Skip to content

Commit

Permalink
Fix: some posts are missing in twitter
Browse files Browse the repository at this point in the history
  • Loading branch information
eight04 committed Jan 22, 2024
1 parent 05dde96 commit a4533ac
Showing 1 changed file with 19 additions and 4 deletions.
23 changes: 19 additions & 4 deletions comiccrawler/mods/twitter.py
Original file line number Diff line number Diff line change
Expand Up @@ -79,19 +79,34 @@ def get_episodes(html, url):

def tweet_result_to_episode(tweet_result):
try:
all_media = tweet_result["legacy"]["entities"]["media"] + tweet_result["legacy"]["extended_entities"]["media"]
legacy = tweet_result["legacy"]
except KeyError:
legacy = tweet_result["tweet"]["legacy"]

try:
all_media = legacy["entities"]["media"] + legacy["extended_entities"]["media"]
except KeyError:
return None
imgs = [find_media_source(m) for m in all_media]
imgs = list(OrderedDict.fromkeys(imgs).keys()) # remove dup
result = None

try:
result = tweet_result["legacy"]["retweeted_status_result"]["result"]
except KeyError:
result = tweet_result

screen_name = result["core"]["user_results"]["result"]["legacy"]["screen_name"]
id_str = result["legacy"]["id_str"]
try:
core = result["core"]
except KeyError:
core = result["tweet"]["core"]
screen_name = core["user_results"]["result"]["legacy"]["screen_name"]

try:
legacy = result["legacy"]
except KeyError:
legacy = result["tweet"]["legacy"]
id_str = legacy["id_str"]

ep_url = f"https://twitter.com/{screen_name}/status/{id_str}"

return Episode(
Expand Down

0 comments on commit a4533ac

Please sign in to comment.