Skip to content

Instantly share code, notes, and snippets.

@chris-x86-64
Last active August 31, 2015 00:51
Show Gist options
  • Star 1 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save chris-x86-64/10098518 to your computer and use it in GitHub Desktop.
Save chris-x86-64/10098518 to your computer and use it in GitHub Desktop.
Recursive batch OCR script (Requires Tesseract-OCR or Cuneiform).
import os
import sys
import json
from optparse import OptionParser
from PIL import Image
from pyocr import pyocr
from pyocr.builders import TextBuilder
def get_opt():
parser = OptionParser()
parser.add_option("-d", "--directory", action = "store", dest = "directory", help = "[Required] Target directory")
parser.add_option("-l", "--lang", action = "store", dest = "lang", help = "Language", default = "jpn")
parser.add_option("-c", "--config", action = "store", dest = "args", help = "Other config args (e.g. digits, only works with Tesseract)")
parser.add_option("-o", "--output", action = "store", dest = "output_file", help = "Output file (Not supported at this time.)")
(options, args) = parser.parse_args()
if not options.directory:
parser.print_help()
sys.exit(1)
else:
print "[DEBUG] Entering directory %s ..." % options.directory
return options
def select_tool():
tools = pyocr.get_available_tools()[:]
if len(tools) == 0:
print "[EMERG] There are no OCR tools available."
sys.exit(1)
print "[DEBUG] Using %s for OCR..." % (tools[0].get_name())
return tools[0]
def process(matches, options):
tool = select_tool()
builder = TextBuilder()
builder.tesseract_configs = [options.args]
for f in matches:
text = tool.image_to_string(Image.open(f), lang = options.lang, builder = builder)
print json.dumps({"filename": f, "text": text}, ensure_ascii=False)
if __name__ == "__main__":
options = get_opt()
matches = []
for root, dirnames, filenames in os.walk(options.directory):
for f in filenames:
if f.endswith(('.png', '.jpg', '.jpeg')):
matches.append(os.path.join(root,f))
print "[DEBUG] Processing %d images..." % len(matches)
process(matches, options)
@tellts
Copy link

tellts commented Jan 22, 2015

Sorry for the bad English. Which operating system and file handling any format is this script?

@chris-x86-64
Copy link
Author

I'm so sorry that I haven't noticed your comment for quite a while.

This script should run on any operating system supporting Python 2.7 and Tesseract OCR.
To use this on Ubuntu 14.04, you should install tesseract-ocr using APT, then PIL and pyocr using pip.

PIL supports many image types but in this script, particularly PNG and JPEG are supported.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment