NewsBlur-viq/apps/analyzer/views.py

import redis
from utils import log as logging
from django.shortcuts import get_object_or_404
from django.views.decorators.http import require_POST
from django.conf import settings
from django.template import RequestContext
from django.shortcuts import render_to_response
from mongoengine.queryset import NotUniqueError
from apps.rss_feeds.models import Feed
from apps.reader.models import UserSubscription
from apps.analyzer.models import MClassifierTitle, MClassifierAuthor, MClassifierFeed, MClassifierTag
from apps.analyzer.models import get_classifiers_for_user, MPopularityQuery
from apps.analyzer.forms import PopularityQueryForm
from apps.social.models import MSocialSubscription
from utils import json_functions as json
from utils.user_functions import get_user
from utils.user_functions import ajax_login_required
from utils.view_functions import render_to

def index(requst):
    pass
    
@require_POST
@ajax_login_required
@json.json_view
def save_classifier(request):
    post = request.POST
    feed_id = post['feed_id']
    feed = None
    social_user_id = None
    if feed_id.startswith('social:'):
        social_user_id = int(feed_id.replace('social:', ''))
        feed_id = None
    else:
        feed_id = int(feed_id)
        feed = get_object_or_404(Feed, pk=feed_id)
    code = 0
    message = 'OK'
    payload = {}

    logging.user(request, "~FGSaving classifier: ~SB%s~SN ~FW%s" % (feed, post))
    
    # Mark subscription as dirty, so unread counts can be recalculated
    usersub = None
    socialsub = None
    if social_user_id:
        socialsub = MSocialSubscription.objects.get(user_id=request.user.pk, subscription_user_id=social_user_id)
        if not socialsub.needs_unread_recalc:
            socialsub.needs_unread_recalc = True
            socialsub.save()
    else:
        try:
            usersub = UserSubscription.objects.get(user=request.user, feed=feed)
        except UserSubscription.DoesNotExist:
            pass
        if usersub and (not usersub.needs_unread_recalc or not usersub.is_trained):
            usersub.needs_unread_recalc = True
            usersub.is_trained = True
            usersub.save()
        
        
    def _save_classifier(ClassifierCls, content_type):
        classifiers = {
            'like_'+content_type: 1, 
            'dislike_'+content_type: -1,
            'remove_like_'+content_type: 0,
            'remove_dislike_'+content_type: 0,
        }
        for opinion, score in classifiers.items():
            if opinion in post:
                post_contents = post.getlist(opinion)
                for post_content in post_contents:
                    if not post_content: continue
                    classifier_dict = {
                        'user_id': request.user.pk,
                        'feed_id': feed_id or 0,
                        'social_user_id': social_user_id or 0,
                    }
                    if content_type in ('author', 'tag', 'title'):
                        classifier_dict.update({content_type: post_content})
                    if content_type == 'feed':
                        if not post_content.startswith('social:'):
                            classifier_dict['feed_id'] = post_content
                    try:
                        classifier = ClassifierCls.objects.get(**classifier_dict)
                    except ClassifierCls.DoesNotExist:
                        classifier_dict.update(dict(score=score))
                        classifier = ClassifierCls.objects.create(**classifier_dict)
                    except NotUniqueError:
                        continue
                    if score == 0:
                        classifier.delete()
                    elif classifier.score != score:
                        if score == 0:
                            if ((classifier.score == 1 and opinion.startswith('remove_like'))
                                or (classifier.score == -1 and opinion.startswith('remove_dislike'))):
                                classifier.delete()
                        else:
                            classifier.score = score
                            classifier.save()
                        
    _save_classifier(MClassifierAuthor, 'author')
    _save_classifier(MClassifierTag, 'tag')
    _save_classifier(MClassifierTitle, 'title')
    _save_classifier(MClassifierFeed, 'feed')

    r = redis.Redis(connection_pool=settings.REDIS_PUBSUB_POOL)
    r.publish(request.user.username, 'feed:%s' % feed_id)

    response = dict(code=code, message=message, payload=payload)
    return response
    
@json.json_view
def get_classifiers_feed(request, feed_id):
    user = get_user(request)
    code = 0
    
    payload = get_classifiers_for_user(user, feed_id=feed_id)
    
    response = dict(code=code, payload=payload)
    
    return response

def popularity_query(request):
    if request.method == 'POST':
        form = PopularityQueryForm(request.POST)
        if form.is_valid():
            logging.user(request.user, "~BC~FRPopularity query: ~SB%s~SN requests \"~SB~FM%s~SN~FR\"" % (request.POST['email'], request.POST['query']))
            query = MPopularityQuery.objects.create(email=request.POST['email'],
                                                    query=request.POST['query'])
            query.queue_email()
            
            response = render_to_response('analyzer/popularity_query.xhtml', {
                'success': True,
                'popularity_query_form': form,
            }, context_instance=RequestContext(request))
            response.set_cookie('newsblur_popularity_query', request.POST['query'])
            
            return response
        else:
            logging.user(request.user, "~BC~FRFailed popularity query: ~SB%s~SN requests \"~SB~FM%s~SN~FR\"" % (request.POST['email'], request.POST['query']))
    else:
        logging.user(request.user, "~BC~FRPopularity query form loading")
        form = PopularityQueryForm(initial={'query': request.COOKIES.get('newsblur_popularity_query', "")})
    
    response = render_to_response('analyzer/popularity_query.xhtml', {
        'popularity_query_form': form,
    }, context_instance=RequestContext(request))

    return response