1
0
mirror of https://github.com/instaloader/instaloader.git synced 2024-07-04 19:05:55 +02:00

Fix downloading comments

Closes issue #2125 by using an iphone endpoint for posts that have more
than 50 comments.
This commit is contained in:
Alexander Graf 2024-01-12 08:07:54 +01:00
parent baaa2c2998
commit 1852447825
2 changed files with 75 additions and 2 deletions

View File

@ -184,6 +184,10 @@ class NodeIterator(Iterator[T]):
"""
return self._node_wrapper(self._first_node) if self._first_node is not None else None
@staticmethod
def page_length() -> int:
return NodeIterator._graphql_page_length
def freeze(self) -> FrozenNodeIterator:
"""Freeze the iterator for later resuming."""
remaining_data = None

View File

@ -545,11 +545,74 @@ class Post:
except KeyError:
return self._field('edge_media_to_comment', 'count')
def _get_comments_via_iphone_endpoint(self) -> Iterable[PostComment]:
"""
Iterate over all comments of the post via an iPhone endpoint.
.. versionadded:: 4.10.3
fallback for :issue:`2125`.
"""
def _query(min_id=None):
pagination_params = {"min_id": min_id} if min_id is not None else {}
return self._context.get_iphone_json(
f"api/v1/media/{self.mediaid}/comments/",
{
"can_support_threading": "true",
"permalink_enabled": "false",
**pagination_params,
},
)
def _answers(comment_node):
def _answer(child_comment):
return PostCommentAnswer(
id=int(child_comment["pk"]),
created_at_utc=datetime.utcfromtimestamp(child_comment["created_at"]),
text=child_comment["text"],
owner=Profile.from_iphone_struct(self._context, child_comment["user"]),
likes_count=child_comment["comment_like_count"],
)
child_comment_count = comment_node["child_comment_count"]
if child_comment_count == 0:
return
preview_child_comments = comment_node["preview_child_comments"]
if child_comment_count == len(preview_child_comments):
yield from (
_answer(child_comment) for child_comment in preview_child_comments
)
return
pk = comment_node["pk"]
answers_json = self._context.get_iphone_json(
f"api/v1/media/{self.mediaid}/comments/{pk}/child_comments/",
{"max_id": ""},
)
yield from (
_answer(child_comment) for child_comment in answers_json["child_comments"]
)
def _paginated_comments(comments_json):
for comment_node in comments_json.get("comments", []):
yield PostComment(
id=int(comment_node["pk"]),
created_at_utc=datetime.utcfromtimestamp(comment_node["created_at"]),
text=comment_node["text"],
owner=Profile.from_iphone_struct(self._context, comment_node["user"]),
likes_count=comment_node["comment_like_count"],
answers=_answers(comment_node),
)
next_min_id = comments_json.get("next_min_id")
if next_min_id:
yield from _paginated_comments(_query(next_min_id))
return _paginated_comments(_query())
def get_comments(self) -> Iterable[PostComment]:
r"""Iterate over all comments of the post.
"""Iterate over all comments of the post.
Each comment is represented by a PostComment NamedTuple with fields text (string), created_at (datetime),
id (int), owner (:class:`Profile`) and answers (:class:`~typing.Iterator`\ [:class:`PostCommentAnswer`])
id (int), owner (:class:`Profile`) and answers (:class:`~typing.Iterator` [:class:`PostCommentAnswer`])
if available.
.. versionchanged:: 4.7
@ -596,6 +659,12 @@ class Post:
if self.comments == len(comment_edges) + answers_count:
# If the Post's metadata already contains all parent comments, don't do GraphQL requests to obtain them
return [_postcomment(comment['node']) for comment in comment_edges]
if self.comments > NodeIterator.page_length():
# comments pagination via our graphql query does not work reliably anymore (issue #2125), fallback to an
# iphone endpoint if needed.
return self._get_comments_via_iphone_endpoint()
return NodeIterator(
self._context,
'97b41c52301f77ce508f55e66d17620e',