User:Lidia.Pereira/SDRII/POPC
< User:Lidia.Pereira | SDRII
Some further developments on the SLASH News project.
What was changed?
The idea with this project is to make a comment on the promiscuity of power relations by making use of absurdity. By mimicking the design of the BBC news website I both try to make the source more visible and underline the fact that the "new meaning" this generated fictions produce was there all along, just not so transparent. Also, more slash fictions were added to the recipe.
What can still change?
Improve visibility of sources, improve text formatting.
#!/usr/bin/env python
#-*- coding:utf-8 -*-
import re
from nltk.probability import FreqDist, LidstoneProbDist
from nltk.probability import ConditionalFreqDist as CFD
from nltk.util import tokenwrap, LazyConcatenation
from nltk.model import NgramModel
from nltk.metrics import f_measure, BigramAssocMeasures
from nltk.collocations import BigramCollocationFinder
from nltk.compat import python_2_unicode_compatible, text_type
import feedparser, pickle, nltk
import random
import nltk.tokenize
import cgi, time
import urlparse, urllib
url = "http://feeds.bbci.co.uk/news/rss.xml"
searchwords= ["president","minister","cameron","presidential","obama","angela","barroso","pm","chancellor"]
rawFeed = feedparser.parse(url)
# thank you mathijs for my own personalized generate function!
def lidia_generate(dumpty, length=100):
if '_trigram_model' not in dumpty.__dict__:
estimator = lambda fdist, bins: LidstoneProbDist(fdist, 0.2)
dumpty._trigram_model = NgramModel(3, dumpty, estimator=estimator)
text = dumpty._trigram_model.generate(length)
return tokenwrap(text) # or just text if you don't want to wrap
def convert (humpty):
tokens = nltk.word_tokenize(humpty)
return nltk.Text(tokens)
superSumario = " "
data = time.strftime("%d"+" "+"%B"+ " "+ "%Y")
for i in rawFeed.entries:
sumario = i["summary"].lower()
sumario = sumario.replace("\"","'")
for w in searchwords:
if w in sumario.strip():
superSumario = superSumario + sumario
out = open("afiltertest.txt", "w")
out.write(superSumario.encode('utf-8'))
textrss = open('afiltertest.txt').read()
slasharchive = ['harrypotterarchive/harrypotterslash.txt','harrypotterarchive/harrypotterslash2.txt',\
'harrypotterarchive/harrypotterslash3.txt','harrypotterarchive/harrypotterslash4.txt',\
'harrypotterarchive/harrypotterslash5.txt']
chosen = random.choice(slasharchive)
textslash = open(chosen).read()
names = re.compile(r"[a-z]\s([A-Z]\w+)")
pi = names.findall(textslash)
lista = []
for name in pi:
if name not in lista:
lista.append(name)
variavel = ""
for name in lista:
variavel = variavel + name + "|"
ze = "(" + variavel.rstrip ("|") + ")"
nomes = re.compile(ze)
office1 = re.compile(r"(president\s\w+\w+)")
office2 = re.compile(r"(minister\s\w+\w+)")
office3 = re.compile(r"(chancellor\s\w+\w+)")
repls = office1.findall(textrss) + office2.findall(textrss) + office3.findall(textrss)
titloptions = ["panties","negligée","wonderbra","satin boxers","dildo"]
def r(m):
return random.choice(repls)
yes = nomes.sub(r,textslash)
lala = re.sub("\"","",yes)
both = lala.strip() + textrss.strip()
puff = convert(both)
#title = lidia_generate(puff,5)
title = random.choice(repls)+"'s"+" "+random.choice(titloptions)
herpDerp = lidia_generate(puff, 300)
herpDerp = str(herpDerp) + "."
#append to a file
dictionary = {}
dictionary["h1"] = title
dictionary["p"] = herpDerp
f = open("slashtest.txt","a")
f.write(urllib.urlencode(dictionary) + "\n")
f.close()
print "Content-Type: text/html"
print
print """
<!DOCTYPE html>
<html>
<head>
<meta charset='UTF-8'>
<title>President Obama's Panties</title>
<link rel='stylesheet' type='text/css' href='/slashtest.css'>
</head>
<body>
<img src='/slashnewsbanner.png' class = 'banner'/>
<img src='/greybanner.png' class = 'greybanner' />
<p class ='data'>"""+data+"""</b>
<div class='description'><b class='what'>WHAT?</b> Slash News generates slash fiction\
from BBC RSS feeds, bringing into play ideas of promiscuity as deeply connected to power structures.</div>
<div class='slashy'>
"""
f = open("slashtest.txt")
lines = f.readlines()
for line in reversed(lines):
d = urlparse.parse_qs(line.rstrip())
print "<h1>"+d["h1"][0]+"</h1>"
print "<p class = 'slash'>"+d["p"][0]+"</p>"
print """ </div>
</body>
</html>"""