Convert scanned pdf to text python

Question

I have a scanned pdf file and I try to extract text from it. I tried to use pypdfocr to make ocr on it but I have error:

"could not found ghostscript in the usual place"

After searching I found this solution Linking Ghostscript to pypdfocr in Windows Platform and I tried to download GhostScript and put it in environment variable but it still has the same error.

How can I searh text in my scanned pdf file using python?

Thanks.

Edit: here is my code sample:

import os
import sys
import re
import json
import shutil
import glob
from pypdfocr import pypdfocr_gs
from pypdfocr import pypdfocr_tesseract 
from PIL import Image

path = PATH_TO_MY_SCANNED_PDF
mainL = []
kk = {}


def new_init(self, kk):
    self.lang = 'heb'   
    self.binary = "tesseract"
    self.msgs = {
            'TS_MISSING': """ 
                Could not execute %s
                Please make sure you have Tesseract installed correctly
                """ % self.binary,
            'TS_VERSION':'Tesseract version is too old',
            'TS_img_MISSING':'Cannot find specified tiff file',
            'TS_FAILED': 'Tesseract-OCR execution failed!',
        }

pypdfocr_tesseract.PyTesseract.__init__ = new_init  

wow = pypdfocr_gs.PyGs(kk)
tt = pypdfocr_tesseract.PyTesseract(kk)


def secFile(filename,oldfilename):
    wow.make_img_from_pdf(filename)


    files = glob.glob("X:/e206333106/ocr-114/balagan/" + '*.jpg')  
    for file in files:
        im = Image.open(file)
        im.save(file + ".tiff") 

    files = glob.glob("PATH" + '*.tiff')  
    for file in files:
        tt.make_hocr_from_pnm(file)
    pdftxt = ""    
    files = glob.glob("PATH" + '*.html') 
    for file in files:
        with open(file) as myfile:
            pdftxt = pdftxt + "#" + "".join(line.rstrip() for line in myfile)
    findNum(pdftxt,oldfilename)

    folder ="PATH"

    for the_file in os.listdir(folder):
        file_path = os.path.join(folder, the_file)
        try:
            if os.path.isfile(file_path):
                os.unlink(file_path)
        except Exception, e:
            print e

def pdf2ocr(filename):
    pdffile = filename
    os.system('pypdfocr -l heb ' + pdffile)

def ocr2txt(filename):  
    pdffile = filename


    output1 = pdffile.replace(".pdf","_ocr.txt")
    output1 = "PATH" + os.path.basename(output1)

    input1 = pdffile.replace(".pdf","_ocr.pdf")

    os.system("pdf2txt" -o  + output1 + " " + input1) 

    with open(output1) as myfile:
        pdftxt="".join(line.rstrip() for line in myfile)
    findNum(pdftxt,filename)


def findNum(pdftxt,pdffile):
    l = re.findall(r'\b\d+\b', pdftxt)


    output = open('PATH' + os.path.basename(pdffile) + '.txt', 'w')
    for i in l:
        output.write(",")
        output.write(i)
    output.close()    

def is_ascii(s):
    return all(ord(c) < 128 for c in s)

i = 0     
files = glob.glob(path + '\\*.pdf') 
print path  
print files 
for file in files:
    if file.endswith(".pdf"):
        if is_ascii(file):
            print file
            pdf2ocr(file)    
            ocr2txt(file)
        else:
            newname = "PATH" + str(i) + ".pdf"
            shutil.copyfile(file, newname)
            print newname
            secFile(newname,file)
        i = i + 1

files = glob.glob(path + '\\' + '*_ocr.pdf')         

for file in files:
    print file
    shutil.copyfile(file, "PATH" + os.path.basename(file))
    os.remove(file)

Could you provide your code sample? – Keeper Aug 03 '17 at 09:37 — Keeper, Aug 03 '17 at 09:37

TRINADH NAGUBADI · Answer 1 · 2019-05-24T13:02:35.620

Take a look at my code it is worked for me.

import os
import io
from PIL import Image
import pytesseract
from wand.image import Image as wi
import gc



pdf=wi(filename=pdf_path,resolution=300)
pdfImg=pdf.convert('jpeg')

imgBlobs=[]
extracted_text=[]

def Get_text_from_image(pdf_path):
    pdf=wi(filename=pdf_path,resolution=300)
    pdfImg=pdf.convert('jpeg')
    imgBlobs=[]
    extracted_text=[]
    for img in pdfImg.sequence:
        page=wi(image=img)
        imgBlobs.append(page.make_blob('jpeg'))

    for imgBlob in imgBlobs:
        im=Image.open(io.BytesIO(imgBlob))
        text=pytesseract.image_to_string(im,lang='eng')
        extracted_text.append(text)

    return (extracted_text)

I fixed it for me by editing the /etc/ImageMagick-6/policy.xml and changed the rights for the pdf line to "read|write":

Open the terminal and change the path

cd /etc/ImageMagick-6
nano policy.xml
<policy domain="coder" rights="read" pattern="PDF" /> 
change to
<policy domain="coder" rights="read|write" pattern="PDF" />
exit

When i was extracting the pdf images to text i faced some issues please go through the below link

https://stackoverflow.com/questions/52699608/wand-policy-error- 
error-constitute-c-readimage-412

https://stackoverflow.com/questions/52861946/imagemagick-not- 
authorized-to-convert-pdf-to-an-image

Increasing the memory limit  please go through the below link
enter code here
https://github.com/phw/peek/issues/112
https://github.com/ImageMagick/ImageMagick/issues/396

is there a method to extract the location, font, size, etc. of the text so that you can create a pdf file with text where it is supposed to be — Nv7, Jun 05 '20 at 18:25

score 11 · Answer 2 · edited Jun 20 '20 at 09:12

Convert pdfs, using pytesseract to do the OCR, and export each page in the pdfs to a text file.

Install these....

conda install -c conda-forge pytesseract

conda install -c conda-forge tesseract

pip install pdf2image

import pytesseract
from pdf2image import convert_from_path
import glob

pdfs = glob.glob(r"yourPath\*.pdf")

for pdf_path in pdfs:
    pages = convert_from_path(pdf_path, 500)

    for pageNum,imgBlob in enumerate(pages):
        text = pytesseract.image_to_string(imgBlob,lang='eng')

        with open(f'{pdf_path[:-4]}_page{pageNum}.txt', 'w') as the_file:
            the_file.write(text)

score 5 · Accepted Answer · edited Jul 12 '18 at 22:23

5

Take a look at this library: https://pypi.python.org/pypi/pypdfocr but a PDF file can have also images in it. You may be able to analyse the page content streams. Some scanners break up the single scanned page into images, so you won't get the text with ghostscript.

edited Jul 12 '18 at 22:23

jeremy

307
4
15

answered Aug 03 '17 at 09:40

ghovat

1,033
1
12
38

still the same error, I wrote **pypdfocr filename.pdf** in the command line and the error: **ERROR: Could not find Ghostscript in the usual place; please specify it using your config file** – Michal Aug 03 '17 at 09:52
which os do you use? – ghovat Aug 03 '17 at 09:56
I'm using windows 64 bit – Michal Aug 03 '17 at 10:06
did you install ghostscript with pip? `pip install ghostscript` – ghovat Aug 03 '17 at 10:18
Possibly its trying to find the 32-bit version of GS, try installing that – KenS Aug 03 '17 at 10:24
I downloaded the 64-bit version – Michal Aug 03 '17 at 11:32
@Michal as kens mentioned download the 32 bit version as well – ghovat Aug 03 '17 at 13:09
Thanks! I downloaded the 32-bit version and it works! – Michal Aug 06 '17 at 05:47

score -1 · Answer 4 · edited Mar 05 '20 at 13:12

-1

PyPDF2 is a python library built as a PDF toolkit. It is capable of:

Extracting document information (title, author, …)
Splitting documents page by page
Merging documents page by page
Cropping pages
Merging multiple pages into a single page
Encrypting and decrypting PDF files
and more!

To install PyPDF2, run following command from command line:

pip install PyPDF2

CODE:

import PyPDF2 

pdfFileObj = open('myPdf.pdf', 'rb') 


pdfReader = PyPDF2.PdfFileReader(pdfFileObj) 

print(pdfReader.numPages) 

pageObj = pdfReader.getPage(0) 

print(pageObj.extractText()) 

pdfFileObj.close()

edited Mar 05 '20 at 13:12

Joey

1,436
2
19
33

answered Mar 05 '20 at 12:51

Ali

15
2

7

I don't think this does OCR though. – Junuxx May 07 '20 at 03:04
6

This works perfectly for text-formatted PDF's. As soon as you try to input a scanned one (e.g. an image) it will not work. – Havard Kleven Jul 24 '20 at 06:56

score -1 · Answer 5 · edited Feb 09 '22 at 05:40

-1

This solution works for Linux operating system(NoelOCR)

Install NoelOCR
```
 pip3 install NoelOCR
```

Use it

 import NoelOCR as nm
 text = nm.processPDF('input.pdf')
 print(text)

After that you should have your plain text from scanned PDF.

edited Feb 09 '22 at 05:40

Suraj Rao

29,388
11
94
103

answered Feb 09 '22 at 05:38

Moses Noel

15
3

Convert scanned pdf to text python

5 Answers5

Linked