jul/lemonde.py

## lemonde.py
#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""Small utility only useful for people able to understand the web front and back
that are bored of actually bloated web pages and only want to see what they
think is useful.
It is not intended as an adblock except if as I, you hate advertisement and
think ads are noise so I tend to not fetch them.
Not intended as an antivirus, but I also tend to not load virus because
there is nothing interesting in them.

"""

from pyquery import PyQuery as P
import requests as req
from functools import wraps
from flask import Flask
from flask.ext.cache import Cache
from random import shuffle, randint
from urlparse import urljoin
from lxml import etree
import re
import json
import sys

tostr = etree.tostring

flask_setting = {}


app = Flask(__name__)
cache = Cache(config={'CACHE_TYPE': 'simple'})

header = u"""
<!doctype html>
  <html lang="fr">
    <head>
      <meta http-equiv="Content-Type" content="text/html; charset=utf8">
      <link href="data:image/x-icon;base64,AAABAAEAEBAAAAAAAABoBAAAFgAAACgAAAAQAAAAIAAAAAEAIAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAA////AAAAAJ4YGAClFxcApRcXAKUWFgCkFRUApBQUAKQTEwCjEhIAoxESAKMREgCjAAAAnv///wAAAAAAAAAAAAAAAJ6zHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf8BAQCfAAAAAAAAAAAREQCjsx0Ju7MdCbuzHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQm7CwsAoQAAAAAAAAAAAAAARLMdCbuzHQm7sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQm7sx0J/wAAAD4AAAAAAAAAAP///wAkJACosx0JlbMdCbuzHQn/sx0J/7MdCf+zHQn/sx0J/7MdCf+zHQn/sx0J/wAAAJ7///8AAAAAAAAAAAD///8AnZ2dCbMdCZWzHQm7sx0Ju7MdCf+zHQn/sx0J/7MdCf+zHQn/sx0JuzAvAqz///8A////AAAAAAAAAAAA////AP///wAAAABWsx0JlbMdCbuzHQm7sx0J/7MdCf+zHQn/sx0Ju7MdCf8AAAA/////AP///wAAAAAAAAAAAP///wD///8A////ACAgAKf8/6Bv/P+TXvz/AC/8/wAs/P8AKvz/llACAgCf////AP///wD///8AAAAAAAAAAAD///8A////AP///wCoqKgJaWsAiP3/oWj8/wAw/P8ALfz/i008PQCn/v7+AP///wD///8A////AAAAAAAAAAAA////AP///wD///8A////AAAAAFbW2ABB+/9rTfz/AC6PkQBRAAAAQP///wD///8A////AP///wAAAAAAAAAAAP///wD///8A////AP///wDt7e0EgYMAZ/v/ak9OTwCWP0AAr////wD///8A////AP///wD///8AAAAAAAAAAAD///8A////AP///wD///8AjY2NBoKEAGf9/2BP/P8AL2xtAGn///8A////AP///wD///8A////AAAAAAAAAAAA////AP///wD///8A////AI2NjQqEhQBn/P9JTfz/ADBpaQBz////AP///wD///8A////AP///wAAAAAAAAAAAP///wD///8A////AP///wCNjY0LhogAZ/z/REv8/wAxYGEAfv///wD///8A////AP///wD///8AAAAAAAAAAAD///8A////AP///wD///8AAAAAPMHEAE37/wA1/P8AMXd4AGLj4+MB////AP///wD///8A////AAAAAAAAAAAA////AP///wD///8A////AB8fAKcdHgCmHBwAphsbAKYaGgClBAUAoP///wD///8A////AP///wAAAAAAwAMAAIABAACAAQAAwAMAAMADAADgBwAA8A8AAPfvAAD73wAA//8AAP8/AAD//wAA//8AAP//AAD//wAA+B8AAA==" rel="icon" type="image/x-icon" />
      <style>
        @import url(http://fonts.googleapis.com/css?family=Crimson+Text);
          body {
            font-family:'Crimson Text', 'Georgia', serif !important;
            font-size: 20px;
            color: #000; }
          a {
            margin-left:.3em; color:#003;
            text-decoration:none;
            border-bottom: 1px dashed black;
            border-radius:0
          }
          a:visited {  color:#123;}
          tr:hover, li:hover,  .entete:hover {
            z-index:4; max-width:100%;transition: all .3s ease-in;
            border-radius:.6em;
            text-shadow: rgba(0,0,0,.6) .1em .1em 2px;
          }
          h3 { font-size:100%; font-weight:normal }
          .numero { visibility:hidden }
          li {  list-style: none;padding-top:.1em; padding-bottom:.1em;  }
          tr:nth-child(odd) > td, li:nth-child(odd) {
            background: #DDD;  border-radius:.6em;
          }
          li:nth-child(even), tr:nth-child(even)> td {
            background: white;  border-radius:.6em;
          }
          .entete:first-child {
            margin-top:1.59em;
          }
          .entete {
            text-shadow: rgba(0,0,0,.6) .08em .08em 1px;
            margin-left:.5em; display:block; margin-top:.1em;
            margin-bottom:3px; border-radius: 15px;background:#46A;color:white;
            padding:.1em;font-size: 1.3em;text-align:center ; border-bottom:0px; border-bottom:0px;
          }
          .footer { z-index:4;margin: auto;margin-left:-.5em;
          background:#46A;color:white;padding:.2em;font-size: .8em;
          text-align:center; position:fixed; top:0px;width:100%; }
        </style>
        <title>127.0.0.1 suss Heim</title>
        <script src='http://code.jquery.com/jquery-latest.min.js' type='text/javascript' ></script>
        <script>
          $( document ).ready( function() {
              $(".entete").next().hide();
              $($(".entete")[0]).css("background-color","#aad")
              $($(".entete")[1]).css("background-color","#bad")
              $($(".entete")[2]).css("background-color","#cad")
              $($(".entete")[3]).css("background-color","#dad")
              $($(".entete")[4]).css("background-color","#ead")
              $(function() {
                  $(".entete").click( function() {
                      $(this).next().toggle();
                  } ) }
              );
              $('[name=webcam]').next().show()

    } )
  </script>
</head>
<body>
  <div class=footer>What The Fuck Public Licence v3.14159 Julien Tayon //<a
        href=http://gist.github.com/jul/6652373>source</a> //<a
        name=top onclick='$(window).scrollTop(0)'>haut
        </a>
  </div>
  <div style="width:90%;margin-top:-.4em;margin-left:auto; margin-right:auto">
"""

footer = u"""</div></body></html>"""

magnet_img = u"""<img alt="" src="data:image/gif;base64,R0lGODlhDAAMALMPAOXl5ewvErW1tebm5oocDkVFRePj47a2ts0WAOTk5MwVAIkcDesuEs0VAEZGRv///yH5BAEAAA8ALAAAAAAMAAwAAARB8MnnqpuzroZYzQvSNMroUeFIjornbK1mVkRzUgQSyPfbFi/dBRdzCAyJoTFhcBQOiYHyAABUDsiCxAFNWj6UbwQAOw==" /> """

def fullpage(*msg):
    return "\n".join([ header ] + list(msg) + [ footer ])

@wraps
def get_or_fail(funct):
    def wraped(*a, **kw):
        try:
            return funct(*a,**kw)
        except Exception as e:
            return str(e)
    return wraped

#@cache.cached(timeout=1000, key_prefix='lemonde')
@get_or_fail
def lemonde_fragment():
    toprint = P(
        'http://www.lemonde.fr')('.plus_partages'
        ).make_links_absolute(
        ).attr('target', '_blank').html()
    toprint = P(toprint).remove(".sociaux_suivre").html()
    return toprint


@get_or_fail
def anime():
    title = u"""<a class="entete" name=anime>animation</a>"""
    that = 'navyvaxm.gi'.decode("rot13")
    site=('http://%s/' % that)
    base_url = urljoin(site, '/fetch.php')
    html = req.post(base_url, headers = {
        'Cache-Control' :   'no-cache',
        'Pragma': 'no-cache',
        'Host': that,
        'Accept': '*/*',
        'Accept-Encoding': 'gzip, deflate',
        'Connection': 'keep-alive',
        'User-Agent' :    'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:26.0) Gecko/20100101 Firefox/26.0',
        'X-Requested-With' : 'XMLHttpRequest','Referer' : 'http://%s/?p=1' % that,
        'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8'},
        data=dict(action='ongoing',page=1,type=0)
    )
    doc = P(html.content)
    for a in doc.find("a"):
        a=P(a)
        rel = a.attr("href")
        a.attr("href", urljoin(site,rel)).attr('target', '_blank')

    return title + "<div><ul>" + doc.html()  + "</ul></div>"

@get_or_fail
def toride(plus="tv"):
    title = u"""<a class="entete" name=anime>A la tele ce soir (%s)</a>""" % plus
    base_url = urljoin('uggcf://xng.pe/'.decode('rot13'), plus)
    script= """<script>
    $("[href^=magnet]").html('%s');
    $('.torrentname > a, .font11px').hide();

    </script>""" % magnet_img
    return title + "<div><ul>" + tostr(P(base_url, verify=False
                                         ).make_links_absolute(
                                         ).attr('target', '_blank'
                                         )("table.data")[0]
                                 )  + "</ul></div>" + script

@get_or_fail
def npr_fragment(channel=1):
    li_list = []
    base_url ="http://www.npr.org/rss/rss.php?id=%d"
    title = "NPR"
    doc = str(P(base_url % channel))
    ### WTF I cannot use lxml to parse fucking RSS O_o
    print doc
    title += P(doc("channel"))("title").html()
    print title
    for item in doc.find('item'):
        el = P(item)
        li_list += [ "<a href=%s >%s</a>" % (P("link").html(),
            P("description").html())  ]
    return title + "<div><ul><li>" + "</li><li>".join(li_list) + "</li></ul></div>"

@get_or_fail
def webcam_fragment():
    MAX_PAGE=100
    MAX_LINE_OFF=2
    title = u"""<a class="entete" name=webcam>Random webcam</a>"""
    base_url = 'uggc://jjj.bcragbcvn.pbz/uvqqrapnz.cuc?fubjzbqr=fgnaqneq&pbhagel=*&frrjung=enaqbz&c=%q'.decode("rot13") % randint(1,MAX_PAGE)
    li_list =  P(P(P(base_url).make_links_absolute()('.camgrid3')[randint(0,MAX_LINE_OFF)])("li"))
    wrapper_o = "<div style='width:100%;text-align:center;margin:auto'>"
    item_o = "<div style='height:100%;width:33%;float:left;margin:auto;display:inline'>"
    cement = "</div>" + item_o
    return title  + wrapper_o + item_o + cement.join(
        [ P(el).css("display", "inline").html() for el in li_list.css("display", "inline").items() ]
        ) + "</div></div></div>"

@cache.cached(timeout=50, key_prefix='cam')
@app.route("/")
def proxy():
    return fullpage(lemonde_fragment(), anime(), toride("tv"), toride("movies"), webcam_fragment())


if __name__ == "__main__":
    app.run(**flask_setting)
	#!/usr/bin/env python
	# -- coding: utf-8 --
	"""Small utility only useful for people able to understand the web front and back
	that are bored of actually bloated web pages and only want to see what they
	think is useful.
	It is not intended as an adblock except if as I, you hate advertisement and
	think ads are noise so I tend to not fetch them.
	Not intended as an antivirus, but I also tend to not load virus because
	there is nothing interesting in them.

	"""

	from pyquery import PyQuery as P
	import requests as req
	from functools import wraps
	from flask import Flask
	from flask.ext.cache import Cache
	from random import shuffle, randint
	from urlparse import urljoin
	from lxml import etree
	import re
	import json
	import sys

	tostr = etree.tostring

	flask_setting = {}



	app = Flask(__name__)
	cache = Cache(config={'CACHE_TYPE': 'simple'})

	header = u"""
	<!doctype html>
	<html lang="fr">
	<head>
	<meta http-equiv="Content-Type" content="text/html; charset=utf8">
	<link href="data:image/x-icon;base64,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" rel="icon" type="image/x-icon" />
	<style>
	@import url(http://fonts.googleapis.com/css?family=Crimson+Text);
	body {
	font-family:'Crimson Text', 'Georgia', serif !important;
	font-size: 20px;
	color: #000; }
	a {
	margin-left:.3em; color:#003;
	text-decoration:none;
	border-bottom: 1px dashed black;
	border-radius:0
	}
	a:visited { color:#123;}
	tr:hover, li:hover, .entete:hover {
	z-index:4; max-width:100%;transition: all .3s ease-in;
	border-radius:.6em;
	text-shadow: rgba(0,0,0,.6) .1em .1em 2px;
	}
	h3 { font-size:100%; font-weight:normal }
	.numero { visibility:hidden }
	li { list-style: none;padding-top:.1em; padding-bottom:.1em; }
	tr:nth-child(odd) > td, li:nth-child(odd) {
	background: #DDD; border-radius:.6em;
	}
	li:nth-child(even), tr:nth-child(even)> td {
	background: white; border-radius:.6em;
	}
	.entete:first-child {
	margin-top:1.59em;
	}
	.entete {
	text-shadow: rgba(0,0,0,.6) .08em .08em 1px;
	margin-left:.5em; display:block; margin-top:.1em;
	margin-bottom:3px; border-radius: 15px;background:#46A;color:white;
	padding:.1em;font-size: 1.3em;text-align:center ; border-bottom:0px; border-bottom:0px;
	}
	.footer { z-index:4;margin: auto;margin-left:-.5em;
	background:#46A;color:white;padding:.2em;font-size: .8em;
	text-align:center; position:fixed; top:0px;width:100%; }
	</style>
	<title>127.0.0.1 suss Heim</title>
	<script src='http://code.jquery.com/jquery-latest.min.js' type='text/javascript' ></script>
	<script>
	$( document ).ready( function() {
	$(".entete").next().hide();
	$($(".entete")[0]).css("background-color","#aad")
	$($(".entete")[1]).css("background-color","#bad")
	$($(".entete")[2]).css("background-color","#cad")
	$($(".entete")[3]).css("background-color","#dad")
	$($(".entete")[4]).css("background-color","#ead")
	$(function() {
	$(".entete").click( function() {
	$(this).next().toggle();
	} ) }
	);
	$('[name=webcam]').next().show()

	} )
	</script>
	</head>
	<body>
	<div class=footer>What The Fuck Public Licence v3.14159 Julien Tayon //<a
	href=http://gist.github.com/jul/6652373>source</a> //<a
	name=top onclick='$(window).scrollTop(0)'>haut
	</a>
	</div>
	<div style="width:90%;margin-top:-.4em;margin-left:auto; margin-right:auto">
	"""

	footer = u"""</div></body></html>"""

	magnet_img = u"""<img alt="" src="data:image/gif;base64,R0lGODlhDAAMALMPAOXl5ewvErW1tebm5oocDkVFRePj47a2ts0WAOTk5MwVAIkcDesuEs0VAEZGRv///yH5BAEAAA8ALAAAAAAMAAwAAARB8MnnqpuzroZYzQvSNMroUeFIjornbK1mVkRzUgQSyPfbFi/dBRdzCAyJoTFhcBQOiYHyAABUDsiCxAFNWj6UbwQAOw==" /> """

	def fullpage(*msg):
	return "\n".join([ header ] + list(msg) + [ footer ])

	@wraps
	def get_or_fail(funct):
	def wraped(a, *kw):
	try:
	return funct(a,*kw)
	except Exception as e:
	return str(e)
	return wraped

	#@cache.cached(timeout=1000, key_prefix='lemonde')
	@get_or_fail
	def lemonde_fragment():
	toprint = P(
	'http://www.lemonde.fr')('.plus_partages'
	).make_links_absolute(
	).attr('target', '_blank').html()
	toprint = P(toprint).remove(".sociaux_suivre").html()
	return toprint


	@get_or_fail
	def anime():
	title = u"""<a class="entete" name=anime>animation</a>"""
	that = 'navyvaxm.gi'.decode("rot13")
	site=('http://%s/' % that)
	base_url = urljoin(site, '/fetch.php')
	html = req.post(base_url, headers = {
	'Cache-Control' : 'no-cache',
	'Pragma': 'no-cache',
	'Host': that,
	'Accept': '/',
	'Accept-Encoding': 'gzip, deflate',
	'Connection': 'keep-alive',
	'User-Agent' : 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:26.0) Gecko/20100101 Firefox/26.0',
	'X-Requested-With' : 'XMLHttpRequest','Referer' : 'http://%s/?p=1' % that,
	'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8'},
	data=dict(action='ongoing',page=1,type=0)
	)
	doc = P(html.content)
	for a in doc.find("a"):
	a=P(a)
	rel = a.attr("href")
	a.attr("href", urljoin(site,rel)).attr('target', '_blank')

	return title + "<div><ul>" + doc.html() + "</ul></div>"

	@get_or_fail
	def toride(plus="tv"):
	title = u"""<a class="entete" name=anime>A la tele ce soir (%s)</a>""" % plus
	base_url = urljoin('uggcf://xng.pe/'.decode('rot13'), plus)
	script= """<script>
	$("[href^=magnet]").html('%s');
	$('.torrentname > a, .font11px').hide();

	</script>""" % magnet_img
	return title + "<div><ul>" + tostr(P(base_url, verify=False
	).make_links_absolute(
	).attr('target', '_blank'
	)("table.data")[0]
	) + "</ul></div>" + script

	@get_or_fail
	def npr_fragment(channel=1):
	li_list = []
	base_url ="http://www.npr.org/rss/rss.php?id=%d"
	title = "NPR"
	doc = str(P(base_url % channel))
	### WTF I cannot use lxml to parse fucking RSS O_o
	print doc
	title += P(doc("channel"))("title").html()
	print title
	for item in doc.find('item'):
	el = P(item)
	li_list += [ "<a href=%s >%s</a>" % (P("link").html(),
	P("description").html()) ]
	return title + "<div><ul><li>" + "</li><li>".join(li_list) + "</li></ul></div>"

	@get_or_fail
	def webcam_fragment():
	MAX_PAGE=100
	MAX_LINE_OFF=2
	title = u"""<a class="entete" name=webcam>Random webcam</a>"""
	base_url = 'uggc://jjj.bcragbcvn.pbz/uvqqrapnz.cuc?fubjzbqr=fgnaqneq&pbhagel=*&frrjung=enaqbz&c=%q'.decode("rot13") % randint(1,MAX_PAGE)
	li_list = P(P(P(base_url).make_links_absolute()('.camgrid3')[randint(0,MAX_LINE_OFF)])("li"))
	wrapper_o = "<div style='width:100%;text-align:center;margin:auto'>"
	item_o = "<div style='height:100%;width:33%;float:left;margin:auto;display:inline'>"
	cement = "</div>" + item_o
	return title + wrapper_o + item_o + cement.join(
	[ P(el).css("display", "inline").html() for el in li_list.css("display", "inline").items() ]
	) + "</div></div></div>"

	@cache.cached(timeout=50, key_prefix='cam')
	@app.route("/")
	def proxy():
	return fullpage(lemonde_fragment(), anime(), toride("tv"), toride("movies"), webcam_fragment())


	if __name__ == "__main__":
	app.run(**flask_setting)