def get_similar_questions(self):
"""
Get 10 similar questions for given one.
Questions with the individual tags will be added to list if above questions are not full.
This function has a limitation that it will
retrieve only 100 records then select 10 most similar
from that list as querying entire database may
be very expensive - this function will benefit from
some sort of optimization
"""
#print datetime.datetime.now()
def get_data():
tags_list = self.tags.all()
similar_questions = self.__class__.objects.filter(
tags__in=self.tags.all()).exclude(
id=self.id, ).exclude(deleted=True).distinct()[:100]
similar_questions = list(similar_questions)
output = list()
for question in similar_questions:
question.similarity = self.get_similarity(
other_question=question)
#sort in reverse order - x and y are interchanged in cmp() call
similar_questions.sort(
lambda x, y: cmp(y.similarity, x.similarity))
if len(similar_questions) > 10:
return similar_questions[:10]
else:
return similar_questions
return LazyList(get_data)
def get_similar_threads(self):
"""
Get 10 similar threads for given one.
Threads with the individual tags will be added to list if above questions are not full.
This function has a limitation that it will
retrieve only 100 records then select 10 most similar
from that list as querying entire database may
be very expensive - this function will benefit from
some sort of optimization
"""
def get_data():
tags_list = self.get_tag_names()
similar_threads = Thread.objects.filter(
tags__name__in=tags_list).exclude(id=self.id).exclude(
posts__post_type='question',
posts__deleted=True).distinct()[:100]
similar_threads = list(similar_threads)
for thread in similar_threads:
thread.similarity = self.get_similarity(other_thread=thread)
similar_threads.sort(key=operator.attrgetter('similarity'),
reverse=True)
similar_threads = similar_threads[:10]
# Denormalize questions to speed up template rendering
thread_map = dict([(thread.id, thread)
for thread in similar_threads])
questions = Post.objects.get_questions()
questions = questions.select_related('thread').filter(
thread__in=similar_threads)
for q in questions:
thread_map[q.thread_id].question_denorm = q
# Postprocess data
similar_threads = [{
'url':
thread.question_denorm.get_absolute_url(),
'title':
thread.get_title(thread.question_denorm)
} for thread in similar_threads]
return similar_threads
def get_cached_data():
"""similar thread data will expire
with the default expiration delay
"""
key = 'similar-threads-%s' % self.id
data = cache.cache.get(key)
if data is None:
data = get_data()
cache.cache.set(key, data)
return data
return LazyList(get_cached_data)
def get_similar_threads(self):
"""
Get 10 similar threads for given one.
Threads with the individual tags will be added to list if above questions are not full.
This function has a limitation that it will
retrieve only 100 records then select 10 most similar
from that list as querying entire database may
be very expensive - this function will benefit from
some sort of optimization
"""
def get_data():
# todo: code in this function would be simpler if
# we had question post id denormalized on the thread
tags_list = self.get_tag_names()
similar_threads = Thread.objects.filter(
tags__name__in=tags_list
).exclude(
id = self.id
).exclude(
posts__post_type='question',
posts__deleted = True
).distinct()[:100]
similar_threads = list(similar_threads)
for thread in similar_threads:
thread.similarity = self.get_similarity(other_thread=thread)
similar_threads.sort(key=operator.attrgetter('similarity'), reverse=True)
similar_threads = similar_threads[:10]
# Denormalize questions to speed up template rendering
# todo: just denormalize question_post_id on the thread!
thread_map = dict([(thread.id, thread) for thread in similar_threads])
questions = Post.objects.get_questions()
questions = questions.select_related('thread').filter(thread__in=similar_threads)
for q in questions:
thread_map[q.thread_id].question_denorm = q
# Postprocess data for the final output
result = list()
for thread in similar_threads:
question_post = getattr(thread, 'question_denorm', None)
# unfortunately the if statement below is necessary due to
# a possible bug
# all this proves that it's wrong to reference threads by
# the question post id in the question page urls!!!
# this is a "legacy" problem inherited from the old models
if question_post:
url = question_post.get_absolute_url()
title = thread.get_title(question_post)
result.append({'url': url, 'title': title})
return result
def get_cached_data():
"""similar thread data will expire
with the default expiration delay
"""
key = 'similar-threads-%s' % self.id
data = cache.cache.get(key)
if data is None:
data = get_data()
cache.cache.set(key, data)
return data
return LazyList(get_cached_data)