From 56c8287c00eab4c3b6673502d5b964a43d70ccb7 Mon Sep 17 00:00:00 2001
From: Vincent Le Gallic <legallic@crans.org>
Date: Fri, 27 Sep 2013 16:36:25 +0200
Subject: [PATCH] =?utf8?q?Mise=20=C3=A0=20jour=20du=20flux=20RSS=20de=20dt?=
 =?utf8?q?c?=
MIME-Version: 1.0
Content-Type: text/plain; charset=utf8
Content-Transfer-Encoding: 8bit

---
 dtc.py       | 33 ++++-----------------------------
 fetch_dtc.py | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 56 insertions(+), 29 deletions(-)
 create mode 100755 fetch_dtc.py

diff --git a/dtc.py b/dtc.py
index 826b2c0..d9cd465 100755
--- a/dtc.py
+++ b/dtc.py
@@ -2,13 +2,12 @@
 # -*- coding:utf-8 -*-
 import sys
 
-import html2text
-import nltk
-import re
-import feedparser
 import psycopg2
 import psycopg2.extras
 import json
+
+import fetch_dtc
+
 VERBOSE = False
 
 def getcursor():
@@ -33,29 +32,6 @@ def last_inserted():
     cur.execute("SELECT MAX(id) AS maxid FROM quotes;")
     return cur.fetchone()["maxid"]
 
-def fetch_rss():
-    """Récupère le flux RSS et le formate"""
-    flux = feedparser.parse('http://feeds.feedburner.com/bashfr-quotes')
-    quotes = []
-    for q in flux["entries"]:
-        try:
-            id = int(q["title"])
-        except ValueError:
-            print "FAILED : " + q["title"]
-            continue
-        date = q["published"]
-        quote = format(q["summary_detail"]["value"])
-        quotes.append({"id" : id, "date" : date, "quote" : quote})
-    return quotes
-
-
-def format(quote):
-    """Dé-HTML-ise la quote"""
-    raw = html2text.unescape(nltk.clean_html(quote))
-    # Fucking garbage
-    raw = re.sub(" Votez !$", "", raw)
-    return raw
-
 def get_quotes(first, last=None):
     """Récupère des quotes dans la base."""
     add = ""
@@ -79,8 +55,7 @@ def display(liste):
 if __name__ == "__main__":
     import sys
     if sys.argv[1] in ["fetch", "update", "pull"]:
-        l = fetch_rss()
-        print l
+        l = fetch_dtc.getquotes()
         last = last_inserted()
         for q in l:
             if q["id"] > last:
diff --git a/fetch_dtc.py b/fetch_dtc.py
new file mode 100755
index 0000000..de84884
--- /dev/null
+++ b/fetch_dtc.py
@@ -0,0 +1,52 @@
+#!/usr/bin/env python
+# -*- encoding: utf-8 -*-
+
+"""Pour récupérer les quotes DTC à partir du flux RSS"""
+
+import BeautifulSoup
+import feedparser
+import re
+
+import html2text
+
+#: Regexp permettant de matcher l'id d'une quote et de virer le garbage qui traîne en bas
+endquote_regexp = ur"\n\[#(\d+)\]\(http://danstonchat.com/\1\) - \[Voir les.commentaires\]\(http://danstonchat.com/\1#c\)"
+endreg = re.compile(endquote_regexp, flags=re.DOTALL)
+
+def properquote(quote):
+    """Prend la quote tirée RSS et la renvoie dans un format lisible."""
+    quote = str(quote).decode("utf-8")
+    quote = html2text.html2text(quote)
+    # Ce truc génère bêtement des lignes vides
+    quote = quote.replace(u"\n\n", u"\n")
+    return quote
+
+def parsequote(quote):
+    """Parse la quote. Renvoie ``(<id>, <quote nettoyée>)`` ou ``None`` si il n'y a pas d'id.
+       (C'est-à-dire, si ce n'est en fait pas une quote)
+       Renvoie aussi ``None`` sur un commentaire ``"_(A propos de)_"``"""
+    ids = endreg.findall(quote)
+    if quote.startswith(u"_(A propos de)_"):
+        return None
+    if ids:
+        return {"id" : int(ids[0]), "quote" : endreg.sub("", quote)}
+    else:
+        return None
+
+def getquotes():
+    """Récupère toutes les quotes, avec date de publication et id."""
+    flux = feedparser.parse("http://danstonchat.com/rss.xml")
+    allquotes = []
+    for entry in flux["entries"]:
+        content = entry["summary"]
+        soup = BeautifulSoup.BeautifulSoup(content)
+        quotes = soup.findChildren(name="p")
+        quotes = [properquote(q) for q in quotes]
+        # Attention, en vrai il n'y a pas que des quotes là-dedans,
+        # mais aussi des top comments et autre random shit
+        # donc on vire les None
+        quotes = [parsequote(q) for q in quotes]
+        quotes = [q for q in quotes if not q is None]
+        [q.update({"date": entry["published"]}) for q in quotes]
+        allquotes.extend(quotes)
+    return allquotes
-- 
2.39.2