add python 3 support!

#114

.gitignore

@@ -5,7 +5,9 @@
 /.eggs/
 /*.egg-info
 /l
+/l3
 /local/
+/local3/
 circleci_token
 datastore.dat
 oauth_client_secret

README.md

@@ -33,6 +33,7 @@ Granary is a library and REST API that fetches and converts between a wide varie
 Here's how to get started:
 
 * Granary is [available on PyPi.](https://pypi.python.org/pypi/granary/) Install with `pip install granary`.
+* Supports Python 2.7+ and 3.4+.
 * [Click here for getting started docs.](#using)
 * [Click here for reference docs.](https://granary.readthedocs.io/en/latest/source/granary.html)
 * The REST API and demo app are deployed at [granary.io](https://granary.io/).
@@ -223,6 +224,7 @@ On the open source side, there are many related projects. [php-mf2-shim](https:/
 Changelog
 ---
 ### 1.11 - unreleased
+* Add Python 3 support!
 * Twitter:
   * Prefer MP4 and other video/... content types to HLS (.m3u8) etc. [Background.](https://twittercommunity.com/t/retiring-mp4-video-output/66093)
   * Prefer HTTPS URLs for media images.

granary/as2.py

@@ -5,6 +5,9 @@
 AS1: http://activitystrea.ms/specs/json/1.0/
      http://activitystrea.ms/specs/json/schema/activity-schema.html
 """
+from __future__ import unicode_literals
+from past.builtins import basestring
+
 import copy
 import logging
 
@@ -105,7 +108,7 @@ def all_from_as1(field, type=None):
     obj['location'] = from_as1(loc, type='Place', context=None)
 
   obj = util.trim_nulls(obj)
-  if obj.keys() == ['url']:
+  if list(obj.keys()) == ['url']:
     return obj['url']
 
   return obj

granary/atom.py

@@ -2,11 +2,16 @@
 
 Atom spec: https://tools.ietf.org/html/rfc4287 (RIP atomenabled.org)
 """
+from __future__ import absolute_import, unicode_literals
+from future import standard_library
+standard_library.install_aliases()
+from builtins import str
+
 import collections
 import mimetypes
 import os
 import re
-import urlparse
+import urllib.parse
 from xml.etree import ElementTree
 import xml.sax.saxutils
 
@@ -16,8 +21,8 @@
 import mf2util
 from oauth_dropins.webutil import util
 
-import microformats2
-import source
+from . import microformats2
+from . import source
 
 FEED_TEMPLATE = 'user_feed.atom'
 ENTRY_TEMPLATE = 'entry.atom'
@@ -60,9 +65,10 @@ def _text(elem, field=None):
     if ':' not in field:
       field = 'atom:' + field
     elem = elem.find(field, NAMESPACES)
+
   if elem is not None and elem.text:
     text = elem.text
-    if not isinstance(elem.text, unicode):
+    if not isinstance(elem.text, str):
       text = text.decode('utf-8')
     return text.strip()
 
@@ -96,7 +102,9 @@ def __init__(self, **kwargs):
       for k, v in kwargs.items()})
 
   def __unicode__(self):
-    return super(Defaulter, self).__unicode__() if self else u''
+    return super(Defaulter, self).__unicode__() if self else ''
+
+  __str__ = __unicode__
 
   def __hash__(self):
     return super(Defaulter, self).__hash__() if self else None.__hash__()
@@ -177,7 +185,7 @@ def atom_to_activities(atom):
   Returns:
     list of ActivityStreams activity dicts
   """
-  assert isinstance(atom, unicode)
+  assert isinstance(atom, str)
   parser = ElementTree.XMLParser(encoding='UTF-8')
   feed = ElementTree.XML(atom.encode('utf-8'), parser=parser)
   if _tag(feed) != 'feed':
@@ -194,7 +202,7 @@ def atom_to_activity(atom):
   Returns:
     dict, ActivityStreams activity
   """
-  assert isinstance(atom, unicode)
+  assert isinstance(atom, str)
   parser = ElementTree.XMLParser(encoding='UTF-8')
   entry = ElementTree.XML(atom.encode('utf-8'), parser=parser)
   if _tag(entry) != 'entry':
@@ -385,10 +393,10 @@ def _prepare_activity(a, reader=True):
     if not image:
       continue
     url = image.get('url')
-    parsed = urlparse.urlparse(url)
+    parsed = urllib.parse.urlparse(url)
     scheme = parsed.scheme
     netloc = parsed.netloc
-    rest = urlparse.urlunparse(('', '') + parsed[2:])
+    rest = urllib.parse.urlunparse(('', '') + parsed[2:])
     img_src_re = re.compile(r"""src *= *['"] *((https?:)?//%s)?%s *['"]""" %
                             (re.escape(netloc), re.escape(rest)))
     if (url and url not in image_urls_seen and
@@ -406,6 +414,6 @@ def _prepare_activity(a, reader=True):
 
 
 def _remove_query_params(url):
-  parsed = list(urlparse.urlparse(url))
+  parsed = list(urllib.parse.urlparse(url))
   parsed[4] = ''
-  return urlparse.urlunparse(parsed)
+  return urllib.parse.urlunparse(parsed)

granary/facebook.py

@@ -10,8 +10,7 @@
 Retrieving @all activities from :meth:`get_activities()` (the default) currently
 returns an incomplete set of activities, ie *NOT* exactly the same set as your
 Facebook News Feed: https://www.facebook.com/help/327131014036297/
-"""
-"""
+
 This is complicated, and I still don't fully understand how or why they differ,
 but based on lots of experimenting and searching, it sounds like the current
 state is that you just can't reproduce the News Feed via Graph API's /me/home,
@@ -42,20 +41,23 @@
 See the fql_stream_to_post() method below for code I used to experiment with the
 FQL stream table.
 """
+from __future__ import absolute_import, division, unicode_literals
+from future import standard_library
+standard_library.install_aliases()
+from builtins import range, str, zip
+
 import collections
 import copy
 import itertools
 import json
 import logging
 import re
-import urllib
-import urllib2
-import urlparse
+import urllib.error, urllib.parse, urllib.request
 import mf2util
 
-import appengine_config
+from . import appengine_config
 from oauth_dropins.webutil import util
-import source
+from . import source
 
 # Since API v2.4, we need to explicitly ask for the fields we want from most API
 # endpoints with ?fields=...
@@ -176,13 +178,13 @@
 }
 # https://developers.facebook.com/docs/graph-api/reference/post/reactions
 REACTION_CONTENT = {
-  'LOVE': u'❤️',
-  'WOW': u'😮',
-  'HAHA': u'😆',
-  'SAD': u'😢',
-  'ANGRY': u'😡',
-  'THANKFUL': u'🌼',  # https://github.com/snarfed/bridgy/issues/748
-  'PRIDE': u'🏳️‍🌈',
+  'LOVE': '❤️',
+  'WOW': '😮',
+  'HAHA': '😆',
+  'SAD': '😢',
+  'ANGRY': '😡',
+  'THANKFUL': '🌼',  # https://github.com/snarfed/bridgy/issues/748
+  'PRIDE': '🏳️‍🌈',
   # nothing for LIKE (it's a like :P) or for NONE
 }
 
@@ -305,7 +307,7 @@ def get_activities_response(self, user_id=None, group_id=None, app_id=None,
         resp = self.urlopen(url, headers=headers, _as=None)
         etag = resp.info().get('ETag')
         posts = self._as(list, source.load_json(resp.read(), url))
-      except urllib2.HTTPError, e:
+      except urllib.error.HTTPError as e:
         if e.code == 304:  # Not Modified, from a matching ETag
           posts = []
         else:
@@ -428,7 +430,7 @@ def _merge_photos(self, posts):
         photo['privacy'] = 'custom'  # ie unknown
 
     return ([p for p in posts if not p.get('object_id')] +
-            posts_by_obj_id.values() + photos)
+            list(posts_by_obj_id.values()) + photos)
 
   def _split_id_requests(self, api_call, ids):
     """Splits an API call into multiple to stay under the MAX_IDS limit per call.
@@ -508,7 +510,7 @@ def get_comment(self, comment_id, activity_id=None, activity_author_id=None,
     """
     try:
       resp = self.urlopen(API_COMMENT % comment_id)
-    except urllib2.HTTPError, e:
+    except urllib.error.HTTPError as e:
       if e.code == 400 and '_' in comment_id:
         # Facebook may want us to ask for this without the other prefixed id(s)
         resp = self.urlopen(API_COMMENT % comment_id.split('_')[-1])
@@ -676,7 +678,7 @@ def _create(self, obj, preview=None, include_link=source.OMIT_LINK,
         '<a href="%s">%s</a>' % (
           tag.get('url'), tag.get('displayName') or 'User %s' % tag['id'])
         for tag in people)
-    msg_data = {'message': content.encode('utf-8')}
+    msg_data = collections.OrderedDict({'message': content.encode('utf-8')})
     if appengine_config.DEBUG:
       msg_data['privacy'] = json.dumps({'value': 'SELF'})
 
@@ -699,7 +701,7 @@ def _create(self, obj, preview=None, include_link=source.OMIT_LINK,
         if image_url:
           msg_data['attachment_url'] = image_url
         resp = self.urlopen(API_PUBLISH_COMMENT % base_id,
-                            data=urllib.urlencode(msg_data))
+                            data=urllib.parse.urlencode(msg_data))
         url = self.comment_url(base_id, resp['id'],
                                post_author_id=base_obj.get('author', {}).get('id'))
         resp.update({'url': url, 'type': 'comment'})
@@ -799,7 +801,7 @@ def _create(self, obj, preview=None, include_link=source.OMIT_LINK,
             # https://developers.facebook.com/docs/graph-api/reference/user/feed#pubfields
             msg_data['tags'] = ','.join(tag['id'] for tag in people)
 
-        resp = self.urlopen(api_call, data=urllib.urlencode(msg_data))
+        resp = self.urlopen(api_call, data=urllib.parse.urlencode(msg_data))
         resp.update({'url': self.post_url(resp), 'type': 'post'})
         if video_url and not resp.get('success', True):
           msg = 'Video upload failed.'
@@ -845,7 +847,7 @@ def _get_person_tags(self, obj):
         tag['id'] = id
         people[id] = tag
 
-    return people.values()
+    return list(people.values())
 
   def create_notification(self, user_id, text, link):
     """Sends the authenticated user a notification.
@@ -871,7 +873,7 @@ def create_notification(self, user_id, text, link):
                                  appengine_config.FACEBOOK_APP_SECRET),
       }
     url = API_BASE + API_NOTIFICATION % user_id
-    resp = util.urlopen(urllib2.Request(url, data=urllib.urlencode(params)))
+    resp = util.urlopen(urllib.request.Request(url, data=urllib.parse.urlencode(params)))
     logging.debug('Response: %s %s', resp.getcode(), resp.read())
 
   def post_url(self, post):
@@ -933,8 +935,8 @@ def base_object(self, obj, verb=None, resolve_numeric_id=False):
         base_obj = self.user_to_actor(self.urlopen(base_id))
 
     try:
-      parsed = urlparse.urlparse(url)
-      params = urlparse.parse_qs(parsed.query)
+      parsed = urllib.parse.urlparse(url)
+      params = urllib.parse.parse_qs(parsed.query)
       assert parsed.path.startswith('/')
       path = parsed.path.strip('/')
       path_parts = path.split('/')
@@ -984,7 +986,7 @@ def base_object(self, obj, verb=None, resolve_numeric_id=False):
         # add author user id prefix. https://github.com/snarfed/bridgy/issues/229
         base_obj['id'] = '%s_%s' % (author['numeric_id'], base_id)
 
-    except BaseException, e:
+    except BaseException as e:
       logging.error(
         "Couldn't parse object URL %s : %s. Falling back to default logic.",
         url, e)
@@ -1721,7 +1723,7 @@ def urlopen(self, url, _as=dict, **kwargs):
     log_url = url
     if self.access_token:
       url = util.add_query_params(url, [('access_token', self.access_token)])
-    resp = util.urlopen(urllib2.Request(url, **kwargs))
+    resp = util.urlopen(urllib.request.Request(url, **kwargs))
 
     if _as is None:
       return resp
@@ -1781,8 +1783,8 @@ def urlopen_batch(self, urls):
     for url, resp in zip(urls, resps):
       code = int(resp.get('code', 0))
       body = resp.get('body')
-      if code / 100 in (4, 5):
-        raise urllib2.HTTPError(url, code, body, resp.get('headers'), None)
+      if code // 100 in (4, 5):
+        raise urllib.error.HTTPError(url, code, body, resp.get('headers'), None)
       bodies.append(body)
 
     return bodies
@@ -1825,7 +1827,7 @@ def urlopen_batch_full(self, requests):
         req['headers'] = [{'name': n, 'value': v}
                           for n, v in req['headers'].items()]
 
-    data = 'batch=' + json.dumps(util.trim_nulls(requests),
+    data = 'batch=' + json.dumps(util.trim_nulls(requests), sort_keys=True,
                                  separators=(',', ':'))  # no whitespace
     resps = self.urlopen('', data=data, _as=list)