Spaces:
Runtime error
Runtime error
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import fitz
|
2 |
+
import io
|
3 |
+
import base64
|
4 |
+
from PIL import Image
|
5 |
+
import gradio as gr
|
6 |
+
import cv2
|
7 |
+
import tempfile
|
8 |
+
import os
|
9 |
+
|
10 |
+
def pdf_to_img(pdf_path):
|
11 |
+
pdf_document = fitz.open(pdf_path)
|
12 |
+
counter = 1
|
13 |
+
img_list = []
|
14 |
+
for page_number in range(len(pdf_document)):
|
15 |
+
page = pdf_document[page_number]
|
16 |
+
image_list = page.get_images()
|
17 |
+
for image in image_list:
|
18 |
+
base_img = pdf_document.extract_image(image[0])
|
19 |
+
image_data = base_img["image"]
|
20 |
+
img = Image.open(io.BytesIO(image_data))
|
21 |
+
extention = base_img['ext']
|
22 |
+
img.save(open(f"image{counter}.{extention}","wb"))
|
23 |
+
img_list.append(f"image{counter}.{extention}")
|
24 |
+
counter += 1
|
25 |
+
return (img_list)
|
26 |
+
|
27 |
+
title = "Extract Image and Text"
|
28 |
+
|
29 |
+
|
30 |
+
with gr.Blocks(theme=gr.themes.Glass(primary_hue=gr.themes.colors.slate)) as demo:
|
31 |
+
gr.Markdown(f'<h1 style="text-align: center;">{title}</h1>')
|
32 |
+
with gr.Row():
|
33 |
+
with gr.Row():
|
34 |
+
with gr.Column():
|
35 |
+
file_input = gr.File(type="filepath", label="Upload .pdf file")
|
36 |
+
upload_button = gr.Button(value="Show Images")
|
37 |
+
img_gallery = gr.Gallery(label="Generated images", show_label=True, elem_id="gallery", object_fit="contain", height="auto",allow_preview=True)
|
38 |
+
|
39 |
+
upload_button.click(pdf_to_img, inputs=file_input, outputs=[img_gallery])
|
40 |
+
demo.launch()
|