samlmd.py

from __future__ import absolute_import, unicode_literals
from datetime import datetime
from .utils import parse_xml, check_signature, root, validate_document, xml_error, \
    schema, iso2datetime, duration2timedelta, filter_lang, url2host, trunc_str, subdomains, \
    has_tag, hash_id, load_callable, rreplace, dumptree, first_text, url_get, img_to_data
from .logs import log
from .constants import config, NS, ATTRS, NF_URI, PLACEHOLDER_ICON
from lxml import etree
from lxml.builder import ElementMaker
from lxml.etree import DocumentInvalid
from itertools import chain
from copy import deepcopy
from .exceptions import *
from six import StringIO


class EntitySet(object):
    def __init__(self, initial=None):
        self._e = dict()
        if initial is not None:
            for e in initial:
                self.add(e)

    def add(self, value):
        self._e[value.get('entityID')] = value

    def discard(self, value):
        entity_id = value.get('entityID')
        if entity_id in self._e:
            del self._e[entity_id]

    def __iter__(self):
        for e in self._e.values():
            yield e

    def __len__(self):
        return len(self._e.keys())

    def __contains__(self, item):
        return item.get('entityID') in self._e.keys()


def find_merge_strategy(strategy_name):
    if '.' not in strategy_name:
        strategy_name = "pyff.merge_strategies:%s" % strategy_name
    if ':' not in strategy_name:
        strategy_name = rreplace(strategy_name, '.', ':')  # backwards compat for old way of specifying these
    return load_callable(strategy_name)


def parse_saml_metadata(source,
                        key=None,
                        base_url=None,
                        fail_on_error=False,
                        filter_invalid=True,
                        cleanup=None,
                        validate=True,
                        validation_errors=None):
    """Parse a piece of XML and return an EntitiesDescriptor element after validation.

:param source: a file-like object containing SAML metadata
:param key: a certificate (file) or a SHA1 fingerprint to use for signature verification
:param base_url: use this base url to resolve relative URLs for XInclude processing
:param fail_on_error: (default: False)
:param filter_invalid: (default True) remove invalid EntityDescriptor elements rather than raise an errror
:param validate: (default: True) set to False to turn off all XML schema validation
:param validation_errors: A dict that will be used to return validation errors to the caller
:param cleanup: A callable that can be used to pre-process parsed metadata before validation. Use as a clue-bat.
(but after xinclude processing and signature validation)
    """

    if validation_errors is None:
        validation_errors = dict()

    try:
        t = parse_xml(source, base_url=base_url)
        t.xinclude()

        expire_time_offset = metadata_expiration(t)

        t = check_signature(t, key)

        if cleanup is not None:
            t = cleanup(t)
        else:  # at least get rid of ID attribute
            for e in iter_entities(t):
                if e.get('ID') is not None:
                    del e.attrib['ID']

        t = root(t)

        if fail_on_error:
            filter_invalid = False

        if validate:
            if filter_invalid:
                t = filter_invalids_from_document(t, base_url=base_url, validation_errors=validation_errors)
            else:  # all or nothing
                try:
                    validate_document(t)
                except DocumentInvalid as ex:
                    validation_errors[base_url] = xml_error(ex.error_log, m=base_url)
                    raise MetadataException("schema validation failed: [{}] '{}': {}"
                                            .format(base_url, source, xml_error(ex.error_log, m=base_url)))

        if t is not None:
            if t.tag == "{%s}EntityDescriptor" % NS['md']:
                t = entitiesdescriptor([t], base_url, copy=False, validate=True, nsmap=t.nsmap)

    except Exception as ex:
        if fail_on_error:
            raise ex
        log.error(ex)
        return None, None

    log.debug("returning %d valid entities" % len(list(iter_entities(t))))

    return t, expire_time_offset


class SAMLMetadataResourceParser():
    def __init__(self):
        pass

    def magic(self, content):
        return "EntitiesDescriptor" in content or "EntityDescriptor" in content

    def parse(self, resource, content):
        info = dict()
        info['Validation Errors'] = dict()
        t, expire_time_offset = parse_saml_metadata(StringIO(content.encode('utf8')),
                                                    key=resource.opts['verify'],
                                                    base_url=resource.url,
                                                    cleanup=resource.opts['cleanup'],
                                                    fail_on_error=resource.opts['fail_on_error'],
                                                    filter_invalid=resource.opts['filter_invalid'],
                                                    validate=resource.opts['validate'],
                                                    validation_errors=info['Validation Errors'])

        if expire_time_offset is not None:
            expire_time = datetime.now() + expire_time_offset
            resource.expire_time = expire_time
            info['Expiration Time'] = str(expire_time)

        if t is not None:
            resource.t = t
            resource.type = "application/samlmetadata+xml"

        return info


from .parse import add_parser

add_parser(SAMLMetadataResourceParser())


def metadata_expiration(t):
    relt = root(t)
    if relt.tag in ('{%s}EntityDescriptor' % NS['md'], '{%s}EntitiesDescriptor' % NS['md']):
        cache_duration = config.default_cache_duration
        valid_until = relt.get('validUntil', None)
        if valid_until is not None:
            now = datetime.utcnow()
            vu = iso2datetime(valid_until)
            now = now.replace(microsecond=0)
            vu = vu.replace(microsecond=0, tzinfo=None)
            return vu - now
        elif config.respect_cache_duration:
            cache_duration = relt.get('cacheDuration', config.default_cache_duration)
            if not cache_duration:
                cache_duration = config.default_cache_duration
            return duration2timedelta(cache_duration)

    return None


def filter_invalids_from_document(t, base_url, validation_errors):
    xsd = schema()
    for e in iter_entities(t):
        if not xsd.validate(e):
            log.debug(etree.tostring(e))
            error = xml_error(xsd.error_log, m=base_url)
            entity_id = e.get("entityID","(Missing entityID)")
            log.warn('removing \'%s\': schema validation failed: %s' % (entity_id, xsd.error_log))
            validation_errors[entity_id] = "{}".format(xsd.error_log)
            if e.getparent() is None:
                return None
            e.getparent().remove(e)
    return t


def entitiesdescriptor(entities,
                       name,
                       lookup_fn=None,
                       cache_duration=None,
                       valid_until=None,
                       validate=True,
                       copy=True,
                       nsmap=None):
    """
:param lookup_fn: a function used to lookup entities by name
:param entities: a set of entities specifiers (lookup is used to find entities from this set)
:param name: the @Name attribute
:param cache_duration: an XML timedelta expression, eg PT1H for 1hr
:param valid_until: a relative time eg 2w 4d 1h for 2 weeks, 4 days and 1hour from now.
:param copy: set to False to avoid making a copy of all the entities in list. This may be dangerous.
:param validate: set to False to skip schema validation of the resulting EntitiesDesciptor element. This is dangerous!

Produce an EntityDescriptors set from a list of entities. Optional Name, cacheDuration and validUntil are affixed.
    """

    if nsmap is None:
        nsmap = dict()

    def _resolve(member, l_fn):
        if hasattr(member, 'tag'):
            return [member]
        else:
            return l_fn(member)

    nsmap.update(NS)
    resolved_entities = set()
    for member in entities:
        for entity in _resolve(member, lookup_fn):
            resolved_entities.add(entity)

    if not resolved_entities:
        return None

    for entity in resolved_entities:
        nsmap.update(entity.nsmap)

    log.debug("selecting %d entities before validation" % len(resolved_entities))

    attrs = dict(Name=name, nsmap=nsmap)
    if cache_duration is not None:
        attrs['cacheDuration'] = cache_duration
    if valid_until is not None:
        attrs['validUntil'] = valid_until
    t = etree.Element("{%s}EntitiesDescriptor" % NS['md'], **attrs)
    for entity in resolved_entities:
        entity_id = entity.get('entityID', None)
        if (entity is not None) and (entity_id is not None):
            ent_insert = entity
            if copy:
                ent_insert = deepcopy(ent_insert)
            t.append(ent_insert)

    if config.devel_write_xml_to_file:
        with open("/tmp/pyff_entities_out.xml", "w") as fd:
            fd.write(dumptree(t))

    if validate:
        try:
            validate_document(t)
        except DocumentInvalid as ex:
            log.debug(xml_error(ex.error_log))
            raise MetadataException("XML schema validation failed: %s" % name)
    return t


def entities_list(t=None):
    """
        :param t: An EntitiesDescriptor or EntityDescriptor element

        Returns the list of contained EntityDescriptor elements
        """
    if t is None:
        return []
    elif root(t).tag == "{%s}EntityDescriptor" % NS['md']:
        return [root(t)]
    else:
        return iter_entities(t)


def iter_entities(t):
    if t is None:
        return []
    return t.iter('{%s}EntityDescriptor' % NS['md'])


def find_entity(t, e_id, attr='entityID'):
    for e in iter_entities(t):
        if e.get(attr) == e_id:
            return e
    return None


# semantics copied from https://github.com/lordal/md-summary/blob/master/md-summary
# many thanks to Anders Lordahl & Scotty Logan for the idea
def guess_entity_software(e):
    for elt in chain(e.findall(".//{%s}SingleSignOnService" % NS['md']),
                     e.findall(".//{%s}AssertionConsumerService" % NS['md'])):
        location = elt.get('Location')
        if location:
            if 'Shibboleth.sso' in location \
                    or 'profile/SAML2/POST/SSO' in location \
                    or 'profile/SAML2/Redirect/SSO' in location \
                    or 'profile/Shibboleth/SSO' in location:
                return 'Shibboleth'
            if location.endswith('saml2/idp/SSOService.php') or 'saml/sp/saml2-acs.php' in location:
                return 'SimpleSAMLphp'
            if location.endswith('user/authenticate'):
                return 'KalturaSSP'
            if location.endswith('adfs/ls') or location.endswith('adfs/ls/'):
                return 'ADFS'
            if '/oala/' in location or 'login.openathens.net' in location:
                return 'OpenAthens'
            if '/idp/SSO.saml2' in location or '/sp/ACS.saml2' in location \
                    or 'sso.connect.pingidentity.com' in location:
                return 'PingFederate'
            if 'idp/saml2/sso' in location:
                return 'Authentic2'
            if 'nidp/saml2/sso' in location:
                return 'Novell Access Manager'
            if 'affwebservices/public/saml2sso' in location:
                return 'CASiteMinder'
            if 'FIM/sps' in location:
                return 'IBMTivoliFIM'
            if 'sso/post' in location \
                    or 'sso/redirect' in location \
                    or 'saml2/sp/acs' in location \
                    or 'saml2/ls' in location \
                    or 'saml2/acs' in location \
                    or 'acs/redirect' in location \
                    or 'acs/post' in location \
                    or 'saml2/sp/ls/' in location:
                return 'PySAML'
            if 'engine.surfconext.nl' in location:
                return 'SURFConext'
            if 'opensso' in location:
                return 'OpenSSO'
            if 'my.salesforce.com' in location:
                return 'Salesforce'

    entity_id = e.get('entityID')
    if '/shibboleth' in entity_id:
        return 'Shibboleth'
    if entity_id.endswith('/metadata.php'):
        return 'SimpleSAMLphp'
    if '/openathens' in entity_id:
        return 'OpenAthens'

    return 'other'


def is_idp(entity):
    return has_tag(entity, "{%s}IDPSSODescriptor" % NS['md'])


def is_sp(entity):
    return has_tag(entity, "{%s}SPSSODescriptor" % NS['md'])


def is_aa(entity):
    return has_tag(entity, "{%s}AttributeAuthorityDescriptor" % NS['md'])


def _domains(entity):
    domains = [url2host(entity.get('entityID'))]
    for d in entity.iter("{%s}DomainHint" % NS['mdui']):
        if d.text not in domains:
            domains.append(d.text)
    return domains


def with_entity_attributes(entity, cb):
    def _stext(e):
        if e.text is not None:
            return e.text.strip()

    for ea in entity.iter("{%s}EntityAttributes" % NS['mdattr']):
        for a in ea.iter("{%s}Attribute" % NS['saml']):
            an = a.get('Name', None)
            if a is not None:
                values = filter(lambda x: x is not None, [_stext(v) for v in a.iter("{%s}AttributeValue" % NS['saml'])])
                cb(an, values)


def _all_domains_and_subdomains(entity):
    dlist = []
    try:
        for dn in _domains(entity):
            for sub in subdomains(dn):
                if len(sub) > 1:  # TLD
                    dlist.append(sub)
    except ValueError:
        pass
    return dlist


def entity_attributes(entity):
    d = {}

    def _u(an, values):
        d[an] = values

    with_entity_attributes(entity, _u)

    return d


def entity_attribute_dict(entity):
    d = {}

    def _u(an, values):
        d[an] = values

    with_entity_attributes(entity, _u)

    d[ATTRS['domain']] = _all_domains_and_subdomains(entity)

    roles = d.setdefault(ATTRS['role'], [])
    if is_idp(entity):
        roles.append('idp')
        eca = ATTRS['entity-category']
        ec = d.setdefault(eca, [])
        if 'http://refeds.org/category/hide-from-discovery' not in ec:
            ec.append('http://pyff.io/category/discoverable')
    if is_sp(entity):
        roles.append('sp')
    if is_aa(entity):
        roles.append('aa')

    if ATTRS['software'] not in d:
        d[ATTRS['software']] = [guess_entity_software(entity)]

    return d

def gen_icon(e):
    scopes = entity_scopes(e)


def entity_icon_url(e, langs=None):
    for ico in filter_lang(e.iter("{%s}Logo" % NS['mdui']), langs=langs):
        return dict(url=ico.text, width=ico.get('width'), height=ico.get('height'))


def privacy_statement_url(entity, langs):
    for url in filter_lang(entity.iter("{%s}PrivacyStatementURL" % NS['mdui']), langs=langs):
        return url.text


def entity_geoloc(entity):
    for loc in entity.iter("{%s}GeolocationHint" % NS['mdui']):
        pos = loc.text[5:].split(",")
        return dict(lat=pos[0], long=pos[1])


def entity_domains(entity):
    domains = []
    for d in entity.iter("{%s}DomainHint" % NS['mdui']):
        if d.text == '.':
            return []
        domains.append(d.text)
    if not domains:
        domains.append(url2host(entity.get('entityID')))
    return domains


def entity_extended_display(entity, langs=None):
    """Utility-method for computing a displayable string for a given entity.

    :param entity: An EntityDescriptor element
    :param langs: The list of languages to search in priority order
    """
    display = entity.get('entityID')
    info = ''

    for organizationName in filter_lang(entity.iter("{%s}OrganizationName" % NS['md']), langs=langs):
        info = display
        display = organizationName.text

    for organizationDisplayName in filter_lang(entity.iter("{%s}OrganizationDisplayName" % NS['md']), langs=langs):
        info = display
        display = organizationDisplayName.text

    for serviceName in filter_lang(entity.iter("{%s}ServiceName" % NS['md']), langs=langs):
        info = display
        display = serviceName.text

    for displayName in filter_lang(entity.iter("{%s}DisplayName" % NS['mdui']), langs=langs):
        info = display
        display = displayName.text

    for organizationUrl in filter_lang(entity.iter("{%s}OrganizationURL" % NS['md']), langs=langs):
        info = organizationUrl.text

    for description in filter_lang(entity.iter("{%s}Description" % NS['mdui']), langs=langs):
        info = description.text

    if info == entity.get('entityID'):
        info = ''

    return trunc_str(display.strip(), 40), trunc_str(info.strip(), 256)


def entity_display_name(entity, langs=None):
    """Utility-method for computing a displayable string for a given entity.

    :param entity: An EntityDescriptor element
    :param langs: The list of languages to search in priority order
    """
    for displayName in filter_lang(entity.iter("{%s}DisplayName" % NS['mdui']), langs=langs):
        return displayName.text.strip()

    for serviceName in filter_lang(entity.iter("{%s}ServiceName" % NS['md']), langs=langs):
        return serviceName.text.strip()

    for organizationDisplayName in filter_lang(entity.iter("{%s}OrganizationDisplayName" % NS['md']), langs=langs):
        return organizationDisplayName.text.strip()

    for organizationName in filter_lang(entity.iter("{%s}OrganizationName" % NS['md']), langs=langs):
        return organizationName.text.strip()

    return entity.get('entityID').strip()


def sub_domains(e):
    lst = []
    domains = entity_domains(e)
    for d in domains:
        for sub in subdomains(d):
            if sub not in lst:
                lst.append(sub)
    return lst


def entity_scopes(e):
    elt = e.findall('.//{%s}IDPSSODescriptor/{%s}Extensions/{%s}Scope' % (NS['md'], NS['md'], NS['shibmd']))
    if elt is None or len(elt) == 0:
        return None
    return [s.text for s in elt]


def discojson(e, langs=None):
    if e is None:
        return dict()

    title, descr = entity_extended_display(e)
    entity_id = e.get('entityID')

    d = dict(title=title,
             descr=descr,
             auth='saml',
             entityID=entity_id)

    eattr = entity_attribute_dict(e)
    if 'idp' in eattr[ATTRS['role']]:
        d['type'] = 'idp'
        d['hidden'] = 'true'
        if 'http://pyff.io/category/discoverable' in eattr[ATTRS['entity-category']]:
            d['hidden'] = 'false'
    elif 'sp' in eattr[ATTRS['role']]:
        d['type'] = 'sp'

    scopes = entity_scopes(e)
    icon_info = entity_icon_url(e)
    urls = []
    if icon_info is not None and 'url' in icon_info:
        url = icon_info['url']
        urls.append(url)
        if scopes is not None and len(scopes) == 1:
            urls.append("https://{}/favico.ico".format(scopes[0]))
            urls.append("https://www.{}/favico.ico".format(scopes[0]))

    d['entity_icon'] = None
    for url in urls:
        if url.startswith("data:"):
            d['entity_icon'] = url
            break

        if '://' in url:
            r = url_get(url)
            if r.ok and r.content:
                d['entity_icon'] = img_to_data(r.content, r.headers.get('Content-Type'))
                break

    if scopes is not None and len(scopes) > 0:
        d['scope'] = ",".join(scopes)
        if len(scopes) == 1:
            d['domain'] = scopes[0]
            d['name_tag'] = (scopes[0].split('.'))[0].upper()

    keywords = filter_lang(e.iter("{%s}Keywords" % NS['mdui']), langs=langs)
    if keywords is not None:
        lst = [elt.text for elt in keywords]
        if len(lst) > 0:
            d['keywords'] = ",".join(lst)
    psu = privacy_statement_url(e, langs)
    if psu:
        d['privacy_statement_url'] = psu
    geo = entity_geoloc(e)
    if geo:
        d['geo'] = geo

    return d


def sha1_id(e):
    return hash_id(e, 'sha1')


def entity_simple_summary(e):
    if e is None:
        return dict()

    title, descr = entity_extended_display(e)
    entity_id = e.get('entityID')
    d = dict(title=title,
             descr=descr,
             entity_id=entity_id,
             entityID=entity_id,
             domains=";".join(sub_domains(e)),
             id=hash_id(e, 'sha1'))

    scopes = entity_scopes(e)
    if scopes is not None and len(scopes) > 0:
        d['scopes'] = " ".join(scopes)

    psu = privacy_statement_url(e, None)
    if psu:
        d['privacy_statement_url'] = psu

    return d

def entity_orgurl(entity, langs=None):
    for organizationUrl in filter_lang(entity.iter("{%s}OrganizationURL" % NS['md']), langs=langs):
        return organizationUrl.text
    return None


def entity_service_name(entity, langs=None):
    for serviceName in filter_lang(entity.iter("{%s}ServiceName" % NS['md']), langs=langs):
        return serviceName.text


def entity_service_description(entity, langs=None):
    for serviceName in filter_lang(entity.iter("{%s}ServiceDescription" % NS['md']), langs=langs):
        return serviceName.text


def entity_requested_attributes(entity, langs=None):
    return [(a.get('Name'), bool(a.get('isRequired'))) for a in
            filter_lang(entity.iter("{%s}RequestedAttribute" % NS['md']), langs=langs)]


def entity_idp(entity):
    for idp in entity.iter("{%s}IDPSSODescriptor" % NS['md']):
        return idp

    return None


def entity_sp(entity):
    for sp in entity.iter("{%s}SPSSODescriptor" % NS['md']):
        return sp

    return None


def entity_contacts(entity):
    def _contact_dict(contact):
        first_name = first_text(contact, "{%s}GivenName" % NS['md'])
        last_name = first_text(contact, "{%s}SurName" % NS['md'])
        org = first_text(entity, "{%s}OrganizationName" % NS['md']) or first_text(entity,
                                                                                  "{%s}OrganizationDisplayName" % NS[
                                                                                      'md'])
        company = first_text(entity, "{%s}Company" % NS['md'])
        mail = first_text(contact, "{%s}EmailAddress" % NS['md'])
        display_name = "Unknown"
        if first_name and last_name:
            display_name = ' '.join([first_name, last_name])
        elif first_name:
            display_name = first_name
        elif last_name:
            display_name = last_name
        elif mail:
            _, _, display_name = mail.partition(':')

        return dict(type=contact.get('contactType'),
                    first_name=first_name,
                    last_name=last_name,
                    company=company or org,
                    display_name=display_name,
                    mail=mail)

    return [_contact_dict(c) for c in entity.iter("{%s}ContactPerson" % NS['md'])]


def entity_nameid_formats(entity):
    return [nif.text for nif in entity.iter("{%s}NameIDFormat" % NS['md'])]


def entity_info(e, langs=None):
    d = entity_simple_summary(e)
    keywords = filter_lang(e.iter("{%s}Keywords" % NS['mdui']), langs=langs)
    if keywords is not None:
        lst = [elt.text for elt in keywords]
        if len(lst) > 0:
            d['keywords'] = ",".join(lst)

    d['privacy_statement_url'] = privacy_statement_url(e, langs)
    d['geo'] = entity_geoloc(e)
    d['orgurl'] = entity_orgurl(e, langs)
    d['scopes'] = entity_scopes(e)
    d['service_name'] = entity_service_name(e, langs)
    d['service_descr'] = entity_service_description(e, langs)
    d['requested_attributes'] = entity_requested_attributes(e, langs)
    d['entity_attributes'] = entity_attributes(e)
    d['contacts'] = entity_contacts(e)
    d['name_id_formats'] = entity_nameid_formats(e)
    d['is_idp'] = is_idp(e)
    d['is_sp'] = is_sp(e)
    d['is_aa'] = is_aa(e)
    d['xml'] = dumptree(e, xml_declaration=False, pretty_print=True).decode('utf8').replace('<', '&lt;').replace('>',
                                                                                                                 '&gt;')
    if d['is_idp']:
        d['protocols'] = entity_idp(e).get('protocolSupportEnumeration', "").split()

    return d


def entity_extensions(e):
    """Return a list of the Extensions elements in the EntityDescriptor

:param e: an EntityDescriptor
:return: a list
    """
    ext = e.find("./{%s}Extensions" % NS['md'])
    if ext is None:
        ext = etree.Element("{%s}Extensions" % NS['md'])
        e.insert(0, ext)
    return ext


def annotate_entity(e, category, title, message, source=None):
    """Add an ATOM annotation to an EntityDescriptor or an EntitiesDescriptor. This is a simple way to
    add non-normative text annotations to metadata, eg for the purpuse of generating reports.

:param e: An EntityDescriptor or an EntitiesDescriptor element
:param category: The ATOM category
:param title: The ATOM title
:param message: The ATOM content
:param source: An optional source URL. It is added as a <link> element with @rel='saml-metadata-source'
    """
    if e.tag != "{%s}EntityDescriptor" % NS['md'] and e.tag != "{%s}EntitiesDescriptor" % NS['md']:
        raise MetadataException('I can only annotate EntityDescriptor or EntitiesDescriptor elements')
    subject = e.get('Name', e.get('entityID', None))
    atom = ElementMaker(nsmap={'atom': 'http://www.w3.org/2005/Atom'}, namespace='http://www.w3.org/2005/Atom')
    args = [atom.published("%s" % datetime.now().isoformat()),
            atom.link(href=subject, rel="saml-metadata-subject")]
    if source is not None:
        args.append(atom.link(href=source, rel="saml-metadata-source"))
    args.extend([atom.title(title),
                 atom.category(term=category),
                 atom.content(message, type="text/plain")])
    entity_extensions(e).append(atom.entry(*args))


def _entity_attributes(e):
    ext = entity_extensions(e)
    ea = ext.find(".//{%s}EntityAttributes" % NS['mdattr'])
    if ea is None:
        ea = etree.Element("{%s}EntityAttributes" % NS['mdattr'])
        ext.append(ea)
    return ea


def _eattribute(e, attr, nf):
    ea = _entity_attributes(e)
    a = ea.xpath(".//saml:Attribute[@NameFormat='%s' and @Name='%s']" % (nf, attr),
                 namespaces=NS,
                 smart_strings=False)
    if a is None or len(a) == 0:
        a = etree.Element("{%s}Attribute" % NS['saml'])
        a.set('NameFormat', nf)
        a.set('Name', attr)
        ea.append(a)
    else:
        a = a[0]
    return a


def set_entity_attributes(e, d, nf=NF_URI):
    """Set an entity attribute on an EntityDescriptor

:param e: The EntityDescriptor element
:param d: A dict of attribute-value pairs that should be added as entity attributes
:param nf: The nameFormat (by default "urn:oasis:names:tc:SAML:2.0:attrname-format:uri") to use.
:raise: MetadataException unless e is an EntityDescriptor element
    """
    if e.tag != "{%s}EntityDescriptor" % NS['md']:
        raise MetadataException("I can only add EntityAttribute(s) to EntityDescriptor elements")

    for attr, value in d.items():
        a = _eattribute(e, attr, nf)
        velt = etree.Element("{%s}AttributeValue" % NS['saml'])
        velt.text = value
        a.append(velt)


def set_pubinfo(e, publisher=None, creation_instant=None):
    if e.tag != "{%s}EntitiesDescriptor" % NS['md']:
        raise MetadataException("I can only set RegistrationAuthority to EntitiesDescriptor elements")
    if publisher is None:
        raise MetadataException("At least publisher must be provided")

    if creation_instant is None:
        now = datetime.utcnow()
        creation_instant = now.strftime("%Y-%m-%dT%H:%M:%SZ")

    ext = entity_extensions(e)
    pi = ext.find(".//{%s}PublicationInfo" % NS['mdrpi'])
    if pi is not None:
        raise MetadataException("A PublicationInfo element is already present")
    pi = etree.Element("{%s}PublicationInfo" % NS['mdrpi'])
    pi.set('publisher', publisher)
    if creation_instant:
        pi.set('creationInstant', creation_instant)
    ext.append(pi)


def set_reginfo(e, policy=None, authority=None):
    if e.tag != "{%s}EntityDescriptor" % NS['md']:
        raise MetadataException("I can only set RegistrationAuthority to EntityDescriptor elements")
    if authority is None:
        raise MetadataException("At least authority must be provided")
    if policy is None:
        policy = dict()

    ext = entity_extensions(e)
    ri = ext.find(".//{%s}RegistrationInfo" % NS['mdrpi'])
    if ri is not None:
        raise MetadataException("A RegistrationInfo element is already present")

    ri = etree.Element("{%s}RegistrationInfo" % NS['mdrpi'])
    ext.append(ri)
    ri.set('registrationAuthority', authority)
    for lang, policy_url in policy.items():
        rp = etree.Element("{%s}RegistrationPolicy" % NS['mdrpi'])
        rp.text = policy_url
        rp.set('{%s}lang' % NS['xml'], lang)
        ri.append(rp)


def expiration(t):
    relt = root(t)
    if relt.tag in ('{%s}EntityDescriptor' % NS['md'], '{%s}EntitiesDescriptor' % NS['md']):
        cache_duration = config.default_cache_duration
        valid_until = relt.get('validUntil', None)
        if valid_until is not None:
            now = datetime.utcnow()
            vu = iso2datetime(valid_until)
            now = now.replace(microsecond=0)
            vu = vu.replace(microsecond=0, tzinfo=None)
            return vu - now
        elif config.respect_cache_duration:
            cache_duration = relt.get('cacheDuration', config.default_cache_duration)
            return duration2timedelta(cache_duration)

    return None


def sort_entities(t, sxp=None):
    """
Sorts the working entities 't' by the value returned by the xpath 'sxp'
By default, entities are sorted by 'entityID' when this method is called without 'sxp', and otherwise as
second criteria.
Entities where no value exists for the given 'sxp' are sorted last.

:param t: An element tree containing the entities to sort
:param sxp: xpath expression selecting the value used for sorting the entities
"""
    def get_key(e):
        eid = e.attrib.get('entityID')
        sv = None
        try:
            sxp_values = e.xpath(sxp, namespaces=NS, smart_strings=False)
            try:
                sv = sxp_values[0]
                try:
                    sv = sv.text
                except AttributeError:
                    pass
            except IndexError:
                log.warn("Sort pipe: unable to sort entity by '%s'. "
                         "Entity '%s' has no such value" % (sxp, eid))
        except TypeError:
            pass

        log.debug("Generated sort key for entityID='%s' and %s='%s'" % (eid, sxp, sv))
        return sv is None, sv, eid

    container = root(t)
    container[:] = sorted(container, key=lambda e: get_key(e))