Spaces:
Running
on
Zero
Running
on
Zero
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from docling.document_converter import DocumentConverter
|
3 |
+
|
4 |
+
def convert_document(file, output_format):
|
5 |
+
# Load document and convert it using Docling
|
6 |
+
converter = DocumentConverter()
|
7 |
+
result = converter.convert(file.name)
|
8 |
+
|
9 |
+
# Choose the output format
|
10 |
+
if output_format == "Markdown":
|
11 |
+
converted_text = result.document.export_to_markdown()
|
12 |
+
elif output_format == "JSON":
|
13 |
+
converted_text = result.document.export_to_json()
|
14 |
+
else:
|
15 |
+
converted_text = "Unsupported format"
|
16 |
+
|
17 |
+
# Extract metadata
|
18 |
+
metadata = {
|
19 |
+
"Title": result.document.title,
|
20 |
+
"Author": result.document.author,
|
21 |
+
"Language": result.document.language,
|
22 |
+
"References": result.document.references
|
23 |
+
}
|
24 |
+
|
25 |
+
return converted_text, metadata
|
26 |
+
|
27 |
+
# Define the Gradio interface
|
28 |
+
input_file = gr.inputs.File(label="Upload Document")
|
29 |
+
output_format = gr.inputs.Radio(["Markdown", "JSON"], label="Choose Output Format")
|
30 |
+
output_text = gr.outputs.Textbox(label="Converted Document")
|
31 |
+
output_metadata = gr.outputs.JSON(label="Metadata")
|
32 |
+
|
33 |
+
app = gr.Interface(
|
34 |
+
fn=convert_document,
|
35 |
+
inputs=[input_file, output_format],
|
36 |
+
outputs=[output_text, output_metadata],
|
37 |
+
title="Document Converter with Docling",
|
38 |
+
description="Upload a document (PDF, DOCX, or image), choose the output format, and get the converted document text along with metadata.",
|
39 |
+
)
|
40 |
+
|
41 |
+
app.launch()
|