From de32c12bfc05b3d04df2686373e84494eefed44f Mon Sep 17 00:00:00 2001 From: Walter Purcaro Date: Tue, 24 Feb 2015 22:22:15 +0100 Subject: captcha -> OCR --- module/plugins/captcha/GigasizeCom.py | 4 +- module/plugins/captcha/LinksaveIn.py | 4 +- module/plugins/captcha/NetloadIn.py | 4 +- module/plugins/captcha/OCR.py | 319 +++++++++++++++++++++++++++++++ module/plugins/captcha/ShareonlineBiz.py | 4 +- module/plugins/captcha/captcha.py | 319 ------------------------------- 6 files changed, 327 insertions(+), 327 deletions(-) create mode 100644 module/plugins/captcha/OCR.py delete mode 100644 module/plugins/captcha/captcha.py diff --git a/module/plugins/captcha/GigasizeCom.py b/module/plugins/captcha/GigasizeCom.py index 244cf6a2a..52c41729b 100644 --- a/module/plugins/captcha/GigasizeCom.py +++ b/module/plugins/captcha/GigasizeCom.py @@ -1,12 +1,12 @@ # -*- coding: utf-8 -*- -from module.plugins.captcha.captcha import OCR +from module.plugins.captcha.OCR import OCR class GigasizeCom(OCR): __name__ = "GigasizeCom" __type__ = "ocr" - __version__ = "0.10" + __version__ = "0.11" __description__ = """Gigasize.com ocr plugin""" __license__ = "GPLv3" diff --git a/module/plugins/captcha/LinksaveIn.py b/module/plugins/captcha/LinksaveIn.py index de6b0e7ff..b5cb0f608 100644 --- a/module/plugins/captcha/LinksaveIn.py +++ b/module/plugins/captcha/LinksaveIn.py @@ -9,13 +9,13 @@ from glob import glob from os import sep from os.path import abspath, dirname -from module.plugins.captcha.captcha import OCR +from module.plugins.captcha.OCR import OCR class LinksaveIn(OCR): __name__ = "LinksaveIn" __type__ = "ocr" - __version__ = "0.10" + __version__ = "0.11" __description__ = """Linksave.in ocr plugin""" __license__ = "GPLv3" diff --git a/module/plugins/captcha/NetloadIn.py b/module/plugins/captcha/NetloadIn.py index 28eb18fb5..1fb258c47 100644 --- a/module/plugins/captcha/NetloadIn.py +++ b/module/plugins/captcha/NetloadIn.py @@ -1,12 +1,12 @@ # -*- coding: utf-8 -*- -from module.plugins.captcha.captcha import OCR +from module.plugins.captcha.OCR import OCR class NetloadIn(OCR): __name__ = "NetloadIn" __type__ = "ocr" - __version__ = "0.10" + __version__ = "0.11" __description__ = """Netload.in ocr plugin""" __license__ = "GPLv3" diff --git a/module/plugins/captcha/OCR.py b/module/plugins/captcha/OCR.py new file mode 100644 index 000000000..1874ba07d --- /dev/null +++ b/module/plugins/captcha/OCR.py @@ -0,0 +1,319 @@ +# -*- coding: utf-8 -*- + +from __future__ import with_statement + +try: + from PIL import Image, GifImagePlugin, JpegImagePlugin, PngImagePlugin, TiffImagePlugin + +except ImportError: + import Image, GifImagePlugin, JpegImagePlugin, PngImagePlugin, TiffImagePlugin + +import logging +import os +import subprocess +#import tempfile + +from module.utils import save_join + + +class OCR(object): + __name__ = "OCR" + __type__ = "ocr" + __version__ = "0.11" + + __description__ = """OCR base plugin""" + __license__ = "GPLv3" + __authors__ = [("pyLoad Team", "admin@pyload.org")] + + + def __init__(self): + self.logger = logging.getLogger("log") + + + def load_image(self, image): + self.image = Image.open(image) + self.pixels = self.image.load() + self.result_captcha = '' + + + def unload(self): + """delete all tmp images""" + pass + + + def threshold(self, value): + self.image = self.image.point(lambda a: a * value + 10) + + + def run(self, command): + """Run a command""" + + popen = subprocess.Popen(command, bufsize = -1, stdout=subprocess.PIPE, stderr=subprocess.PIPE) + popen.wait() + output = popen.stdout.read() +" | "+ popen.stderr.read() + popen.stdout.close() + popen.stderr.close() + self.logger.debug("Tesseract ReturnCode %s Output: %s" % (popen.returncode, output)) + + + def run_tesser(self, subset=False, digits=True, lowercase=True, uppercase=True): + #tmpTif = tempfile.NamedTemporaryFile(suffix=".tif") + try: + tmpTif = open(save_join("tmp", "tmpTif_%s.tif" % self.__name__), "wb") + tmpTif.close() + + #tmpTxt = tempfile.NamedTemporaryFile(suffix=".txt") + tmpTxt = open(save_join("tmp", "tmpTxt_%s.txt" % self.__name__), "wb") + tmpTxt.close() + + except IOError, e: + self.logError(e) + return + + self.logger.debug("save tiff") + self.image.save(tmpTif.name, 'TIFF') + + if os.name == "nt": + tessparams = [os.path.join(pypath, "tesseract", "tesseract.exe")] + else: + tessparams = ["tesseract"] + + tessparams.extend( [os.path.abspath(tmpTif.name), os.path.abspath(tmpTxt.name).replace(".txt", "")] ) + + if subset and (digits or lowercase or uppercase): + #tmpSub = tempfile.NamedTemporaryFile(suffix=".subset") + with open(save_join("tmp", "tmpSub_%s.subset" % self.__name__), "wb") as tmpSub: + tmpSub.write("tessedit_char_whitelist ") + + if digits: + tmpSub.write("0123456789") + if lowercase: + tmpSub.write("abcdefghijklmnopqrstuvwxyz") + if uppercase: + tmpSub.write("ABCDEFGHIJKLMNOPQRSTUVWXYZ") + + tmpSub.write("\n") + tessparams.append("nobatch") + tessparams.append(os.path.abspath(tmpSub.name)) + + self.logger.debug("run tesseract") + self.run(tessparams) + self.logger.debug("read txt") + + try: + with open(tmpTxt.name, 'r') as f: + self.result_captcha = f.read().replace("\n", "") + except Exception: + self.result_captcha = "" + + self.logger.debug(self.result_captcha) + try: + os.remove(tmpTif.name) + os.remove(tmpTxt.name) + if subset and (digits or lowercase or uppercase): + os.remove(tmpSub.name) + except Exception: + pass + + + def get_captcha(self, name): + raise NotImplementedError + + + def to_greyscale(self): + if self.image.mode != 'L': + self.image = self.image.convert('L') + + self.pixels = self.image.load() + + + def eval_black_white(self, limit): + self.pixels = self.image.load() + w, h = self.image.size + for x in xrange(w): + for y in xrange(h): + if self.pixels[x, y] > limit: + self.pixels[x, y] = 255 + else: + self.pixels[x, y] = 0 + + + def clean(self, allowed): + pixels = self.pixels + + w, h = self.image.size + + for x in xrange(w): + for y in xrange(h): + if pixels[x, y] == 255: + continue + # No point in processing white pixels since we only want to remove black pixel + count = 0 + + try: + if pixels[x-1, y-1] != 255: + count += 1 + if pixels[x-1, y] != 255: + count += 1 + if pixels[x-1, y + 1] != 255: + count += 1 + if pixels[x, y + 1] != 255: + count += 1 + if pixels[x + 1, y + 1] != 255: + count += 1 + if pixels[x + 1, y] != 255: + count += 1 + if pixels[x + 1, y-1] != 255: + count += 1 + if pixels[x, y-1] != 255: + count += 1 + except Exception: + pass + + # not enough neighbors are dark pixels so mark this pixel + # to be changed to white + if count < allowed: + pixels[x, y] = 1 + + # second pass: this time set all 1's to 255 (white) + for x in xrange(w): + for y in xrange(h): + if pixels[x, y] == 1: + pixels[x, y] = 255 + + self.pixels = pixels + + + def derotate_by_average(self): + """rotate by checking each angle and guess most suitable""" + + w, h = self.image.size + pixels = self.pixels + + for x in xrange(w): + for y in xrange(h): + if pixels[x, y] == 0: + pixels[x, y] = 155 + + highest = {} + counts = {} + + for angle in xrange(-45, 45): + + tmpimage = self.image.rotate(angle) + + pixels = tmpimage.load() + + w, h = self.image.size + + for x in xrange(w): + for y in xrange(h): + if pixels[x, y] == 0: + pixels[x, y] = 255 + + + count = {} + + for x in xrange(w): + count[x] = 0 + for y in xrange(h): + if pixels[x, y] == 155: + count[x] += 1 + + sum = 0 + cnt = 0 + + for x in count.values(): + if x != 0: + sum += x + cnt += 1 + + avg = sum / cnt + counts[angle] = cnt + highest[angle] = 0 + for x in count.values(): + if x > highest[angle]: + highest[angle] = x + + highest[angle] = highest[angle] - avg + + hkey = 0 + hvalue = 0 + + for key, value in highest.iteritems(): + if value > hvalue: + hkey = key + hvalue = value + + self.image = self.image.rotate(hkey) + pixels = self.image.load() + + for x in xrange(w): + for y in xrange(h): + if pixels[x, y] == 0: + pixels[x, y] = 255 + + if pixels[x, y] == 155: + pixels[x, y] = 0 + + self.pixels = pixels + + + def split_captcha_letters(self): + captcha = self.image + started = False + letters = [] + width, height = captcha.size + bottomY, topY = 0, height + pixels = captcha.load() + + for x in xrange(width): + black_pixel_in_col = False + for y in xrange(height): + if pixels[x, y] != 255: + if not started: + started = True + firstX = x + lastX = x + + if y > bottomY: + bottomY = y + if y < topY: + topY = y + if x > lastX: + lastX = x + + black_pixel_in_col = True + + if black_pixel_in_col is False and started is True: + rect = (firstX, topY, lastX, bottomY) + new_captcha = captcha.crop(rect) + + w, h = new_captcha.size + if w > 5 and h > 5: + letters.append(new_captcha) + + started = False + bottomY, topY = 0, height + + return letters + + + def correct(self, values, var=None): + if var: + result = var + else: + result = self.result_captcha + + for key, item in values.iteritems(): + + if key.__class__ == str: + result = result.replace(key, item) + else: + for expr in key: + result = result.replace(expr, item) + + if var: + return result + else: + self.result_captcha = result diff --git a/module/plugins/captcha/ShareonlineBiz.py b/module/plugins/captcha/ShareonlineBiz.py index 8210e8859..6fad66600 100644 --- a/module/plugins/captcha/ShareonlineBiz.py +++ b/module/plugins/captcha/ShareonlineBiz.py @@ -1,12 +1,12 @@ # -*- coding: utf-8 -*- -from module.plugins.captcha.captcha import OCR +from module.plugins.captcha.OCR import OCR class ShareonlineBiz(OCR): __name__ = "ShareonlineBiz" __type__ = "ocr" - __version__ = "0.10" + __version__ = "0.11" __description__ = """Shareonline.biz ocr plugin""" __license__ = "GPLv3" diff --git a/module/plugins/captcha/captcha.py b/module/plugins/captcha/captcha.py deleted file mode 100644 index 1874ba07d..000000000 --- a/module/plugins/captcha/captcha.py +++ /dev/null @@ -1,319 +0,0 @@ -# -*- coding: utf-8 -*- - -from __future__ import with_statement - -try: - from PIL import Image, GifImagePlugin, JpegImagePlugin, PngImagePlugin, TiffImagePlugin - -except ImportError: - import Image, GifImagePlugin, JpegImagePlugin, PngImagePlugin, TiffImagePlugin - -import logging -import os -import subprocess -#import tempfile - -from module.utils import save_join - - -class OCR(object): - __name__ = "OCR" - __type__ = "ocr" - __version__ = "0.11" - - __description__ = """OCR base plugin""" - __license__ = "GPLv3" - __authors__ = [("pyLoad Team", "admin@pyload.org")] - - - def __init__(self): - self.logger = logging.getLogger("log") - - - def load_image(self, image): - self.image = Image.open(image) - self.pixels = self.image.load() - self.result_captcha = '' - - - def unload(self): - """delete all tmp images""" - pass - - - def threshold(self, value): - self.image = self.image.point(lambda a: a * value + 10) - - - def run(self, command): - """Run a command""" - - popen = subprocess.Popen(command, bufsize = -1, stdout=subprocess.PIPE, stderr=subprocess.PIPE) - popen.wait() - output = popen.stdout.read() +" | "+ popen.stderr.read() - popen.stdout.close() - popen.stderr.close() - self.logger.debug("Tesseract ReturnCode %s Output: %s" % (popen.returncode, output)) - - - def run_tesser(self, subset=False, digits=True, lowercase=True, uppercase=True): - #tmpTif = tempfile.NamedTemporaryFile(suffix=".tif") - try: - tmpTif = open(save_join("tmp", "tmpTif_%s.tif" % self.__name__), "wb") - tmpTif.close() - - #tmpTxt = tempfile.NamedTemporaryFile(suffix=".txt") - tmpTxt = open(save_join("tmp", "tmpTxt_%s.txt" % self.__name__), "wb") - tmpTxt.close() - - except IOError, e: - self.logError(e) - return - - self.logger.debug("save tiff") - self.image.save(tmpTif.name, 'TIFF') - - if os.name == "nt": - tessparams = [os.path.join(pypath, "tesseract", "tesseract.exe")] - else: - tessparams = ["tesseract"] - - tessparams.extend( [os.path.abspath(tmpTif.name), os.path.abspath(tmpTxt.name).replace(".txt", "")] ) - - if subset and (digits or lowercase or uppercase): - #tmpSub = tempfile.NamedTemporaryFile(suffix=".subset") - with open(save_join("tmp", "tmpSub_%s.subset" % self.__name__), "wb") as tmpSub: - tmpSub.write("tessedit_char_whitelist ") - - if digits: - tmpSub.write("0123456789") - if lowercase: - tmpSub.write("abcdefghijklmnopqrstuvwxyz") - if uppercase: - tmpSub.write("ABCDEFGHIJKLMNOPQRSTUVWXYZ") - - tmpSub.write("\n") - tessparams.append("nobatch") - tessparams.append(os.path.abspath(tmpSub.name)) - - self.logger.debug("run tesseract") - self.run(tessparams) - self.logger.debug("read txt") - - try: - with open(tmpTxt.name, 'r') as f: - self.result_captcha = f.read().replace("\n", "") - except Exception: - self.result_captcha = "" - - self.logger.debug(self.result_captcha) - try: - os.remove(tmpTif.name) - os.remove(tmpTxt.name) - if subset and (digits or lowercase or uppercase): - os.remove(tmpSub.name) - except Exception: - pass - - - def get_captcha(self, name): - raise NotImplementedError - - - def to_greyscale(self): - if self.image.mode != 'L': - self.image = self.image.convert('L') - - self.pixels = self.image.load() - - - def eval_black_white(self, limit): - self.pixels = self.image.load() - w, h = self.image.size - for x in xrange(w): - for y in xrange(h): - if self.pixels[x, y] > limit: - self.pixels[x, y] = 255 - else: - self.pixels[x, y] = 0 - - - def clean(self, allowed): - pixels = self.pixels - - w, h = self.image.size - - for x in xrange(w): - for y in xrange(h): - if pixels[x, y] == 255: - continue - # No point in processing white pixels since we only want to remove black pixel - count = 0 - - try: - if pixels[x-1, y-1] != 255: - count += 1 - if pixels[x-1, y] != 255: - count += 1 - if pixels[x-1, y + 1] != 255: - count += 1 - if pixels[x, y + 1] != 255: - count += 1 - if pixels[x + 1, y + 1] != 255: - count += 1 - if pixels[x + 1, y] != 255: - count += 1 - if pixels[x + 1, y-1] != 255: - count += 1 - if pixels[x, y-1] != 255: - count += 1 - except Exception: - pass - - # not enough neighbors are dark pixels so mark this pixel - # to be changed to white - if count < allowed: - pixels[x, y] = 1 - - # second pass: this time set all 1's to 255 (white) - for x in xrange(w): - for y in xrange(h): - if pixels[x, y] == 1: - pixels[x, y] = 255 - - self.pixels = pixels - - - def derotate_by_average(self): - """rotate by checking each angle and guess most suitable""" - - w, h = self.image.size - pixels = self.pixels - - for x in xrange(w): - for y in xrange(h): - if pixels[x, y] == 0: - pixels[x, y] = 155 - - highest = {} - counts = {} - - for angle in xrange(-45, 45): - - tmpimage = self.image.rotate(angle) - - pixels = tmpimage.load() - - w, h = self.image.size - - for x in xrange(w): - for y in xrange(h): - if pixels[x, y] == 0: - pixels[x, y] = 255 - - - count = {} - - for x in xrange(w): - count[x] = 0 - for y in xrange(h): - if pixels[x, y] == 155: - count[x] += 1 - - sum = 0 - cnt = 0 - - for x in count.values(): - if x != 0: - sum += x - cnt += 1 - - avg = sum / cnt - counts[angle] = cnt - highest[angle] = 0 - for x in count.values(): - if x > highest[angle]: - highest[angle] = x - - highest[angle] = highest[angle] - avg - - hkey = 0 - hvalue = 0 - - for key, value in highest.iteritems(): - if value > hvalue: - hkey = key - hvalue = value - - self.image = self.image.rotate(hkey) - pixels = self.image.load() - - for x in xrange(w): - for y in xrange(h): - if pixels[x, y] == 0: - pixels[x, y] = 255 - - if pixels[x, y] == 155: - pixels[x, y] = 0 - - self.pixels = pixels - - - def split_captcha_letters(self): - captcha = self.image - started = False - letters = [] - width, height = captcha.size - bottomY, topY = 0, height - pixels = captcha.load() - - for x in xrange(width): - black_pixel_in_col = False - for y in xrange(height): - if pixels[x, y] != 255: - if not started: - started = True - firstX = x - lastX = x - - if y > bottomY: - bottomY = y - if y < topY: - topY = y - if x > lastX: - lastX = x - - black_pixel_in_col = True - - if black_pixel_in_col is False and started is True: - rect = (firstX, topY, lastX, bottomY) - new_captcha = captcha.crop(rect) - - w, h = new_captcha.size - if w > 5 and h > 5: - letters.append(new_captcha) - - started = False - bottomY, topY = 0, height - - return letters - - - def correct(self, values, var=None): - if var: - result = var - else: - result = self.result_captcha - - for key, item in values.iteritems(): - - if key.__class__ == str: - result = result.replace(key, item) - else: - for expr in key: - result = result.replace(expr, item) - - if var: - return result - else: - self.result_captcha = result -- cgit v1.2.3