use crate::command;
use std::{
borrow::Cow,
collections::{
BTreeMap,
HashSet,
},
num::TryFromIntError,
sync::{
Arc,
Mutex
},
};
use anyhow::{
anyhow,
bail,
Result,
};
use async_std::task;
use chrono::DateTime;
use frankenstein::{
client_reqwest::Bot,
methods::{
GetUpdatesParams,
SendMessageParams
},
types::{
AllowedUpdate,
MessageEntityType,
User,
},
updates::UpdateContent,
AsyncTelegramApi,
ParseMode,
};
use sqlx::postgres::PgPoolOptions;
use thiserror::Error;
#[derive(Error, Debug)]
pub enum RssError {
// #[error(transparent)]
// Tg(#[from] TgError),
#[error(transparent)]
Int(#[from] TryFromIntError),
}
#[derive(Clone)]
pub struct Core {
owner_chat: i64,
pub tg: Bot,
pub me: User,
pool: sqlx::Pool<sqlx::Postgres>,
sources: Arc<Mutex<HashSet<Arc<i32>>>>,
http_client: reqwest::Client,
}
impl Core {
pub async fn new(settings: config::Config) -> Result<Arc<Core>> {
let owner_chat = settings.get_int("owner")?;
let api_key = settings.get_string("api_key")?;
let tg = Bot::new(&api_key);
let mut client = reqwest::Client::builder();
if let Ok(proxy) = settings.get_string("proxy") {
let proxy = reqwest::Proxy::all(proxy)?;
client = client.proxy(proxy);
}
let http_client = client.build()?;
let me = tg.get_me().await?;
let me = me.result;
let core = Arc::new(Core {
tg,
me,
owner_chat,
pool: PgPoolOptions::new()
.max_connections(5)
.acquire_timeout(std::time::Duration::new(300, 0))
.idle_timeout(std::time::Duration::new(60, 0))
.connect_lazy(&settings.get_string("pg")?)?,
sources: Arc::new(Mutex::new(HashSet::new())),
http_client,
});
let clone = core.clone();
task::spawn(async move {
loop {
let delay = match &clone.autofetch().await {
Err(err) => {
if let Err(err) = clone.send(format!("š {err:?}"), None, None).await {
eprintln!("Autofetch error: {err:?}");
};
std::time::Duration::from_secs(60)
},
Ok(time) => *time,
};
task::sleep(delay).await;
}
});
Ok(core)
}
pub async fn stream(&self) -> Result<()> {
let mut offset: i64 = 0;
let mut params = GetUpdatesParams {
offset: None,
limit: Some(100),
timeout: Some(300),
allowed_updates: Some(vec![AllowedUpdate::Message]),
};
loop {
let updates = self.tg.get_updates(¶ms).await?.result;
if updates.is_empty() {
offset = 0;
params.offset = None;
continue;
}
for update in updates {
if i64::from(update.update_id) >= offset {
offset = i64::from(update.update_id) + 1;
params.offset = Some(offset);
}
if let UpdateContent::Message(msg) = update.content {
if let Some(text) = msg.text {
if let Some(entities) = msg.entities {
let chars: Vec<u16> = text.encode_utf16().collect();
for entity in entities {
if entity.type_field == MessageEntityType::BotCommand && entity.offset != 0 {
bail!("commands should be at message start");
};
let cmd = String::from_utf16_lossy(&chars[entity.offset as usize..entity.length as usize]);
let words: Vec<&str> = text.split_whitespace().collect();
match cmd.as_ref() {
"/check" | "/clean" | "/enable" | "/delete" | "/disable" => { command::command(self, msg.chat.id, words).await? },
"/start" => { command::start(self, msg.chat.id).await?; },
"/list" => { command::list(self, msg.chat.id).await?; },
"/add" | "/update" => { command::update(self, msg.chat.id, words).await?; },
any => {
self.send(format!("\\#error\n```\nUnknown command: {any}\n```"),
Some(msg.chat.id),
Some(ParseMode::MarkdownV2)
).await?;
},
};
};
};
};
};
}
}
}
pub async fn send <S>(&self, msg: S, target: Option<i64>, mode: Option<ParseMode>) -> Result<()>
where S: Into<String> {
let msg = msg.into();
let mode = mode.unwrap_or(ParseMode::Html);
let target = target.unwrap_or(self.owner_chat);
let send = SendMessageParams::builder()
.chat_id(target)
.text(msg)
.parse_mode(mode)
.build();
self.tg.send_message(&send).await?;
Ok(())
}
pub async fn check (&self, id: &i32, owner: i64, real: bool) -> Result<Cow<'_, str>> {
let mut posted: i32 = 0;
let mut conn = self.pool.acquire().await?;
let id = {
let mut set = self.sources.lock().unwrap();
match set.get(id) {
Some(id) => id.clone(),
None => {
let id = Arc::new(*id);
set.insert(id.clone());
id.clone()
},
}
};
let count = Arc::strong_count(&id);
if count == 2 {
let source = sqlx::query!("select source_id, channel_id, url, iv_hash, owner, url_re from rsstg_source where source_id = $1 and owner = $2",
*id, owner).fetch_one(&mut *conn).await?;
let destination = match real {
true => source.channel_id,
false => source.owner,
};
let mut this_fetch: Option<DateTime<chrono::FixedOffset>> = None;
let mut posts: BTreeMap<DateTime<chrono::FixedOffset>, String> = BTreeMap::new();
let response = self.http_client.get(&source.url).send().await?;
let status = response.status();
let content = response.bytes().await?;
match rss::Channel::read_from(&content[..]) {
Ok(feed) => {
for item in feed.items() {
if let Some(link) = item.link() {
let date = match item.pub_date() {
Some(feed_date) => DateTime::parse_from_rfc2822(feed_date),
None => DateTime::parse_from_rfc3339(&item.dublin_core_ext().unwrap().dates()[0]),
}?;
let url = link;
posts.insert(date, url.to_string());
}
};
},
Err(err) => match err {
rss::Error::InvalidStartTag => {
match atom_syndication::Feed::read_from(&content[..]) {
Ok(feed) => {
for item in feed.entries() {
let date = item.published().unwrap();
let url = item.links()[0].href();
posts.insert(*date, url.to_string());
};
},
Err(err) => {
bail!("Unsupported or mangled content:\n{:?}\n{err:#?}\n{status:#?}\n", &source.url)
},
}
},
rss::Error::Eof => (),
_ => bail!("Unsupported or mangled content:\n{:?}\n{err:#?}\n{status:#?}\n", &source.url)
}
};
for (date, url) in posts.iter() {
let post_url: Cow<str> = match source.url_re {
Some(ref x) => sedregex::ReplaceCommand::new(x)?.execute(url),
None => url.into(),
};
if let Some(exists) = sqlx::query!("select exists(select true from rsstg_post where url = $1 and source_id = $2) as exists;",
&post_url, *id).fetch_one(&mut *conn).await?.exists {
if ! exists {
if this_fetch.is_none() || *date > this_fetch.unwrap() {
this_fetch = Some(*date);
};
self.send( match &source.iv_hash {
Some(hash) => format!("<a href=\"https://t.me/iv?url={post_url}&rhash={hash}\"> </a>{post_url}"),
None => format!("{post_url}"),
}, Some(destination), Some(ParseMode::Html)).await?;
sqlx::query!("insert into rsstg_post (source_id, posted, url) values ($1, $2, $3);",
*id, date, &post_url).execute(&mut *conn).await?;
};
};
posted += 1;
};
posts.clear();
};
sqlx::query!("update rsstg_source set last_scrape = now() where source_id = $1;",
*id).execute(&mut *conn).await?;
Ok(format!("Posted: {posted}").into())
}
pub async fn delete (&self, source_id: &i32, owner: i64) -> Result<Cow<'_, str>> {
match sqlx::query!("delete from rsstg_source where source_id = $1 and owner = $2;",
source_id, owner).execute(&mut *self.pool.acquire().await?).await?.rows_affected() {
0 => { Ok("No data found found.".into()) },
x => { Ok(format!("{} sources removed.", x).into()) },
}
}
pub async fn clean (&self, source_id: &i32, owner: i64) -> Result<Cow<'_, str>> {
match sqlx::query!("delete from rsstg_post p using rsstg_source s where p.source_id = $1 and owner = $2 and p.source_id = s.source_id;",
source_id, owner).execute(&mut *self.pool.acquire().await?).await?.rows_affected() {
0 => { Ok("No data found found.".into()) },
x => { Ok(format!("{x} posts purged.").into()) },
}
}
pub async fn enable (&self, source_id: &i32, owner: i64) -> Result<&str> {
match sqlx::query!("update rsstg_source set enabled = true where source_id = $1 and owner = $2",
source_id, owner).execute(&mut *self.pool.acquire().await?).await?.rows_affected() {
1 => { Ok("Source enabled.") },
0 => { Ok("Source not found.") },
_ => { Err(anyhow!("Database error.")) },
}
}
pub async fn disable (&self, source_id: &i32, owner: i64) -> Result<&str> {
match sqlx::query!("update rsstg_source set enabled = false where source_id = $1 and owner = $2",
source_id, owner).execute(&mut *self.pool.acquire().await?).await?.rows_affected() {
1 => { Ok("Source disabled.") },
0 => { Ok("Source not found.") },
_ => { Err(anyhow!("Database error.")) },
}
}
pub async fn update (&self, update: Option<i32>, channel: &str, channel_id: i64, url: &str, iv_hash: Option<&str>, url_re: Option<&str>, owner: i64) -> Result<&str> {
let mut conn = self.pool.acquire().await?;
match match update {
Some(id) => {
sqlx::query!("update rsstg_source set channel_id = $2, url = $3, iv_hash = $4, owner = $5, channel = $6, url_re = $7 where source_id = $1",
id, channel_id, url, iv_hash, owner, channel, url_re).execute(&mut *conn).await
},
None => {
sqlx::query!("insert into rsstg_source (channel_id, url, iv_hash, owner, channel, url_re) values ($1, $2, $3, $4, $5, $6)",
channel_id, url, iv_hash, owner, channel, url_re).execute(&mut *conn).await
},
} {
Ok(_) => Ok(match update {
Some(_) => "Channel updated.",
None => "Channel added.",
}),
Err(sqlx::Error::Database(err)) => {
match err.downcast::<sqlx::postgres::PgDatabaseError>().routine() {
Some("_bt_check_unique", ) => {
Ok("Duplicate key.")
},
Some(_) => {
Ok("Database error.")
},
None => {
Ok("No database error extracted.")
},
}
},
Err(err) => {
bail!("Sorry, unknown error:\n{err:#?}\n");
},
}
}
async fn autofetch(&self) -> Result<std::time::Duration> {
let mut delay = chrono::Duration::minutes(1);
let now = chrono::Local::now();
let mut queue = sqlx::query!(r#"select source_id, next_fetch as "next_fetch: DateTime<chrono::Local>", owner from rsstg_order natural left join rsstg_source where next_fetch < now() + interval '1 minute';"#)
.fetch_all(&mut *self.pool.acquire().await?).await?;
for row in queue.iter() {
if let Some(next_fetch) = row.next_fetch {
if next_fetch < now {
if let (Some(owner), Some(source_id)) = (row.owner, row.source_id) {
let clone = Core {
owner_chat: owner,
..self.clone()
};
task::spawn(async move {
if let Err(err) = clone.check(&source_id, owner, true).await {
if let Err(err) = clone.send(&format!("š {err:?}"), None, None).await {
eprintln!("Check error: {err:?}");
// clone.disable(&source_id, owner).await.unwrap();
};
};
});
}
} else if next_fetch - now < delay {
delay = next_fetch - now;
}
}
};
queue.clear();
Ok(delay.to_std()?)
}
pub async fn list (&self, owner: i64) -> Result<String> {
let mut reply: Vec<Cow<str>> = vec![];
reply.push("Channels:".into());
let rows = sqlx::query!("select source_id, channel, enabled, url, iv_hash, url_re from rsstg_source where owner = $1 order by source_id",
owner).fetch_all(&mut *self.pool.acquire().await?).await?;
for row in rows.iter() {
reply.push(format!("\n\\#ļøā£ {} \\*ļøā£ `{}` {}\nš `{}`", row.source_id, row.channel,
match row.enabled {
true => "š enabled",
false => "ā disabled",
}, row.url).into());
if let Some(hash) = &row.iv_hash {
reply.push(format!("IV: `{hash}`").into());
}
if let Some(re) = &row.url_re {
reply.push(format!("RE: `{re}`").into());
}
};
Ok(reply.join("\n"))
}
}