summaryrefslogtreecommitdiffstats
path: root/module/plugins/captcha
diff options
context:
space:
mode:
Diffstat (limited to 'module/plugins/captcha')
-rw-r--r--module/plugins/captcha/AdYouLike.py107
-rw-r--r--module/plugins/captcha/AdsCaptcha.py77
-rw-r--r--module/plugins/captcha/GigasizeCom.py24
-rw-r--r--module/plugins/captcha/LinksaveIn.py158
-rw-r--r--module/plugins/captcha/NetloadIn.py29
-rw-r--r--module/plugins/captcha/ReCaptcha.py73
-rw-r--r--module/plugins/captcha/ShareonlineBiz.py39
-rw-r--r--module/plugins/captcha/SolveMedia.py50
-rw-r--r--module/plugins/captcha/__init__.py1
-rw-r--r--module/plugins/captcha/captcha.py319
10 files changed, 308 insertions, 569 deletions
diff --git a/module/plugins/captcha/AdYouLike.py b/module/plugins/captcha/AdYouLike.py
new file mode 100644
index 000000000..bfb8b5ccb
--- /dev/null
+++ b/module/plugins/captcha/AdYouLike.py
@@ -0,0 +1,107 @@
+# -*- coding: utf-8 -*-
+
+import re
+
+from pyload.plugin.Captcha import Captcha
+from pyload.utils import json_loads
+
+
+class AdYouLike(Captcha):
+ __name__ = "AdYouLike"
+ __type__ = "captcha"
+ __version__ = "0.02"
+
+ __description__ = """AdYouLike captcha service plugin"""
+ __license__ = "GPLv3"
+ __authors__ = [("Walter Purcaro", "vuolter@gmail.com")]
+
+
+ AYL_PATTERN = r'Adyoulike\.create\s*\((.+?)\)'
+ CALLBACK_PATTERN = r'(Adyoulike\.g\._jsonp_\d+)'
+
+
+ def detect_key(self, html=None):
+ if not html:
+ if hasattr(self.plugin, "html") and self.plugin.html:
+ html = self.plugin.html
+ else:
+ errmsg = _("AdYouLike html not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ m = re.search(self.AYL_PATTERN, html)
+ n = re.search(self.CALLBACK_PATTERN, html)
+ if m and n:
+ self.key = (m.group(1).strip(), n.group(1).strip())
+ self.plugin.logDebug("AdYouLike ayl|callback: %s | %s" % self.key)
+ return self.key #: key is the tuple(ayl, callback)
+ else:
+ self.plugin.logDebug("AdYouLike ayl or callback not found")
+ return None
+
+
+ def challenge(self, key=None):
+ if not key:
+ if self.detect_key():
+ key = self.key
+ else:
+ errmsg = _("AdYouLike key not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ ayl, callback = key
+
+ # {"adyoulike":{"key":"P~zQ~O0zV0WTiAzC-iw0navWQpCLoYEP"},
+ # "all":{"element_id":"ayl_private_cap_92300","lang":"fr","env":"prod"}}
+ ayl = json_loads(ayl)
+
+ html = self.plugin.req.load("http://api-ayl.appspot.com/challenge",
+ get={'key' : ayl['adyoulike']['key'],
+ 'env' : ayl['all']['env'],
+ 'callback': callback})
+ try:
+ challenge = json_loads(re.search(callback + r'\s*\((.+?)\)', html).group(1))
+ except Exception:
+ errmsg = _("AdYouLike challenge pattern not found")
+ self.plugin.error(errmsg)
+ raise ValueError(errmsg)
+
+ self.plugin.logDebug("AdYouLike challenge: %s" % challenge)
+
+ return self.result(ayl, challenge)
+
+
+ def result(self, server, challenge):
+ # Adyoulike.g._jsonp_5579316662423138
+ # ({"translations":{"fr":{"instructions_visual":"Recopiez « Soonnight » ci-dessous :"}},
+ # "site_under":true,"clickable":true,"pixels":{"VIDEO_050":[],"DISPLAY":[],"VIDEO_000":[],"VIDEO_100":[],
+ # "VIDEO_025":[],"VIDEO_075":[]},"medium_type":"image/adyoulike",
+ # "iframes":{"big":"<iframe src=\"http://www.soonnight.com/campagn.html\" scrolling=\"no\"
+ # height=\"250\" width=\"300\" frameborder=\"0\"></iframe>"},"shares":{},"id":256,
+ # "token":"e6QuI4aRSnbIZJg02IsV6cp4JQ9~MjA1","formats":{"small":{"y":300,"x":0,"w":300,"h":60},
+ # "big":{"y":0,"x":0,"w":300,"h":250},"hover":{"y":440,"x":0,"w":300,"h":60}},
+ # "tid":"SqwuAdxT1EZoi4B5q0T63LN2AkiCJBg5"})
+
+ if isinstance(server, basestring):
+ server = json_loads(server)
+
+ if isinstance(challenge, basestring):
+ challenge = json_loads(challenge)
+
+ try:
+ instructions_visual = challenge['translations'][server['all']['lang']]['instructions_visual']
+ result = re.search(u'«(.+?)»', instructions_visual).group(1).strip()
+ except Exception:
+ errmsg = _("AdYouLike result not found")
+ self.plugin.error(errmsg)
+ raise ValueError(errmsg)
+
+ result = {'_ayl_captcha_engine' : "adyoulike",
+ '_ayl_env' : server['all']['env'],
+ '_ayl_tid' : challenge['tid'],
+ '_ayl_token_challenge': challenge['token'],
+ '_ayl_response' : response}
+
+ self.plugin.logDebug("AdYouLike result: %s" % result)
+
+ return result
diff --git a/module/plugins/captcha/AdsCaptcha.py b/module/plugins/captcha/AdsCaptcha.py
new file mode 100644
index 000000000..5c01a422e
--- /dev/null
+++ b/module/plugins/captcha/AdsCaptcha.py
@@ -0,0 +1,77 @@
+# -*- coding: utf-8 -*-
+
+import re
+
+from random import random
+
+from pyload.plugin.Captcha import Captcha
+
+
+class AdsCaptcha(Captcha):
+ __name__ = "AdsCaptcha"
+ __type__ = "captcha"
+ __version__ = "0.06"
+
+ __description__ = """AdsCaptcha captcha service plugin"""
+ __license__ = "GPLv3"
+ __authors__ = [("pyLoad Team", "admin@pyload.org")]
+
+
+ CAPTCHAID_PATTERN = r'api\.adscaptcha\.com/Get\.aspx\?[^"\']*CaptchaId=(\d+)'
+ PUBLICKEY_PATTERN = r'api\.adscaptcha\.com/Get\.aspx\?[^"\']*PublicKey=([\w-]+)'
+
+
+ def detect_key(self, html=None):
+ if not html:
+ if hasattr(self.plugin, "html") and self.plugin.html:
+ html = self.plugin.html
+ else:
+ errmsg = _("AdsCaptcha html not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ m = re.search(self.PUBLICKEY_PATTERN, html)
+ n = re.search(self.CAPTCHAID_PATTERN, html)
+ if m and n:
+ self.key = (m.group(1).strip(), n.group(1).strip()) #: key is the tuple(PublicKey, CaptchaId)
+ self.plugin.logDebug("AdsCaptcha key|id: %s | %s" % self.key)
+ return self.key
+ else:
+ self.plugin.logDebug("AdsCaptcha key or id not found")
+ return None
+
+
+ def challenge(self, key=None):
+ if not key:
+ if self.detect_key():
+ key = self.key
+ else:
+ errmsg = _("AdsCaptcha key not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ PublicKey, CaptchaId = key
+
+ html = self.plugin.req.load("http://api.adscaptcha.com/Get.aspx", get={'CaptchaId': CaptchaId, 'PublicKey': PublicKey})
+ try:
+ challenge = re.search("challenge: '(.+?)',", html).group(1)
+ server = re.search("server: '(.+?)',", html).group(1)
+ except Exception:
+ errmsg = _("AdsCaptcha challenge pattern not found")
+ self.plugin.error(errmsg)
+ raise ValueError(errmsg)
+
+ self.plugin.logDebug("AdsCaptcha challenge: %s" % challenge)
+
+ return challenge, self.result(server, challenge)
+
+
+ def result(self, server, challenge):
+ result = self.plugin.decryptCaptcha("%sChallenge.aspx" % server,
+ get={'cid': challenge, 'dummy': random()},
+ cookies=True,
+ imgtype="jpg")
+
+ self.plugin.logDebug("AdsCaptcha result: %s" % result)
+
+ return result
diff --git a/module/plugins/captcha/GigasizeCom.py b/module/plugins/captcha/GigasizeCom.py
deleted file mode 100644
index 244cf6a2a..000000000
--- a/module/plugins/captcha/GigasizeCom.py
+++ /dev/null
@@ -1,24 +0,0 @@
-# -*- coding: utf-8 -*-
-
-from module.plugins.captcha.captcha import OCR
-
-
-class GigasizeCom(OCR):
- __name__ = "GigasizeCom"
- __type__ = "ocr"
- __version__ = "0.10"
-
- __description__ = """Gigasize.com ocr plugin"""
- __license__ = "GPLv3"
- __authors__ = [("pyLoad Team", "admin@pyload.org")]
-
-
- def __init__(self):
- OCR.__init__(self)
-
-
- def get_captcha(self, image):
- self.load_image(image)
- self.threshold(2.8)
- self.run_tesser(True, False, False, True)
- return self.result_captcha
diff --git a/module/plugins/captcha/LinksaveIn.py b/module/plugins/captcha/LinksaveIn.py
deleted file mode 100644
index de6b0e7ff..000000000
--- a/module/plugins/captcha/LinksaveIn.py
+++ /dev/null
@@ -1,158 +0,0 @@
-# -*- coding: utf-8 -*-
-
-try:
- from PIL import Image
-except ImportError:
- import Image
-
-from glob import glob
-from os import sep
-from os.path import abspath, dirname
-
-from module.plugins.captcha.captcha import OCR
-
-
-class LinksaveIn(OCR):
- __name__ = "LinksaveIn"
- __type__ = "ocr"
- __version__ = "0.10"
-
- __description__ = """Linksave.in ocr plugin"""
- __license__ = "GPLv3"
- __authors__ = [("pyLoad Team", "admin@pyload.org")]
-
-
- def __init__(self):
- OCR.__init__(self)
- self.data_dir = dirname(abspath(__file__)) + sep + "LinksaveIn" + sep
-
-
- def load_image(self, image):
- im = Image.open(image)
- frame_nr = 0
-
- lut = im.resize((256, 1))
- lut.putdata(range(256))
- lut = list(lut.convert("RGB").getdata())
-
- new = Image.new("RGB", im.size)
- npix = new.load()
- while True:
- try:
- im.seek(frame_nr)
- except EOFError:
- break
- frame = im.copy()
- pix = frame.load()
- for x in xrange(frame.size[0]):
- for y in xrange(frame.size[1]):
- if lut[pix[x, y]] != (0,0,0):
- npix[x, y] = lut[pix[x, y]]
- frame_nr += 1
- new.save(self.data_dir+"unblacked.png")
- self.image = new.copy()
- self.pixels = self.image.load()
- self.result_captcha = ''
-
-
- def get_bg(self):
- stat = {}
- cstat = {}
- img = self.image.convert("P")
- for bgpath in glob(self.data_dir+"bg/*.gif"):
- stat[bgpath] = 0
- bg = Image.open(bgpath)
-
- bglut = bg.resize((256, 1))
- bglut.putdata(range(256))
- bglut = list(bglut.convert("RGB").getdata())
-
- lut = img.resize((256, 1))
- lut.putdata(range(256))
- lut = list(lut.convert("RGB").getdata())
-
- bgpix = bg.load()
- pix = img.load()
- for x in xrange(bg.size[0]):
- for y in xrange(bg.size[1]):
- rgb_bg = bglut[bgpix[x, y]]
- rgb_c = lut[pix[x, y]]
- try:
- cstat[rgb_c] += 1
- except Exception:
- cstat[rgb_c] = 1
- if rgb_bg == rgb_c:
- stat[bgpath] += 1
- max_p = 0
- bg = ""
- for bgpath, value in stat.iteritems():
- if max_p < value:
- bg = bgpath
- max_p = value
- return bg
-
-
- def substract_bg(self, bgpath):
- bg = Image.open(bgpath)
- img = self.image.convert("P")
-
- bglut = bg.resize((256, 1))
- bglut.putdata(range(256))
- bglut = list(bglut.convert("RGB").getdata())
-
- lut = img.resize((256, 1))
- lut.putdata(range(256))
- lut = list(lut.convert("RGB").getdata())
-
- bgpix = bg.load()
- pix = img.load()
- orgpix = self.image.load()
- for x in xrange(bg.size[0]):
- for y in xrange(bg.size[1]):
- rgb_bg = bglut[bgpix[x, y]]
- rgb_c = lut[pix[x, y]]
- if rgb_c == rgb_bg:
- orgpix[x, y] = (255,255,255)
-
-
- def eval_black_white(self):
- new = Image.new("RGB", (140, 75))
- pix = new.load()
- orgpix = self.image.load()
- thresh = 4
- for x in xrange(new.size[0]):
- for y in xrange(new.size[1]):
- rgb = orgpix[x, y]
- r, g, b = rgb
- pix[x, y] = (255,255,255)
- if r > max(b, g)+thresh:
- pix[x, y] = (0,0,0)
- if g < min(r, b):
- pix[x, y] = (0,0,0)
- if g > max(r, b)+thresh:
- pix[x, y] = (0,0,0)
- if b > max(r, g)+thresh:
- pix[x, y] = (0,0,0)
- self.image = new
- self.pixels = self.image.load()
-
-
- def get_captcha(self, image):
- self.load_image(image)
- bg = self.get_bg()
- self.substract_bg(bg)
- self.eval_black_white()
- self.to_greyscale()
- self.image.save(self.data_dir+"cleaned_pass1.png")
- self.clean(4)
- self.clean(4)
- self.image.save(self.data_dir+"cleaned_pass2.png")
- letters = self.split_captcha_letters()
- final = ""
- for n, letter in enumerate(letters):
- self.image = letter
- self.image.save(ocr.data_dir+"letter%d.png" % n)
- self.run_tesser(True, True, False, False)
- final += self.result_captcha
-
- return final
diff --git a/module/plugins/captcha/NetloadIn.py b/module/plugins/captcha/NetloadIn.py
deleted file mode 100644
index 28eb18fb5..000000000
--- a/module/plugins/captcha/NetloadIn.py
+++ /dev/null
@@ -1,29 +0,0 @@
-# -*- coding: utf-8 -*-
-
-from module.plugins.captcha.captcha import OCR
-
-
-class NetloadIn(OCR):
- __name__ = "NetloadIn"
- __type__ = "ocr"
- __version__ = "0.10"
-
- __description__ = """Netload.in ocr plugin"""
- __license__ = "GPLv3"
- __authors__ = [("pyLoad Team", "admin@pyload.org")]
-
-
- def __init__(self):
- OCR.__init__(self)
-
-
- def get_captcha(self, image):
- self.load_image(image)
- self.to_greyscale()
- self.clean(3)
- self.clean(3)
- self.run_tesser(True, True, False, False)
-
- self.result_captcha = self.result_captcha.replace(" ", "")[:4] # cut to 4 numbers
-
- return self.result_captcha
diff --git a/module/plugins/captcha/ReCaptcha.py b/module/plugins/captcha/ReCaptcha.py
new file mode 100644
index 000000000..4b900c57b
--- /dev/null
+++ b/module/plugins/captcha/ReCaptcha.py
@@ -0,0 +1,73 @@
+# -*- coding: utf-8 -*-
+
+import re
+
+from pyload.plugin.Captcha import Captcha
+
+
+class ReCaptcha(Captcha):
+ __name__ = "ReCaptcha"
+ __type__ = "captcha"
+ __version__ = "0.08"
+
+ __description__ = """ReCaptcha captcha service plugin"""
+ __license__ = "GPLv3"
+ __authors__ = [("pyLoad Team", "admin@pyload.org")]
+
+
+ KEY_PATTERN = r'recaptcha(?:/api|\.net)/(?:challenge|noscript)\?k=([\w-]+)'
+ KEY_AJAX_PATTERN = r'Recaptcha\.create\s*\(\s*["\']([\w-]+)'
+
+
+ def detect_key(self, html=None):
+ if not html:
+ if hasattr(self.plugin, "html") and self.plugin.html:
+ html = self.plugin.html
+ else:
+ errmsg = _("ReCaptcha html not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ m = re.search(self.KEY_PATTERN, html) or re.search(self.KEY_AJAX_PATTERN, html)
+ if m:
+ self.key = m.group(1).strip()
+ self.plugin.logDebug("ReCaptcha key: %s" % self.key)
+ return self.key
+ else:
+ self.plugin.logDebug("ReCaptcha key not found")
+ return None
+
+
+ def challenge(self, key=None):
+ if not key:
+ if self.detect_key():
+ key = self.key
+ else:
+ errmsg = _("ReCaptcha key not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ html = self.plugin.req.load("http://www.google.com/recaptcha/api/challenge", get={'k': key})
+ try:
+ challenge = re.search("challenge : '(.+?)',", html).group(1)
+ server = re.search("server : '(.+?)',", html).group(1)
+ except Exception:
+ errmsg = _("ReCaptcha challenge pattern not found")
+ self.plugin.error(errmsg)
+ raise ValueError(errmsg)
+
+ self.plugin.logDebug("ReCaptcha challenge: %s" % challenge)
+
+ return challenge, self.result(server, challenge)
+
+
+ def result(self, server, challenge):
+ result = self.plugin.decryptCaptcha("%simage" % server,
+ get={'c': challenge},
+ cookies=True,
+ forceUser=True,
+ imgtype="jpg")
+
+ self.plugin.logDebug("ReCaptcha result: %s" % result)
+
+ return result
diff --git a/module/plugins/captcha/ShareonlineBiz.py b/module/plugins/captcha/ShareonlineBiz.py
deleted file mode 100644
index 8210e8859..000000000
--- a/module/plugins/captcha/ShareonlineBiz.py
+++ /dev/null
@@ -1,39 +0,0 @@
-# -*- coding: utf-8 -*-
-
-from module.plugins.captcha.captcha import OCR
-
-
-class ShareonlineBiz(OCR):
- __name__ = "ShareonlineBiz"
- __type__ = "ocr"
- __version__ = "0.10"
-
- __description__ = """Shareonline.biz ocr plugin"""
- __license__ = "GPLv3"
- __authors__ = [("RaNaN", "RaNaN@pyload.org")]
-
-
- def __init__(self):
- OCR.__init__(self)
-
-
- def get_captcha(self, image):
- self.load_image(image)
- self.to_greyscale()
- self.image = self.image.resize((160, 50))
- self.pixels = self.image.load()
- self.threshold(1.85)
- #self.eval_black_white(240)
- #self.derotate_by_average()
-
- letters = self.split_captcha_letters()
-
- final = ""
- for letter in letters:
- self.image = letter
- self.run_tesser(True, True, False, False)
- final += self.result_captcha
-
- return final
-
- #tesseract at 60%
diff --git a/module/plugins/captcha/SolveMedia.py b/module/plugins/captcha/SolveMedia.py
new file mode 100644
index 000000000..6e2a6c362
--- /dev/null
+++ b/module/plugins/captcha/SolveMedia.py
@@ -0,0 +1,50 @@
+# -*- coding: utf-8 -*-
+
+import re
+
+from pyload.plugin.Captcha import Captcha
+
+
+class SolveMedia(Captcha):
+ __name__ = "SolveMedia"
+ __type__ = "captcha"
+ __version__ = "0.06"
+
+ __description__ = """SolveMedia captcha service plugin"""
+ __license__ = "GPLv3"
+ __authors__ = [("pyLoad Team", "admin@pyload.org")]
+
+
+ KEY_PATTERN = r'api\.solvemedia\.com/papi/challenge\.(?:no)?script\?k=(.+?)["\']'
+
+
+ def challenge(self, key=None):
+ if not key:
+ if self.detect_key():
+ key = self.key
+ else:
+ errmsg = _("SolveMedia key not found")
+ self.plugin.fail(errmsg)
+ raise TypeError(errmsg)
+
+ html = self.plugin.req.load("http://api.solvemedia.com/papi/challenge.noscript", get={'k': key})
+ try:
+ challenge = re.search(r'<input type=hidden name="adcopy_challenge" id="adcopy_challenge" value="([^"]+)">',
+ html).group(1)
+ server = "http://api.solvemedia.com/papi/media"
+ except Exception:
+ errmsg = _("SolveMedia challenge pattern not found")
+ self.plugin.error(errmsg)
+ raise ValueError(errmsg)
+
+ self.plugin.logDebug("SolveMedia challenge: %s" % challenge)
+
+ return challenge, self.result(server, challenge)
+
+
+ def result(self, server, challenge):
+ result = self.plugin.decryptCaptcha(server, get={'c': challenge}, imgtype="gif")
+
+ self.plugin.logDebug("SolveMedia result: %s" % result)
+
+ return result
diff --git a/module/plugins/captcha/__init__.py b/module/plugins/captcha/__init__.py
index e69de29bb..40a96afc6 100644
--- a/module/plugins/captcha/__init__.py
+++ b/module/plugins/captcha/__init__.py
@@ -0,0 +1 @@
+# -*- coding: utf-8 -*-
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
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-24 04:50:57 +0100