erjemin/instaloader
erjemin/instaloader
1
0
Fork 0
Code Issues Wiki Activity

Make {Post,StoryItem}.date_local timezone aware

Browse Source 
This commit is made up of 21365ec6688b55c91803040693c3b67cc8820cb8 555c86633c61b106c5c3201e1d5f030a04a4f801 d864ce08ff43a2d86ac6123459bb9fa32b924657

Discussed in #1305, #1316, #1372.

Co-Authored-By: Misael <misael.om@gmail.com>
Co-Authored-By: Eduardo Kalinowski <eduardo@kalinowski.com.br>
This commit is contained in:
Alexander Graf 2022-01-15 17:29:28 +01:00
parent 8c752532e1
commit 2d4f3b8537
2 changed files with 24 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
instaloader/instaloader.py
View File

@ -816,7 +816,7 @@ class Instaloader:
last_scraped = latest_stamps.get_last_story_timestamp(name) last_scraped = latest_stamps.get_last_story_timestamp(name)
scraped_timestamp = datetime.now().astimezone() scraped_timestamp = datetime.now().astimezone()
for item in user_story.get_items(): 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 break
if storyitem_filter is not None and not storyitem_filter(item): if storyitem_filter is not None and not storyitem_filter(item):
self.context.log("<{} skipped>".format(item), flush=True) self.context.log("<{} skipped>".format(item), flush=True)
@ -1205,7 +1205,7 @@ class Instaloader:
posts_takewhile: Optional[Callable[[Post], bool]] = None posts_takewhile: Optional[Callable[[Post], bool]] = None
if latest_stamps is not None: if latest_stamps is not None:
last_scraped = latest_stamps.get_last_tagged_timestamp(profile.username) 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() tagged_posts = profile.get_tagged_posts()
self.posts_download_loop(tagged_posts, self.posts_download_loop(tagged_posts,
target if target target if target
@ -1213,7 +1213,7 @@ class Instaloader:
_PostPathFormatter.sanitize_path(':tagged')), _PostPathFormatter.sanitize_path(':tagged')),
fast_update, post_filter, takewhile=posts_takewhile) fast_update, post_filter, takewhile=posts_takewhile)
if latest_stamps is not None and tagged_posts.first_item is not None: 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, def download_igtv(self, profile: Profile, fast_update: bool = False,
post_filter: Optional[Callable[[Post], bool]] = None, post_filter: Optional[Callable[[Post], bool]] = None,
@ -1228,12 +1228,12 @@ class Instaloader:
posts_takewhile: Optional[Callable[[Post], bool]] = None posts_takewhile: Optional[Callable[[Post], bool]] = None
if latest_stamps is not None: if latest_stamps is not None:
last_scraped = latest_stamps.get_last_igtv_timestamp(profile.username) 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() igtv_posts = profile.get_igtv_posts()
self.posts_download_loop(igtv_posts, profile.username, fast_update, post_filter, self.posts_download_loop(igtv_posts, profile.username, fast_update, post_filter,
total_count=profile.igtvcount, owner_profile=profile, takewhile=posts_takewhile) total_count=profile.igtvcount, owner_profile=profile, takewhile=posts_takewhile)
if latest_stamps is not None and igtv_posts.first_item is not None: 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: def _get_id_filename(self, profile_name: str) -> str:
if ((format_string_contains_key(self.dirname_pattern, 'profile') or if ((format_string_contains_key(self.dirname_pattern, 'profile') or
@ -1426,14 +1426,14 @@ class Instaloader:
if latest_stamps is not None: if latest_stamps is not None:
# pylint:disable=cell-var-from-loop # pylint:disable=cell-var-from-loop
last_scraped = latest_stamps.get_last_post_timestamp(profile_name) 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() posts_to_download = profile.get_posts()
self.posts_download_loop(posts_to_download, profile_name, fast_update, post_filter, self.posts_download_loop(posts_to_download, profile_name, fast_update, post_filter,
total_count=profile.mediacount, owner_profile=profile, total_count=profile.mediacount, owner_profile=profile,
takewhile=posts_takewhile) takewhile=posts_takewhile)
if latest_stamps is not None and posts_to_download.first_item is not None: if latest_stamps is not None and posts_to_download.first_item is not None:
latest_stamps.set_last_post_timestamp(profile_name, 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: if stories and profiles:
with self.context.error_catcher("Download stories"): with self.context.error_catcher("Download stories"):

26
instaloader/structures.py
View File

@ -262,17 +262,16 @@ class Post:
@property @property
def date_local(self) -> datetime: def date_local(self) -> datetime:
"""Timestamp when the post was created (local time zone).""" """Timestamp when the post was created (local time zone).
return datetime.fromtimestamp(self._node["date"]
if "date" in self._node .. versionchanged:: 4.9
else self._node["taken_at_timestamp"]) Return timezone aware datetime object."""
return datetime.fromtimestamp(self._get_timestamp_date_created()).astimezone()
@property @property
def date_utc(self) -> datetime: def date_utc(self) -> datetime:
"""Timestamp when the post was created (UTC).""" """Timestamp when the post was created (UTC)."""
return datetime.utcfromtimestamp(self._node["date"] return datetime.utcfromtimestamp(self._get_timestamp_date_created())
if "date" in self._node
else self._node["taken_at_timestamp"])
@property @property
def date(self) -> datetime: def date(self) -> datetime:
@ -313,6 +312,12 @@ class Post:
return len(edges) return len(edges)
return 1 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]: def get_is_videos(self) -> List[bool]:
""" """
Return a list containing the ``is_video`` property for each media in the post. Return a list containing the ``is_video`` property for each media in the post.
@ -1140,8 +1145,11 @@ class StoryItem:
@property @property
def date_local(self) -> datetime: def date_local(self) -> datetime:
"""Timestamp when the StoryItem was created (local time zone).""" """Timestamp when the StoryItem was created (local time zone).
return datetime.fromtimestamp(self._node['taken_at_timestamp'])
.. versionchanged:: 4.9
Return timezone aware datetime object."""
return datetime.fromtimestamp(self._node['taken_at_timestamp']).astimezone()
@property @property
def date_utc(self) -> datetime: def date_utc(self) -> datetime: