PDF OCR console version or python?

Question

Python

Is there something suitable for PDF recognition (interested in text) via the command line or to fence your own in python, then can you tell me a more or less adequate library?

Answer 1

cuneiform from the command line normally recognizes and saves. And so - tesseract and pyocr. PDF does not directly know how, but rasterizing it to PNG is business something ...
Something like this:

from wand.image import Image as Img
from wand.color import Color
from PIL import Image
import pyocr
import pyocr.builders
import os
from timeit import default_timer as timer

pdf_name = '1.pdf'
pdf_path = os.path.join(os.getcwd(), pdf_name)

img_name = 'pdf_1'
image = f'{img_name}.png'

with Img(filename=pdf_path, resolution=300) as img:
    img.format = 'png'
    img.background_color = Color('white')
    img.alpha_channel = 'remove'    
    img.save(filename=image)

tools = pyocr.get_available_tools()[1]
lang = tools.get_available_languages()

builder = pyocr.builders.TextBuilder()

start = timer()
text = tools.image_to_string(Image.open(image), lang='rus', 
                             builder=builder)
end = timer()
print(f"{end - start} \n\n")

print(text)

PyOCR - there is normal documentation