Python util.get_webmention_target函数代码示例

def get_webmention_targets(source, activity):
  """Returns a set of string target URLs to attempt to send webmentions to.

  Side effect: runs the original post discovery algorithm on the activity and
  adds the resulting URLs to the activity as tags, in place.

  Args:
   source: models.Source subclass
   activity: activity dict
  """
  original_post_discovery.discover(source, activity)

  targets = set()
  obj = activity.get('object') or activity

  for tag in obj.get('tags', []):
    url = tag.get('url')
    if url and tag.get('objectType') == 'article':
      url, domain, send = util.get_webmention_target(url)
      tag['url'] = url
      if send:
        targets.add(url)

  for url in obj.get('upstreamDuplicates', []):
    url, domain, send = util.get_webmention_target(url)
    if send:
      targets.add(url)

  return targets

  def test_get_webmention_target_blacklisted_urls(self):
    gwt = util.get_webmention_target
    for bad in ('http://facebook.com/x', 'https://www.facebook.com/y',
                'http://sub.dom.ain.facebook.com/z'):
      self.assertFalse(util.get_webmention_target(bad)[2], bad)

    self.assertTrue(util.get_webmention_target('http://good.com/a')[2])

 def test_get_webmention_target_too_big(self):
   self.expect_requests_head('http://orig', response_headers={
     'Content-Length': str(util.MAX_HTTP_RESPONSE_SIZE + 1),
   })
   self.mox.ReplayAll()
   self.assert_equals(('http://orig', 'orig', False),
                      util.get_webmention_target('http://orig'))

  def resolve_profile_url(url, resolve=True):
    """Resolves a profile URL to be added to a source.

    Args:
      url: string
      resolve: boolean, whether to make HTTP requests to follow redirects, etc.

    Returns: string, resolved URL, or None
    """
    final, _, ok = util.get_webmention_target(url, resolve=resolve)
    if not ok:
      return None

    final = final.lower()
    if util.schemeless(final).startswith(util.schemeless(url.lower())):
      # redirected to a deeper path. use the original higher level URL. #652
      final = url

    # If final has a path segment check if root has a matching rel=me.
    match = re.match(r'^(https?://[^/]+)/.+', final)
    if match and resolve:
      root = match.group(1)
      try:
        resp = util.requests_get(root)
        resp.raise_for_status()
        data = util.mf2py_parse(resp.text, root)
        me_urls = data.get('rels', {}).get('me', [])
        if final in me_urls:
          final = root
      except requests.RequestException:
        logging.warning("Couldn't fetch %s, preserving path in %s",
                        root, final, exc_info=True)

    return final

  def _urls_and_domains(self, auth_entity, user_url):
    """Returns this user's valid (not webmention-blacklisted) URLs and domains.

    Converts the auth entity's user_json to an ActivityStreams actor and uses
    its 'urls' and 'url' fields. May be overridden by subclasses.

    Args:
      auth_entity: oauth_dropins.models.BaseAuth
      user_url: string, optional URL passed in when authorizing

    Returns: ([string url, ...], [string domain, ...])
    """
    actor = self.gr_source.user_to_actor(json.loads(auth_entity.user_json))
    logging.debug('Converted to actor: %s', json.dumps(actor, indent=2))

    candidates = util.trim_nulls(util.uniquify(
        [user_url] + microformats2.object_urls(actor)))

    if len(candidates) > MAX_AUTHOR_URLS:
      logging.warning('Too many profile links! Only resolving the first %s: %s',
                      MAX_AUTHOR_URLS, candidates)

    urls = []
    for i, url in enumerate(candidates):
      url, domain, send = util.get_webmention_target(url, resolve=i < MAX_AUTHOR_URLS)
      if send:
        urls.append(url)

    urls = util.dedupe_urls(urls)  # normalizes domains to lower case
    domains = [util.domain_from_link(url) for url in urls]
    return urls, domains

  def _url_and_domain(self, auth_entity):
    """Returns this source's URL and domain.

    Uses the auth entity user_json 'url' field by default. May be overridden
    by subclasses.

    Args:
      auth_entity: oauth_dropins.models.BaseAuth

    Returns: (string url, string domain, boolean ok) tuple
    """
    user_json = json.loads(auth_entity.user_json)
    actor = self.as_source.user_to_actor(user_json)
    urls = util.trim_nulls([actor.get('url')] +
                           # also look at G+'s urls field
                           [u.get('value') for u in user_json.get('urls', [])])

    first_url = first_domain = None
    for url in urls:
      # TODO: fully support multiple urls
      for url in url.split():
        url, domain, ok = util.get_webmention_target(url)
        if ok:
          domain = domain.lower()
          return url, domain, True
        elif not first_url:
          first_url = url
          first_domain = domain

    return first_url, first_domain, False

    def test_get_webmention_middle_redirect_blacklisted(self):
        """We should allow blacklisted domains in the middle of a redirect chain.

    ...e.g. Google's redirector https://www.google.com/url?...
    """
        self.expect_requests_head("http://orig", redirected_url=["https://www.google.com/url?xyz", "https://end"])
        self.mox.ReplayAll()
        self.assert_equals(("https://end", "end", True), util.get_webmention_target("http://orig", resolve=True))

 def resolve(urls):
   resolved = set()
   for url in urls:
     final, _, send = util.get_webmention_target(url)
     if send:
       resolved.add(final)
       if include_redirect_sources:
         resolved.add(url)
   return resolved

  def add_original_post_urls(self, post_id, obj, prop):
    """Extracts original post URLs and adds them to an object, in place.

    If the post object has upstreamDuplicates, *only* they are considered
    original post URLs and added as tags with objectType 'article', and the
    post's own links and 'article' tags are added with objectType 'mention'.

    Args:
      post_id: string post id
      obj: ActivityStreams post object
      prop: string property name in obj to add the original post URLs to
    """
    post = None
    try:
      post = self.source.get_post(post_id)
    except:
      logging.warning('Error fetching source post %s', post_id, exc_info=True)
      return
    if not post:
      logging.warning('Source post %s not found', post_id)
      return

    original_post_discovery.discover(self.source, post, fetch_hfeed=False)
    tags = [tag for tag in post['object'].get('tags', [])
            if 'url' in tag and tag['objectType'] == 'article']
    upstreams = post['object'].get('upstreamDuplicates', [])

    if not isinstance(obj.setdefault(prop, []), list):
      obj[prop] = [obj[prop]]
    if upstreams:
      obj[prop] += [{'url': url, 'objectType': 'article'} for url in upstreams]
      obj.setdefault('tags', []).extend(
        [{'url': tag.get('url'), 'objectType': 'mention'} for tag in tags])
    else:
      obj[prop] += tags

    # check for redirects, and if there are any follow them and add final urls
    # in addition to the initial urls.
    seen = set()
    for url_list in obj[prop], obj.get('tags', []):
      for url_obj in url_list:
        url = util.clean_webmention_url(url_obj.get('url', ''))
        if not url or url in seen:
          continue
        seen.add(url)
        # when debugging locally, replace my (snarfed.org) URLs with localhost
        url_obj['url'] = url = util.replace_test_domains_with_localhost(url)
        resolved, _, send = util.get_webmention_target(url)
        if send and resolved != url and resolved not in seen:
          seen.add(resolved)
          url_list.append({'url': resolved, 'objectType': url_obj.get('objectType')})

    logging.info('After original post discovery, urls are: %s', seen)

  def post(self):
    logging.debug('Params: %s', self.request.params)

    if self.lease(ndb.Key(urlsafe=self.request.params['key'])):
      source_domains = self.entity.source.get().domains
      to_send = set()
      for url in self.entity.unsent:
        url, domain, ok = util.get_webmention_target(url)
        # skip "self" links to this blog's domain
        if ok and domain not in source_domains:
          to_send.add(url)

      self.entity.unsent = list(to_send)
      self.send_webmentions()

  def _urls_and_domains(self, auth_entity, user_url):
    """Returns this user's valid (not webmention-blacklisted) URLs and domains.

    Converts the auth entity's user_json to an ActivityStreams actor and uses
    its 'urls' and 'url' fields. May be overridden by subclasses.

    Args:
      auth_entity: :class:`oauth_dropins.models.BaseAuth`
      user_url: string, optional URL passed in when authorizing

    Returns:
      ([string url, ...], [string domain, ...])
    """
    actor = self.gr_source.user_to_actor(json.loads(auth_entity.user_json))
    logging.debug('Converted to actor: %s', json.dumps(actor, indent=2))

    candidates = util.trim_nulls(util.uniquify(
        [user_url] + microformats2.object_urls(actor)))

    if len(candidates) > MAX_AUTHOR_URLS:
      logging.info('Too many profile links! Only resolving the first %s: %s',
                   MAX_AUTHOR_URLS, candidates)

    urls = []
    for i, url in enumerate(candidates):
      final, domain, ok = util.get_webmention_target(url, resolve=i < MAX_AUTHOR_URLS)
      if ok:
        final = final.lower()
        if util.schemeless(final).startswith(util.schemeless(url.lower())):
          # redirected to a deeper path. use the original higher level URL. #652
          final = url
        # If final has a path segment check if root has a matching rel=me.
        match = re.match(r'^(https?://[^/]+)/.+', final)
        if match and i < MAX_AUTHOR_URLS:
          root = match.group(1)
          resp = util.requests_get(root)
          resp.raise_for_status()
          data = util.mf2py_parse(resp.text, root)
          me_urls = data.get('rels', {}).get('me', [])
          if final in me_urls:
            final = root
        urls.append(final)

    urls = util.dedupe_urls(urls)  # normalizes domains to lower case
    domains = [util.domain_from_link(url) for url in urls]
    return urls, domains

  def do_send_webmentions(self):
    urls = self.entity.unsent + self.entity.error + self.entity.failed
    unsent = set()
    self.entity.error = []
    self.entity.failed = []

    for orig_url in urls:
      # recheck the url here since the checks may have failed during the poll
      # or streaming add.
      url, domain, ok = util.get_webmention_target(orig_url)
      if ok:
        if len(url) <= _MAX_STRING_LENGTH:
          unsent.add(url)
        else:
          logging.warning('Giving up on target URL over %s chars! %s',
                          _MAX_STRING_LENGTH, url)
          self.entity.failed.append(orig_url)
    self.entity.unsent = sorted(unsent)

    while self.entity.unsent:
      target = self.entity.unsent.pop(0)
      source_url = self.source_url(target)
      logging.info('Webmention from %s to %s', source_url, target)

      # see if we've cached webmention discovery for this domain. the cache
      # value is a string URL endpoint if discovery succeeded, a
      # WebmentionSend error dict if it failed (semi-)permanently, or None.
      cache_key = util.webmention_endpoint_cache_key(target)
      cached = memcache.get(cache_key)
      if cached:
        logging.info('Using cached webmention endpoint %r: %s', cache_key, cached)

      # send! and handle response or error
      error = None
      if isinstance(cached, dict):
        error = cached
      else:
        mention = send.WebmentionSend(source_url, target, endpoint=cached)
        logging.info('Sending...')
        try:
          if not mention.send(timeout=999, headers=util.USER_AGENT_HEADER):
            error = mention.error
        except BaseException, e:
          logging.warning('', exc_info=True)
          error = getattr(mention, 'error')
          if not error:
            error = ({'code': 'BAD_TARGET_URL', 'http_status': 499}
                     if 'DNS lookup failed for URL:' in str(e)
                     else {'code': 'EXCEPTION'})

      if not cached:
        val = (error if error and error['code'] in ('NO_ENDPOINT', 'BAD_TARGET_URL')
               else mention.receiver_endpoint)
        memcache.set(cache_key, val, time=WEBMENTION_DISCOVERY_CACHE_TIME)

      if error is None:
        logging.info('Sent! %s', mention.response)
        self.record_source_webmention(mention)
        self.entity.sent.append(target)
      else:
        code = error['code']
        status = error.get('http_status', 0)
        if (code == 'NO_ENDPOINT' or
            (code == 'BAD_TARGET_URL' and status == 204)):  # 204 is No Content
          logging.info('Giving up this target. %s', error)
          self.entity.skipped.append(target)
        elif status // 100 == 4:
          # Give up on 4XX errors; we don't expect later retries to succeed.
          logging.info('Giving up this target. %s', error)
          self.entity.failed.append(target)
        else:
          self.fail('Error sending to endpoint: %s' % error)
          self.entity.error.append(target)

      if target in self.entity.unsent:
        self.entity.unsent.remove(target)

  def post(self):
    logging.info('Params: %self', self.request.params.items())
    self.source_url = util.get_required_param(self, 'source')
    self.target_url = util.get_required_param(self, 'target')
    assert self.PREVIEW in (True, False)

    # parse and validate target URL
    try:
      parsed = urlparse.urlparse(self.target_url)
    except BaseException:
      return self.error(msg, 'Could not parse target URL %s' % self.target_url)

    domain = parsed.netloc
    path_parts = parsed.path.rsplit('/', 1)
    source_cls = SOURCE_NAMES.get(path_parts[-1])
    if (domain not in ('brid.gy', 'www.brid.gy', 'localhost:8080') or
        len(path_parts) != 2 or path_parts[0] != '/publish' or not source_cls):
      return self.error('Target must be brid.gy/publish/{facebook,twitter}')
    elif source_cls in (Instagram, GooglePlusPage):
      return self.error('Sorry, %s is not yet supported.' %
                        source_cls.AS_CLASS.NAME)

    # resolve source URL
    url, domain, ok = util.get_webmention_target(self.source_url)
    # show nice error message if they're trying to publish a silo post
    if domain in SOURCE_DOMAINS:
      return self.error(
        "Looks like that's a %s URL. Try one from your web site instead!" %
        SOURCE_DOMAINS[domain].AS_CLASS.NAME)
    elif not ok:
      return self.error('Unsupported source URL %s' % url)
    elif not domain:
      return self.error('Could not parse source URL %s' % url)

    # When debugging locally, use snarfed.org for localhost webmentions
    if appengine_config.DEBUG and domain == 'localhost':
      domain = 'snarfed.org'

    # look up source by domain
    domain = domain.lower()
    sources = source_cls.query().filter(source_cls.domains == domain).fetch(100)
    if not sources:
      return self.error("Could not find <b>%(type)s</b> account for <b>%(domain)s</b>. Check that your %(type)s profile has %(domain)s in its <em>web site</em> or <em>link</em> field, then try signing up again." %
        {'type': source_cls.AS_CLASS.NAME, 'domain': domain})

    for source in sources:
      logging.info('Source: %s , features %s, status %s' %
                   (source.bridgy_url(self), source.features, source.status))
      if source.status == 'enabled' and 'publish' in source.features:
        self.source = source
        break
    else:
      return self.error(
        'Publish is not enabled for your account(s). Please visit %s and sign up!' %
        ' or '.join(s.bridgy_url(self) for s in sources))

    # show nice error message if they're trying to publish their home page
    for domain_url in self.source.domain_urls:
      domain_url_parts = urlparse.urlparse(domain_url)
      source_url_parts = urlparse.urlparse(self.source_url)
      if (source_url_parts.netloc == domain_url_parts.netloc and
          source_url_parts.path.strip('/') == domain_url_parts.path.strip('/') and
          not source_url_parts.query):
        return self.error(
          "Looks like that's your home page. Try one of your posts instead!")

    # done with the sanity checks, ready to fetch the source url. create the
    # Publish entity so we can store the result.
    entity = self.get_or_add_publish_entity(url)
    if (entity.status == 'complete' and entity.type != 'preview' and
        not self.PREVIEW and not appengine_config.DEBUG):
      return self.error("Sorry, you've already published that page, and Bridgy Publish doesn't yet support updating or deleting existing posts. Ping Ryan if you want that feature!")
    self.entity = entity

    # fetch source page
    resp = self.fetch_mf2(url)
    if not resp:
      return
    self.fetched, data = resp

    # loop through each item and its children and try to preview/create it. if
    # it fails, try the next one. break after the first one that works.
    resp = None
    types = set()
    queue = collections.deque(data.get('items', []))
    while queue:
      item = queue.popleft()
      item_types = set(item.get('type'))
      if 'h-feed' in item_types and 'h-entry' not in item_types:
        queue.extend(item.get('children', []))
        continue

      try:
        resp = self.attempt_single_item(item)
        if self.entity.published:
          break
        if resp.abort:
          return self.error(resp.error_plain, html=resp.error_html, data=item)
        # try the next item
        for embedded in ('rsvp', 'invitee', 'repost', 'repost-of', 'like',
#.........这里部分代码省略.........

    def do_send_webmentions(self):
        urls = self.entity.unsent + self.entity.error + self.entity.failed
        unsent = set()
        self.entity.error = []
        self.entity.failed = []

        for orig_url in urls:
            # recheck the url here since the checks may have failed during the poll
            # or streaming add.
            url, domain, ok = util.get_webmention_target(orig_url)
            if ok:
                if len(url) <= _MAX_STRING_LENGTH:
                    unsent.add(url)
                else:
                    logging.warning("Giving up on target URL over %s chars! %s", _MAX_STRING_LENGTH, url)
                    self.entity.failed.append(orig_url)
        self.entity.unsent = sorted(unsent)

        while self.entity.unsent:
            target = self.entity.unsent.pop(0)
            source_url = self.source_url(target)
            logging.info("Webmention from %s to %s", source_url, target)

            # see if we've cached webmention discovery for this domain. the cache
            # value is a string URL endpoint if discovery succeeded, a
            # WebmentionSend error dict if it failed (semi-)permanently, or None.
            cache_key = util.webmention_endpoint_cache_key(target)
            cached = memcache.get(cache_key)
            if cached:
                logging.info("Using cached webmention endpoint %r: %s", cache_key, cached)

            # send! and handle response or error
            error = None
            if isinstance(cached, dict):
                error = cached
            else:
                mention = send.WebmentionSend(source_url, target, endpoint=cached)
                logging.info("Sending...")
                try:
                    if not mention.send(timeout=999, headers=util.REQUEST_HEADERS):
                        error = mention.error
                except BaseException, e:
                    logging.warning("", exc_info=True)
                    error = getattr(mention, "error")
                    if not error:
                        error = (
                            {"code": "BAD_TARGET_URL", "http_status": 499}
                            if "DNS lookup failed for URL:" in str(e)
                            else {"code": "EXCEPTION"}
                        )

            error_code = error["code"] if error else None
            if error_code != "BAD_TARGET_URL" and not cached:
                val = error if error_code == "NO_ENDPOINT" else mention.receiver_endpoint
                memcache.set(cache_key, val, time=WEBMENTION_DISCOVERY_CACHE_TIME)

            if error is None:
                logging.info("Sent! %s", mention.response)
                self.record_source_webmention(mention)
                self.entity.sent.append(target)
            else:
                status = error.get("http_status", 0)
                if error_code == "NO_ENDPOINT" or (error_code == "BAD_TARGET_URL" and status == 204):  # No Content
                    logging.info("Giving up this target. %s", error)
                    self.entity.skipped.append(target)
                elif status // 100 == 4:
                    # Give up on 4XX errors; we don't expect later retries to succeed.
                    logging.info("Giving up this target. %s", error)
                    self.entity.failed.append(target)
                else:
                    self.fail("Error sending to endpoint: %s" % error)
                    self.entity.error.append(target)

            if target in self.entity.unsent:
                self.entity.unsent.remove(target)

  def do_send_webmentions(self):
    unsent = set()
    for url in self.entity.unsent + self.entity.error:
      # recheck the url here since the checks may have failed during the poll
      # or streaming add.
      url, domain, ok = util.get_webmention_target(url)
      if ok:
        # When debugging locally, redirect our own webmentions to localhost
        if appengine_config.DEBUG and domain in util.LOCALHOST_TEST_DOMAINS:
            url = url.replace(domain, 'localhost')
        unsent.add(url)
    self.entity.unsent = sorted(unsent)
    self.entity.error = []

    while self.entity.unsent:
      target = self.entity.unsent.pop(0)
      source_url = self.source_url(target)
      logging.info('Webmention from %s to %s', source_url, target)

      # see if we've cached webmention discovery for this domain. the cache
      # value is a string URL endpoint if discovery succeeded, a
      # WebmentionSend error dict if it failed (semi-)permanently, or None.
      domain = util.domain_from_link(target)
      cache_key = 'W ' + domain
      cached = memcache.get(cache_key)
      if cached:
        logging.info('Using cached webmention endpoint for %s: %s',
                     domain, cached)

      # send! and handle response or error
      error = None
      if isinstance(cached, dict):
        error = cached
      else:
        mention = send.WebmentionSend(source_url, target, endpoint=cached)
        logging.info('Sending...')
        try:
          if not mention.send(timeout=999):
            error = mention.error
        except:
          logging.warning('', exc_info=True)
          error = getattr(mention, 'error', None)
          if not error:
            error = {'code': 'EXCEPTION'}

      if error is None:
        logging.info('Sent! %s', mention.response)
        if not self.entity.sent:
          self.set_last_webmention_sent()
        self.entity.sent.append(target)
        memcache.set(cache_key, mention.receiver_endpoint,
                     time=WEBMENTION_DISCOVERY_CACHE_TIME)
      else:
        if error['code'] == 'NO_ENDPOINT':
          logging.info('Giving up this target. %s', error)
          self.entity.skipped.append(target)
          memcache.set(cache_key, error, time=WEBMENTION_DISCOVERY_CACHE_TIME)
        elif (error['code'] == 'BAD_TARGET_URL' and
              error['http_status'] / 100 == 4):
          # Give up on 4XX errors; we don't expect later retries to succeed.
          logging.info('Giving up this target. %s', error)
          self.entity.failed.append(target)
        else:
          self.fail('Error sending to endpoint: %s' % error)
          self.entity.error.append(target)

      if target in self.entity.unsent:
        self.entity.unsent.remove(target)

    if self.entity.error:
      logging.warning('Propagate task failed')
      self.release('error')
    else:
      self.complete()