Step 7 Complete: Docker image with Flask API and Gradio UI working locally

.ipynb_checkpoints/Dockerfile-checkpoint

@@ -0,0 +1,30 @@
+# 1. Utiliser Python 3.10 (stable pour PyTorch et Gradio)
+FROM python:3.10-slim
+
+# 2. Définir le dossier de travail dans le conteneur
+WORKDIR /app
+
+# 3. Installer les dépendances système nécessaires (OpenCV, Pillow, etc.)
+RUN apt-get update && apt-get install -y \
+    libgl1 \
+    libglib2.0-0 \
+    && rm -rf /var/lib/apt/lists/*
+
+# 4. Copier et installer les bibliothèques Python
+COPY requirements-api.txt .
+RUN pip install --no-cache-dir -r requirements-api.txt
+# On force l'installation de Gradio et Requests au cas où
+RUN pip install --no-cache-dir gradio requests 
+
+# 5. Copier tout ton code (le modèle .pth doit être dans un dossier 'weights/')
+COPY . .
+
+# 6. Exposer les ports pour l'extérieur
+# 5075 = Flask (API) | 7860 = Gradio (Interface)
+EXPOSE 5075 7860
+
+# 7. La commande de lancement
+# On utilise 'sh -c' pour lancer deux processus en même temps :
+# 'python movieposter_api.py &' lance l'API en tâche de fond.
+# 'python app_gradio.py' lance l'interface au premier plan.
+CMD sh -c "python movieposter_api.py --model_path weights/movieposter_net.pth & python app_gradio.py"

.ipynb_checkpoints/README-checkpoint.md

@@ -0,0 +1,6 @@
+# projet_AIF
+projet avancé jusqu' au point 3. de la partie 1. Il faut faire le 4. Build Gradio interface
+Docker commun
+#commandes dans terminal :
+docker build -t movie-poster-app .
+docker run -p 5075:5075 -p 7860:7860 movie-poster-app

.ipynb_checkpoints/app_gradio-checkpoint.py

@@ -0,0 +1,39 @@
+import gradio as gr
+import requests
+import io
+from PIL import Image
+
+def predict_movie_genre(image):
+    # 1. Configuration (basée sur le test de ta collègue)
+    API_URL = "http://127.0.0.1:5075/predict"
+    
+    # 2. Conversion de l'image Gradio (PIL) en bytes pour l'API
+    img_byte_arr = io.BytesIO()
+    image.save(img_byte_arr, format='JPEG')
+    img_data = img_byte_arr.getvalue()
+
+    try:
+        # 3. Envoi de la requête (format data brut comme dans son test)
+        response = requests.post(API_URL, data=img_data)
+        
+        if response.status_code == 200:
+            prediction = response.json().get('label', 'Genre inconnu')
+            return f"🎬 Genre prédit : {prediction}"
+        else:
+            return f"⚠️ Erreur API : Code {response.status_code}"
+            
+    except Exception as e:
+        return f"Impossible de contacter l'API. Est-elle lancée sur le port 5075 ? ({e})"
+
+# 4. Création de l'interface visuelle
+demo = gr.Interface(
+    fn=predict_movie_genre,
+    inputs=gr.Image(type="pil", label="Déposez un poster ici"),
+    outputs=gr.Text(label="Résultat de l'analyse"),
+    title="Analyseur de Posters de Films",
+    description="Cette interface utilise une API Flask et un modèle Deep Learning pour prédire le genre d'un film."
+)
+
+if __name__ == "__main__":
+#CRUCIAL pour Docker : server_name="0.0.0.0" permet l'accès extérieur
+    demo.launch(server_name="0.0.0.0", server_port=7860)

.ipynb_checkpoints/movieposter_api-checkpoint.py

@@ -0,0 +1,74 @@
+import argparse
+import torch
+import torchvision.transforms as transforms
+from flask import Flask, jsonify, request
+from PIL import Image
+import io
+from model import MovieposterNet
+ 
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+ 
+app = Flask(__name__)
+ 
+# Liste des classes pour le mapping
+CLASSES = ['action', 'animation', 'comedy', 'documentary', 'drama', 'fantasy', 'horror', 'romance', 'science Fiction', 'thriller']
+
+
+
+parser = argparse.ArgumentParser()
+parser.add_argument('--model_path', type=str, default = 'weights/movieposter_net.pth', help='model path')
+args = parser.parse_args()
+model_path = args.model_path
+ 
+model = MovieposterNet().to(device)
+model.load_state_dict(torch.load(model_path, map_location=device))
+model.eval()
+ 
+# Resizing des images  
+transform = transforms.Compose([
+    transforms.Resize((224, 224)),
+    transforms.ToTensor(),
+    transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
+])
+ 
+@app.route('/predict', methods=['POST'])
+def predict():
+    img_binary = request.data
+    img_pil = Image.open(io.BytesIO(img_binary))
+ 
+    # Transform the PIL image
+    tensor = transform(img_pil).to(device)
+    tensor = tensor.unsqueeze(0)  # Add batch dimension
+    
+    # Make prediction
+    with torch.no_grad():
+        outputs = model(tensor)
+        _, predicted = outputs.max(1)
+ 
+    return jsonify({"prediction": int(predicted[0]), "label": CLASSES[int(predicted[0])]})
+ 
+@app.route('/batch_predict', methods=['POST'])
+def batch_predict():
+    # Get the image data from the request
+    images_binary = request.files.getlist("images[]")
+ 
+    tensors = []
+ 
+    for img_binary in images_binary:
+        img_pil = Image.open(img_binary.stream)
+        tensor = transform(img_pil)
+        tensors.append(tensor)
+ 
+    # Stack tensors to form a batch tensor
+    batch_tensor = torch.stack(tensors, dim=0).to(device)
+
+    # Make prediction
+    with torch.no_grad():
+        outputs = model(batch_tensor)
+        _, predictions = outputs.max(1)
+ 
+    res = [CLASSES[idx] for idx in predictions.tolist()]
+    return jsonify({"predictions": res})
+ 
+if __name__ == "__main__":
+    app.run(host='0.0.0.0', port=5075, debug=True)

.ipynb_checkpoints/requirements-api-checkpoint.txt

@@ -0,0 +1,7 @@
+torch==2.0.1
+torchvision==0.15.2
+flask==2.3.2
+pillow==10.0.0
+numpy==1.24.4
+gradio
+requests

.ipynb_checkpoints/test_api-checkpoint.ipynb

@@ -0,0 +1,106 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "cc94cf7f",
+   "metadata": {},
+   "outputs": [
+    {
+     "ename": "ValueError",
+     "evalue": "Sample larger than population or is negative",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mValueError\u001b[0m                                Traceback (most recent call last)",
+      "Cell \u001b[1;32mIn[1], line 23\u001b[0m\n\u001b[0;32m     20\u001b[0m             all_images\u001b[38;5;241m.\u001b[39mappend((full_path, genre_reel))\n\u001b[0;32m     22\u001b[0m \u001b[38;5;66;03m# 3. Sélection de 10 posters au hasard\u001b[39;00m\n\u001b[1;32m---> 23\u001b[0m test_samples \u001b[38;5;241m=\u001b[39m random\u001b[38;5;241m.\u001b[39msample(all_images, \u001b[38;5;241m10\u001b[39m)\n\u001b[0;32m     25\u001b[0m \u001b[38;5;66;03m# 4. Affichage des résultats\u001b[39;00m\n\u001b[0;32m     26\u001b[0m plt\u001b[38;5;241m.\u001b[39mfigure(figsize\u001b[38;5;241m=\u001b[39m(\u001b[38;5;241m20\u001b[39m, \u001b[38;5;241m10\u001b[39m))\n",
+      "File \u001b[1;32m~\\anaconda3\\Lib\\random.py:430\u001b[0m, in \u001b[0;36mRandom.sample\u001b[1;34m(self, population, k, counts)\u001b[0m\n\u001b[0;32m    428\u001b[0m randbelow \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_randbelow\n\u001b[0;32m    429\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;241m0\u001b[39m \u001b[38;5;241m<\u001b[39m\u001b[38;5;241m=\u001b[39m k \u001b[38;5;241m<\u001b[39m\u001b[38;5;241m=\u001b[39m n:\n\u001b[1;32m--> 430\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mSample larger than population or is negative\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[0;32m    431\u001b[0m result \u001b[38;5;241m=\u001b[39m [\u001b[38;5;28;01mNone\u001b[39;00m] \u001b[38;5;241m*\u001b[39m k\n\u001b[0;32m    432\u001b[0m setsize \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m21\u001b[39m        \u001b[38;5;66;03m# size of a small set minus size of an empty list\u001b[39;00m\n",
+      "\u001b[1;31mValueError\u001b[0m: Sample larger than population or is negative"
+     ]
+    }
+   ],
+   "source": [
+    "import requests\n",
+    "import os\n",
+    "import random\n",
+    "import matplotlib.pyplot as plt\n",
+    "from PIL import Image\n",
+    "import io\n",
+    "\n",
+    "# 1. Configuration\n",
+    "API_URL = \"http://localhost:5075/predict\"\n",
+    "DATASET_PATH = \"../sorted_movie_posters_paligema\"\n",
+    "\n",
+    "# 2. Récupération de tous les chemins d'images\n",
+    "all_images = []\n",
+    "for root, dirs, files in os.walk(DATASET_PATH):\n",
+    "    for file in files:\n",
+    "        if file.lower().endswith(('.png', '.jpg', '.jpeg')):\n",
+    "            full_path = os.path.join(root, file)\n",
+    "            # On extrait le genre à partir du nom du dossier parent\n",
+    "            genre_reel = os.path.basename(root)\n",
+    "            all_images.append((full_path, genre_reel))\n",
+    "\n",
+    "# 3. Sélection de 10 posters au hasard\n",
+    "test_samples = random.sample(all_images, 10)\n",
+    "\n",
+    "# 4. Affichage des résultats\n",
+    "plt.figure(figsize=(20, 10))\n",
+    "\n",
+    "for i, (img_path, ground_truth) in enumerate(test_samples):\n",
+    "    # Lecture et envoi de l'image à l'API\n",
+    "    with open(img_path, \"rb\") as f:\n",
+    "        img_data = f.read()\n",
+    "    \n",
+    "    try:\n",
+    "        response = requests.post(API_URL, data=img_data)\n",
+    "        prediction = response.json().get('label', 'Erreur')\n",
+    "    except Exception as e:\n",
+    "        prediction = \"API Down\"\n",
+    "\n",
+    "    # Affichage\n",
+    "    img = Image.open(img_path)\n",
+    "    plt.subplot(2, 5, i + 1)\n",
+    "    plt.imshow(img)\n",
+    "    \n",
+    "    # Couleur du titre : vert si correct, rouge si erreur\n",
+    "    color = 'green' if prediction == ground_truth else 'red'\n",
+    "    \n",
+    "    plt.title(f\"Réel: {ground_truth}\\nPred: {prediction}\", color=color, fontsize=10)\n",
+    "    plt.axis('off')\n",
+    "\n",
+    "plt.tight_layout()\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7ab72703",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

Dockerfile

@@ -0,0 +1,30 @@
+# 1. Utiliser Python 3.10 (stable pour PyTorch et Gradio)
+FROM python:3.10-slim
+
+# 2. Définir le dossier de travail dans le conteneur
+WORKDIR /app
+
+# 3. Installer les dépendances système nécessaires (OpenCV, Pillow, etc.)
+RUN apt-get update && apt-get install -y \
+    libgl1 \
+    libglib2.0-0 \
+    && rm -rf /var/lib/apt/lists/*
+
+# 4. Copier et installer les bibliothèques Python
+COPY requirements-api.txt .
+RUN pip install --no-cache-dir -r requirements-api.txt
+# On force l'installation de Gradio et Requests au cas où
+RUN pip install --no-cache-dir gradio requests 
+
+# 5. Copier tout ton code (le modèle .pth doit être dans un dossier 'weights/')
+COPY . .
+
+# 6. Exposer les ports pour l'extérieur
+# 5075 = Flask (API) | 7860 = Gradio (Interface)
+EXPOSE 5075 7860
+
+# 7. La commande de lancement
+# On utilise 'sh -c' pour lancer deux processus en même temps :
+# 'python movieposter_api.py &' lance l'API en tâche de fond.
+# 'python app_gradio.py' lance l'interface au premier plan.
+CMD sh -c "python movieposter_api.py --model_path weights/movieposter_net.pth & python app_gradio.py"

Dockerfile-api

@@ -1,20 +0,0 @@
-# Use an official Python runtime as the parent image
-FROM python:3.10-slim
-
-# Set the working directory in the container to /app
-WORKDIR /app
-
-# Copy the current directory contents into the container at /app
-COPY . /app
-
-# Install any needed packages specified in requirements.txt
-RUN pip install --trusted-host pypi.python.org -r requirements-api.txt
-
-# Make port 5075 available to the world outside this container
-EXPOSE 5075
-
-# Define environment variable for Flask to run in production mode
-ENV FLASK_ENV=production
-
-# Run mnist_api.py when the container launches
-CMD ["python", "movieposter_api.py", "--model_path", "weights/movieposter_net.pth"]

README.md

@@ -1,3 +1,6 @@
 # projet_AIF
 projet avancé jusqu' au point 3. de la partie 1. Il faut faire le 4. Build Gradio interface
-
+Docker commun
+#commandes dans terminal :
+docker build -t movie-poster-app .
+docker run -p 5075:5075 -p 7860:7860 movie-poster-app

app_gradio.py

@@ -5,7 +5,7 @@ from PIL import Image
 
 def predict_movie_genre(image):
     # 1. Configuration (basée sur le test de ta collègue)
-    API_URL = "http://localhost:5075/predict"
+    API_URL = "http://127.0.0.1:5075/predict"
     
     # 2. Conversion de l'image Gradio (PIL) en bytes pour l'API
     img_byte_arr = io.BytesIO()
@@ -35,4 +35,5 @@ demo = gr.Interface(
 )
 
 if __name__ == "__main__":
-    demo.launch()
+#CRUCIAL pour Docker : server_name="0.0.0.0" permet l'accès extérieur
+    demo.launch(server_name="0.0.0.0", server_port=7860)

requirements-api.txt

@@ -2,4 +2,6 @@ torch==2.0.1
 torchvision==0.15.2
 flask==2.3.2
 pillow==10.0.0
-numpy==1.24.4
+numpy==1.24.4
+gradio
+requests