Lines of
src/core.rs
from check-in 3dc9cddd4d
that are changed by the sequence of edits moving toward
check-in 79c91a5357:
1: use anyhow::{anyhow, bail, Context, Result};
2: use async_std::task;
3: use chrono::DateTime;
4: use sqlx::postgres::PgPoolOptions;
5: use std::{
6: borrow::Cow,
7: collections::{
8: BTreeMap,
9: HashSet,
10: },
11: sync::{Arc, Mutex},
12: };
13:
14: #[derive(Clone)]
15: pub struct Core {
16: owner_chat: telegram_bot::UserId,
17: pub tg: telegram_bot::Api,
18: pub my: telegram_bot::User,
19: pool: sqlx::Pool<sqlx::Postgres>,
20: sources: Arc<Mutex<HashSet<Arc<i32>>>>,
21: http_client: reqwest::Client,
22: }
23:
24: impl Core {
25: pub fn new(settings: config::Config) -> Result<Arc<Core>> {
26: let owner = settings.get_int("owner")?;
27: let api_key = settings.get_string("api_key")?;
28: let tg = telegram_bot::Api::new(api_key);
29: let tg_cloned = tg.clone();
30:
31: let mut client = reqwest::Client::builder();
32: if let Ok(proxy) = settings.get_string("proxy") {
33: let proxy = reqwest::Proxy::all(proxy)?;
34: client = client.proxy(proxy);
35: }
36: let http_client = client.build()?;
37: let core = Arc::new(Core {
38: tg,
39: my: task::block_on(async {
40: tg_cloned.send(telegram_bot::GetMe).await
41: })?,
42: owner_chat: telegram_bot::UserId::new(owner),
43: pool: PgPoolOptions::new()
44: .max_connections(5)
45: .acquire_timeout(std::time::Duration::new(300, 0))
46: .idle_timeout(std::time::Duration::new(60, 0))
47: .connect_lazy(&settings.get_string("pg")?)?,
48: sources: Arc::new(Mutex::new(HashSet::new())),
49: http_client,
50: });
51: let clone = core.clone();
52: task::spawn(async move {
53: loop {
54: let delay = match &clone.autofetch().await {
55: Err(err) => {
56: if let Err(err) = clone.send(format!("š {:?}", err), None, None).await {
57: eprintln!("Autofetch error: {}", err);
58: };
59: std::time::Duration::from_secs(60)
60: },
61: Ok(time) => *time,
62: };
63: task::sleep(delay).await;
64: }
65: });
66: Ok(core)
67: }
68:
69: pub fn stream(&self) -> telegram_bot::UpdatesStream {
70: self.tg.stream()
71: }
72:
73: pub async fn send<'a, S>(&self, msg: S, target: Option<telegram_bot::UserId>, mode: Option<telegram_bot::types::ParseMode>) -> Result<()>
74: where S: Into<Cow<'a, str>> {
75: let mode = mode.unwrap_or(telegram_bot::types::ParseMode::Html);
76: let target = target.unwrap_or(self.owner_chat);
3dc9cddd4d 2023-08-04 77: self.tg.send(telegram_bot::SendMessage::new(target, msg).parse_mode(mode)).await?;
78: Ok(())
79: }
80:
81: pub async fn check<S>(&self, id: &i32, owner: S, real: bool) -> Result<Cow<'_, str>>
82: where S: Into<i64> {
83: let owner = owner.into();
84:
85: let mut posted: i32 = 0;
86: let id = {
87: let mut set = self.sources.lock().unwrap();
88: match set.get(id) {
89: Some(id) => id.clone(),
90: None => {
91: let id = Arc::new(*id);
92: set.insert(id.clone());
93: id.clone()
94: },
95: }
96: };
97: let count = Arc::strong_count(&id);
98: if count == 2 {
99: 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",
100: *id, owner).fetch_one(&mut self.pool.acquire().await?).await?;
101: let destination = match real {
102: true => telegram_bot::UserId::new(source.channel_id),
103: false => telegram_bot::UserId::new(source.owner),
104: };
105: let mut this_fetch: Option<DateTime<chrono::FixedOffset>> = None;
106: let mut posts: BTreeMap<DateTime<chrono::FixedOffset>, String> = BTreeMap::new();
107:
108: let response = self.http_client.get(&source.url).send().await?;
109: let status = response.status();
110: let content = response.bytes().await?;
111: match rss::Channel::read_from(&content[..]) {
112: Ok(feed) => {
113: for item in feed.items() {
114: if let Some(link) = item.link() {
115: let date = match item.pub_date() {
116: Some(feed_date) => DateTime::parse_from_rfc2822(feed_date),
117: None => DateTime::parse_from_rfc3339(&item.dublin_core_ext().unwrap().dates()[0]),
118: }?;
119: let url = link;
120: posts.insert(date, url.to_string());
121: }
122: };
123: },
124: Err(err) => match err {
125: rss::Error::InvalidStartTag => {
126: let feed = atom_syndication::Feed::read_from(&content[..])
127: .with_context(|| format!("Problem opening feed url:\n{}\n{}", &source.url, status))?;
128: for item in feed.entries() {
129: let date = item.published().unwrap();
130: let url = item.links()[0].href();
131: posts.insert(*date, url.to_string());
132: };
133: },
134: rss::Error::Eof => (),
135: _ => bail!("Unsupported or mangled content:\n{:?}\n{:#?}\n{:#?}\n", &source.url, err, status)
136: }
137: };
138: for (date, url) in posts.iter() {
139: let post_url: Cow<str> = match source.url_re {
140: Some(ref x) => sedregex::ReplaceCommand::new(x)?.execute(url),
141: None => url.into(),
142: };
143: if let Some(exists) = sqlx::query!("select exists(select true from rsstg_post where url = $1 and source_id = $2) as exists;",
144: &post_url, *id).fetch_one(&mut self.pool.acquire().await?).await?.exists {
145: if ! exists {
146: if this_fetch.is_none() || *date > this_fetch.unwrap() {
147: this_fetch = Some(*date);
148: };
3dc9cddd4d 2023-08-04 149: self.tg.send( match &source.iv_hash {
150: Some(hash) => telegram_bot::SendMessage::new(destination, format!("<a href=\"https://t.me/iv?url={}&rhash={}\"> </a>{0}", &post_url, hash)),
151: None => telegram_bot::SendMessage::new(destination, format!("{}", post_url)),
152: }.parse_mode(telegram_bot::types::ParseMode::Html)).await
153: .context("Can't post message:")?;
154: sqlx::query!("insert into rsstg_post (source_id, posted, url) values ($1, $2, $3);",
155: *id, date, &post_url).execute(&mut self.pool.acquire().await?).await?;
3dc9cddd4d 2023-08-04 156: task::sleep(std::time::Duration::new(4, 0)).await;
157: };
158: };
159: posted += 1;
160: };
161: posts.clear();
162: };
163: sqlx::query!("update rsstg_source set last_scrape = now() where source_id = $1;",
164: *id).execute(&mut self.pool.acquire().await?).await?;
165: Ok(format!("Posted: {}", &posted).into())
166: }
167:
168: pub async fn delete<S>(&self, source_id: &i32, owner: S) -> Result<Cow<'_, str>>
169: where S: Into<i64> {
170: let owner = owner.into();
171:
172: match sqlx::query!("delete from rsstg_source where source_id = $1 and owner = $2;",
173: source_id, owner).execute(&mut self.pool.acquire().await?).await?.rows_affected() {
174: 0 => { Ok("No data found found.".into()) },
175: x => { Ok(format!("{} sources removed.", x).into()) },
176: }
177: }
178:
179: pub async fn clean<S>(&self, source_id: &i32, owner: S) -> Result<Cow<'_, str>>
180: where S: Into<i64> {
181: let owner = owner.into();
182:
183: 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;",
184: source_id, owner).execute(&mut self.pool.acquire().await?).await?.rows_affected() {
185: 0 => { Ok("No data found found.".into()) },
186: x => { Ok(format!("{} posts purged.", x).into()) },
187: }
188: }
189:
190: pub async fn enable<S>(&self, source_id: &i32, owner: S) -> Result<&str>
191: where S: Into<i64> {
192: let owner = owner.into();
193:
194: match sqlx::query!("update rsstg_source set enabled = true where source_id = $1 and owner = $2",
195: source_id, owner).execute(&mut self.pool.acquire().await?).await?.rows_affected() {
196: 1 => { Ok("Source enabled.") },
197: 0 => { Ok("Source not found.") },
198: _ => { Err(anyhow!("Database error.")) },
199: }
200: }
201:
202: pub async fn disable<S>(&self, source_id: &i32, owner: S) -> Result<&str>
203: where S: Into<i64> {
204: let owner = owner.into();
205:
206: match sqlx::query!("update rsstg_source set enabled = false where source_id = $1 and owner = $2",
207: source_id, owner).execute(&mut self.pool.acquire().await?).await?.rows_affected() {
208: 1 => { Ok("Source disabled.") },
209: 0 => { Ok("Source not found.") },
210: _ => { Err(anyhow!("Database error.")) },
211: }
212: }
213:
214: pub async fn update<S>(&self, update: Option<i32>, channel: &str, channel_id: i64, url: &str, iv_hash: Option<&str>, url_re: Option<&str>, owner: S) -> Result<&str>
215: where S: Into<i64> {
216: let owner = owner.into();
217:
218: match match update {
219: Some(id) => {
220: 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",
221: id, channel_id, url, iv_hash, owner, channel, url_re).execute(&mut self.pool.acquire().await?).await
222: },
223: None => {
224: sqlx::query!("insert into rsstg_source (channel_id, url, iv_hash, owner, channel, url_re) values ($1, $2, $3, $4, $5, $6)",
225: channel_id, url, iv_hash, owner, channel, url_re).execute(&mut self.pool.acquire().await?).await
226: },
227: } {
228: Ok(_) => Ok(match update {
229: Some(_) => "Channel updated.",
230: None => "Channel added.",
231: }),
232: Err(sqlx::Error::Database(err)) => {
233: match err.downcast::<sqlx::postgres::PgDatabaseError>().routine() {
234: Some("_bt_check_unique", ) => {
235: Ok("Duplicate key.")
236: },
237: Some(_) => {
238: Ok("Database error.")
239: },
240: None => {
241: Ok("No database error extracted.")
242: },
243: }
244: },
245: Err(err) => {
246: bail!("Sorry, unknown error:\n{:#?}\n", err);
247: },
248: }
249: }
250:
251: async fn autofetch(&self) -> Result<std::time::Duration> {
252: let mut delay = chrono::Duration::minutes(1);
253: let now = chrono::Local::now();
254: 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';"#)
255: .fetch_all(&mut self.pool.acquire().await?).await?;
256: for row in queue.iter() {
257: if let Some(next_fetch) = row.next_fetch {
258: if next_fetch < now {
259: if let (Some(owner), Some(source_id)) = (row.owner, row.source_id) {
260: let clone = Core {
261: owner_chat: telegram_bot::UserId::new(owner),
262: ..self.clone()
263: };
264: task::spawn(async move {
265: if let Err(err) = clone.check(&source_id, owner, true).await {
266: if let Err(err) = clone.send(&format!("š {:?}", err), None, None).await {
3dc9cddd4d 2023-08-04 267: eprintln!("Check error: {}", err);
3dc9cddd4d 2023-08-04 268: clone.disable(&source_id, owner).await.unwrap();
269: };
270: };
271: });
272: }
273: } else if next_fetch - now < delay {
274: delay = next_fetch - now;
275: }
276: }
277: };
278: queue.clear();
279: Ok(delay.to_std()?)
280: }
281:
282: pub async fn list<S>(&self, owner: S) -> Result<String>
283: where S: Into<i64> {
284: let owner = owner.into();
285:
286: let mut reply: Vec<Cow<str>> = vec![];
287: reply.push("Channels:".into());
288: let rows = sqlx::query!("select source_id, channel, enabled, url, iv_hash, url_re from rsstg_source where owner = $1 order by source_id",
3dc9cddd4d 2023-08-04 289: owner).fetch_all(&mut self.pool.acquire().await?).await?;
290: for row in rows.iter() {
3dc9cddd4d 2023-08-04 291: reply.push(format!("\n\\#ļøā£ {} \\*ļøā£ `{}` {}\nš `{}`", row.source_id, row.channel,
292: match row.enabled {
293: true => "š enabled",
294: false => "ā disabled",
295: }, row.url).into());
296: if let Some(hash) = &row.iv_hash {
297: reply.push(format!("IV: `{}`", hash).into());
298: }
299: if let Some(re) = &row.url_re {
300: reply.push(format!("RE: `{}`", re).into());
301: }
302: };
303: Ok(reply.join("\n"))
304: }
305: }