# copyright 2010-2015 LOGILAB S.A. (Paris, FRANCE), all rights reserved.

# contact http://www.logilab.fr/ -- mailto:contact@logilab.fr

#

# This file is part of CubicWeb.

#

# CubicWeb is free software: you can redistribute it and/or modify it under the

# terms of the GNU Lesser General Public License as published by the Free

# Software Foundation, either version 2.1 of the License, or (at your option)

# any later version.

#

# CubicWeb is distributed in the hope that it will be useful, but WITHOUT

# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS

# FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public License for more

# details.

#

# You should have received a copy of the GNU Lesser General Public License along

# with CubicWeb.  If not, see <http://www.gnu.org/licenses/>.

"""datafeed parser for xml generated by cubicweb

Example of mapping for CWEntityXMLParser::

  {u'CWUser': {                                        # EntityType

      (u'in_group', u'subject', u'link'): [            # (rtype, role, action)

          (u'CWGroup', {u'linkattr': u'name'})],       #   -> rules = [(EntityType, options), ...]

      (u'tags', u'object', u'link-or-create'): [       # (...)

          (u'Tag', {u'linkattr': u'name'})],           #   -> ...

      (u'use_email', u'subject', u'copy'): [           # (...)

          (u'EmailAddress', {})]                       #   -> ...

      }

   }

"""

from datetime import datetime, time

import urllib

from six import text_type

from six.moves.urllib.parse import urlparse, urlunparse, parse_qs, urlencode

import pytz

from logilab.common.date import todate, totime

from logilab.common.textutils import splitstrip, text_to_dict

from logilab.common.decorators import classproperty

from yams.constraints import BASE_CONVERTERS

from yams.schema import role_name as rn

from cubicweb import ValidationError, RegistryException

from cubicweb.view import Component

from cubicweb.server.sources import datafeed

from cubicweb.server.hook import match_rtype

# XXX see cubicweb.cwvreg.YAMS_TO_PY

# XXX see cubicweb.web.views.xmlrss.SERIALIZERS

DEFAULT_CONVERTERS = BASE_CONVERTERS.copy()

DEFAULT_CONVERTERS['String'] = text_type

DEFAULT_CONVERTERS['Password'] = lambda x: x.encode('utf8')

def convert_date(ustr):

    return todate(datetime.strptime(ustr, '%Y-%m-%d'))

DEFAULT_CONVERTERS['Date'] = convert_date

def convert_datetime(ustr):

    if '.' in ustr: # assume %Y-%m-%d %H:%M:%S.mmmmmm

        ustr = ustr.split('.', 1)[0]

    return datetime.strptime(ustr, '%Y-%m-%d %H:%M:%S')

DEFAULT_CONVERTERS['Datetime'] = convert_datetime

# XXX handle timezone, though this will be enough as TZDatetime are

# serialized without time zone by default (UTC time). See

# cw.web.views.xmlrss.SERIALIZERS.

def convert_tzdatetime(ustr):

    date = convert_datetime(ustr)

    date = date.replace(tzinfo=pytz.utc)

    return date

DEFAULT_CONVERTERS['TZDatetime'] = convert_tzdatetime

def convert_time(ustr):

    return totime(datetime.strptime(ustr, '%H:%M:%S'))

DEFAULT_CONVERTERS['Time'] = convert_time

DEFAULT_CONVERTERS['TZTime'] = convert_time

def convert_interval(ustr):

    return time(seconds=int(ustr))

DEFAULT_CONVERTERS['Interval'] = convert_interval

def extract_typed_attrs(eschema, stringdict, converters=DEFAULT_CONVERTERS):

    typeddict = {}

    for rschema in eschema.subject_relations():

        if rschema.final and rschema in stringdict:

            if rschema in ('eid', 'cwuri'):  # XXX really omit cwuri?

                continue

            attrtype = eschema.destination(rschema)

            value = stringdict[rschema]

            if value is not None:

                value = converters[attrtype](value)

            typeddict[rschema.type] = value

    return typeddict

def rtype_role_rql(rtype, role):

    if role == 'object':

        return 'Y %s X WHERE X eid %%(x)s' % rtype

    else:

        return 'X %s Y WHERE X eid %%(x)s' % rtype

class CWEntityXMLParser(datafeed.DataFeedXMLParser):

    """datafeed parser for the 'xml' entity view

    Most of the logic is delegated to the following components:

    * an "item builder" component, turning an etree xml node into a specific

      python dictionary representing an entity

    * "action" components, selected given an entity, a relation and its role in

      the relation, and responsible to link the entity to given related items

      (eg dictionary)

    So the parser is only doing the gluing service and the connection to the

    source.

    """

    __regid__ = 'cw.entityxml'

    def __init__(self, *args, **kwargs):

        super(CWEntityXMLParser, self).__init__(*args, **kwargs)

        self._parsed_urls = {}

        self._processed_entities = set()

    def select_linker(self, action, rtype, role, entity=None):

        try:

            return self._cw.vreg['components'].select(

                'cw.entityxml.action.%s' % action, self._cw, entity=entity,

                rtype=rtype, role=role, parser=self)

        except RegistryException:

            raise RegistryException('Unknown action %s' % action)

    def list_actions(self):

        reg = self._cw.vreg['components']

        return sorted(clss[0].action for rid, clss in reg.items()

                      if rid.startswith('cw.entityxml.action.'))

    # mapping handling #########################################################

    def add_schema_config(self, schemacfg, checkonly=False):

        """added CWSourceSchemaConfig, modify mapping accordingly"""

        _ = self._cw._

        try:

            rtype = schemacfg.schema.rtype.name

        except AttributeError:

            msg = _("entity and relation types can't be mapped, only attributes "

                    "or relations")

            raise ValidationError(schemacfg.eid, {rn('cw_for_schema', 'subject'): msg})

        if schemacfg.options:

            options = text_to_dict(schemacfg.options)

        else:

            options = {}

        try:

            role = options.pop('role')

            if role not in ('subject', 'object'):

                raise KeyError

        except KeyError:

            msg = _('"role=subject" or "role=object" must be specified in options')

            raise ValidationError(schemacfg.eid, {rn('options', 'subject'): msg})

        try:

            action = options.pop('action')

            linker = self.select_linker(action, rtype, role)

            linker.check_options(options, schemacfg.eid)

        except KeyError:

            msg = _('"action" must be specified in options; allowed values are '

                    '%s') % ', '.join(self.list_actions())

            raise ValidationError(schemacfg.eid, {rn('options', 'subject'): msg})

        except RegistryException:

            msg = _('allowed values for "action" are %s') % ', '.join(self.list_actions())

            raise ValidationError(schemacfg.eid, {rn('options', 'subject'): msg})

        if not checkonly:

            if role == 'subject':

                etype = schemacfg.schema.stype.name

                ttype = schemacfg.schema.otype.name

            else:

                etype = schemacfg.schema.otype.name

                ttype = schemacfg.schema.stype.name

            etyperules = self.source.mapping.setdefault(etype, {})

            etyperules.setdefault((rtype, role, action), []).append(

                (ttype, options))

            self.source.mapping_idx[schemacfg.eid] = (

                etype, rtype, role, action, ttype)

    def del_schema_config(self, schemacfg, checkonly=False):

        """deleted CWSourceSchemaConfig, modify mapping accordingly"""

        etype, rtype, role, action, ttype = self.source.mapping_idx[schemacfg.eid]

        rules = self.source.mapping[etype][(rtype, role, action)]

        rules = [x for x in rules if not x[0] == ttype]

        if not rules:

            del self.source.mapping[etype][(rtype, role, action)]

    # import handling ##########################################################

    def process(self, url, raise_on_error=False):

        """IDataFeedParser main entry point"""

        if url.startswith('http'): # XXX similar loose test as in parse of sources.datafeed

            url = self.complete_url(url)

        super(CWEntityXMLParser, self).process(url, raise_on_error)

    def parse_etree(self, parent):

        """Overriden from :class:`DataFeedXMLParser` to use a builder component."""

        for node in list(parent):

            builder = self._cw.vreg['components'].select(

                'cw.entityxml.item-builder', self._cw, node=node,

                parser=self)

            yield builder.build_item()

    def process_item(self, item, rels, raise_on_error=False):

        """

        item and rels are what's returned by the item builder `build_item` method:

        * `item` is an {attribute: value} dictionary

        * `rels` is for relations and structured as

           {role: {relation: [(related item, related rels)...]}

        """

        entity = self.extid2entity(item['cwuri'].encode('ascii'), item['cwtype'],

                                   cwsource=item['cwsource'], item=item,

                                   raise_on_error=raise_on_error)

        if entity is None:

            return None

        if entity.eid in self._processed_entities:

            return entity

        self._processed_entities.add(entity.eid)

        if not (self.created_during_pull(entity) or self.updated_during_pull(entity)):

            attrs = extract_typed_attrs(entity.e_schema, item)

            self.update_if_necessary(entity, attrs)

        self.process_relations(entity, rels)

        return entity

    def process_relations(self, entity, rels):

        etype = entity.cw_etype

        for (rtype, role, action), rules in self.source.mapping.get(etype, {}).items():

            try:

                related_items = rels[role][rtype]

            except KeyError:

                self.import_log.record_error('relation %s-%s not found in xml export of %s'

                                             % (rtype, role, etype))

                continue

            try:

                linker = self.select_linker(action, rtype, role, entity)

            except RegistryException:

                self.import_log.record_error('no linker for action %s' % action)

            else:

                linker.link_items(related_items, rules)

    def before_entity_copy(self, entity, sourceparams):

        """IDataFeedParser callback"""

        attrs = extract_typed_attrs(entity.e_schema, sourceparams['item'])

        entity.cw_edited.update(attrs)

    def normalize_url(self, url):

        """overridden to add vid=xml if vid is not set in the qs"""

        url = super(CWEntityXMLParser, self).normalize_url(url)

        purl = urlparse(url)

        if purl.scheme in ('http', 'https'):

            params = parse_qs(purl.query)

            if 'vid' not in params:

                params['vid'] = ['xml']

                purl = list(purl)

                purl[4] = urlencode(params, doseq=True)

                return urlunparse(purl)

        return url

    def complete_url(self, url, etype=None, known_relations=None):

        """append to the url's query string information about relation that should

        be included in the resulting xml, according to source mapping.

        If etype is not specified, try to guess it using the last path part of

        the url, i.e. the format used by default in cubicweb to map all entities

        of a given type as in 'http://mysite.org/EntityType'.

        If `known_relations` is given, it should be a dictionary of already

        known relations, so they don't get queried again.

        """

        purl = urlparse(url)

        params = parse_qs(purl.query)

        if etype is None:

            etype = purl.path.split('/')[-1]

        try:

            etype = self._cw.vreg.case_insensitive_etypes[etype.lower()]

        except KeyError:

            return url

        relations = params['relation'] = set(params.get('relation', ()))

        for rtype, role, _ in self.source.mapping.get(etype, ()):

            if known_relations and rtype in known_relations.get('role', ()):

                continue

            relations.add('%s-%s' % (rtype, role))

        purl = list(purl)

        purl[4] = urlencode(params, doseq=True)

        return urlunparse(purl)

    def complete_item(self, item, rels):

        try:

            return self._parsed_urls[item['cwuri']]

        except KeyError:

            itemurl = self.complete_url(item['cwuri'], item['cwtype'], rels)

            item_rels = list(self.parse(itemurl))

            assert len(item_rels) == 1, 'url %s expected to bring back one '\

                'and only one entity, got %s' % (itemurl, len(item_rels))

            self._parsed_urls[item['cwuri']] = item_rels[0]

            if rels:

                # XXX (do it better) merge relations

                new_rels = item_rels[0][1]

                new_rels.get('subject', {}).update(rels.get('subject', {}))

                new_rels.get('object', {}).update(rels.get('object', {}))

            return item_rels[0]

class CWEntityXMLItemBuilder(Component):

    __regid__ = 'cw.entityxml.item-builder'

    def __init__(self, _cw, parser, node, **kwargs):

        super(CWEntityXMLItemBuilder, self).__init__(_cw, **kwargs)

        self.parser = parser

        self.node = node

    def build_item(self):

        """parse a XML document node and return two dictionaries defining (part

        of) an entity:

        - {attribute: value}

        - {role: {relation: [(related item, related rels)...]}

        """

        node = self.node

        item = dict(node.attrib.items())

        item['cwtype'] = text_type(node.tag)

        item.setdefault('cwsource', None)

        item['eid'] = int(item['eid'])

        rels = {}

        for child in node:

            role = child.get('role')

            if role:

                # relation

                related = rels.setdefault(role, {}).setdefault(child.tag, [])

                related += self.parser.parse_etree(child)

            elif child.text:

                # attribute

                item[child.tag] = text_type(child.text)

            else:

                # None attribute (empty tag)

                item[child.tag] = None

        return item, rels

class CWEntityXMLActionCopy(Component):

    """implementation of cubicweb entity xml parser's'copy' action

    Takes no option.

    """

    __regid__ = 'cw.entityxml.action.copy'

    def __init__(self, _cw, parser, rtype, role, entity=None, **kwargs):

        super(CWEntityXMLActionCopy, self).__init__(_cw, **kwargs)

        self.parser = parser

        self.rtype = rtype

        self.role = role

        self.entity = entity

    @classproperty

    def action(cls):

        return cls.__regid__.rsplit('.', 1)[-1]

    def check_options(self, options, eid):

        self._check_no_options(options, eid)

    def _check_no_options(self, options, eid, msg=None):

        if options:

            if msg is None:

                msg = self._cw._("'%s' action doesn't take any options") % self.action