Compare commits
No commits in common. "7d379448ff463dd21ca864624f48b9cf80eae953" and "0e9692a8efafe8137879ad39d3f750a6059caf58" have entirely different histories.
7d379448ff
...
0e9692a8ef
3 changed files with 16 additions and 35 deletions
|
@ -1,6 +1,6 @@
|
||||||
[package]
|
[package]
|
||||||
name = "frogbot"
|
name = "frogbot"
|
||||||
version = "0.1.1"
|
version = "0.1.0"
|
||||||
edition = "2021"
|
edition = "2021"
|
||||||
|
|
||||||
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
|
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
|
||||||
|
|
|
@ -15,7 +15,6 @@ use regex::Regex;
|
||||||
use scraper::{Html, Selector};
|
use scraper::{Html, Selector};
|
||||||
|
|
||||||
/// Represents an Embed in the chat
|
/// Represents an Embed in the chat
|
||||||
#[derive(Default)]
|
|
||||||
pub struct Embed {
|
pub struct Embed {
|
||||||
/// The title of the embed
|
/// The title of the embed
|
||||||
pub title: String,
|
pub title: String,
|
||||||
|
@ -31,7 +30,7 @@ impl Embed {
|
||||||
}
|
}
|
||||||
|
|
||||||
/// Scrapes the HTML of a webpage and generates an [`Embed`] with the scraped information.
|
/// Scrapes the HTML of a webpage and generates an [`Embed`] with the scraped information.
|
||||||
pub fn parse_metadata(page: &str) -> Option<Embed> {
|
pub fn parse_metadata(page: &str) -> Embed {
|
||||||
let doc_body = Html::parse_document(page);
|
let doc_body = Html::parse_document(page);
|
||||||
|
|
||||||
// Selectors used to get metadata are defined here
|
// Selectors used to get metadata are defined here
|
||||||
|
@ -45,23 +44,19 @@ pub fn parse_metadata(page: &str) -> Option<Embed> {
|
||||||
let mut meta_title = String::default();
|
let mut meta_title = String::default();
|
||||||
let mut meta_description = String::default();
|
let mut meta_description = String::default();
|
||||||
|
|
||||||
match (title, desc) {
|
if let Some(title) = title {
|
||||||
// If both title and description aren't found return None
|
meta_title = title.text().collect();
|
||||||
(None, None) => {
|
} else {
|
||||||
warn!("Couldn't parse any metadata for URL");
|
warn!("Failed to parse title HTML");
|
||||||
return None;
|
|
||||||
}
|
|
||||||
// Otherwise set the title/description to whatever we find
|
|
||||||
(Some(title), Some(desc)) => {
|
|
||||||
meta_title = title.text().collect();
|
|
||||||
meta_description = desc.value().attr("content").unwrap().to_string();
|
|
||||||
}
|
|
||||||
// Handle logging of parse failures
|
|
||||||
(Some(_), None) => warn!("Failed to parse description HTML"),
|
|
||||||
(None, Some(_)) => warn!("Failed to parse title HTML"),
|
|
||||||
}
|
}
|
||||||
|
|
||||||
Some(Embed::new(meta_title, meta_description))
|
if let Some(desc) = desc {
|
||||||
|
meta_description = desc.value().attr("content").unwrap().to_string();
|
||||||
|
} else {
|
||||||
|
warn!("Failed to parse description HTML");
|
||||||
|
}
|
||||||
|
|
||||||
|
Embed::new(meta_title, meta_description)
|
||||||
}
|
}
|
||||||
|
|
||||||
/// Check if the message has any urls in it and get them if it does
|
/// Check if the message has any urls in it and get them if it does
|
||||||
|
@ -124,23 +119,15 @@ pub async fn embed_handler(event: OriginalSyncRoomMessageEvent, room: Room, clie
|
||||||
|
|
||||||
let urls = get_urls_from_message(&text_content.body);
|
let urls = get_urls_from_message(&text_content.body);
|
||||||
|
|
||||||
let reqwest_client = reqwest::Client::builder().user_agent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/118.0.0.0 Safari/537.36").build().unwrap();
|
let reqwest_client = reqwest::Client::builder().user_agent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36").build().unwrap();
|
||||||
|
|
||||||
for mut url in urls {
|
for url in urls {
|
||||||
if let Ok(req) = reqwest_client.get(url).send().await {
|
if let Ok(req) = reqwest_client.get(url).send().await {
|
||||||
if let Ok(res) = req.text().await {
|
if let Ok(res) = req.text().await {
|
||||||
// beware, dirty HTML parsing code
|
// beware, dirty HTML parsing code
|
||||||
let metadata = parse_metadata(&res);
|
let embed = parse_metadata(&res);
|
||||||
|
|
||||||
// If we didn't get any metadata set URL to nothing so it won't get repeated
|
|
||||||
// With no other embed data in the bot's embed message
|
|
||||||
if metadata.is_none() {
|
|
||||||
url = "";
|
|
||||||
}
|
|
||||||
|
|
||||||
// Build our message reply
|
// Build our message reply
|
||||||
let embed = metadata
|
|
||||||
.unwrap_or(Embed::new("No metadata found".to_string(), "".to_string()));
|
|
||||||
let bot_reply = RoomMessageEventContent::text_html(
|
let bot_reply = RoomMessageEventContent::text_html(
|
||||||
&embed.title,
|
&embed.title,
|
||||||
format!(
|
format!(
|
||||||
|
|
|
@ -141,12 +141,6 @@ pub async fn run(config: Config) -> anyhow::Result<()> {
|
||||||
.await
|
.await
|
||||||
.expect("frogbot couldn't log into it's account.");
|
.expect("frogbot couldn't log into it's account.");
|
||||||
|
|
||||||
// Set the bot account's display name according to config
|
|
||||||
client
|
|
||||||
.account()
|
|
||||||
.set_display_name(Some(&config.display_name))
|
|
||||||
.await?;
|
|
||||||
|
|
||||||
warn!("Logged in successfully!");
|
warn!("Logged in successfully!");
|
||||||
warn!(
|
warn!(
|
||||||
"server: '{}', username: '{}', display name: '{}'",
|
"server: '{}', username: '{}', display name: '{}'",
|
||||||
|
|
Loading…
Reference in a new issue