Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
from docling.document_converter import DocumentConverter
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
def pdf_to_json(source):
|
6 |
+
converter = DocumentConverter()
|
7 |
+
result = converter.convert(source)
|
8 |
+
|
9 |
+
try:
|
10 |
+
json_output = result.document.export_to_json()
|
11 |
+
except AttributeError:
|
12 |
+
content = {
|
13 |
+
"title": result.document.title if hasattr(result.document, 'title') else "Untitled",
|
14 |
+
"sections": [section.text for section in result.document.sections]
|
15 |
+
}
|
16 |
+
json_output = json.dumps(content, indent=2)
|
17 |
+
|
18 |
+
return json_output
|
19 |
+
|
20 |
+
# Gradio interface
|
21 |
+
iface = gr.Interface(
|
22 |
+
fn=pdf_to_json,
|
23 |
+
inputs=gr.Textbox(label="PDF URL"),
|
24 |
+
outputs="text",
|
25 |
+
title="PDF to JSON Converter"
|
26 |
+
)
|
27 |
+
|
28 |
+
iface.launch()
|