Utility: Grep in OCR (tesseract) output

ocr_grep.py

"""
To be used when you want to grep in OCR output.

python ocr_grep.py Screenshot\*.png -i werkzaamheden
                   <glob pattern>   <grep arguments>
                   
Each file is run through tesseract, its output is then piped through
 `grep <grep_arguments>`

#futurework: Cache OCR output
"""
import argparse
import glob
import os

from subprocess import Popen, PIPE

def tesseract_text(name, grep_pattern):
    with Popen(["tesseract", name, "-"], stdout=PIPE, stderr=PIPE) as proc:
        with Popen(["grep"] + grep_pattern,
                   stdin=proc.stdout, stdout=PIPE) as grep_proc:
            grep_out = grep_proc.stdout.read().decode('utf8').strip()
            for line in grep_out.split("\n"):
                if (line):
                    print("{} {}".format(name, line))



def for_globbed_files(pattern, grep_pattern):
    for file_name in glob.glob(pattern):
        tesseract_text(os.path.abspath(file_name), grep_pattern)

parser = argparse.ArgumentParser(description='Run OCR (tesseract) on all '
    'matching files')
parser.add_argument('pattern', default='*.png', type=str, help='Glob pattern')
parser.add_argument('args', nargs=argparse.REMAINDER,
                    help='Args passed to grep')


if __name__ == "__main__":
    args = parser.parse_args()
    for_globbed_files(args.pattern, args.args)