Mercurial > libervia-web
view src/pages/blog_new/page_meta.py @ 985:64826e69f365
pages: cache mechanism, first draft:
a cache mechanism has been implemented to retrieve pages with a complexe rendering and/or calling expensive methods (e.g. network calls).
For now it's is done only for Pubsub and with service profile (i.e. profile used when user is not logged in).
When a LiberviaPage use cache, node is subscribed, and as long as no event is received (even can be item update, item retraction, or node deletion), the cached page is returned.
This is a first draft, it is planed to handle in the future logged users (which can be tricky as we must not let (un)subscribed node if user is not willing to), multi-nodes pages (e.g.: item + comments) and cache for page not depending on pubsub (e.g. chat).
author | Goffi <goffi@goffi.org> |
---|---|
date | Sun, 19 Nov 2017 17:18:14 +0100 |
parents | 96a56856d357 |
children | d042f194624a |
line wrap: on
line source
#!/usr/bin/env python2.7 # -*- coding: utf-8 -*- from libervia.server.constants import Const as C from twisted.internet import defer from twisted.words.protocols.jabber import jid from sat.tools.common import data_objects access = C.PAGES_ACCESS_PUBLIC # can be a callable template = u"blog/articles.html" @defer.inlineCallbacks def parse_url(self, request): try: prof_requested = self.nextPath(request) except IndexError: self.pageError(request) template_data = request.template_data target_profile = yield self.host.bridge.profileNameGet(prof_requested) template_data[u'target_profile'] = target_profile target_jid = yield self.host.bridge.asyncGetParamA('JabberID', 'Connection', 'value', profile_key=target_profile) target_jid = jid.JID(target_jid).userhost() template_data[u'target_jid'] = target_jid @defer.inlineCallbacks def prepare_render(self, request): target_jid = request.template_data[u'target_jid'] blog_data = yield self.host.bridge.mbGet( target_jid, "", 10, [], {}, C.SERVICE_PROFILE) request.template_data[u'items'] = data_objects.BlogItems(blog_data)