Create app.py

app.py

@@ -0,0 +1,42 @@
+import gradio as gr
+from PIL import Image
+from mineru_vl_utils import MinerUClient
+from transformers import AutoProcessor, Qwen2VLForConditionalGeneration
+
+# Charger le modèle MinerU
+model_path = "opendatalab/MinerU2.5-2509-1.2B"
+model = Qwen2VLForConditionalGeneration.from_pretrained(
+    model_path,
+    torch_dtype="auto",
+    device_map="auto"
+)
+processor = AutoProcessor.from_pretrained(model_path, use_fast=True)
+
+client = MinerUClient(
+    backend="transformers",
+    model=model,
+    processor=processor
+)
+
+def extract_from_image(image):
+    # Conversion si nécessaire
+    if not isinstance(image, Image.Image):
+        image = Image.fromarray(image)
+
+    # Extraction
+    blocks = client.two_step_extract(image)
+    # On retourne le texte concaténé
+    extracted_text = "\n".join([b.text for b in blocks if hasattr(b, "text")])
+    return extracted_text
+
+# Interface Gradio
+demo = gr.Interface(
+    fn=extract_from_image,
+    inputs=gr.Image(type="pil"),
+    outputs="text",
+    title="MinerU2.5 - Document Extract",
+    description="Upload an image or PDF page and extract structured text with MinerU2.5"
+)
+
+if __name__ == "__main__":
+    demo.launch()