g = ThreadNetwork(CommonAuthor(m))
g(forum)
return g
def TextSimilarityNetwork(forum, n=0.3):
g = ThreadNetwork(TextSimilarity(n, cosine_similarity))
g(forum)
return g
def PostAfterNetwork(forum, dist=1, count=3):
g = DirectedAuthorNetwork(PostAfter(dist=dist, count=count))
g(forum)
return g
def ThreadParticipationNetwork(forum, k=5):
g = AuthorNetwork(ThreadParticipation(k))
g(forum)
return g
########
# Analysis
########
if __name__ == '__main__':
from DataModel import load_xml, rbp_tokenize
#f = load_xml('adcs.xml')
f = load_xml('nabble.xml')
#f.run_tokenizer(rbp_tokenize)
g = ThreadParticipationNetwork(f, k=1)
import pdb;pdb.set_trace()
def wanas_Post_features(post):
# tokenize if not yet done.
if post.thread.token_index is None:
post.thread.forum.run_tokenizer(rbp_tokenize)
features = dict()
features['onThreadTopic'] = onThreadTopic(post)
features['overlapPrevious'] = overlapPrevious(post)
features['overlapDistance'] = overlapDistance(post)
features['timeliness'] = timeliness(post)
features['lengthiness'] = lengthiness(post)
features['formatEmoticons'] = formatEmoticons(post)
features['formatCapitals'] = formatCapitals(post)
features['weblinks'] = weblinks(post)
assert all((v >= 0.0 for v in features.values())), "Ended up with a negative feature!"
return features
# From wanas testing
#f.run_tokenizer(rbp_tokenize)
#for p in f.posts[:100]:
# #print p, len(p), onThreadTopic(p), overlapPrevious(p), overlapDistance(p), timeliness(p), lengthiness(p), formatEmoticons(p)
# print p, len(p), formatEmoticons(p), formatCapitals(p), weblinks(p)
from common import user_post_aggregate
if __name__ == '__main__':
from DataModel import load_xml, rbp_tokenize
f = load_xml('adcs.xml')
#f = load_xml('nabble.xml')
user_post_aggregate(f, wanas_Post_features)
import pdb;pdb.set_trace()