From 1cab671a130aa7de9ee191be6e35e1fc8419a8bb Mon Sep 17 00:00:00 2001 From: Vxrtrauter Date: Sat, 18 Oct 2025 01:08:36 +0200 Subject: [PATCH] add more testing stuff --- core/data/scrapers/monkrus.py | 11 ++++++----- core/utils/data/state.py | 5 +---- 2 files changed, 7 insertions(+), 9 deletions(-) diff --git a/core/data/scrapers/monkrus.py b/core/data/scrapers/monkrus.py index 3b60645..ea4bf8c 100644 --- a/core/data/scrapers/monkrus.py +++ b/core/data/scrapers/monkrus.py @@ -12,14 +12,15 @@ def scrape_monkrus_telegram(): print(response.status_code) # rm later soup = BeautifulSoup(response.text, "html.parser") - posts = soup.find_all("div", class_=["tgme_widget_message_wrap js-widget_message_wrap", "tgme_widget_message_wrap js-widget_message_wrap date_visible"]) + bubbles = soup.find_all("div", class_="tgme_widget_message_bubble") - for post in posts: - posts_txt = post.find("div", class_="tgme_widget_message_text js-message_text") # this doesnt work + for bubble in bubbles: + post_txt = bubble.find("div", class_="tgme_widget_message_text js-message_text") # this doesnt work + if post_txt.b: + posts.append(Post(text=post_txt.get_text())) - - print(posts_txt) + print(post_txt.b.text) diff --git a/core/utils/data/state.py b/core/utils/data/state.py index deb9558..ebedd07 100644 --- a/core/utils/data/state.py +++ b/core/utils/data/state.py @@ -23,7 +23,4 @@ state = AppState(posts=[], post_titles=[], post_urls=[], post_author=[], downloa @dataclass class Post: - id: int - title: str - url: str - author: str + text: str