Add app.py and others

app.py

@@ -0,0 +1,77 @@
+from transformers import NougatProcessor, VisionEncoderDecoderModel
+import gradio as gr
+from pdf2image import convert_from_path
+
+# Load the model and processor
+processor = NougatProcessor.from_pretrained("MohamedRashad/arabic-small-nougat")
+model = VisionEncoderDecoderModel.from_pretrained("MohamedRashad/arabic-small-nougat")
+device = "cpu"
+
+context_length = 2048
+
+def extract_text_from_image(image):
+    """
+    Extract text from PIL image
+
+    Args:
+    image (PIL.Image): Input image
+
+    Returns:
+    str: Extracted text from the image
+    """
+
+    # prepare PDF image for the model
+    pixel_values = processor(image, return_tensors="pt").pixel_values
+
+    # generate transcription
+    outputs = model.generate(
+        pixel_values.to(device),
+        min_length=1,
+        max_new_tokens=context_length,
+        bad_words_ids=[[processor.tokenizer.unk_token_id]],
+    )
+    page_sequence = processor.batch_decode(outputs, skip_special_tokens=True)[0]
+    page_sequence = processor.post_process_generation(page_sequence, fix_markdown=False)
+    return page_sequence
+
+def extract_text_from_pdf(pdf_path, progress=gr.Progress()):
+    """
+    Extract text from PDF
+    
+    Args:
+    pdf_path (str): Path to the PDF file
+    progress (gr.Progress): Progress bar
+    
+    Returns:
+    str: Extracted text from the PDF
+    """
+    
+    progress(0, desc="Starting...")
+    images = convert_from_path(pdf_path)
+    texts = []
+    for image in progress.tqdm(images):
+        extracted_text = extract_text_from_image(image)
+        texts.append(extracted_text)
+
+    return "\n".join(texts)
+
+with gr.Blocks(title="Arabic Small Nougat") as demo:
+    gr.HTML("<h1 style='text-align: center'>Arabic End-to-End Structured OCR for textbooks</h1>")
+
+    with gr.Tab("Extract Text from Image"):
+        with gr.Row():
+            with gr.Column():
+                image = gr.Image(label="Input Image", type="pil")
+                image_submit_button = gr.Button(value="Submit", variant="primary")
+            output = gr.Markdown(label="Output Markdown", rtl=True)    
+        image_submit_button.click(extract_text_from_image, inputs=[image], outputs=output)
+    
+    with gr.Tab("Extract Text from PDF"):
+        with gr.Row():
+            with gr.Column():
+                pdf = gr.File(label="Input PDF", type="filepath")
+                pdf_submit_button = gr.Button(value="Submit", variant="primary")
+            output = gr.Markdown(label="Output Markdown", rtl=True)    
+        pdf_submit_button.click(extract_text_from_pdf, inputs=[pdf], outputs=output)
+
+demo.queue().launch(share=False)

requirements.txt

@@ -0,0 +1 @@
+pdf2image