  logger.info(f"Model saved to {SETTING.MODEL_SETTING.MODEL_PATH}")
 
  def plot_history(self, *args, **kwargs):
@@ -351,11 +351,11 @@ def predict(self, *args, **kwargs):
  if model_path is None:
  logger.info(f"model_path from SETTING is was used - {model_path}")
 
- test_folder_dir = kwargs.get(
- "test_folder_dir", SETTING.DATA_ADDRESS_SETTING.TEST_DIR_ADDRESS
- )
- if test_folder_dir is None:
- raise ValueError("test_folder_address is None")
+ # test_folder_dir = kwargs.get(
+ #  "test_folder_dir", SETTING.DATA_ADDRESS_SETTING.TEST_DIR_ADDRESS
+ # )
+ # if test_folder_dir is None:
+ #  raise ValueError("test_folder_address is None")
 
  model = keras.models.load_model(model_path)
  logger.info(f"Model loaded from {model_path}")
@@ -365,20 +365,20 @@ def predict(self, *args, **kwargs):
  number_of_cols = SETTING.FIGURE_SETTING.NUM_COLS_IN_PRED_MODEL
  number_of_rows = SETTING.FIGURE_SETTING.NUM_ROWS_IN_PRED_MODEL
  number_of_test_to_pred = SETTING.MODEL_SETTING.NUMBER_OF_TEST_TO_PRED
- if test_folder_dir:
- train_test_val_dir = (
- test_folder_dir
- or SETTING.PREPROCESSING_SETTING.TRAIN_TEST_VAL_SPLIT_DIR_ADDRESS
- )
- else:
- train_test_val_dir = (
- self.train_test_val_dir
- or SETTING.PREPROCESSING_SETTING.TRAIN_TEST_VAL_SPLIT_DIR_ADDRESS
- )
+ # if test_folder_dir:
+ #  train_test_val_dir = (
+ #  test_folder_dir
+ #  or SETTING.PREPROCESSING_SETTING.TRAIN_TEST_VAL_SPLIT_DIR_ADDRESS
+ #  )
+ # else:
+ #  train_test_val_dir = (
+ #  self.train_test_val_dir
+ #  or SETTING.PREPROCESSING_SETTING.TRAIN_TEST_VAL_SPLIT_DIR_ADDRESS
+ #  )
 
  # get the list of test images
  test_images_list = os.listdir(
- train_test_val_dir
+ self.train_test_val_dir
  / SETTING.PREPROCESSING_SETTING.TRAIN_TEST_SPLIT_DIR_NAMES[1]
  / category
  )
@@ -387,7 +387,7 @@ def predict(self, *args, **kwargs):
 
  for i, img in enumerate(test_images_list[0:number_of_test_to_pred]):
  path_to_img = (
- train_test_val_dir
+ self.train_test_val_dir
  / SETTING.PREPROCESSING_SETTING.TRAIN_TEST_SPLIT_DIR_NAMES[1]
  / category
  / str(img)
@@ -432,7 +432,7 @@ def predict(self, *args, **kwargs):
 
  datagen = image.ImageDataGenerator(SETTING.DATA_GEN_SETTING.RESCALE)
  DoubleCheck_generator = datagen.flow_from_directory(
- directory=test_folder_dir / "test",
+ directory=self.train_test_val_dir / "test",
  target_size=SETTING.FLOW_FROM_DIRECTORY_SETTING.TARGET_SIZE,
  color_mode=SETTING.FLOW_FROM_DIRECTORY_SETTING.COLOR_MODE,
  classes=None,
@@ -603,7 +603,7 @@ def _save_and_display_gradcam(
 
  @staticmethod
  def return_best_model_name(
- directory: str = SETTING.MODEL_SETTING.MODEL_PATH,
+ directory: Path = SETTING.MODEL_SETTING.MODEL_PATH,
  ) -> str:
  """
  Return the best model name

neural_network_model/script_run_all.py

@@ -0,0 +1,58 @@
+from neural_network_model.bit_vision import BitVision
+from neural_network_model.process_data import Preprocessing
+from pathlib import Path
+
+
+def main():
+
+ parent_dir = Path(__file__).resolve().parent.parent
+ # download the images
+ obj = Preprocessing(dataset_address=parent_dir / "dataset")
+ obj.download_images(limit=8)
+ print(obj.image_dict)
+ obj.augment_data(
+ number_of_images_tobe_gen=10,
+ augment_data_address=parent_dir / "dataset_augmented",
+ )
+ obj.train_test_split(
+ augmented_data_address=parent_dir / "dataset_augmented",
+ train_test_val_split_dir_address=parent_dir / "dataset_train_test_val"
+ )
+
+ obj = BitVision(train_test_val_dir=parent_dir / "dataset_train_test_val")
+ print(obj.categories)
+ print(obj.data_details)
+ obj.plot_image_category()
+ obj.compile_model()
+ #
+ model_name = "model_epoch_{epoch:02d}_loss_{loss:.2f}_acc_{accuracy:.2f}_val_acc_{val_accuracy:.2f}_.h5"
+ obj.train_model(
+ epochs=8,
+ model_save_address=parent_dir / "deep_model",
+ model_name=model_name
+ )
+ obj.plot_history(fig_folder_address=parent_dir / "figures")
+
+ best_model = obj.return_best_model_name(directory=parent_dir / "deep_model")
+
+ obj.predict(
+ fig_save_address=parent_dir / "figures",
+ model_path=parent_dir / "deep_model" / best_model,
+ test_folder_address=parent_dir / "dataset_train_test_val" / "test"
+ )
+
+ # find list of images in the parent_dir / "dataset_train_test_val" / "test" / "pdc_bit"
+ directory_path = parent_dir / "dataset_train_test_val" / "test" / "pdc_bit"
+ list_of_images = [str(x) for x in directory_path.glob("*.jpeg")]
+
+ obj.grad_cam_viz(
+ model_path=parent_dir / "deep_model" / best_model,
+ fig_to_save_address=parent_dir / "figures",
+ img_to_be_applied_path=parent_dir / "dataset_train_test_val" / "test" / "pdc_bit" / list_of_images[
+ 0],
+ output_gradcam_fig_name="test.png"
+ )
+
+
+if __name__ == "__main__":
+ main()

tests/test_run.py

@@ -1,4 +1,3 @@
-import argparse
 import os
 from pathlib import Path
 import sys
@@ -9,86 +8,9 @@
 main_dir = current_dir.parent
 # Add the main directory to the Python path
 sys.path.append(str(main_dir))
-# skip this test TODO: fix this test later
-import pytest
 
-from neural_network_model.bit_vision import BitVision
-from neural_network_model.process_data import Preprocessing
 
-
-@pytest.mark.skip(reason="no way of currently testing this")
-def test_run():
- # check if a dir name test_resources exists if not create one
- if not os.path.exists(Path(__file__).parent / "test_resources"):
- os.mkdir(Path(__file__).parent / "test_resources")
-
- # download the images
- obj = Preprocessing(
- dataset_address=Path(__file__).parent / "test_resources" / "dataset"
- )
- obj.download_images(limit=10)
- print(obj.image_dict)
- obj.augment_data(
- number_of_images_tobe_gen=10,
- augment_data_address=Path(__file__).parent
- / "test_resources"
- / "augmented_dataset",
- )
- obj.train_test_split(
- augmented_data_address=Path(__file__).parent
- / "test_resources"
- / "augmented_dataset",
- train_test_val_split_dir_address=Path(__file__).parent
- / "test_resources"
- / "dataset_train_test_val",
- )
-
- obj = BitVision(
- train_test_val_dir=Path(__file__).parent
- / "test_resources"
- / "dataset_train_test_val"
- )
- print(obj.categories)
- print(obj.data_details)
- obj.plot_image_category()
- obj.compile_model()
-
- model_name = "model_epoch_{epoch:02d}_loss_{loss:.2f}_acc_{accuracy:.2f}_val_acc_{val_accuracy:.2f}_.h5"
- obj.train_model(
- model_save_address=Path(__file__).parent / "test_resources" / "deep_model",
- model_name=model_name,
- epochs=10,
- )
- obj.plot_history(
- fig_folder_address=Path(__file__).parent / "test_resources" / "figures"
- )
-
- obj.predict(
- fig_save_address=Path(__file__).parent / "test_resources" / "figures",
- model_path=Path(__file__).parent / "test_resources" / "deep_model" / model_name,
- test_folder_address=Path(__file__).parent
- / "test_resources"
- / "dataset_train_test_val"
- / "test",
- )
-
- # find list of images in the Path(__file__).parent / "dataset_train_test_val" / "test" / "pdc_bit"
- directory_path = (
- Path(__file__).parent
- / "test_resources"
- / "dataset_train_test_val"
- / "test"
- / "pdc_bit"
- )
- list_of_images = [str(x) for x in directory_path.glob("*.jpeg")]
- obj.grad_cam_viz(
- model_path=Path(__file__).parent / "test_resources" / "deep_model" / model_name,
- fig_to_save_address=Path(__file__).parent / "test_resources" / "figures",
- img_to_be_applied_path=Path(__file__).parent
- / "test_resources"
- / "dataset_train_test_val"
- / "test"
- / "pdc_bit"
- / list_of_images[0],
- output_gradcam_fig_name="gradcam.png",
- )
+def test_run_all():
+ script_path = Path(__file__).parent / ".." / "neural_network_model" / "script_run_all.py"
+ os.system(f'python {script_path}')
+ assert True