Amélioration du support Poppler pour la conversion PDF et la prévisualisation

Dockerfile

@@ -2,12 +2,24 @@ FROM python:3.9
 
 WORKDIR /code
 
-# Installer poppler-utils pour pdf2image
-RUN apt-get update && apt-get install -y poppler-utils && rm -rf /var/lib/apt/lists/*
+# Installer poppler-utils complet pour pdf2image
+RUN apt-get update && apt-get install -y \
+    poppler-utils \
+    libpoppler-cpp-dev \
+    libpoppler-private-dev \
+    pkg-config \
+    && rm -rf /var/lib/apt/lists/* \
+    && which pdftoppm && echo "Poppler correctement installé!" \
+    || echo "ERREUR: Poppler non installé correctement!"
 
 COPY requirements.txt /code/requirements.txt
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 
+# Vérification de l'installation de pdf2image
+RUN python -c "from pdf2image import convert_from_bytes; print('pdf2image est correctement installé!')" \
+    || echo "ERREUR: pdf2image n'est pas installé correctement!"
+
 COPY . /code
 
-CMD ["python", "app.py"]
+# Commande de démarrage de l'application Gradio
+CMD ["python", "app.py"] 

app.py

@@ -4,9 +4,48 @@ import os
 import json
 import time
 import base64
+import subprocess
 from PIL import Image
 from pdf2image import convert_from_path
 
+# Vérification de la disponibilité de Poppler
+def check_poppler():
+    poppler_path = None
+    potential_paths = [
+        '/usr/bin', 
+        '/usr/local/bin', 
+        '/opt/homebrew/bin',
+        '/app/bin'
+    ]
+    
+    for path in potential_paths:
+        if os.path.exists(os.path.join(path, 'pdftoppm')):
+            poppler_path = path
+            print(f"✅ Poppler trouvé dans: {poppler_path}")
+            return poppler_path
+            
+    # Essayer de localiser avec la commande which
+    try:
+        which_result = subprocess.run(['which', 'pdftoppm'], capture_output=True, text=True)
+        if which_result.returncode == 0:
+            poppler_bin = which_result.stdout.strip()
+            poppler_path = os.path.dirname(poppler_bin)
+            print(f"✅ Poppler trouvé via 'which' dans: {poppler_path}")
+            return poppler_path
+    except:
+        pass
+        
+    print("⚠️ AVERTISSEMENT: Poppler non trouvé dans les chemins standards!")
+    print("⚠️ Les fonctionnalités de prévisualisation et de traitement PDF pourraient ne pas fonctionner.")
+    print("⚠️ Veuillez installer Poppler:")
+    print("   - Linux: apt-get install poppler-utils")
+    print("   - macOS: brew install poppler")
+    print("   - Windows: Téléchargez depuis https://github.com/oschwartz10612/poppler-windows/")
+    return None
+
+# Vérifier Poppler au démarrage
+POPPLER_PATH = check_poppler()
+
 # Configuration
 GOOGLE_API_KEY = "AIzaSyA4ma5pE1pPCzHHn-i9tDWuKqQEgSltMtI"
 genai.configure(api_key=GOOGLE_API_KEY)
@@ -919,7 +958,21 @@ def process_document(file, progress=gr.Progress()):
         
     try:
         if file.name.lower().endswith('.pdf'):
-            images = convert_from_path(file.name)
+            # Utiliser le chemin Poppler détecté au démarrage
+            if POPPLER_PATH:
+                images = convert_from_path(
+                    file.name, 
+                    poppler_path=POPPLER_PATH,
+                    use_pdftocairo=True,
+                    dpi=150
+                )
+            else:
+                print("Trying without poppler_path")
+                images = convert_from_path(
+                    file.name, 
+                    use_pdftocairo=True,
+                    dpi=150
+                )
             
             if len(images) > 10:
                 return {"error": TEXT["error"]["too_many_pages"]}
@@ -954,8 +1007,27 @@ def update_preview(file):
         
     if file.name.lower().endswith('.pdf'):
         try:
-            # Just show first 3 pages
-            images = convert_from_path(file.name, first_page=1, last_page=3)
+            # Utiliser le chemin Poppler détecté au démarrage
+            if POPPLER_PATH:
+                images = convert_from_path(
+                    file.name, 
+                    first_page=1, 
+                    last_page=3, 
+                    poppler_path=POPPLER_PATH,
+                    use_pdftocairo=True,
+                    dpi=150
+                )
+            else:
+                # Essayer sans spécifier le chemin, en utilisant des options simplifiées
+                print("Trying without poppler_path")
+                images = convert_from_path(
+                    file.name, 
+                    first_page=1, 
+                    last_page=3,
+                    use_pdftocairo=True,
+                    dpi=150
+                )
+            
             image_paths = []
             
             for i, img in enumerate(images):
@@ -963,8 +1035,11 @@ def update_preview(file):
                 img.save(temp_filename)
                 image_paths.append(temp_filename)
                 
+            print(f"Successfully created {len(image_paths)} preview images")
             return image_paths
-        except:
+        except Exception as e:
+            print(f"Error converting PDF to images: {str(e)}")
+            # En cas d'erreur, retourner une image d'erreur qui sera affichée
             return []
     elif file.name.lower().endswith(('.png', '.jpg', '.jpeg')):
         return [file.name]