Amrrs commited on
Commit
cf393ba
1 Parent(s): ed3b6df

Upload app.py

Browse files
Files changed (1) hide show
  1. app.py +62 -0
app.py ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import streamlit as st # data app development
2
+ import subprocess # process in the os
3
+ from subprocess import STDOUT, check_call #os process manipuation
4
+ import os #os process manipuation
5
+ import base64 # byte object into a pdf file
6
+ import camelot as cam # extracting tables from PDFs
7
+
8
+ # to run this only once and it's cached
9
+ @st.cache
10
+ def gh():
11
+ """install ghostscript on the linux machine"""
12
+ proc = subprocess.Popen('apt-get install -y ghostscript', shell=True, stdin=None, stdout=open(os.devnull,"wb"), stderr=STDOUT, executable="/bin/bash")
13
+ proc.wait()
14
+
15
+ gh()
16
+
17
+
18
+
19
+ st.title("PDF Table Extractor")
20
+ st.subheader("with `Camelot` Python library")
21
+
22
+ st.image("https://raw.githubusercontent.com/camelot-dev/camelot/master/docs/_static/camelot.png", width=200)
23
+
24
+
25
+ # file uploader on streamlit
26
+
27
+ input_pdf = st.file_uploader(label = "upload your pdf here", type = 'pdf')
28
+
29
+ st.markdown("### Page Number")
30
+
31
+ page_number = st.text_input("Enter the page # from where you want to extract the PDF eg: 3", value = 1)
32
+
33
+ # run this only when a PDF is uploaded
34
+
35
+ if input_pdf is not None:
36
+ # byte object into a PDF file
37
+ with open("input.pdf", "wb") as f:
38
+ base64_pdf = base64.b64encode(input_pdf.read()).decode('utf-8')
39
+ f.write(base64.b64decode(base64_pdf))
40
+ f.close()
41
+
42
+ # read the pdf and parse it using stream
43
+ table = cam.read_pdf("input.pdf", pages = page_number, flavor = 'stream')
44
+
45
+ st.markdown("### Number of Tables")
46
+
47
+ # display the output after parsing
48
+ st.write(table)
49
+
50
+ # display the table
51
+
52
+ if len(table) > 0:
53
+
54
+ # extract the index value of the table
55
+
56
+ option = st.selectbox(label = "Select the Table to be displayed", options = range(len(table) + 1))
57
+
58
+ st.markdown('### Output Table')
59
+
60
+ # display the dataframe
61
+
62
+ st.dataframe(table[int(option)-1].df)