abhinavsarkar
commited on
Commit
•
77f71a3
1
Parent(s):
1a4e7ec
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import pymupdf4llm
|
3 |
+
import tempfile
|
4 |
+
import pathlib
|
5 |
+
import markdown2
|
6 |
+
from docx import Document
|
7 |
+
from bs4 import BeautifulSoup
|
8 |
+
|
9 |
+
def pdf_to_markdown(pdf_file):
|
10 |
+
# Create a temporary file to save the uploaded PDF
|
11 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as temp_file:
|
12 |
+
temp_file.write(pdf_file.read())
|
13 |
+
temp_file_path = temp_file.name
|
14 |
+
|
15 |
+
# Convert PDF to Markdown
|
16 |
+
pdf_text = pymupdf4llm.to_markdown(temp_file_path)
|
17 |
+
|
18 |
+
# Save the Markdown content to a file
|
19 |
+
md_file_path = pathlib.Path("Output.md")
|
20 |
+
md_file_path.write_bytes(pdf_text.encode())
|
21 |
+
|
22 |
+
return pdf_text, md_file_path
|
23 |
+
|
24 |
+
def create_docx_from_markdown(md_content):
|
25 |
+
# Convert Markdown to HTML
|
26 |
+
html_content = markdown2.markdown(md_content)
|
27 |
+
|
28 |
+
# Create a new Document
|
29 |
+
doc = Document()
|
30 |
+
doc.add_heading('Converted PDF Content', level=1)
|
31 |
+
|
32 |
+
# Use BeautifulSoup to parse the HTML and extract text
|
33 |
+
soup = BeautifulSoup(html_content, "html.parser")
|
34 |
+
|
35 |
+
# Add the rendered HTML content to the document
|
36 |
+
for element in soup:
|
37 |
+
if element.name == 'h1':
|
38 |
+
doc.add_heading(element.get_text(), level=1)
|
39 |
+
elif element.name == 'h2':
|
40 |
+
doc.add_heading(element.get_text(), level=2)
|
41 |
+
elif element.name == 'h3':
|
42 |
+
doc.add_heading(element.get_text(), level=3)
|
43 |
+
elif element.name == 'p':
|
44 |
+
doc.add_paragraph(element.get_text())
|
45 |
+
elif element.name == 'strong':
|
46 |
+
doc.add_paragraph(element.get_text(), style='IntenseQuote')
|
47 |
+
elif element.name == 'em':
|
48 |
+
p = doc.add_paragraph()
|
49 |
+
p.add_run(element.get_text()).italic = True
|
50 |
+
|
51 |
+
# Save the document
|
52 |
+
docx_file_path = pathlib.Path("Output.docx")
|
53 |
+
doc.save(docx_file_path)
|
54 |
+
|
55 |
+
return docx_file_path
|
56 |
+
|
57 |
+
# Streamlit application
|
58 |
+
st.title("📄 Structured PDF Data Extractor")
|
59 |
+
st.subheader("Upload a PDF file, preview the structured content, and download it if needed.")
|
60 |
+
|
61 |
+
# File uploader for PDF
|
62 |
+
pdf_input = st.file_uploader("Upload PDF", type="pdf")
|
63 |
+
|
64 |
+
if pdf_input is not None:
|
65 |
+
# Convert to Markdown when the PDF is uploaded
|
66 |
+
with st.spinner("Converting PDF to Markdown..."):
|
67 |
+
try:
|
68 |
+
pdf_text, md_file_path = pdf_to_markdown(pdf_input)
|
69 |
+
|
70 |
+
# Display the Markdown content
|
71 |
+
st.markdown("### Markdown Content Preview:", unsafe_allow_html=True)
|
72 |
+
st.markdown(pdf_text, unsafe_allow_html=True)
|
73 |
+
|
74 |
+
# Create a download button for the Markdown file
|
75 |
+
st.markdown("### Download Markdown File:")
|
76 |
+
with open(md_file_path, "rb") as file:
|
77 |
+
st.download_button(
|
78 |
+
label="Download Markdown",
|
79 |
+
data=file,
|
80 |
+
file_name=md_file_path.name,
|
81 |
+
mime="text/markdown"
|
82 |
+
)
|
83 |
+
|
84 |
+
# Create the .docx file from rendered Markdown content
|
85 |
+
docx_file_path = create_docx_from_markdown(pdf_text)
|
86 |
+
|
87 |
+
# Create a download button for the .docx file
|
88 |
+
st.markdown("### Download Word Document:")
|
89 |
+
with open(docx_file_path, "rb") as file:
|
90 |
+
st.download_button(
|
91 |
+
label="Download Word Document",
|
92 |
+
data=file,
|
93 |
+
file_name=docx_file_path.name,
|
94 |
+
mime="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
|
95 |
+
)
|
96 |
+
|
97 |
+
except Exception as e:
|
98 |
+
st.error(f"An error occurred during conversion: {e}")
|
99 |
+
|
100 |
+
# Add some styling to make it visually appealing
|
101 |
+
st.markdown(
|
102 |
+
"""
|
103 |
+
<style>
|
104 |
+
body {
|
105 |
+
font-family: 'Arial', sans-serif;
|
106 |
+
line-height: 1.6;
|
107 |
+
font-size: 16px;
|
108 |
+
color: #333;
|
109 |
+
}
|
110 |
+
.stButton>button {
|
111 |
+
background-color: #4CAF50; /* Green */
|
112 |
+
border: none;
|
113 |
+
color: white;
|
114 |
+
padding: 10px 20px;
|
115 |
+
text-align: center;
|
116 |
+
text-decoration: none;
|
117 |
+
display: inline-block;
|
118 |
+
font-size: 16px;
|
119 |
+
margin: 4px 2px;
|
120 |
+
cursor: pointer;
|
121 |
+
border-radius: 5px;
|
122 |
+
transition: background-color 0.3s;
|
123 |
+
}
|
124 |
+
.stButton>button:hover {
|
125 |
+
background-color: #45a049; /* Darker green */
|
126 |
+
}
|
127 |
+
</style>
|
128 |
+
""",
|
129 |
+
unsafe_allow_html=True
|
130 |
+
)
|