[utils] change extract_attributes to work in python 2

[youtube-dl] / youtube_dl / utils.py

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index ae813099dded05c15ed57d27e51be7704e778520..518cea98bd52639403caaf16dc7a5137651e5e2f 100644 (file)
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -139,21 +139,24 @@ def write_json_file(obj, fn):
 
 
 if sys.version_info >= (2, 7):
 
 
 if sys.version_info >= (2, 7):

-    def find_xpath_attr(node, xpath, key, val):
+    def find_xpath_attr(node, xpath, key, val=None):

         """ Find the xpath xpath[@key=val] """
         """ Find the xpath xpath[@key=val] """

-        assert re.match(r'^[a-zA-Z-]+$', key)
-        assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
-        expr = xpath + "[@%s='%s']" % (key, val)
+        assert re.match(r'^[a-zA-Z_-]+$', key)
+        if val:
+            assert re.match(r'^[a-zA-Z0-9@\s:._-]*$', val)
+        expr = xpath + ('[@%s]' % key if val is None else "[@%s='%s']" % (key, val))

         return node.find(expr)
 else:
         return node.find(expr)
 else:

-    def find_xpath_attr(node, xpath, key, val):
+    def find_xpath_attr(node, xpath, key, val=None):

         # Here comes the crazy part: In 2.6, if the xpath is a unicode,
         # .//node does not match if a node is a direct child of . !
         if isinstance(xpath, compat_str):
             xpath = xpath.encode('ascii')
 
         for f in node.findall(xpath):
         # Here comes the crazy part: In 2.6, if the xpath is a unicode,
         # .//node does not match if a node is a direct child of . !
         if isinstance(xpath, compat_str):
             xpath = xpath.encode('ascii')
 
         for f in node.findall(xpath):

-            if f.attrib.get(key) == val:
+            if key not in f.attrib:
+                continue
+            if val is None or f.attrib.get(key) == val:

                 return f
         return None
 
                 return f
         return None
 


@@ -173,12 +176,12 @@ def xpath_with_ns(path, ns_map):
     return '/'.join(replaced)
 
 
     return '/'.join(replaced)
 
 

-def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
+def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):

     if sys.version_info < (2, 7):  # Crazy 2.6
         xpath = xpath.encode('ascii')
 
     n = node.find(xpath)
     if sys.version_info < (2, 7):  # Crazy 2.6
         xpath = xpath.encode('ascii')
 
     n = node.find(xpath)

-    if n is None or n.text is None:
+    if n is None:

         if default is not NO_DEFAULT:
             return default
         elif fatal:
         if default is not NO_DEFAULT:
             return default
         elif fatal:


@@ -186,9 +189,37 @@ def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
             raise ExtractorError('Could not find XML element %s' % name)
         else:
             return None
             raise ExtractorError('Could not find XML element %s' % name)
         else:
             return None

+    return n
+
+
+def xpath_text(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
+    n = xpath_element(node, xpath, name, fatal=fatal, default=default)
+    if n is None or n == default:
+        return n
+    if n.text is None:
+        if default is not NO_DEFAULT:
+            return default
+        elif fatal:
+            name = xpath if name is None else name
+            raise ExtractorError('Could not find XML element\'s text %s' % name)
+        else:
+            return None

     return n.text
 
 
     return n.text
 
 

+def xpath_attr(node, xpath, key, name=None, fatal=False, default=NO_DEFAULT):
+    n = find_xpath_attr(node, xpath, key)
+    if n is None:
+        if default is not NO_DEFAULT:
+            return default
+        elif fatal:
+            name = '%s[@%s]' % (xpath, key) if name is None else name
+            raise ExtractorError('Could not find XML attribute %s' % name)
+        else:
+            return None
+    return n.attrib[key]
+
+

 def get_element_by_id(id, html):
     """Return the content of the tag with the specified ID in the passed HTML document"""
     return get_element_by_attribute("id", id, html)
 def get_element_by_id(id, html):
     """Return the content of the tag with the specified ID in the passed HTML document"""
     return get_element_by_attribute("id", id, html)


@@ -217,6 +248,15 @@ def get_element_by_attribute(attribute, value, html):
     return unescapeHTML(res)
 
 
     return unescapeHTML(res)
 
 

+def extract_attributes(attributes_str, attributes_regex=r'(?s)\s*([^\s=]+)\s*=\s*["\']([^"\']+)["\']'):
+    attributes = re.findall(attributes_regex, attributes_str)
+    attributes_dict = {}
+    if attributes:
+        for (attribute_name, attribute_value) in attributes:
+            attributes_dict[attribute_name] = attribute_value
+    return attributes_dict
+
+

 def clean_html(html):
     """Clean an HTML snippet into a readable string"""
 
 def clean_html(html):
     """Clean an HTML snippet into a readable string"""
 


@@ -576,16 +616,19 @@ class ContentTooShortError(Exception):
     download is too small for what the server announced first, indicating
     the connection was probably interrupted.
     """
     download is too small for what the server announced first, indicating
     the connection was probably interrupted.
     """

-    # Both in bytes
-    downloaded = None
-    expected = None

 
     def __init__(self, downloaded, expected):
 
     def __init__(self, downloaded, expected):

+        # Both in bytes

         self.downloaded = downloaded
         self.expected = expected
 
 
 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):
         self.downloaded = downloaded
         self.expected = expected
 
 
 def _create_http_connection(ydl_handler, http_class, is_https, *args, **kwargs):

+    # Working around python 2 bug (see http://bugs.python.org/issue17849) by limiting
+    # expected HTTP responses to meet HTTP/1.0 or later (see also
+    # https://github.com/rg3/youtube-dl/issues/6727)
+    if sys.version_info < (3, 0):
+        kwargs['strict'] = True

     hc = http_class(*args, **kwargs)
     source_address = ydl_handler._params.get('source_address')
     if source_address is not None:
     hc = http_class(*args, **kwargs)
     source_address = ydl_handler._params.get('source_address')
     if source_address is not None:


@@ -650,6 +693,26 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
         return ret
 
     def http_request(self, req):
         return ret
 
     def http_request(self, req):

+        # According to RFC 3986, URLs can not contain non-ASCII characters, however this is not
+        # always respected by websites, some tend to give out URLs with non percent-encoded
+        # non-ASCII characters (see telemb.py, ard.py [#3412])
+        # urllib chokes on URLs with non-ASCII characters (see http://bugs.python.org/issue3991)
+        # To work around aforementioned issue we will replace request's original URL with
+        # percent-encoded one
+        # Since redirects are also affected (e.g. http://www.southpark.de/alle-episoden/s18e09)
+        # the code of this workaround has been moved here from YoutubeDL.urlopen()
+        url = req.get_full_url()
+        url_escaped = escape_url(url)
+
+        # Substitute URL if any change after escaping
+        if url != url_escaped:
+            req_type = HEADRequest if req.get_method() == 'HEAD' else compat_urllib_request.Request
+            new_req = req_type(
+                url_escaped, data=req.data, headers=req.headers,
+                origin_req_host=req.origin_req_host, unverifiable=req.unverifiable)
+            new_req.timeout = req.timeout
+            req = new_req
+

         for h, v in std_headers.items():
             # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
             # The dict keys are capitalized because of this bug by urllib
         for h, v in std_headers.items():
             # Capitalize is needed because of Python bug 2275: http://bugs.python.org/issue2275
             # The dict keys are capitalized because of this bug by urllib


@@ -694,6 +757,18 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
             gz = io.BytesIO(self.deflate(resp.read()))
             resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
             resp.msg = old_resp.msg
             gz = io.BytesIO(self.deflate(resp.read()))
             resp = self.addinfourl_wrapper(gz, old_resp.headers, old_resp.url, old_resp.code)
             resp.msg = old_resp.msg

+        # Percent-encode redirect URL of Location HTTP header to satisfy RFC 3986 (see
+        # https://github.com/rg3/youtube-dl/issues/6457).
+        if 300 <= resp.code < 400:
+            location = resp.headers.get('Location')
+            if location:
+                # As of RFC 2616 default charset is iso-8859-1 that is respected by python 3
+                if sys.version_info >= (3, 0):
+                    location = location.encode('iso-8859-1').decode('utf-8')
+                location_escaped = escape_url(location)
+                if location != location_escaped:
+                    del resp.headers['Location']
+                    resp.headers['Location'] = location_escaped

         return resp
 
     https_request = http_request
         return resp
 
     https_request = http_request


@@ -717,6 +792,30 @@ class YoutubeDLHTTPSHandler(compat_urllib_request.HTTPSHandler):
             req, **kwargs)
 
 
             req, **kwargs)
 
 

+class YoutubeDLCookieProcessor(compat_urllib_request.HTTPCookieProcessor):
+    def __init__(self, cookiejar=None):
+        compat_urllib_request.HTTPCookieProcessor.__init__(self, cookiejar)
+
+    def http_response(self, request, response):
+        # Python 2 will choke on next HTTP request in row if there are non-ASCII
+        # characters in Set-Cookie HTTP header of last response (see
+        # https://github.com/rg3/youtube-dl/issues/6769).
+        # In order to at least prevent crashing we will percent encode Set-Cookie
+        # header before HTTPCookieProcessor starts processing it.
+        # if sys.version_info < (3, 0) and response.headers:
+        #     for set_cookie_header in ('Set-Cookie', 'Set-Cookie2'):
+        #         set_cookie = response.headers.get(set_cookie_header)
+        #         if set_cookie:
+        #             set_cookie_escaped = compat_urllib_parse.quote(set_cookie, b"%/;:@&=+$,!~*'()?#[] ")
+        #             if set_cookie != set_cookie_escaped:
+        #                 del response.headers[set_cookie_header]
+        #                 response.headers[set_cookie_header] = set_cookie_escaped
+        return compat_urllib_request.HTTPCookieProcessor.http_response(self, request, response)
+
+    https_request = compat_urllib_request.HTTPCookieProcessor.http_request
+    https_response = http_response
+
+

 def parse_iso8601(date_str, delimiter='T', timezone=None):
     """ Return a UNIX timestamp from the given date """
 
 def parse_iso8601(date_str, delimiter='T', timezone=None):
     """ Return a UNIX timestamp from the given date """
 


@@ -1546,6 +1645,10 @@ def urlencode_postdata(*args, **kargs):
     return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
 
 
     return compat_urllib_parse.urlencode(*args, **kargs).encode('ascii')
 
 

+def encode_dict(d, encoding='utf-8'):
+    return dict((k.encode(encoding), v.encode(encoding)) for k, v in d.items())
+
+

 try:
     etree_iter = xml.etree.ElementTree.Element.iter
 except AttributeError:  # Python <=2.6
 try:
     etree_iter = xml.etree.ElementTree.Element.iter
 except AttributeError:  # Python <=2.6


@@ -1886,6 +1989,32 @@ def dfxp2srt(dfxp_data):
     return ''.join(out)
 
 
     return ''.join(out)
 
 

+def cli_option(params, command_option, param):
+    param = params.get(param)
+    return [command_option, param] if param is not None else []
+
+
+def cli_bool_option(params, command_option, param, true_value='true', false_value='false', separator=None):
+    param = params.get(param)
+    assert isinstance(param, bool)
+    if separator:
+        return [command_option + separator + (true_value if param else false_value)]
+    return [command_option, true_value if param else false_value]
+
+
+def cli_valueless_option(params, command_option, param, expected_value=True):
+    param = params.get(param)
+    return [command_option] if param == expected_value else []
+
+
+def cli_configuration_args(params, param, default=[]):
+    ex_args = params.get(param)
+    if ex_args is None:
+        return default
+    assert isinstance(ex_args, list)
+    return ex_args
+
+

 class ISO639Utils(object):
     # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
     _lang_map = {
 class ISO639Utils(object):
     # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
     _lang_map = {