Update app.py

app.py

@@ -1,96 +1,270 @@
-import gradio as gr
+from fastapi import FastAPI, File, UploadFile
+from fastapi.responses import HTMLResponse
 from transformers import pipeline
-import torch
-from PIL import Image, ImageDraw
-import numpy as np
+from PIL import Image
+import io
+import uvicorn
+import base64
+
+app = FastAPI()
 
 # Chargement des modèles
-classifier = pipeline("image-classification", model="abhishek/chest-xray-classification")
-detector = pipeline("object-detection", model="nickysam/detect-thorax-anomaly-75acc")
+def load_models():
+    return {
+        "chest_classifier": pipeline("image-classification", model="codewithdark/vit-chest-xray")
+    }
 
-def draw_boxes(image, predictions):
-    # Convertir l'image numpy en PIL si nécessaire
-    if isinstance(image, np.ndarray):
-        image = Image.fromarray(np.uint8(image))
-    
-    draw = ImageDraw.Draw(image)
-    
-    # Dessiner les boîtes de détection
-    for pred in predictions:
-        box = pred['box']
-        score = pred['score']
-        label = pred['label']
-        
-        # Coordonnées de la boîte
-        x1, y1 = box['xmin'], box['ymin']
-        x2, y2 = box['xmax'], box['ymax']
-        
-        # Couleur en fonction du score
-        color = (255, 0, 0) if score > 0.7 else (255, 165, 0)
-        
-        # Dessiner le rectangle
-        draw.rectangle([x1, y1, x2, y2], outline=color, width=2)
-        
-        # Ajouter le label et le score
-        label_text = f"{label}: {score:.1%}"
-        draw.text((x1, y1-15), label_text, fill=color)
-    
-    return image
+models = load_models()
+
+def translate_label(label):
+    translations = {
+        'Cardiomegaly': 'Kardiomegalie',
+        'Edema': 'Ödem',
+        'Consolidation': 'Konsolidierung',
+        'Pneumonia': 'Lungenentzündung',
+        'No Finding': 'Kein Befund'
+    }
+    return translations.get(label, label)
 
-def analyze_xray(image):
+def image_to_base64(image):
+    buffered = io.BytesIO()
+    image.save(buffered, format="PNG")
+    img_str = base64.b64encode(buffered.getvalue()).decode()
+    return f"data:image/png;base64,{img_str}"
+
+COMMON_STYLES = """
+    body {
+        font-family: system-ui, -apple-system, sans-serif;
+        background: #f0f2f5;
+        margin: 0;
+        padding: 20px;
+        color: #1a1a1a;
+    }
+    .container {
+        max-width: 1200px;
+        margin: 0 auto;
+        background: white;
+        padding: 20px;
+        border-radius: 10px;
+        box-shadow: 0 2px 4px rgba(0,0,0,0.1);
+    }
+    .button {
+        background: #2d2d2d;
+        color: white;
+        border: none;
+        padding: 12px 30px;
+        border-radius: 8px;
+        cursor: pointer;
+        font-size: 1.1em;
+        transition: all 0.3s ease;
+        position: relative;
+    }
+    .button:hover {
+        background: #404040;
+    }
+    @keyframes blink {
+        0% { opacity: 1; }
+        50% { opacity: 0; }
+        100% { opacity: 1; }
+    }
+    #loading {
+        display: none;
+        color: white;
+        margin-top: 10px;
+        animation: blink 1s infinite;
+        text-align: center;
+    }
+    .upload-section {
+        background: #2d2d2d;
+        padding: 40px;
+        border-radius: 12px;
+        margin: 20px 0;
+        text-align: center;
+        border: 2px dashed #404040;
+        transition: all 0.3s ease;
+        color: white;
+    }
+    .upload-section:hover {
+        border-color: #555;
+    }
+    input[type="file"] {
+        font-size: 1.1em;
+        margin: 20px 0;
+        color: white;
+    }
+    input[type="file"]::file-selector-button {
+        font-size: 1em;
+        padding: 10px 20px;
+        border-radius: 8px;
+        border: 1px solid #404040;
+        background: #2d2d2d;
+        color: white;
+        transition: all 0.3s ease;
+        cursor: pointer;
+    }
+    input[type="file"]::file-selector-button:hover {
+        background: #404040;
+    }
+    .preview-image {
+        max-width: 300px;
+        margin: 20px auto;
+        display: none;
+    }
+    .results-grid {
+        display: grid;
+        grid-template-columns: 1fr 1fr;
+        gap: 20px;
+        margin-top: 20px;
+    }
+    .result-box {
+        background: white;
+        padding: 20px;
+        border-radius: 12px;
+        margin: 10px 0;
+        border: 1px solid #e9ecef;
+    }
+    .analyzed-image {
+        max-width: 400px;
+        margin: 0 auto;
+    }
+    .score-high {
+        color: #0066cc;
+        font-weight: bold;
+    }
+    .score-medium {
+        color: #ffa500;
+        font-weight: bold;
+    }
+    h3 {
+        color: #0066cc;
+        margin-top: 0;
+    }
+    @media (max-width: 768px) {
+        .results-grid {
+            grid-template-columns: 1fr;
+        }
+    }
+"""
+
+@app.get("/", response_class=HTMLResponse)
+async def main():
+    content = f"""
+    <!DOCTYPE html>
+    <html>
+    <head>
+        <title>Röntgenbild-Analyse</title>
+        <meta name="viewport" content="width=device-width, initial-scale=1.0">
+        <style>
+            {COMMON_STYLES}
+        </style>
+    </head>
+    <body>
+        <div class="container">
+            <div class="upload-section">
+                <form action="/analyze" method="post" enctype="multipart/form-data" 
+                      onsubmit="document.getElementById('loading').style.display = 'block';">
+                    <div>
+                        <input type="file" name="file" accept="image/*" required 
+                               onchange="document.getElementById('preview').src = window.URL.createObjectURL(this.files[0]); 
+                                        document.getElementById('preview').style.display = 'block';">
+                    </div>
+                    <img id="preview" class="preview-image" src="" alt="Vorschau">
+                    <button type="submit" class="button">
+                        Analysieren
+                    </button>
+                    <div id="loading">Wird geladen...</div>
+                </form>
+            </div>
+        </div>
+    </body>
+    </html>
+    """
+    return content
+
+@app.post("/analyze", response_class=HTMLResponse)
+async def analyze_file(file: UploadFile = File(...)):
     try:
-        # Classification générale
-        classifications = classifier(image)
+        contents = await file.read()
+        image = Image.open(io.BytesIO(contents))
         
-        # Détection des anomalies
-        detections = detector(image)
+        predictions = models["chest_classifier"](image)
+        result_image_b64 = image_to_base64(image)
         
-        # Dessiner les boîtes sur l'image
-        annotated_image = draw_boxes(image, detections)
+        results_html = f"""
+        <!DOCTYPE html>
+        <html>
+        <head>
+            <title>Ergebnisse</title>
+            <meta name="viewport" content="width=device-width, initial-scale=1.0">
+            <style>
+                {COMMON_STYLES}
+            </style>
+        </head>
+        <body>
+            <div class="container">
+                <div class="results-grid">
+                    <div class="result-box">
+                        <h3>Analyse-Ergebnisse</h3>
+        """
         
-        # Préparer les résultats
-        results = "Classifications:\n"
-        for pred in classifications:
-            results += f"{pred['label']}: {pred['score']:.1%}\n"
+        for pred in predictions:
+            confidence_class = "score-high" if pred['score'] > 0.7 else "score-medium"
+            results_html += f"""
+                <div>
+                    <span class="{confidence_class}">{pred['score']:.1%}</span> - 
+                    {translate_label(pred['label'])}
+                </div>
+            """
         
-        results += "\nDetected Anomalies:\n"
-        for det in detections:
-            results += f"{det['label']}: {det['score']:.1%}\n"
+        results_html += f"""
+                    </div>
+                    <div class="result-box">
+                        <h3>Röntgenbild</h3>
+                        <img src="{result_image_b64}" alt="Analysiertes Röntgenbild" class="analyzed-image">
+                    </div>
+                </div>
+                
+                <a href="/" class="button back-button">
+                    ← Zurück
+                </a>
+            </div>
+        </body>
+        </html>
+        """
         
-        return annotated_image, results
-    except Exception as e:
-        return image, f"Error: {str(e)}"
-
-# Interface Gradio
-with gr.Blocks(theme=gr.themes.Soft(
-    primary_hue="gray",
-    secondary_hue="gray",
-)) as demo:
-    gr.Markdown("""
-    # Chest X-Ray Analysis
-    This application analyzes chest X-rays to:
-    1. Classify general conditions
-    2. Detect and locate specific anomalies
-    """)
-    
-    with gr.Row():
-        with gr.Column():
-            input_image = gr.Image(label="Upload X-Ray Image", type="pil")
-            analyze_btn = gr.Button("Analyze", variant="primary")
+        return results_html
         
-        with gr.Column():
-            output_image = gr.Image(label="Analyzed Image")
-            output_text = gr.Textbox(label="Results", lines=10)
-    
-    analyze_btn.click(
-        fn=analyze_xray,
-        inputs=[input_image],
-        outputs=[output_image, output_text]
-    )
-    
-    gr.Markdown("""
-    Note: This tool is for demonstration purposes only and should not be used for medical diagnosis.
-    """)
+    except Exception as e:
+        return f"""
+        <!DOCTYPE html>
+        <html>
+        <head>
+            <title>Fehler</title>
+            <meta name="viewport" content="width=device-width, initial-scale=1.0">
+            <style>
+                {COMMON_STYLES}
+                .error-box {{
+                    background: #fee2e2;
+                    border: 1px solid #ef4444;
+                    padding: 20px;
+                    border-radius: 8px;
+                    margin: 20px 0;
+                }}
+            </style>
+        </head>
+        <body>
+            <div class="container">
+                <div class="error-box">
+                    <h3>Fehler</h3>
+                    <p>{str(e)}</p>
+                </div>
+                <a href="/" class="button back-button">
+                    ← Zurück
+                </a>
+            </div>
+        </body>
+        </html>
+        """
 
-# Lancement de l'application
-demo.launch()
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=7860)