Add image processing functionality and update dependencies

app.py

@@ -1,6 +1,8 @@
 from transformers import NougatProcessor, VisionEncoderDecoderModel
 import gradio as gr
 import torch
+from PIL import Image
+from pathlib import Path
 from pdf2image import convert_from_path
 
 # Load the model and processor
@@ -67,6 +69,8 @@ This is a demo for the Arabic Small Nougat model. It is an end-to-end OCR model
 **Note**: The model is a prototype in my book and may not work well on all types of images and PDFs. **Check the output carefully before using it for any serious work.**
 """
 
+example_images = [Image.open(Path(__file__).parent / "book_page.jpeg")]
+
 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>")
     gr.Markdown(model_description)
@@ -74,10 +78,11 @@ with gr.Blocks(title="Arabic Small Nougat") as demo:
     with gr.Tab("Extract Text from Image"):
         with gr.Row():
             with gr.Column():
-                image = gr.Image(label="Input Image", type="pil")
+                input_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)
+        image_submit_button.click(extract_text_from_image, inputs=[input_image], outputs=output)
+        gr.Examples(example_images, [input_image], output, extract_text_from_image, cache_examples=True)
     
     with gr.Tab("Extract Text from PDF"):
         with gr.Row():

requirements.txt

@@ -2,4 +2,7 @@ pdf2image
 torch
 transformers
 gradio
-python-Levenshtein
+python-Levenshtein
+pillow
+pathlib
+nltk