diff options
author | 2015-03-26 17:04:38 +0100 | |
---|---|---|
committer | 2015-03-26 17:04:38 +0100 | |
commit | ccb6aaee9d987b56faf19fb48dd06a04ff8ca721 (patch) | |
tree | 6a385417ff77f3ed17329e2d606f80b673bdb3c0 /module/plugins/captcha | |
parent | Prepare to merging (diff) | |
parent | Tiny code cosmetics (diff) | |
download | pyload-ccb6aaee9d987b56faf19fb48dd06a04ff8ca721.tar.xz |
Merge branch 'stable' into 0.4.10
Conflicts:
module/plugins/accounts/Keep2ShareCc.py
module/plugins/captcha/OCR.py
module/plugins/captcha/captcha.py
module/plugins/crypter/DailymotionBatch.py
module/plugins/crypter/DailymotionComFolder.py
module/plugins/crypter/YoutubeBatch.py
module/plugins/crypter/YoutubeComFolder.py
module/plugins/internal/CaptchaService.py
pyload/plugin/Extractor.py
pyload/plugin/OCR.py
pyload/plugin/account/DebridItaliaCom.py
pyload/plugin/account/MegaRapidCz.py
pyload/plugin/account/NoPremiumPl.py
pyload/plugin/account/RapideoPl.py
pyload/plugin/addon/AndroidPhoneNotify.py
pyload/plugin/addon/Checksum.py
pyload/plugin/addon/ClickAndLoad.py
pyload/plugin/addon/DeleteFinished.py
pyload/plugin/addon/DownloadScheduler.py
pyload/plugin/addon/ExternalScripts.py
pyload/plugin/addon/ExtractArchive.py
pyload/plugin/addon/HotFolder.py
pyload/plugin/addon/IRCInterface.py
pyload/plugin/addon/JustPremium.py
pyload/plugin/addon/MergeFiles.py
pyload/plugin/addon/MultiHome.py
pyload/plugin/addon/RestartFailed.py
pyload/plugin/addon/RestartSlow.py
pyload/plugin/addon/SkipRev.py
pyload/plugin/addon/UnSkipOnFail.py
pyload/plugin/addon/UpdateManager.py
pyload/plugin/addon/WindowsPhoneNotify.py
pyload/plugin/crypter/DDLMusicOrg.py
pyload/plugin/crypter/DailymotionBatch.py
pyload/plugin/crypter/DevhostSt.py
pyload/plugin/crypter/EmbeduploadCom.py
pyload/plugin/crypter/MultiloadCz.py
pyload/plugin/crypter/YoutubeBatch.py
pyload/plugin/extractor/SevenZip.py
pyload/plugin/extractor/UnRar.py
pyload/plugin/extractor/UnZip.py
pyload/plugin/hook/BypassCaptcha.py
pyload/plugin/hook/Captcha9Kw.py
pyload/plugin/hook/CaptchaBrotherhood.py
pyload/plugin/hook/DeathByCaptcha.py
pyload/plugin/hook/ExpertDecoders.py
pyload/plugin/hook/ImageTyperz.py
pyload/plugin/hook/XFileSharingPro.py
pyload/plugin/hoster/AlldebridCom.py
pyload/plugin/hoster/CzshareCom.py
pyload/plugin/hoster/EuroshareEu.py
pyload/plugin/hoster/FastixRu.py
pyload/plugin/hoster/FastshareCz.py
pyload/plugin/hoster/GooIm.py
pyload/plugin/hoster/MediafireCom.py
pyload/plugin/hoster/MegaDebridEu.py
pyload/plugin/hoster/NitroflareCom.py
pyload/plugin/hoster/OverLoadMe.py
pyload/plugin/hoster/PremiumTo.py
pyload/plugin/hoster/PremiumizeMe.py
pyload/plugin/hoster/RapidgatorNet.py
pyload/plugin/hoster/RealdebridCom.py
pyload/plugin/hoster/SimplyPremiumCom.py
pyload/plugin/hoster/SimplydebridCom.py
pyload/plugin/hoster/SmoozedCom.py
pyload/plugin/hoster/SoundcloudCom.py
pyload/plugin/hoster/UlozTo.py
pyload/plugin/hoster/UploadableCh.py
pyload/plugin/hoster/UploadedTo.py
pyload/plugin/hoster/UploadheroCom.py
pyload/plugin/hoster/VeehdCom.py
pyload/plugin/hoster/VimeoCom.py
pyload/plugin/hoster/ZeveraCom.py
pyload/plugin/hoster/ZippyshareCom.py
pyload/plugin/internal/BasePlugin.py
pyload/plugin/internal/MultiHoster.py
pyload/plugin/internal/SimpleDereferer.py
pyload/plugin/internal/SimpleHoster.py
pyload/plugin/internal/XFSHoster.py
pyload/plugin/ocr/GigasizeCom.py
pyload/plugin/ocr/LinksaveIn.py
pyload/plugin/ocr/NetloadIn.py
pyload/plugin/ocr/ShareonlineBiz.py
Diffstat (limited to 'module/plugins/captcha')
-rw-r--r-- | module/plugins/captcha/OCR.py | 319 |
1 files changed, 319 insertions, 0 deletions
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 |