wip
This commit is contained in:
76
sentiments.py
Normal file
76
sentiments.py
Normal file
@@ -0,0 +1,76 @@
|
||||
import os
|
||||
import sys
|
||||
from collections import defaultdict
|
||||
|
||||
from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer
|
||||
|
||||
from loader import load, cms, dmt
|
||||
|
||||
printnoln = lambda text: print(text, end='', flush=True)
|
||||
rprint = lambda text: print('\r' + text)
|
||||
|
||||
DAYS_NEW_USER = 7
|
||||
OLD_USER_YEAR = 3
|
||||
|
||||
analyser = SentimentIntensityAnalyzer()
|
||||
colors = ['red', 'green', 'blue', 'orange', 'deeppink']
|
||||
|
||||
|
||||
def main(folder):
|
||||
users, posts, firstcontrib, sumcontrib = load(folder)
|
||||
|
||||
outfolder = folder + "/output/"
|
||||
os.system("mkdir -p " + outfolder)
|
||||
outfilename = outfolder + "sentiments"
|
||||
|
||||
# computer toxic levels
|
||||
# start = cms()
|
||||
# printnoln("computing toxic levels: filtering")
|
||||
|
||||
# toxlevels = defaultdict(list)
|
||||
# for (i, post) in enumerate(posts):
|
||||
# if (i + 1) % 100 == 0:
|
||||
# printnoln("\rcomputing toxic levels: post #" + str(i + 1) + "/" + str(len(posts)))
|
||||
# if (i + 1) == len(posts):
|
||||
# printnoln("\rcomputing toxic levels: post #" + str(i + 1) + "/" + str(len(posts)))
|
||||
# for a in post['Answers']:
|
||||
# toxlevel = computeToxLevel(a['Body'])
|
||||
# toxlevels[post['Id']].append(toxlevel)
|
||||
# rprint("computing toxic levels: post #" + str(len(posts)) + "/" + str(len(posts)) + " ... took " + str(cms() - start) + "ms")
|
||||
toxlevels = dmt(posts).map(lambda p: (p['Id'], computeSentimentForPost(p)), "calculating sentiments").getresults()
|
||||
toxlevels = {id: p for (id, p) in toxlevels}
|
||||
|
||||
dumptoxlevels(toxlevels, outfilename + ".py")
|
||||
|
||||
|
||||
def computeSentimentForPost(post):
|
||||
anwsers = {a['Id']: computeToxLevel(a['Body']) for a in post['Answers']}
|
||||
return anwsers
|
||||
|
||||
|
||||
def computeToxLevel(text):
|
||||
return analyser.polarity_scores(text)
|
||||
|
||||
|
||||
def dumptoxlevels(lvls, filename):
|
||||
answers = dict()
|
||||
for p in lvls.values():
|
||||
for id, a in p.items():
|
||||
answers[id] = a
|
||||
with open(filename, "w") as file:
|
||||
file.write("posts = " + str(lvls) + "\n")
|
||||
file.write("answers = " + str(answers) + "\n")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
# execute only if run as a script
|
||||
usage = sys.argv[0] + " <folder>"
|
||||
if len(sys.argv) < 2:
|
||||
print(usage)
|
||||
sys.exit(1)
|
||||
folder = sys.argv[1]
|
||||
if not os.path.isdir(folder):
|
||||
print(folder + " is not a folder")
|
||||
sys.exit(1)
|
||||
|
||||
main(folder)
|
||||
Reference in New Issue
Block a user