Switching to requests from urllib2/httplib. Bring on the page errors.

apps/rss_feeds/management/commands/refresh_feeds.py

@@ -5,6 +5,7 @@ from apps.rss_feeds.models import Feed
 from optparse import make_option
 from utils import feed_fetcher
 from utils.management_functions import daemonize
+import django
 import socket
 import datetime
 import redis
@@ -75,12 +76,14 @@ class Command(BaseCommand):
         feeds_queue = []
         for _ in range(num_workers):
             feeds_queue.append([])
-            
+        
         i = 0
         for feed in feeds:
             feeds_queue[i%num_workers].append(feed.pk)
             i += 1
         disp.add_jobs(feeds_queue, i)
         
+        django.db.connection.close()
+        
         print " ---> Fetching %s feeds..." % feeds.count()
         disp.run_jobs()

apps/rss_feeds/page_importer.py

@@ -1,18 +1,14 @@
-import urllib2, httplib
+import requests
 import re
 import urlparse
 import traceback
 import feedparser
 import time
+from django.conf import settings
 from utils import log as logging
 from apps.rss_feeds.models import MFeedPage
 from utils.feed_functions import timelimit, mail_feed_error_to_admin
 
-HEADERS = {
-    'User-Agent': 'NewsBlur Page Fetcher - http://www.newsblur.com',
-    'Connection': 'close',
-}
-
 BROKEN_PAGES = [
     'tag:', 
     'info:', 
@@ -26,6 +22,18 @@ class PageImporter(object):
     def __init__(self, url, feed):
         self.url = url
         self.feed = feed
+        self.setup_headers()
+        
+    def setup_headers(self):
+        s = requests.session()
+        s.config['keep_alive'] = False
+        self.headers = {
+            'User-Agent': 'NewsBlur Page Fetcher (%s subscriber%s) - %s (Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_1) AppleWebKit/534.48.3 (KHTML, like Gecko) Version/5.1 Safari/534.48.3)' % (
+                self.feed.num_subscribers,
+                's' if self.feed.num_subscribers != 1 else '',
+                settings.NEWSBLUR_URL
+            ),
+        }
     
     @timelimit(15)
     def fetch_page(self):
@@ -35,10 +43,9 @@ class PageImporter(object):
         
         try:
             if self.url.startswith('http'):
-                request = urllib2.Request(self.url, headers=HEADERS)
-                response = urllib2.urlopen(request)
+                response = requests.get(self.url, headers=self.headers)
                 time.sleep(0.01) # Grrr, GIL.
-                data = response.read()
+                data = response.content
             elif any(self.url.startswith(s) for s in BROKEN_PAGES):
                 self.save_no_page()
                 return
@@ -46,17 +53,17 @@ class PageImporter(object):
                 data = open(self.url, 'r').read()
             html = self.rewrite_page(data)
             self.save_page(html)
-        except (ValueError, urllib2.URLError, httplib.BadStatusLine, httplib.InvalidURL), e:
-            self.feed.save_page_history(401, "Bad URL", e)
-            fp = feedparser.parse(self.feed.feed_address)
-            self.feed.feed_link = fp.feed.get('link', "")
-            self.feed.save()
-        except (urllib2.HTTPError), e:
-            self.feed.save_page_history(e.code, e.msg, e.fp.read())
-            return
-        except (httplib.IncompleteRead), e:
-            self.feed.save_page_history(500, "IncompleteRead", e)
-            return
+        # except (ValueError, urllib2.URLError, httplib.BadStatusLine, httplib.InvalidURL), e:
+        #     self.feed.save_page_history(401, "Bad URL", e)
+        #     fp = feedparser.parse(self.feed.feed_address)
+        #     self.feed.feed_link = fp.feed.get('link', "")
+        #     self.feed.save()
+        # except (urllib2.HTTPError), e:
+        #     self.feed.save_page_history(e.code, e.msg, e.fp.read())
+        #     return
+        # except (httplib.IncompleteRead), e:
+        #     self.feed.save_page_history(500, "IncompleteRead", e)
+        #     return
         except Exception, e:
             logging.debug('[%d] ! -------------------------' % (self.feed.id,))
             tb = traceback.format_exc()

local_settings.py.template

@@ -1,5 +1,46 @@
 import logging
 
+# ===================
+# = Server Settings =
+# ===================
+
+ADMINS                = (
+    ('Samuel Clay', 'samuel@ofbrooklyn.com'),
+)
+
+SERVER_EMAIL          = 'server@newsblur.com'
+HELLO_EMAIL           = 'hello@newsblur.com'
+NEWSBLUR_URL          = 'http://www.newsblur.com'
+
+# ==================
+# = Global Settngs =
+# ==================
+
+DEBUG = True
+MEDIA_URL = '/media/'
+SECRET_KEY = 'YOUR SECRET KEY'
+
+CACHE_BACKEND = 'dummy:///'
+# CACHE_BACKEND = 'locmem:///'
+# CACHE_BACKEND = 'memcached://127.0.0.1:11211'
+
+EMAIL_BACKEND = 'django.core.mail.backends.console.EmailBackend'
+
+# Set this to the username that is shown on the homepage to unauthenticated users.
+HOMEPAGE_USERNAME = 'conesus'
+
+# Google Reader OAuth API Keys
+OAUTH_KEY = 'www.example.com'
+OAUTH_SECRET = 'SECRET_KEY_FROM_GOOGLE'
+
+S3_ACCESS_KEY = 'XXX'
+S3_SECRET = 'SECRET'
+S3_BACKUP_BUCKET = 'newsblur_backups'
+
+# =============
+# = Databases =
+# =============
+
 DATABASES = {
     'default': {
         'NAME': 'newsblur',
@@ -20,28 +61,13 @@ MONGODB_SLAVE = {
     'host': '127.0.0.1'
 }
 
-DEBUG = True
-
-MEDIA_URL = '/media/'
-
-SECRET_KEY = 'YOUR SECRET KEY'
-
-CACHE_BACKEND = 'dummy:///'
-# CACHE_BACKEND = 'locmem:///'
-# CACHE_BACKEND = 'memcached://127.0.0.1:11211'
-
-EMAIL_BACKEND = 'django.core.mail.backends.console.EmailBackend'
-
-# Set this to the username that is shown on the homepage to unauthenticated users.
-HOMEPAGE_USERNAME = 'conesus'
-
-# Google Reader OAuth API Keys
-OAUTH_KEY = 'www.example.com'
-OAUTH_SECRET = 'SECRET_KEY_FROM_GOOGLE'
-
 # Celery RabbitMQ Broker
 BROKER_HOST = "127.0.0.1"
 
+# ===========
+# = Logging =
+# ===========
+
 # Logging (setup for development)
 LOG_TO_STREAM = True
 
@@ -51,7 +77,3 @@ if len(logging._handlerList) < 1:
                             format='%(asctime)-12s: %(message)s',
                             datefmt='%b %d %H:%M:%S',
                             handler=logging.StreamHandler)
-
-S3_ACCESS_KEY = 'XXX'
-S3_SECRET = 'SECRET'
-S3_BACKUP_BUCKET = 'newsblur_backups'

settings.py

@@ -4,6 +4,18 @@ import os
 from mongoengine import connect
 import redis
 
+# ===================
+# = Server Settings =
+# ===================
+
+ADMINS       = (
+    ('Samuel Clay', 'samuel@ofbrooklyn.com'),
+)
+
+SERVER_EMAIL = 'server@newsblur.com'
+HELLO_EMAIL  = 'hello@newsblur.com'
+NEWSBLUR_URL = 'http://www.newsblur.com'
+
 # ===========================
 # = Directory Declaractions =
 # ===========================
@@ -25,14 +37,10 @@ if '/utils' not in ' '.join(sys.path):
     sys.path.append(UTILS_ROOT)
 if '/vendor' not in ' '.join(sys.path):
     sys.path.append(VENDOR_ROOT)
-
 # ===================
 # = Global Settings =
 # ===================
 
-ADMINS                = (
-    ('Samuel Clay', 'samuel@ofbrooklyn.com'),
-)
 TEST_DEBUG            = False
 SEND_BROKEN_LINK_EMAILS = False
 MANAGERS              = ADMINS
@@ -305,8 +313,6 @@ SESSION_ENGINE          = "django.contrib.sessions.backends.db"
 TEST_RUNNER             = "utils.testrunner.TestRunner"
 SESSION_COOKIE_NAME     = 'newsblur_sessionid'
 SESSION_COOKIE_AGE      = 60*60*24*365*2 # 2 years
-SERVER_EMAIL            = 'server@newsblur.com'
-HELLO_EMAIL             = 'hello@newsblur.com'
 
 # ===========
 # = Logging =

utils/feed_fetcher.py

@@ -23,7 +23,6 @@ import redis
 # Refresh feed code adapted from Feedjack.
 # http://feedjack.googlecode.com
 
-URL = 'http://www.newsblur.com/'
 SLOWFEED_WARNING = 10
 ENTRY_NEW, ENTRY_UPDATED, ENTRY_SAME, ENTRY_ERR = range(4)
 FEED_OK, FEED_SAME, FEED_ERRPARSE, FEED_ERRHTTP, FEED_ERREXC = range(5)
@@ -63,7 +62,7 @@ class FetchFeed:
         USER_AGENT = 'NewsBlur Feed Fetcher (%s subscriber%s) - %s (Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_1) AppleWebKit/534.48.3 (KHTML, like Gecko) Version/5.1 Safari/534.48.3)' % (
             self.feed.num_subscribers,
             's' if self.feed.num_subscribers != 1 else '',
-            URL
+            settings.NEWSBLUR_URL
         )
 
         self.fpf = feedparser.parse(self.feed.feed_address,