nougat / app.py
ysharma's picture
ysharma HF staff
Update app.py
4b3f925
import gradio as gr
import subprocess
css = """
.mkd {
height: 500px;
overflow: auto;
border: 1px solid #ccc;
}
"""
def nougat_ocr(file_name):
print('******* inside nougat_ocr *******')
# CLI Command to run
cli_command = [
'nougat',
'--out', 'output',
'pdf', f'{file_name}',
'--checkpoint', 'nougat'
]
# Run the command and get .mmd file in an output folder
subprocess.run(cli_command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
return
def predict(pdf_file):
print('******* inside predict *******')
print(f"temporary file - {pdf_file.name}")
pdf_name = pdf_file.name.split('/')[-1].split('.')[0]
print(f"pdf file name - {pdf_name}")
#! Get prediction for a PDF using nougat
nougat_ocr(pdf_file.name)
print("BAACCKKK")
# Open the multimarkdown (.mmd) file for reading
with open(f'output/{pdf_name}.mmd', 'r') as file:
content = file.read()
return content
with gr.Blocks(css=css) as demo:
gr.HTML("<h1><center>Nougat: Neural Optical Understanding for Academic Documents<center><h1>")
gr.HTML("<h3><center>Lukas Blecher et al. <a href='https://arxiv.org/pdf/2308.13418.pdf' target='_blank'>Paper</a>, <a href='https://facebookresearch.github.io/nougat/'>Project</a><center></h3>")
with gr.Row():
pdf_file = gr.File(label='Upload a PDF', scale=1)
mkd = gr.Markdown('<h2><center><i>OR</i></center></h2>',scale=1)
pdf_link = gr.Textbox(placeholder='Enter an arxiv link here', label='Provide a link', scale=1)
btn = gr.Button()
parsed_output = gr.Markdown(elem_id='mkd')
btn.click(predict, pdf_file, parsed_output )
demo.queue()
demo.launch(debug=True)