Beispiel #1
0
def __init__(self):
    HTMLParser.__init__(self)
    self.recording = 0
    self.data = []

    def handle_starttag(self, tag, attributes):
        if tag != 'td':
            return
        if self.recording:
            self.recording += 1
            return
        for name, value in attributes:
            if name == 'class' and value == 'cen':
                break
            else:
                return
        self.recording = 1

    def handle_endtag(self, tag):
        if tag == 'td' and self.recording:
            self.recording -= 1

    def handle_data(self, data):
        if 'TM' in data and self.recording:
            self.data.append(data)
Beispiel #2
0
	def __init__(self):
		HTMLParser.__init__(self)
		self.lasttag = None
		self.title = ""
		self.pagedata = StringIO()
		self.links = []
		self.inbody = False
Beispiel #3
0
    def __init__(self):
        HTMLParser.__init__(self)
        self.trouve=False
        self.encours=False
        self.reponse={}
        self.correspondance={
            'CVSS Score':'cvss_score',
            'Confidentiality Impact':'confidentialite',
            'Integrity Impact':'integrite',
            'Availability Impact':'disponibilite',
            'Access Complexity':'complexite',
            'Authentication':'authentification',
            'Vulnerability Type(s)':'type',
            'CWE ID':None,
            'Gained Access':'acces_obtention'
        }

        self.reponse={
            'cvss_score':None,
            'confidentialite':None,
            'integrite':None,
            'disponibilite':None,
            'complexite':None,
            'authentification':None,
            'type':None,
            'acces_obtention':None
        }

        self.precedent=None
Beispiel #4
0
    def __init__(self, day='tomorrow'):
        HTMLParser.__init__(self)
        if day == 'today':
            self.theDay = date.today()
        elif day == 'yesterday':
            self.theDay = date.today() - timedelta(1)
        elif day == 'tomorrow':
            self.theDay = date.today() + timedelta(1)
        else:
            self.theDay = None

        self.in_episode = False
        self.in_date = False
        self.in_dateA = False
        self.in_validep = False
        self.in_showname = False
        self.in_number = False
        self.date_format = "%d-%b-%Y"
        self.shows = []
        self.buffshow = {}

        if self.theDay is not None:
            one_day_ago = (self.theDay - timedelta(1)).strftime(self.date_format)
            two_day_ago = (self.theDay - timedelta(2)).strftime(self.date_format)
            three_day_ago = (self.theDay - timedelta(3)).strftime(self.date_format)
            four_day_ago = (self.theDay - timedelta(4)).strftime(self.date_format)

            self.theDay = self.theDay.strftime(self.date_format)
            self.previous_lookup = [self.theDay, one_day_ago, two_day_ago, three_day_ago, four_day_ago]
Beispiel #5
0
    def __init__(self, site = None):
        HTMLParser.__init__(self)
        dict.__init__(self, ())
        self.in_form = False
        self.select = None

        if site: self.load(site)
Beispiel #6
0
 def __init__(self):
     """ Constructor; initializes washer """
     HTMLParser.__init__(self)
     self.result = ''
     self.nb = 0
     self.previous_nbs = []
     self.previous_type_lists = []
     self.url = ''
     self.render_unallowed_tags = False
     self.allowed_tag_whitelist = \
             CFG_HTML_BUFFER_ALLOWED_TAG_WHITELIST
     self.allowed_attribute_whitelist = \
             CFG_HTML_BUFFER_ALLOWED_ATTRIBUTE_WHITELIST
     # javascript:
     self.re_js = re.compile( ".*(j|j|J)"\
                             "\s*(a|a|A)"\
                             "\s*(v|v|V)"\
                             "\s*(a|a|A)"\
                             "\s*(s|s|S)"\
                             "\s*(c|c|C)"\
                             "\s*(r|r|R)"\
                             "\s*(i|Ã|I)"\
                             "\s*(p|p|P)"\
                             "\s*(t|p|&#84)"\
                             "\s*(:|:).*", re.IGNORECASE | re.DOTALL)
     # vbscript:
     self.re_vb = re.compile( ".*(v|v|V)"\
                             "\s*(b|b|B)"\
                             "\s*(s|s|S)"\
                             "\s*(c|c|C)"\
                             "\s*(r|r|R)"\
                             "\s*(i|Ã|I)"\
                             "\s*(p|p|P)"\
                             "\s*(t|p|T)"\
                             "\s*(:|:).*", re.IGNORECASE | re.DOTALL)
Beispiel #7
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.url = None
     self.params = {}
     self.in_form = False
     self.form_parsed = False
     self.method = "GET"
 def __init__(self):
     HTMLParser.__init__(self)
     self.in_records_table = False
     self.record = -1
     self.column = -1
     self.data_row = False
     self.data = []
Beispiel #9
0
 def __init__(self, properties):
     HTMLParser.__init__(self)
     self.properties = dict((key, value) for key, value in (prop.split(',')
                            for prop in properties.split(';') if prop.find(',') > -1))
     self.data = []
     self.in_td = 0
     self.tr_name = None
Beispiel #10
0
	def __init__(self):
		HTMLParser.__init__(self)
		self.subjectList = {}
		self.tagi = 0
		self.tdi = 0
		self.dataFlag = 0
		self.subName = ""
    def __init__(self, new_path, filename, reference_support_info, host=Host(), convert_test_harness_links=True):
        HTMLParser.__init__(self)

        self._host = host
        self._filesystem = self._host.filesystem
        self._webkit_root = WebKitFinder(self._filesystem).webkit_base()

        self.converted_data = []
        self.converted_properties = []
        self.converted_property_values = []
        self.in_style_tag = False
        self.style_data = []
        self.filename = filename
        self.reference_support_info = reference_support_info

        resources_path = self.path_from_webkit_root('LayoutTests', 'resources')
        resources_relpath = self._filesystem.relpath(resources_path, new_path)
        self.new_test_harness_path = resources_relpath
        self.convert_test_harness_links = convert_test_harness_links

        # These settings might vary between WebKit and Blink
        self._css_property_file = self.path_from_webkit_root('Source', 'WebCore', 'css', 'CSSPropertyNames.in')
        self._css_property_value_file = self.path_from_webkit_root('Source', 'WebCore', 'css', 'CSSValueKeywords.in')

        self.test_harness_re = re.compile('/resources/testharness')

        self.prefixed_properties = self.read_webkit_prefixed_css_property_list(self._css_property_file)
        prop_regex = '([\s{]|^)(' + "|".join(prop.replace('-webkit-', '') for prop in self.prefixed_properties) + ')(\s+:|:)'
        self.prop_re = re.compile(prop_regex)

        self.prefixed_property_values = self.read_webkit_prefixed_css_property_list(self._css_property_value_file)
        prop_value_regex = '(:\s*|^\s*)(' + "|".join(value.replace('-webkit-', '') for value in self.prefixed_property_values) + ')(\s*;|\s*}|\s*$)'
        self.prop_value_re = re.compile(prop_value_regex)
Beispiel #12
0
    def __init__(self, tag="a", attr="href", process=None, unique=False):
        HTMLParser.__init__(self)

        self.scan_tag = tag if callable(tag) else lambda t: t == tag
        self.scan_attr = attr if callable(attr) else lambda a: a == attr
        self.process_attr = process if callable(process) else lambda v: v
        self.unique = unique
Beispiel #13
0
	def __init__(self, feed_data):
		HTMLParser.__init__(self)
		self.courses = tuple()
		self.is_course = False
		self.is_coursename = False
		self.is_homework = False
		self.feed(feed_data)
 def __init__(self):
     HTMLParser.__init__(self)
     self.in_div = False
     self.in_a = False
     self.pattern = re.compile(r'(.*)\((.*)\)')
     self.tangshi_list = []
     self.current_poem = {}
Beispiel #15
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.glink = False
     self.elink = False
     self.ingroup = []
     self.href = ''
     self.name = ''
Beispiel #16
0
	def __init__(self):
		HTMLParser.__init__(self)
		self.foundName = False
		self.foundDescription = False
		self.foundPrice = False
		self.foundScore = True
		self.gameInfo = {}
Beispiel #17
0
 def __init__(self, builder=None, encoding=None):
     self.__stack = []
     if builder is None:
         builder = ElementTree.TreeBuilder()
     self.__builder = builder
     self.encoding = encoding or "iso-8859-1"
     HTMLParser.__init__(self)
Beispiel #18
0
 def __init__(self, *a, **kw):
     self.indent = '';
     HTMLParser.__init__(self, *a, **kw)
     self.processed_text = ''
     self.tagtracker = []
     self.error_line = 0
     self.line_number = 1
 def __init__(self):
     #super(formParser, self).__init__()
     HTMLParser.__init__(self)
     self.dict = {}
     self.stack = []
     self.post=""
     pass
Beispiel #20
0
 def __init__(self, pdf):
     HTMLParser.__init__(self)
     self.style = {}
     self.pre = False
     self.href = ''
     self.align = ''
     self.page_links = {}
     self.font_list = ("times","courier", "helvetica")
     self.font = None
     self.font_stack = [] 
     self.pdf = pdf
     self.r = self.g = self.b = 0
     self.indent = 0
     self.bullet = []
     self.set_font("times", 12)
     self.font_face = "times"    # initialize font
     self.color = 0              #initialize font color
     self.table = None           # table attributes
     self.table_col_width = None # column (header) widths
     self.table_col_index = None # current column index
     self.td = None              # cell attributes
     self.th = False             # header enabled
     self.tr = None
     self.theader = None           # table header cells
     self.tfooter = None           # table footer cells
     self.thead = None
     self.tfoot = None
     self.theader_out = self.tfooter_out = False
	def __init__(self):
		HTMLParser.__init__(self)
		self.title = False
		self.updated = False
		self.titlestr = ''
		self.updatedstr = ''
		self.list = []
Beispiel #22
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.tracks = []
     self.endDate = None
     self.curData = ''
     self.curTrack = {}
     self.recording = None
 def __init__(self, translationFoundCallback):
     HTMLParser.__init__(self)
     self.translationFoundCallback = translationFoundCallback
     self.recording = False
     self.js_translations_script_found = False
     self.js_rogerthat_script_found = False
     self._js_rogerthat_script_regex = re.compile('rogerthat/rogerthat-(\d+\.)+js')
Beispiel #24
0
 def __init__(self, text, closed=("input", "link")):
     HTMLParser.__init__(self)
     self.tree = self.parent = TAG[""]()
     self.closed = closed
     self.tags = [x for x in __all__ if isinstance(eval(x), DIV)]
     self.last = None
     self.feed(text)
Beispiel #25
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.flag = 0
     self.links = []
     self.title = ""
     self.img= ""
     self.content = ""
Beispiel #26
0
	def __init__(self):
		HTMLParser.__init__(self)
		self.tag_count = 0
		self.token_count = 0
		self.in_script = False
		self.slope = []
		self.slope.append((0,0))
Beispiel #27
0
 def __init__(self, fh):
     """
     {fh} must be an input stream returned by open() or urllib2.urlopen()
     """
     HTMLParser.__init__(self)
     self.fileids = []
     self.feed(fh.read())
Beispiel #28
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.InRow=0
     self.InEntry=0
     self.table =  []
     self.tmpRow = []
     self.hyperlinks = []
     self.RunNumber = 0
     self.TriggerRates = []
     self.Nevts = []
     self.LiveLumiByLS = []
     self.DeliveredLumiByLS = []
     self.FirstLS = -1
     self.LastLS = -1
     self.AvLiveLumi = []
     self.AvDeliveredLumi = []
     self.AvDeadtime = []
     self.DeadTime = []#grant
     self.L1Prescales=[]
     self.RunPage = ''
     self.RatePage = ''
     self.LumiPage = ''
     self.L1Page=''
     self.L1_LS_Page = ''#grant
     self.PrescaleColumn=[]
     self.PrescaleColumnString = ''
Beispiel #29
0
 def __init__(self, html=None):
     self._conn = make_db_connection()
     self._sort_order = 0
     HTMLParser.__init__(self)
     self._tag_tree = []
     if html is not None:
         self.feed(html)
    def __init__(self, url):
        """Returns new Sequence object with specified url

        url: link to mp3.zing.vn web page
        """
        HTMLParser.__init__(self)
        self.song_name = []
        self.song_artist = []
        self.song_link = []
        self.song_type = []
        req = urlopen(url)  # open connection to web page
        data = None
        if req.info().get('Content-Encoding') == "gzip":
            buf = StringIO( req.read())
            f = gzip.GzipFile(fileobj=buf)
            data = f.read().split("\n")
        else:
            data = req.read().split("\n")  # split web page with \n
        feed_data = None
        for param in data:
            if (param.find('<param name="flashvars" value="') > -1):
                """Find line to get xml url
                """
                feed_data = param
                break
        self.feed(feed_data)  # parser html data
Beispiel #31
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.links = []
Beispiel #32
0
 def __init__(self, sanitizationLevel=_defaultSanitizationLevel):
     HTMLParser.__init__(self)
     if sanitizationLevel not in range(0, 3):
         sanitizationLevel = self._defaultSanitizationLevel
     self._sanitizationLevel = sanitizationLevel
 def __init__(self):
     HTMLParser.__init__(self)
     self.ck = None
 def __init__(self):
     HTMLParser.__init__(self)
     self.captcha_id = None
     self.captcha_url = None
Beispiel #35
0
 def __init__(self, *args, **kwargs):
     HTMLParser.__init__(self, *args, **kwargs)
     self._section = None
     self.translations = []
Beispiel #36
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.terms = []
Beispiel #37
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.boxes = []
     self.__line_text = None
     self.__char_positions = None
Beispiel #38
0
 def __init__(self, *args, **kwargs):
     HTMLParser.__init__(self, *args, **kwargs)
     self.elements = set()
Beispiel #39
0
 def __init__(self, scale):
     HTMLParser.__init__(self)
     self.__scale = scale
     self.__sio = StringIO()
     self.map_id = ''
Beispiel #40
0
 def __init__(self, placeToStadium):
     HTMLParser.__init__(self)
     self.m_rawSchedule = []
     self.m_placeToStadium = placeToStadium
Beispiel #41
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.worksheets = []
Beispiel #42
0
    def __init__(self):

        HTMLParser.__init__(self)

        self.data = list()
        self.recording = False
Beispiel #43
0
 def __init__(self, *args, **kwargs):
     HTMLParser.__init__(self)
     super(HTMLRewriter, self).__init__(*args, **kwargs)
Beispiel #44
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.links = {}
     self.tag_now = ""
     self.endtag = ['br', 'img']
Beispiel #45
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.intag = ''
Beispiel #46
0
 def __init__(self, folder):
     HTMLParser.__init__(self)
     self.folder = folder
     self.num_tests = 0
     self.testcase = None
     self.start_copy = False
Beispiel #47
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.flag = False
     self.peacekeeper_score = ''
Beispiel #48
0
 def __init__(self):
     # Old style class in py2 :(
     HTMLParser.__init__(self)
     self.hrefs = []
Beispiel #49
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.data = []
Beispiel #50
0
 def __init__(self, parent_path, *args, **kwargs):
     self.parent_path_parts = parent_path.split("/")[:-1]
     #SGMLParser is an old style classe
     HTMLParser.__init__(self, *args, **kwargs)
Beispiel #51
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.count = 0
     self.price = ''
Beispiel #52
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.__text = []
Beispiel #53
0
 def __init__(self, request_data):
     HTMLParser.__init__(self)
     self.pre_flag = False
     #		self.data_format=data_format
     self.record_tag = "pre"
     self.request_data = request_data
Beispiel #54
0
 def __init__(self):
     self.strlst = []
     self.a = False
     HTMLParser.__init__(self)
Beispiel #55
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.pdf = ''
Beispiel #56
0
 def __init__(self, baseHref):
     HTMLParser.__init__(self)
     self.baseHref = baseHref
     self.staticContent = []
     self.externalLinks = []
     self.internalLinks = []
Beispiel #57
0
 def __init__(self, out):
     HTMLParser.__init__(self)
     self.out = out
     self.in_javascript = False
Beispiel #58
0
 def __init__(self, count):
     HTMLParser.__init__(self)
     self.count = count
     self.summary = u''
Beispiel #59
0
 def __init__(self):
     HTMLParser.__init__(self)
Beispiel #60
0
 def __init__(self):
     HTMLParser.__init__(self)
     self.papers = []
     self.start = False