Update app.py

app.py

@@ -1,41 +1,48 @@
 import gradio as gr
 import random
 import re
+import json
 from pathlib import Path
 from huggingface_hub import InferenceClient, snapshot_download
-from datasets import load_dataset
 import shutil
 
-# Descargar y preparar dataset de tarot
 def prepare_tarot_dataset():
+    repo_id = "multimodalart/1920-raider-waite-tarot-public-domain"
     dataset_path = Path("tarot_dataset")
     image_dir = dataset_path / "images"
     image_dir.mkdir(parents=True, exist_ok=True)
     
-    # Descargar dataset
-    dataset = load_dataset("multimodalart/1920-raider-waite-tarot-public-domain", split="train")
+    # Descargar repositorio completo
+    repo_path = snapshot_download(repo_id=repo_id, repo_type="dataset")
     
-    # Procesar metadatos y organizar imágenes
+    # Procesar archivo JSONL manualmente
+    jsonl_path = Path(repo_path) / "output_file.jsonl"
     card_map = {}
-    pattern = re.compile(r'"([^"]+)"$')  # Para extraer el nombre de la carta
-    
-    for item in dataset:
-        text = item["text"]
-        match = pattern.search(text)
-        if match:
-            card_name = match.group(1).lower()
-            src_path = Path(item["file_name"])
-            dest_path = image_dir / f"{card_name.replace(' ', '_')}.jpg"
-            
-            if not dest_path.exists():
-                try:
-                    # Copiar imagen desde el dataset descargado
-                    shutil.copy(src_path, dest_path)
-                except:
-                    continue
-            
-            card_map[card_name] = str(dest_path)
-    
+    pattern = re.compile(r'"([^"]+)"$')
+
+    with open(jsonl_path, "r") as f:
+        for line in f:
+            try:
+                data = json.loads(line)
+                text = data["text"]
+                file_name = data["file_name"]
+                
+                # Extraer nombre de la carta
+                match = pattern.search(text)
+                if match:
+                    card_name = match.group(1).lower().replace('"', '')
+                    src_image = Path(repo_path) / file_name
+                    dest_image = image_dir / f"{card_name.replace(' ', '_')}.jpg"
+                    
+                    if not dest_image.exists():
+                        shutil.copy(src_image, dest_image)
+                    
+                    card_map[card_name] = str(dest_image)
+                    
+            except (KeyError, json.JSONDecodeError) as e:
+                print(f"Error procesando línea: {e}")
+                continue
+
     return card_map
 
 # Preparar dataset y obtener mapeo de cartas