fix(all): clean

.github/workflows/main.yml

@@ -10,10 +10,10 @@ env:
   PNPM_VERSION: 8.5.1
   PYTHON_VERSION: 3.10.15
   NODE_VERSION: 18
-  AZURE_RESOURCE_GROUP_NAME: "azure-ml-gui"
+  AZURE_RESOURCE_GROUP_NAME: "azure-ml-guil"
   AZURE_LOCATION: "northeurope"
-  AZURE_ML_WORKSPACE_NAME: "cats-dogs-gui"
-  AZURE_WEBAPP_NAME: "cats-dogs-gui"
+  AZURE_ML_WORKSPACE_NAME: "cats-dogs-guil"
+  AZURE_API_NAME: "cats-dogs-guil"
   DELETE_WEBAPP: "false"
   DOCKER_API_IMAGE_NAME: "mlopspython-api"
   DOCKER_REPOSITORY: ${{ github.repository_owner }}
@@ -276,7 +276,7 @@ jobs:
     - name: Deploy container
       run: |
         # https://learn.microsoft.com/en-us/cli/azure/container?view=azure-cli-latest#az-container-create()
-        az container create --resource-group ${{ env.AZURE_RESOURCE_GROUP_NAME}} --name ${{ env.AZURE_WEBAPP_NAME }} --dns-name-label ${{ env.AZURE_WEBAPP_NAME }} --image ${{ env.DOCKER_REGISTRY }}/${{ env.DOCKER_REPOSITORY }}/${{ env.DOCKER_API_IMAGE_NAME }}:${{ needs.tags.outputs.new_version }} --ports 5000
+        az container create --resource-group ${{ env.AZURE_RESOURCE_GROUP_NAME}} --name ${{ env.AZURE_API_NAME }} --dns-name-label ${{ env.AZURE_API_NAME }} --image ${{ env.DOCKER_REGISTRY }}/${{ env.DOCKER_REPOSITORY }}/${{ env.DOCKER_API_IMAGE_NAME }}:${{ needs.tags.outputs.new_version }} --ports 5000
   integration_tests:
     environment: MLOpsPython
     runs-on: ubuntu-latest
@@ -307,12 +307,12 @@ jobs:
             --experiment_id ${{ needs.train.outputs.EXPERIMENT_ID }} \
             --integration_dataset_name "cats-dogs-others-integration" \
             --integration_dataset_version ${{ needs.train.outputs.INTEGRATION_DATASET_VERSION }} \
-            --url "http://${{ env.AZURE_WEBAPP_NAME }}.${{ env.AZURE_LOCATION }}.azurecontainer.io:5000/upload" > integration_output.txt 
+            --url "http://${{ env.AZURE_API_NAME }}.${{ env.AZURE_LOCATION }}.azurecontainer.io:5000/upload" > integration_output.txt 
         cat integration_output.txt
     - name: Delete Environment
       run: |
           if [ "${{ env.DELETE_WEBAPP }}" = "true" ]; then
-            az container delete --resource-group ${{ env.AZURE_RESOURCE_GROUP_NAME}} --name ${{ env.AZURE_WEBAPP_NAME }} --yes
+            az container delete --resource-group ${{ env.AZURE_RESOURCE_GROUP_NAME}} --name ${{ env.AZURE_API_NAME }} --yes
           fi
   webapp_unit_tests:
     runs-on: ubuntu-latest

packages/mlopspython-inference/mlopspython_inference/inference_pillow.py

@@ -1,32 +1,61 @@
+from abc import ABC
 from io import BytesIO
 
 import numpy as np
-from tensorflow.keras.preprocessing.image import load_img, img_to_array
+from PIL import Image
 from keras.models import load_model
 from pathlib import Path
 
 
-# load and prepare the image
-def load_image(filename: str|BytesIO):
-    # load the image
-    img = load_img(filename, target_size=(224, 224))
-    # convert to array
-    img = img_to_array(img)
-    # reshape into a single sample with 3 channels
-    img = img.reshape(1, 224, 224, 3)
-    # center pixel data
-    img = img.astype('float32')
-    img = img - [123.68, 116.779, 103.939]
-    return img
+def load_image(filename: str | BytesIO):
+    # Open the image
+    if isinstance(filename, BytesIO):
+        img = Image.open(filename)
+    else:
+        img = Image.open(Path(filename))
+
+    # Resize the image to 224x224
+    img = img.resize((224, 224))
+
+    # Convert image to NumPy array (shape: height x width x channels)
+    img_array = np.array(img)
+
+    # Ensure it's 3 channels (RGB)
+    if img_array.shape[-1] != 3:
+        img_array = img_array[:, :, :3]  # Convert to RGB if it has an alpha channel
+
+    # Convert the array to float32 and reshape into (1, 224, 224, 3)
+    img_array = img_array.astype('float32')
+    img_array = img_array.reshape((1, 224, 224, 3))
+
+    # Center pixel values (subtract mean pixel values)
+    img_array -= [123.68, 116.779, 103.939]
+
+    return img_array
 
 
 BASE_PATH = Path(__file__).resolve().parent
 
+class IModel(ABC):
+    def predict(self, img: np.ndarray) -> np.ndarray:
+        pass
+
+class Model(IModel):
+    def __init__(self, model_path: str):
+        self.model = load_model(model_path)
+
+    def predict(self, img: np.ndarray) -> np.ndarray:
+        return self.model.predict(img)
+
+class ModelMock(IModel):
+    def predict(self, img: np.ndarray) -> np.ndarray:
+        return np.array([[0.1, 0.2, 0.7]])
+
 
 class Inference:
-    def __init__(self, logging, model_path: str):
+    def __init__(self, logging, model: IModel):
         self.logger = logging.getLogger(__name__)
-        self.model = load_model(model_path)
+        self.model = model
 
     def execute(self, filepath:str|BytesIO):
         img = load_image(filepath)