Second Beta Release for Version 4.9

This commit is made up of 21365ec668 555c86633c d864ce08ff

Discussed in #1305, #1316, #1372.

Co-Authored-By: Misael <misael.om@gmail.com>
Co-Authored-By: Eduardo Kalinowski <eduardo@kalinowski.com.br>

instaloader/__init__.py

@@ -1,7 +1,7 @@
 """Download pictures (or videos) along with their captions and other metadata from Instagram."""
 
 
-__version__ = '4.9b1'
+__version__ = '4.9b2'
 
 
 try:

instaloader/instaloader.py

@@ -458,9 +458,12 @@ class Instaloader:
     def save_location(self, filename: str, location: PostLocation, mtime: datetime) -> None:
         """Save post location name and Google Maps link."""
         filename += '_location.txt'
-        location_string = (location.name + "\n" +
-                           "https://maps.google.com/maps?q={0},{1}&ll={0},{1}\n".format(location.lat,
-                                                                                        location.lng))
+        if location.lat is not None and location.lng is not None:
+            location_string = (location.name + "\n" +
+                               "https://maps.google.com/maps?q={0},{1}&ll={0},{1}\n".format(location.lat,
+                                                                                            location.lng))
+        else:
+            location_string = location.name
         with open(filename, 'wb') as text_file:
             with BytesIO(location_string.encode()) as bio:
                 shutil.copyfileobj(cast(IO, bio), text_file)
@@ -813,7 +816,7 @@ class Instaloader:
                 last_scraped = latest_stamps.get_last_story_timestamp(name)
                 scraped_timestamp = datetime.now().astimezone()
             for item in user_story.get_items():
-                if latest_stamps is not None and item.date_utc.replace(tzinfo=timezone.utc) <= last_scraped:
+                if latest_stamps is not None and item.date_local <= last_scraped:
                     break
                 if storyitem_filter is not None and not storyitem_filter(item):
                     self.context.log("<{} skipped>".format(item), flush=True)
@@ -1202,7 +1205,7 @@ class Instaloader:
         posts_takewhile: Optional[Callable[[Post], bool]] = None
         if latest_stamps is not None:
             last_scraped = latest_stamps.get_last_tagged_timestamp(profile.username)
-            posts_takewhile = lambda p: p.date_utc.replace(tzinfo=timezone.utc) > last_scraped
+            posts_takewhile = lambda p: p.date_local > last_scraped
         tagged_posts = profile.get_tagged_posts()
         self.posts_download_loop(tagged_posts,
                                  target if target
@@ -1210,7 +1213,7 @@ class Instaloader:
                                        _PostPathFormatter.sanitize_path(':tagged')),
                                  fast_update, post_filter, takewhile=posts_takewhile)
         if latest_stamps is not None and tagged_posts.first_item is not None:
-            latest_stamps.set_last_tagged_timestamp(profile.username, tagged_posts.first_item.date_local.astimezone())
+            latest_stamps.set_last_tagged_timestamp(profile.username, tagged_posts.first_item.date_local)
 
     def download_igtv(self, profile: Profile, fast_update: bool = False,
                       post_filter: Optional[Callable[[Post], bool]] = None,
@@ -1225,12 +1228,12 @@ class Instaloader:
         posts_takewhile: Optional[Callable[[Post], bool]] = None
         if latest_stamps is not None:
             last_scraped = latest_stamps.get_last_igtv_timestamp(profile.username)
-            posts_takewhile = lambda p: p.date_utc.replace(tzinfo=timezone.utc) > last_scraped
+            posts_takewhile = lambda p: p.date_local > last_scraped
         igtv_posts = profile.get_igtv_posts()
         self.posts_download_loop(igtv_posts, profile.username, fast_update, post_filter,
                                  total_count=profile.igtvcount, owner_profile=profile, takewhile=posts_takewhile)
         if latest_stamps is not None and igtv_posts.first_item is not None:
-            latest_stamps.set_last_igtv_timestamp(profile.username, igtv_posts.first_item.date_local.astimezone())
+            latest_stamps.set_last_igtv_timestamp(profile.username, igtv_posts.first_item.date_local)
 
     def _get_id_filename(self, profile_name: str) -> str:
         if ((format_string_contains_key(self.dirname_pattern, 'profile') or
@@ -1423,14 +1426,14 @@ class Instaloader:
                     if latest_stamps is not None:
                         # pylint:disable=cell-var-from-loop
                         last_scraped = latest_stamps.get_last_post_timestamp(profile_name)
-                        posts_takewhile = lambda p: p.date_utc.replace(tzinfo=timezone.utc) > last_scraped
+                        posts_takewhile = lambda p: p.date_local > last_scraped
                     posts_to_download = profile.get_posts()
                     self.posts_download_loop(posts_to_download, profile_name, fast_update, post_filter,
                                              total_count=profile.mediacount, owner_profile=profile,
                                              takewhile=posts_takewhile)
                     if latest_stamps is not None and posts_to_download.first_item is not None:
                         latest_stamps.set_last_post_timestamp(profile_name,
-                                                              posts_to_download.first_item.date_local.astimezone())
+                                                              posts_to_download.first_item.date_local)
 
         if stories and profiles:
             with self.context.error_catcher("Download stories"):

instaloader/nodeiterator.py

@@ -133,14 +133,15 @@ class NodeIterator(Iterator[T]):
             return item
         if self._data['page_info']['has_next_page']:
             query_response = self._query(self._data['page_info']['end_cursor'])
-            page_index, data = self._page_index, self._data
-            try:
-                self._page_index = 0
-                self._data = query_response
-            except KeyboardInterrupt:
-                self._page_index, self._data = page_index, data
-                raise
-            return self.__next__()
+            if self._data['edges'] != query_response['edges']:
+                page_index, data = self._page_index, self._data
+                try:
+                    self._page_index = 0
+                    self._data = query_response
+                except KeyboardInterrupt:
+                    self._page_index, self._data = page_index, data
+                    raise
+                return self.__next__()
         raise StopIteration()
 
     @property

instaloader/structures.py

@@ -38,8 +38,8 @@ PostLocation.id.__doc__ = "ID number of location."
 PostLocation.name.__doc__ = "Location name."
 PostLocation.slug.__doc__ = "URL friendly variant of location name."
 PostLocation.has_public_page.__doc__ = "Whether location has a public page."
-PostLocation.lat.__doc__ = "Latitude (:class:`float`)."
-PostLocation.lng.__doc__ = "Longitude (:class:`float`)."
+PostLocation.lat.__doc__ = "Latitude (:class:`float` or None)."
+PostLocation.lng.__doc__ = "Longitude (:class:`float` or None)."
 
 
 class Post:
@@ -262,17 +262,16 @@ class Post:
 
     @property
     def date_local(self) -> datetime:
-        """Timestamp when the post was created (local time zone)."""
-        return datetime.fromtimestamp(self._node["date"]
-                                      if "date" in self._node
-                                      else self._node["taken_at_timestamp"])
+        """Timestamp when the post was created (local time zone).
+
+        .. versionchanged:: 4.9
+           Return timezone aware datetime object."""
+        return datetime.fromtimestamp(self._get_timestamp_date_created()).astimezone()
 
     @property
     def date_utc(self) -> datetime:
         """Timestamp when the post was created (UTC)."""
-        return datetime.utcfromtimestamp(self._node["date"]
-                                         if "date" in self._node
-                                         else self._node["taken_at_timestamp"])
+        return datetime.utcfromtimestamp(self._get_timestamp_date_created())
 
     @property
     def date(self) -> datetime:
@@ -313,6 +312,12 @@ class Post:
             return len(edges)
         return 1
 
+    def _get_timestamp_date_created(self) -> float:
+        """Timestamp when the post was created"""
+        return (self._node["date"]
+                if "date" in self._node
+                else self._node["taken_at_timestamp"])
+
     def get_is_videos(self) -> List[bool]:
         """
         Return a list containing the ``is_video`` property for each media in the post.
@@ -629,7 +634,7 @@ class Post:
             loc.update(self._context.get_json("explore/locations/{0}/".format(location_id),
                                               params={'__a': 1})['native_location_data']['location_info'])
         self._location = PostLocation(location_id, loc['name'], loc['slug'], loc['has_public_page'],
-                                      loc['lat'], loc['lng'])
+                                      loc.get('lat'), loc.get('lng'))
         return self._location
 
 
@@ -1140,8 +1145,11 @@ class StoryItem:
 
     @property
     def date_local(self) -> datetime:
-        """Timestamp when the StoryItem was created (local time zone)."""
-        return datetime.fromtimestamp(self._node['taken_at_timestamp'])
+        """Timestamp when the StoryItem was created (local time zone).
+
+        .. versionchanged:: 4.9
+           Return timezone aware datetime object."""
+        return datetime.fromtimestamp(self._node['taken_at_timestamp']).astimezone()
 
     @property
     def date_utc(self) -> datetime:
@@ -1655,7 +1663,7 @@ class TopSearchResults:
             place = location.get('place', {})
             slug = place.get('slug')
             loc = place.get('location', {})
-            yield PostLocation(int(loc['pk']), loc['name'], slug, None, loc['lat'], loc['lng'])
+            yield PostLocation(int(loc['pk']), loc['name'], slug, None, loc.get('lat'), loc.get('lng'))
 
     def get_hashtag_strings(self) -> Iterator[str]:
         """