1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926
// Thanks to the author of shellcaster(https://github.com/jeff-hughes/shellcaster). Most parts of following code are taken from it.
#[allow(unused)]
pub mod db;
use crate::config::Settings;
use crate::types::{Msg, PCMsg};
use crate::utils::StringUtils;
use anyhow::{anyhow, Context, Result};
use chrono::{DateTime, Utc};
use db::Database;
use lazy_static::lazy_static;
use opml::{Body, Head, Outline, OPML};
use regex::{Match, Regex};
use reqwest::blocking::ClientBuilder;
use rfc822_sanitizer::parse_from_rfc2822_with_fallback;
use rss::{Channel, Item};
use sanitize_filename::{sanitize_with_options, Options};
use std::cmp::Ordering;
use std::fs::File;
use std::io::{Read, Write};
use std::path::{Path, PathBuf};
use std::sync::{
mpsc::{self, Sender},
Arc, Mutex,
};
use std::thread;
use std::time::Duration;
// How many columns we need, minimum, before we display the
// (unplayed/total) after the podcast title
pub const PODCAST_UNPLAYED_TOTALS_LENGTH: usize = 25;
// How many columns we need, minimum, before we display the duration of
// the episode
pub const EPISODE_DURATION_LENGTH: usize = 45;
// How many columns we need, minimum, before we display the pubdate
// of the episode
pub const EPISODE_PUBDATE_LENGTH: usize = 60;
lazy_static! {
/// Regex for parsing an episode "duration", which could take the form
/// of HH:MM:SS, MM:SS, or SS.
static ref RE_DURATION: Regex = Regex::new(r"(\d+)(?::(\d+))?(?::(\d+))?").expect("Regex error");
/// Regex for removing "A", "An", and "The" from the beginning of
/// podcast titles
static ref RE_ARTICLES: Regex = Regex::new(r"^(a|an|the) ").expect("Regex error");
}
/// Defines interface used for both podcasts and episodes, to be
/// used and displayed in menus.
pub trait Menuable {
fn get_id(&self) -> i64;
fn get_title(&self, length: usize) -> String;
fn is_played(&self) -> bool;
}
/// Struct holding data about an individual podcast feed. This includes a
/// (possibly empty) vector of episodes.
#[derive(Debug, Clone)]
pub struct Podcast {
pub id: i64,
pub title: String,
pub sort_title: String,
pub url: String,
pub description: Option<String>,
pub author: Option<String>,
pub explicit: Option<bool>,
pub last_checked: DateTime<Utc>,
pub episodes: Vec<Episode>,
pub image_url: Option<String>,
}
impl Podcast {
// Counts and returns the number of unplayed episodes in the podcast.
pub fn num_unplayed(&self) -> usize {
self.episodes
.iter()
.map(|ep| usize::from(!ep.is_played()))
.sum()
}
}
impl Menuable for Podcast {
/// Returns the database ID for the podcast.
fn get_id(&self) -> i64 {
self.id
}
/// Returns the title for the podcast, up to length characters.
fn get_title(&self, length: usize) -> String {
let mut title_length = length;
// if the size available is big enough, we add the unplayed data
// to the end
if length > PODCAST_UNPLAYED_TOTALS_LENGTH {
let meta_str = format!("({}/{})", self.num_unplayed(), self.episodes.len());
title_length = length - meta_str.chars().count() - 3;
let out = self.title.substr(0, title_length);
format!(
" {out} {meta_str:>width$} ",
width = length - out.grapheme_len() - 3
) // this pads spaces between title and totals
} else {
format!(" {} ", self.title.substr(0, title_length - 2))
}
}
fn is_played(&self) -> bool {
self.num_unplayed() == 0
}
}
impl PartialEq for Podcast {
fn eq(&self, other: &Self) -> bool {
self.sort_title == other.sort_title
}
}
impl Eq for Podcast {}
impl PartialOrd for Podcast {
fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
Some(self.cmp(other))
}
}
impl Ord for Podcast {
fn cmp(&self, other: &Self) -> Ordering {
self.sort_title.cmp(&other.sort_title)
}
}
/// Struct holding data about an individual podcast episode. Most of this
/// is metadata, but if the episode has been downloaded to the local
/// machine, the filepath will be included here as well. `played`
/// indicates whether the podcast has been marked as played or unplayed.
#[derive(Debug, Clone, Default)]
pub struct Episode {
pub id: i64,
pub pod_id: i64,
pub title: String,
pub url: String,
pub guid: String,
pub description: String,
pub pubdate: Option<DateTime<Utc>>,
pub duration: Option<i64>,
pub path: Option<PathBuf>,
pub played: bool,
pub last_position: Option<i64>,
pub image_url: Option<String>,
}
impl Episode {
/// Formats the duration in seconds into an HH:MM:SS format.
pub fn format_duration(&self) -> String {
match self.duration {
Some(dur) => {
let mut seconds = dur;
let hours = seconds / 3600;
seconds -= hours * 3600;
let minutes = seconds / 60;
seconds -= minutes * 60;
format!("{hours:02}:{minutes:02}:{seconds:02}")
}
None => "--:--:--".to_string(),
}
}
}
impl Menuable for Episode {
/// Returns the database ID for the episode.
fn get_id(&self) -> i64 {
self.id
}
/// Returns the title for the episode, up to length characters.
fn get_title(&self, length: usize) -> String {
let out = match self.path {
Some(_) => {
let title = self.title.substr(0, length - 4);
format!("[D] {title}")
}
None => self.title.substr(0, length),
};
if length > EPISODE_PUBDATE_LENGTH {
let dur = self.format_duration();
let meta_dur = format!("[{dur}]");
if let Some(pubdate) = self.pubdate {
// print pubdate and duration
let pd = pubdate.format("%F");
let meta_str = format!("({pd}) {meta_dur}");
let added_len = meta_str.chars().count();
let out_added = out.substr(0, length - added_len - 3);
format!(
" {out_added} {meta_str:>width$} ",
width = length - out_added.grapheme_len() - 3
)
} else {
// just print duration
let out_added = out.substr(0, length - meta_dur.chars().count() - 3);
format!(
" {out_added} {meta_dur:>width$} ",
width = length - out_added.grapheme_len() - 3
)
}
} else if length > EPISODE_DURATION_LENGTH {
let dur = self.format_duration();
let meta_dur = format!("[{dur}]");
let out_added = out.substr(0, length - meta_dur.chars().count() - 3);
format!(
" {out_added} {meta_dur:>width$} ",
width = length - out_added.grapheme_len() - 3
)
} else {
format!(" {} ", out.substr(0, length - 2))
}
}
fn is_played(&self) -> bool {
self.played
}
}
/// Struct holding data about an individual podcast feed, before it has
/// been inserted into the database. This includes a
/// (possibly empty) vector of episodes.
#[derive(Debug, Clone, Eq, PartialEq)]
#[allow(clippy::module_name_repetitions)]
pub struct PodcastNoId {
pub title: String,
pub url: String,
pub description: Option<String>,
pub author: Option<String>,
pub explicit: Option<bool>,
pub last_checked: DateTime<Utc>,
pub episodes: Vec<EpisodeNoId>,
pub image_url: Option<String>,
}
/// Struct holding data about an individual podcast episode, before it
/// has been inserted into the database.
#[derive(Debug, Clone, Eq, PartialEq)]
pub struct EpisodeNoId {
pub title: String,
pub url: String,
pub guid: String,
pub description: String,
pub pubdate: Option<DateTime<Utc>>,
pub duration: Option<i64>,
pub image_url: Option<String>,
}
/// Struct holding data about an individual podcast episode, specifically
/// for the popup window that asks users which new episodes they wish to
/// download.
#[derive(Debug, Clone)]
pub struct NewEpisode {
pub id: i64,
pub pod_id: i64,
pub title: String,
pub pod_title: String,
pub selected: bool,
}
#[derive(Debug, Clone, Eq, PartialEq)]
#[allow(clippy::module_name_repetitions)]
pub struct PodcastFeed {
pub id: Option<i64>,
pub url: String,
pub title: Option<String>,
}
impl PodcastFeed {
pub fn new(id: Option<i64>, url: &str, title: Option<String>) -> Self {
Self {
id,
url: url.to_string(),
title,
}
}
}
/// Spawns a new thread to check a feed and retrieve podcast data.
///
/// # Panics
///
/// if sending commands to `tx_to_main` fails
pub fn check_feed(
feed: PodcastFeed,
max_retries: usize,
threadpool: &Threadpool,
tx_to_main: Sender<Msg>,
) {
threadpool.execute(move || {
tx_to_main
.send(Msg::Podcast(PCMsg::FetchPodcastStart(feed.url.clone())))
.expect("thread messaging error in fetch start");
match get_feed_data(&feed.url, max_retries) {
Ok(pod) => match feed.id {
Some(id) => {
tx_to_main
.send(Msg::Podcast(PCMsg::SyncData((id, pod))))
.expect("Thread messaging error when sync old");
}
None => tx_to_main
.send(Msg::Podcast(PCMsg::NewData(pod)))
.expect("Thread messaging error when add new"),
},
Err(_err) => tx_to_main
.send(Msg::Podcast(PCMsg::Error(feed.url.to_string(), feed)))
.expect("Thread messaging error when get feed"),
}
});
}
/// Given a URL, this attempts to pull the data about a podcast and its
/// episodes from an RSS feed.
fn get_feed_data(url: &str, mut max_retries: usize) -> Result<PodcastNoId> {
let agent = ClientBuilder::new()
.connect_timeout(Duration::from_secs(5))
.build()?;
let request: Result<reqwest::blocking::Response> = loop {
let response = agent.get(url).send();
if let Ok(resp) = response {
break Ok(resp);
}
max_retries -= 1;
if max_retries == 0 {
break Err(anyhow!("No response from feed"));
}
};
match request {
Ok(mut resp) => {
// let mut reader = resp.read();
let mut resp_data = Vec::new();
// reader.read_to_end(&mut resp_data)?;
resp.read_to_end(&mut resp_data)?;
let channel = Channel::read_from(&resp_data[..])?;
Ok(parse_feed_data(channel, url))
}
Err(err) => Err(err),
}
}
/// Given a Channel with the RSS feed data, this parses the data about a
/// podcast and its episodes and returns a Podcast. There are existing
/// specifications for podcast RSS feeds that a feed should adhere to, but
/// this does try to make some attempt to account for the possibility that
/// a feed might not be valid according to the spec.
fn parse_feed_data(channel: Channel, url: &str) -> PodcastNoId {
let title = channel.title().to_string();
let url = url.to_string();
let description = Some(channel.description().to_string());
let last_checked = Utc::now();
let mut author = None;
let mut explicit = None;
let mut image_url = None;
if let Some(itunes) = channel.itunes_ext() {
author = itunes.author().map(std::string::ToString::to_string);
explicit = match itunes.explicit() {
None => None,
Some(s) => {
let ss = s.to_lowercase();
match &ss[..] {
"yes" | "explicit" | "true" => Some(true),
"no" | "clean" | "false" => Some(false),
_ => None,
}
}
};
image_url = itunes.image().map(std::string::ToString::to_string);
}
let mut episodes = Vec::new();
let items = channel.into_items();
if !items.is_empty() {
for item in &items {
episodes.push(parse_episode_data(item));
}
}
PodcastNoId {
title,
url,
description,
author,
explicit,
last_checked,
episodes,
image_url,
}
}
/// For an item (episode) in an RSS feed, this pulls data about the item
/// and converts it to an Episode. There are existing specifications for
/// podcast RSS feeds that a feed should adhere to, but this does try to
/// make some attempt to account for the possibility that a feed might
/// not be valid according to the spec.
fn parse_episode_data(item: &Item) -> EpisodeNoId {
let title = match item.title() {
Some(s) => s.to_string(),
None => String::new(),
};
let url = match item.enclosure() {
Some(enc) => enc.url().to_string(),
None => String::new(),
};
let guid = match item.guid() {
Some(guid) => guid.value().to_string(),
None => String::new(),
};
let description = match item.description() {
Some(dsc) => dsc.to_string(),
None => String::new(),
};
let pubdate = item
.pub_date()
.and_then(|pd| parse_from_rfc2822_with_fallback(pd).ok())
.map(std::convert::Into::into);
let mut duration = None;
let mut image_url = None;
if let Some(itunes) = item.itunes_ext() {
duration = duration_to_int(itunes.duration()).map(i64::from);
image_url = itunes.image().map(std::string::ToString::to_string);
}
EpisodeNoId {
title,
url,
guid,
description,
pubdate,
duration,
image_url,
}
}
/// Given a string representing an episode duration, this attempts to
/// convert to an integer representing the duration in seconds. Covers
/// formats HH:MM:SS, MM:SS, and SS. If the duration cannot be converted
/// (covering numerous reasons), it will return None.
fn duration_to_int(duration: Option<&str>) -> Option<i32> {
match duration {
Some(dur) => {
match RE_DURATION.captures(dur) {
Some(cap) => {
/*
* Provided that the regex succeeds, we should have
* 4 capture groups (with 0th being the full match).
* Depending on the string format, however, some of
* these may return None. We first loop through the
* capture groups and push Some results to an array.
* This will fail on the first non-numeric value,
* so the duration is parsed only if all components
* of it were successfully converted to integers.
* Finally, we convert hours, minutes, and seconds
* into a total duration in seconds and return.
*/
let mut times = [None; 3];
let mut counter = 0;
// cap[0] is always full match
for c in cap.iter().skip(1).flatten() {
if let Ok(intval) = regex_to_int(c) {
times[counter] = Some(intval);
counter += 1;
} else {
return None;
}
}
match counter {
// HH:MM:SS
3 => Some(
times[0].unwrap() * 60 * 60
+ times[1].unwrap() * 60
+ times[2].unwrap(),
),
// MM:SS
2 => Some(times[0].unwrap() * 60 + times[1].unwrap()),
// SS
1 => times[0],
_ => None,
}
}
None => None,
}
}
None => None,
}
}
/// Helper function converting a match from a regex capture group into an
/// integer.
fn regex_to_int(re_match: Match<'_>) -> Result<i32, std::num::ParseIntError> {
let mstr = re_match.as_str();
mstr.parse::<i32>()
}
// Much of the threadpool implementation here was taken directly from
// the Rust Book: https://doc.rust-lang.org/book/ch20-02-multithreaded.html
// and https://doc.rust-lang.org/book/ch20-03-graceful-shutdown-and-cleanup.html
/// Manages a threadpool of a given size, sending jobs to workers as
/// necessary. Implements Drop trait to allow threads to complete
/// their current jobs before being stopped.
pub struct Threadpool {
workers: Vec<Worker>,
sender: mpsc::Sender<JobMessage>,
}
impl Threadpool {
/// Creates a new Threadpool of a given size.
pub fn new(n_threads: usize) -> Threadpool {
let (sender, receiver) = mpsc::channel();
let receiver_lock = Arc::new(Mutex::new(receiver));
let mut workers = Vec::with_capacity(n_threads);
for _ in 0..n_threads {
workers.push(Worker::new(Arc::clone(&receiver_lock)));
}
Threadpool { workers, sender }
}
/// Adds a new job to the threadpool, passing closure to first
/// available worker.
///
/// # Panics
///
/// if sending commands to the sender fails
pub fn execute<F>(&self, func: F)
where
F: FnOnce() + Send + 'static,
{
let job = Box::new(func);
self.sender
.send(JobMessage::NewJob(job))
.expect("Thread messaging error");
}
}
impl Drop for Threadpool {
/// Upon going out of scope, Threadpool sends terminate message to
/// all workers but allows them to complete current jobs.
fn drop(&mut self) {
for _ in &self.workers {
self.sender
.send(JobMessage::Terminate)
.expect("Thread messaging error");
}
for worker in &mut self.workers {
if let Some(thread) = worker.thread.take() {
// joins to ensure threads finish job before stopping
thread.join().expect("Error dropping threads");
}
}
}
}
type Job = Box<dyn FnOnce() + Send + 'static>;
/// Messages used by Threadpool to communicate with Workers.
enum JobMessage {
NewJob(Job),
Terminate,
}
/// Used by Threadpool to complete jobs. Each Worker manages a single
/// thread.
struct Worker {
thread: Option<thread::JoinHandle<()>>,
}
impl Worker {
/// Creates a new Worker, which waits for Jobs to be passed by the
/// Threadpool.
#[allow(clippy::missing_panics_doc)]
fn new(receiver: Arc<Mutex<mpsc::Receiver<JobMessage>>>) -> Worker {
let thread = thread::spawn(move || loop {
let message = receiver
.lock()
.expect("Threadpool error")
.recv()
.expect("Thread messaging error");
match message {
JobMessage::NewJob(job) => job(),
JobMessage::Terminate => break,
}
});
Worker {
thread: Some(thread),
}
}
}
/// Imports a list of podcasts from OPML format, either reading from a
/// file or from stdin. If the `replace` flag is set, this replaces all
/// existing data in the database.
pub fn import_from_opml(db_path: &Path, config: &Settings, filepath: &str) -> Result<()> {
// read from file or from stdin
let mut f =
File::open(filepath).with_context(|| format!("Could not open OPML file: {filepath}"))?;
let mut contents = String::new();
f.read_to_string(&mut contents)
.with_context(|| format!("Failed to read from OPML file: {filepath}"))?;
let xml = contents;
let mut podcast_list = import_opml_feeds(&xml).with_context(|| {
"Could not properly parse OPML file -- file may be formatted improperly or corrupted."
})?;
if podcast_list.is_empty() {
println!("No podcasts to import.");
return Ok(());
}
let db_inst = db::Database::connect(db_path)?;
// delete database if we are replacing the data
// if args.is_present("replace") {
// db_inst
// .clear_db()
// .with_context(|| "Error clearing database")?;
// } else {
let old_podcasts = db_inst.get_podcasts()?;
// if URL is already in database, remove it from import
podcast_list.retain(|pod| {
for op in &old_podcasts {
if pod.url == op.url {
return false;
}
}
true
});
// }
// check again, now that we may have removed feeds after looking at
// the database
if podcast_list.is_empty() {
println!("No podcasts to import.");
return Ok(());
}
println!("Importing {} podcasts...", podcast_list.len());
let threadpool = Threadpool::new(config.podcast_simultanious_download);
let (tx_to_main, rx_to_main) = mpsc::channel();
for pod in &podcast_list {
check_feed(
pod.clone(),
config.podcast_max_retries,
&threadpool,
tx_to_main.clone(),
);
}
let mut msg_counter: usize = 0;
let mut failure = false;
while let Some(message) = rx_to_main.iter().next() {
match message {
Msg::Podcast(PCMsg::NewData(pod)) => {
msg_counter += 1;
let title = pod.title.clone();
let db_result = db_inst.insert_podcast(&pod);
match db_result {
Ok(_) => {
println!("Added {title}");
}
Err(_err) => {
failure = true;
error!("Error adding {title}");
}
}
}
Msg::Podcast(PCMsg::Error(_, feed)) => {
msg_counter += 1;
failure = true;
if let Some(t) = feed.title {
error!("Error retrieving RSS feed: {t}");
} else {
error!("Error retrieving RSS feed");
}
}
Msg::Podcast(PCMsg::SyncData((_id, _pod))) => {
msg_counter += 1;
}
_ => {}
}
if msg_counter >= podcast_list.len() {
break;
}
}
if failure {
return Err(anyhow!("Process finished with errors."));
}
println!("Import successful.");
Ok(())
}
/// Exports all podcasts to OPML format, either printing to stdout or
/// exporting to a file.
pub fn export_to_opml(db_path: &Path, file: &str) -> Result<()> {
let db_inst = Database::connect(db_path)?;
let podcast_list = db_inst.get_podcasts()?;
let opml = export_opml_feeds(&podcast_list);
let xml = opml
.to_string()
.map_err(|err| anyhow!(err))
.with_context(|| "Could not create OPML format")?;
let mut dst =
File::create(file).with_context(|| format!("Could not create output file: {file}"))?;
dst.write_all(xml.as_bytes())
.with_context(|| format!("Could not copy OPML data to output file: {file}"))?;
Ok(())
}
/// Import a list of podcast feeds from an OPML file. Supports
/// v1.0, v1.1, and v2.0 OPML files.
fn import_opml_feeds(xml: &str) -> Result<Vec<PodcastFeed>> {
match OPML::from_str(xml) {
Err(err) => Err(anyhow!(err)),
Ok(opml) => {
let mut feeds = Vec::new();
for pod in opml.body.outlines {
if pod.xml_url.is_some() {
// match against title attribute first -- if this is
// not set or empty, then match against the text
// attribute; this must be set, but can be empty
let temp_title = pod.title.filter(|t| !t.is_empty());
let title = match temp_title {
Some(t) => Some(t),
None => {
if pod.text.is_empty() {
None
} else {
Some(pod.text)
}
}
};
feeds.push(PodcastFeed::new(None, &pod.xml_url.unwrap(), title));
}
}
Ok(feeds)
}
}
}
/// Converts the current set of podcast feeds to the OPML format
fn export_opml_feeds(podcasts: &[Podcast]) -> OPML {
let date = Utc::now();
let mut opml = OPML {
head: Some(Head {
title: Some("Termusic Podcast Feeds".to_string()),
date_created: Some(date.to_rfc2822()),
..Head::default()
}),
..Default::default()
};
let mut outlines = Vec::new();
for pod in podcasts {
// opml.add_feed(&pod.title, &pod.url);
outlines.push(Outline {
text: pod.title.clone(),
r#type: Some("rss".to_string()),
xml_url: Some(pod.url.clone()),
title: Some(pod.title.clone()),
..Outline::default()
});
}
opml.body = Body { outlines };
opml
}
/// Enum used to communicate relevant data to the threadpool.
#[derive(Debug, Clone, Eq, PartialEq)]
pub struct EpData {
pub id: i64,
pub pod_id: i64,
pub title: String,
pub url: String,
pub pubdate: Option<DateTime<Utc>>,
pub file_path: Option<PathBuf>,
}
/// This is the function the main controller uses to indicate new
/// files to download. It uses the threadpool to start jobs
/// for every episode to be downloaded. New jobs can be requested
/// by the user while there are still ongoing jobs.
#[allow(clippy::missing_panics_doc)]
pub fn download_list(
episodes: Vec<EpData>,
dest: &Path,
max_retries: usize,
threadpool: &Threadpool,
tx_to_main: &Sender<Msg>,
) {
// parse episode details and push to queue
for ep in episodes {
let tx = tx_to_main.clone();
let dest2 = dest.to_path_buf();
threadpool.execute(move || download_job(&tx, ep, dest2, max_retries));
}
}
/// # Panics
///
/// if sending command via `tx_to_main` fails
fn download_job(tx_to_main: &Sender<Msg>, ep: EpData, dest: PathBuf, max_retries: usize) {
tx_to_main
.send(Msg::Podcast(PCMsg::DLStart(ep.clone())))
.expect("Thread messaging error when start download");
let result = download_file(ep, dest, max_retries);
tx_to_main
.send(Msg::Podcast(result))
.expect("Thread messaging error");
}
/// Downloads a file to a local filepath, returning `DownloadMsg` variant
/// indicating success or failure.
#[allow(clippy::single_match_else)]
fn download_file(mut ep_data: EpData, destination_path: PathBuf, mut max_retries: usize) -> PCMsg {
let agent = ClientBuilder::new()
.connect_timeout(Duration::from_secs(10))
.build()
.expect("reqwest client build failed");
let request: Result<reqwest::blocking::Response, ()> = loop {
let response = agent.get(&ep_data.url).send();
match response {
Ok(resp) => break Ok(resp),
Err(_) => {
max_retries -= 1;
if max_retries == 0 {
break Err(());
}
}
}
};
if request.is_err() {
return PCMsg::DLResponseError(ep_data);
};
let mut response = request.unwrap();
// figure out the file type
let mut ext = "mp3";
match response.headers().get("content-type") {
Some(content_type) => match content_type.to_str() {
Ok("audio/x-m4a") => ext = "m4a",
// Ok("audio/mpeg") => ext = "mp3",
Ok("video/quicktime") => ext = "mov",
Ok("video/mp4") => ext = "mp4",
Ok("video/x-m4v") => ext = "m4v",
Ok(_) | Err(_) => ext = "mp3",
},
None => error!("The response doesn't contain a content type"),
// Some("audio/x-m4a") => "m4a",
// // Some("audio/mpeg") => "mp3",
// Some("video/quicktime") => "mov",
// Some("video/mp4") => "mp4",
// Some("video/x-m4v") => "m4v",
// _ => "mp3", // assume .mp3 unless we figure out otherwise
};
let mut file_name = sanitize_with_options(
&ep_data.title,
Options {
truncate: true,
windows: true, // for simplicity, we'll just use Windows-friendly paths for everyone
replacement: "",
},
);
if let Some(pubdate) = ep_data.pubdate {
file_name = format!("{file_name}_{}", pubdate.format("%Y%m%d_%H%M%S"));
}
let mut file_path = destination_path;
file_path.push(format!("{file_name}.{ext}"));
let dst = File::create(&file_path);
if dst.is_err() {
return PCMsg::DLFileCreateError(ep_data);
};
ep_data.file_path = Some(file_path);
let mut resp_data = Vec::new();
let _drop = response.read_to_end(&mut resp_data);
match dst.unwrap().write_all(&resp_data) {
// match std::io::copy(&mut resp_data, &mut dst.unwrap()) {
Ok(()) => PCMsg::DLComplete(ep_data),
Err(_) => PCMsg::DLFileWriteError(ep_data),
}
}