751 lines
26 KiB
Rust
751 lines
26 KiB
Rust
use crate::api::ClientVersion;
|
|
use crate::providers::Provider;
|
|
use crate::status::*;
|
|
use crate::util::cache::VideoCache;
|
|
use crate::util::requester::Requester;
|
|
use crate::videos::{ServerOptions, VideoFormat, VideoItem};
|
|
use crate::{DbPool, db};
|
|
|
|
use async_trait::async_trait;
|
|
use chrono::NaiveDateTime;
|
|
use error_chain::error_chain;
|
|
use futures::stream::{self, StreamExt};
|
|
use htmlentity::entity::{ICodedDataTrait, decode};
|
|
use regex::Regex;
|
|
use serde::Deserialize;
|
|
use std::collections::{HashMap, HashSet};
|
|
use std::sync::{Arc, Mutex, OnceLock, RwLock};
|
|
use std::vec;
|
|
use wreq::Version;
|
|
|
|
// How long a cached listing/search entry is considered usable at all.
|
|
const HARD_TTL_SECS: u64 = 60 * 60 * 24;
|
|
// Past this age we still answer instantly from cache/DB but trigger a
|
|
// background refresh so the next request gets fresh data.
|
|
const SOFT_TTL_SECS: u64 = 60 * 60;
|
|
|
|
const SITE: &str = "https://hentaihaven.xxx";
|
|
// Poster art lives on a separate image host; the catalogue only stores the path.
|
|
const IMG_BASE: &str = "https://img.hentaihaven.xxx/";
|
|
// The catalogue API rejects the request outright above this page size.
|
|
const MAX_PER_PAGE: usize = 40;
|
|
|
|
pub const CHANNEL_METADATA: crate::providers::ProviderChannelMetadata =
|
|
crate::providers::ProviderChannelMetadata {
|
|
group_id: "hentai-animation",
|
|
tags: &["hentai", "anime", "curated"],
|
|
};
|
|
|
|
error_chain! {
|
|
foreign_links {
|
|
Io(std::io::Error);
|
|
HttpRequest(wreq::Error);
|
|
Json(serde_json::Error);
|
|
}
|
|
errors {
|
|
Parse(msg: String) {
|
|
description("parse error")
|
|
display("parse error: {}", msg)
|
|
}
|
|
}
|
|
}
|
|
|
|
#[derive(Debug, Deserialize, Default)]
|
|
struct ApiTitle {
|
|
#[serde(default)]
|
|
rendered: String,
|
|
}
|
|
|
|
#[derive(Debug, Deserialize, Default)]
|
|
struct ApiMeta {
|
|
#[serde(default)]
|
|
vraven_remote_thumbnail: String,
|
|
}
|
|
|
|
#[derive(Debug, Deserialize)]
|
|
struct ApiEntry {
|
|
id: u64,
|
|
#[serde(default)]
|
|
date: String,
|
|
slug: String,
|
|
#[serde(default)]
|
|
title: ApiTitle,
|
|
#[serde(default)]
|
|
meta: ApiMeta,
|
|
}
|
|
|
|
#[derive(Debug, Deserialize, Default)]
|
|
struct ApiEngagement {
|
|
#[serde(default)]
|
|
rating: Option<f32>,
|
|
#[serde(default)]
|
|
views: Option<u32>,
|
|
}
|
|
|
|
#[derive(Debug, Deserialize)]
|
|
struct ApiResponse {
|
|
#[serde(default)]
|
|
data: Vec<ApiEntry>,
|
|
#[serde(default)]
|
|
engagement: HashMap<String, ApiEngagement>,
|
|
}
|
|
|
|
/// Everything the catalogue API already tells us about a series, so resolving
|
|
/// an item only has to go looking for its episodes.
|
|
#[derive(Debug, Clone)]
|
|
struct ListingEntry {
|
|
id: u64,
|
|
slug: String,
|
|
url: String,
|
|
title: String,
|
|
thumb: String,
|
|
views: u32,
|
|
rating: f32,
|
|
uploaded_at: u64,
|
|
}
|
|
|
|
#[derive(Debug, Clone)]
|
|
pub struct HentaihavenProvider {
|
|
categories: Arc<RwLock<Vec<FilterOption>>>,
|
|
}
|
|
|
|
impl HentaihavenProvider {
|
|
pub fn new() -> Self {
|
|
Self {
|
|
categories: Arc::new(RwLock::new(vec![])),
|
|
}
|
|
}
|
|
|
|
fn build_channel(&self, clientversion: ClientVersion) -> Channel {
|
|
let _ = clientversion;
|
|
Channel {
|
|
id: "hentaihaven".to_string(),
|
|
name: "Hentai Haven".to_string(),
|
|
description: "Watch Free Hentai Videos HD!".to_string(),
|
|
premium: false,
|
|
favicon: "https://www.google.com/s2/favicons?sz=64&domain=hentaihaven.xxx".to_string(),
|
|
status: "active".to_string(),
|
|
categories: self
|
|
.categories
|
|
.read()
|
|
.map(|categories| categories.iter().map(|c| c.title.clone()).collect())
|
|
.unwrap_or_else(|e| {
|
|
crate::providers::report_provider_error_background(
|
|
"hentaihaven",
|
|
"build_channel.categories_read",
|
|
&e.to_string(),
|
|
);
|
|
vec![]
|
|
}),
|
|
options: vec![],
|
|
nsfw: true,
|
|
cacheDuration: None,
|
|
ytdlpCommand: Some("yt-dlp --impersonate chrome-120".to_string()),
|
|
}
|
|
}
|
|
|
|
fn push_unique(target: &Arc<RwLock<Vec<FilterOption>>>, item: FilterOption) {
|
|
if let Ok(mut vec) = target.write() {
|
|
if !vec.iter().any(|x| x.id == item.id) {
|
|
vec.push(item);
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Build the catalogue API URL backing a listing or search request. The URL
|
|
/// doubles as the cache key, so every parameter that changes the result set
|
|
/// has to be part of it.
|
|
fn api_url(page: u8, per_page: usize, sort: &str, query: Option<&str>) -> String {
|
|
let per_page = per_page.clamp(1, MAX_PER_PAGE);
|
|
let mut serializer = url::form_urlencoded::Serializer::new(String::new());
|
|
serializer
|
|
.append_pair("per_page", &per_page.to_string())
|
|
.append_pair("page", &page.max(1).to_string());
|
|
if let Some(query) = query {
|
|
serializer.append_pair("search", query);
|
|
} else if matches!(
|
|
sort,
|
|
"views" | "popular" | "most-viewed" | "trending" | "hot" | "rating"
|
|
) {
|
|
// The catalogue's ranked chart, which is what the site itself serves
|
|
// on /browse/trending/.
|
|
serializer
|
|
.append_pair("trending_period", "monthly")
|
|
.append_pair("live", "1");
|
|
}
|
|
format!("{SITE}/api/manga/?{}", serializer.finish())
|
|
}
|
|
|
|
async fn list(
|
|
&self,
|
|
cache: VideoCache,
|
|
page: u8,
|
|
per_page: usize,
|
|
sort: &str,
|
|
query: Option<&str>,
|
|
options: ServerOptions,
|
|
pool: DbPool,
|
|
) -> Result<Vec<VideoItem>> {
|
|
let api_url = Self::api_url(page, per_page, sort, query);
|
|
|
|
// Fast path: a usable in-memory entry exists. Answer immediately; once it
|
|
// is older than the soft TTL, kick a background refresh so the next caller
|
|
// sees fresher data without anyone waiting on it now.
|
|
if let Some((time, items)) = cache.get(&api_url) {
|
|
let age = time.elapsed().unwrap_or_default().as_secs();
|
|
if age < HARD_TTL_SECS && !items.is_empty() {
|
|
if age >= SOFT_TTL_SECS {
|
|
let requester = crate::providers::requester_or_default(
|
|
&options,
|
|
module_path!(),
|
|
"missing_requester",
|
|
);
|
|
self.spawn_refresh(requester, pool, cache, api_url, None);
|
|
}
|
|
return Ok(items);
|
|
}
|
|
}
|
|
|
|
// One cheap JSON request gives us the whole page of series with titles,
|
|
// posters and engagement counts already attached.
|
|
let mut requester =
|
|
crate::providers::requester_or_default(&options, module_path!(), "missing_requester");
|
|
let entries = match Self::fetch_listing(&mut requester, &api_url).await {
|
|
Ok(entries) => entries,
|
|
Err(e) => {
|
|
crate::providers::report_provider_error(
|
|
"hentaihaven",
|
|
"list.request",
|
|
&format!("url={api_url}; error={e}"),
|
|
)
|
|
.await;
|
|
return Ok(Self::cached(&cache, &api_url));
|
|
}
|
|
};
|
|
if entries.is_empty() {
|
|
return Ok(Self::cached(&cache, &api_url));
|
|
}
|
|
|
|
// Serve the page straight from the DB when it covers *every* series on it,
|
|
// then refresh the whole listing in the background. Partial coverage is
|
|
// not good enough: returning only the handful of already-resolved series
|
|
// would silently hand the client a short page.
|
|
let db_items = Self::items_from_db(&entries, &pool);
|
|
if db_items.len() == entries.len() {
|
|
cache.insert(api_url.clone(), db_items.clone());
|
|
self.spawn_refresh(requester, pool, cache, api_url, Some(entries));
|
|
return Ok(db_items);
|
|
}
|
|
|
|
// Cold start (or a page with newly published series): resolve now so the
|
|
// caller gets the full page.
|
|
let items = self.resolve_entries(entries, &requester, pool).await;
|
|
if !items.is_empty() {
|
|
cache.insert(api_url.clone(), items.clone());
|
|
return Ok(items);
|
|
}
|
|
Ok(Self::cached(&cache, &api_url))
|
|
}
|
|
|
|
fn cached(cache: &VideoCache, key: &str) -> Vec<VideoItem> {
|
|
cache.get(key).map(|(_, items)| items).unwrap_or_default()
|
|
}
|
|
|
|
/// Fetch and flatten one page of the catalogue API.
|
|
async fn fetch_listing(
|
|
requester: &mut Requester,
|
|
api_url: &str,
|
|
) -> std::result::Result<Vec<ListingEntry>, String> {
|
|
let body = Self::get_with_retry(requester, api_url, 3).await?;
|
|
let response: ApiResponse = serde_json::from_str(&body).map_err(|e| {
|
|
format!(
|
|
"failed to parse catalogue json: {e}; body={}",
|
|
body.chars().take(200).collect::<String>()
|
|
)
|
|
})?;
|
|
|
|
Ok(response
|
|
.data
|
|
.into_iter()
|
|
.map(|entry| {
|
|
let engagement = response.engagement.get(&entry.id.to_string());
|
|
let title = decode(entry.title.rendered.as_bytes())
|
|
.to_string()
|
|
.unwrap_or_else(|_| entry.title.rendered.clone());
|
|
ListingEntry {
|
|
id: entry.id,
|
|
url: format!("{SITE}/watch/{}/", entry.slug),
|
|
slug: entry.slug,
|
|
title,
|
|
thumb: Self::thumb_url(&entry.meta.vraven_remote_thumbnail),
|
|
views: engagement.and_then(|e| e.views).unwrap_or(0),
|
|
rating: engagement.and_then(|e| e.rating).unwrap_or(0.0),
|
|
uploaded_at: Self::parse_timestamp(&entry.date),
|
|
}
|
|
})
|
|
.collect())
|
|
}
|
|
|
|
/// Poster paths are stored raw and routinely contain spaces, so they have to
|
|
/// go through the URL parser rather than plain concatenation.
|
|
fn thumb_url(path: &str) -> String {
|
|
let path = path.trim();
|
|
if path.is_empty() {
|
|
return String::new();
|
|
}
|
|
if path.starts_with("http") {
|
|
return path.to_string();
|
|
}
|
|
match url::Url::parse(&format!("{IMG_BASE}{}", path.trim_start_matches('/'))) {
|
|
Ok(url) => url.to_string(),
|
|
Err(_) => String::new(),
|
|
}
|
|
}
|
|
|
|
/// Catalogue dates are naive local timestamps like `2026-09-01T01:27:10`.
|
|
fn parse_timestamp(date: &str) -> u64 {
|
|
NaiveDateTime::parse_from_str(date.trim(), "%Y-%m-%dT%H:%M:%S")
|
|
.map(|dt| dt.and_utc().timestamp().max(0) as u64)
|
|
.unwrap_or(0)
|
|
}
|
|
|
|
/// Build a response from already-resolved items stored in the DB, preserving
|
|
/// the order of `entries`. Items not yet in the DB are simply skipped.
|
|
fn items_from_db(entries: &[ListingEntry], pool: &DbPool) -> Vec<VideoItem> {
|
|
let mut conn = match pool.get() {
|
|
Ok(conn) => conn,
|
|
Err(_) => return vec![],
|
|
};
|
|
entries
|
|
.iter()
|
|
.filter_map(|entry| match db::get_video(&mut conn, entry.url.clone()) {
|
|
Ok(Some(json)) => VideoItem::from(json).ok(),
|
|
_ => None,
|
|
})
|
|
.collect()
|
|
}
|
|
|
|
/// Resolve each series into a full `VideoItem`, persisting every success to
|
|
/// the DB. On failure we fall back to any stored copy so a transient error
|
|
/// does not drop the item from the page.
|
|
async fn resolve_entries(
|
|
&self,
|
|
entries: Vec<ListingEntry>,
|
|
requester: &Requester,
|
|
pool: DbPool,
|
|
) -> Vec<VideoItem> {
|
|
stream::iter(entries.into_iter().map(|entry| {
|
|
let provider = self.clone();
|
|
let mut req = requester.clone();
|
|
let pool = pool.clone();
|
|
async move {
|
|
let url = entry.url.clone();
|
|
match provider.fetch_video_item(&entry, &mut req).await {
|
|
Ok((item, complete)) => {
|
|
if let Ok(mut conn) = pool.get() {
|
|
let old_item = db::get_video(&mut conn, url.clone())
|
|
.ok()
|
|
.flatten()
|
|
.and_then(|json| VideoItem::from(json).ok());
|
|
let old_len = old_item
|
|
.as_ref()
|
|
.and_then(|o| o.formats.as_ref())
|
|
.map_or(0, |f| f.len());
|
|
let new_len = item.formats.as_ref().map_or(0, |f| f.len());
|
|
if complete || new_len >= old_len {
|
|
// Every episode the series page advertises resolved,
|
|
// so this is authoritative even when it is shorter
|
|
// than what we stored earlier.
|
|
let _ = db::upsert_video(
|
|
&mut conn,
|
|
&url,
|
|
&serde_json::to_string(&item).unwrap_or_default(),
|
|
);
|
|
Some(item)
|
|
} else {
|
|
// Some episode pages failed this round — keep the
|
|
// richer stored copy rather than degrading it.
|
|
old_item.or(Some(item))
|
|
}
|
|
} else {
|
|
Some(item)
|
|
}
|
|
}
|
|
Err(e) => {
|
|
eprintln!("Hentai Haven Provider: Failed to resolve {url}: {e}");
|
|
if let Ok(mut conn) = pool.get() {
|
|
if let Ok(Some(cached)) = db::get_video(&mut conn, url.clone()) {
|
|
if let Ok(item) = VideoItem::from(cached) {
|
|
return Some(item);
|
|
}
|
|
}
|
|
}
|
|
None
|
|
}
|
|
}
|
|
}
|
|
}))
|
|
.buffered(2)
|
|
.filter_map(|item| async move { item })
|
|
.collect::<Vec<_>>()
|
|
.await
|
|
}
|
|
|
|
/// Per-listing in-flight guard so we never run two background refreshes for
|
|
/// the same page concurrently.
|
|
fn refresh_in_flight() -> &'static Mutex<HashSet<String>> {
|
|
static SET: OnceLock<Mutex<HashSet<String>>> = OnceLock::new();
|
|
SET.get_or_init(|| Mutex::new(HashSet::new()))
|
|
}
|
|
|
|
fn try_begin_refresh(key: &str) -> bool {
|
|
match Self::refresh_in_flight().lock() {
|
|
Ok(mut set) => set.insert(key.to_string()),
|
|
Err(_) => false,
|
|
}
|
|
}
|
|
|
|
fn end_refresh(key: &str) {
|
|
if let Ok(mut set) = Self::refresh_in_flight().lock() {
|
|
set.remove(key);
|
|
}
|
|
}
|
|
|
|
/// Spawn a non-blocking refresh of a listing/search page. `entries` may be
|
|
/// supplied when the caller already fetched the catalogue page; otherwise the
|
|
/// refresh re-fetches it itself.
|
|
fn spawn_refresh(
|
|
&self,
|
|
requester: Requester,
|
|
pool: DbPool,
|
|
cache: VideoCache,
|
|
key: String,
|
|
entries: Option<Vec<ListingEntry>>,
|
|
) {
|
|
if !Self::try_begin_refresh(&key) {
|
|
return;
|
|
}
|
|
let provider = self.clone();
|
|
tokio::spawn(async move {
|
|
provider
|
|
.refresh(requester, pool, cache, key.clone(), entries)
|
|
.await;
|
|
Self::end_refresh(&key);
|
|
});
|
|
}
|
|
|
|
async fn refresh(
|
|
&self,
|
|
mut requester: Requester,
|
|
pool: DbPool,
|
|
cache: VideoCache,
|
|
key: String,
|
|
entries: Option<Vec<ListingEntry>>,
|
|
) {
|
|
let entries = match entries {
|
|
Some(entries) => entries,
|
|
None => match Self::fetch_listing(&mut requester, &key).await {
|
|
Ok(entries) => entries,
|
|
Err(e) => {
|
|
crate::providers::report_provider_error_background(
|
|
"hentaihaven",
|
|
"refresh.request",
|
|
&format!("url={key}; error={e}"),
|
|
);
|
|
return;
|
|
}
|
|
},
|
|
};
|
|
if entries.is_empty() {
|
|
return;
|
|
}
|
|
let items = self.resolve_entries(entries, &requester, pool).await;
|
|
if !items.is_empty() {
|
|
cache.insert(key, items);
|
|
}
|
|
}
|
|
|
|
fn episode_regex(slug: &str) -> Result<Regex> {
|
|
Regex::new(&format!(
|
|
r#"href="(/watch/{}/([^"/]+)/)""#,
|
|
regex::escape(slug)
|
|
))
|
|
.map_err(|e| Error::from(format!("failed to build episode regex: {e}")))
|
|
}
|
|
|
|
fn genre_regex() -> &'static Regex {
|
|
static RE: OnceLock<Regex> = OnceLock::new();
|
|
RE.get_or_init(|| {
|
|
Regex::new(r#"href="/series/([a-z0-9-]+)/"[^>]*>([^<]+)<"#)
|
|
.expect("valid hentaihaven genre regex")
|
|
})
|
|
}
|
|
|
|
fn content_url_regex() -> &'static Regex {
|
|
static RE: OnceLock<Regex> = OnceLock::new();
|
|
RE.get_or_init(|| {
|
|
Regex::new(r#""contentUrl"\s*:\s*"([^"]+)""#)
|
|
.expect("valid hentaihaven contentUrl regex")
|
|
})
|
|
}
|
|
|
|
fn duration_regex() -> &'static Regex {
|
|
static RE: OnceLock<Regex> = OnceLock::new();
|
|
RE.get_or_init(|| {
|
|
Regex::new(r#""duration"\s*:\s*"(PT[^"]+)""#)
|
|
.expect("valid hentaihaven duration regex")
|
|
})
|
|
}
|
|
|
|
/// The series page also renders sidebars full of *other* series' episodes, so
|
|
/// every candidate link is scoped to this series' own slug.
|
|
fn parse_episode_paths(html: &str, slug: &str) -> Result<Vec<(u32, String)>> {
|
|
let regex = Self::episode_regex(slug)?;
|
|
let mut seen: HashSet<String> = HashSet::new();
|
|
let mut episodes: Vec<(u32, String)> = vec![];
|
|
for capture in regex.captures_iter(html) {
|
|
let path = capture[1].to_string();
|
|
let segment = capture[2].to_string();
|
|
if !segment.starts_with("episode") {
|
|
continue;
|
|
}
|
|
if !seen.insert(path.clone()) {
|
|
continue;
|
|
}
|
|
let number = segment
|
|
.rsplit('-')
|
|
.next()
|
|
.and_then(|s| s.parse::<u32>().ok())
|
|
.unwrap_or(u32::MAX);
|
|
episodes.push((number, format!("{SITE}{path}")));
|
|
}
|
|
episodes.sort_by_key(|(number, _)| *number);
|
|
Ok(episodes)
|
|
}
|
|
|
|
fn parse_genres(html: &str) -> Vec<FilterOption> {
|
|
let mut seen: HashSet<String> = HashSet::new();
|
|
Self::genre_regex()
|
|
.captures_iter(html)
|
|
.filter_map(|capture| {
|
|
let id = capture[1].to_string();
|
|
let title = decode(capture[2].as_bytes())
|
|
.to_string()
|
|
.unwrap_or_else(|_| capture[2].to_string())
|
|
.trim()
|
|
.to_string();
|
|
if title.is_empty() || !seen.insert(id.clone()) {
|
|
return None;
|
|
}
|
|
Some(FilterOption { id, title })
|
|
})
|
|
.collect()
|
|
}
|
|
|
|
/// Parse an ISO-8601 duration such as `PT6M32S` into seconds.
|
|
fn parse_iso8601_duration(value: &str) -> u32 {
|
|
let mut seconds = 0u32;
|
|
let mut number = 0u32;
|
|
for c in value.trim_start_matches("PT").chars() {
|
|
match c {
|
|
'0'..='9' => number = number.saturating_mul(10).saturating_add(c as u32 - '0' as u32),
|
|
'H' => {
|
|
seconds = seconds.saturating_add(number.saturating_mul(3600));
|
|
number = 0;
|
|
}
|
|
'M' => {
|
|
seconds = seconds.saturating_add(number.saturating_mul(60));
|
|
number = 0;
|
|
}
|
|
'S' => {
|
|
seconds = seconds.saturating_add(number);
|
|
number = 0;
|
|
}
|
|
_ => number = 0,
|
|
}
|
|
}
|
|
seconds
|
|
}
|
|
|
|
/// Resolve one series into a `VideoItem`. The boolean is `true` when every
|
|
/// episode the series page advertises produced a playable format, which lets
|
|
/// the caller tell a genuinely short series from a partly failed fetch.
|
|
async fn fetch_video_item(
|
|
&self,
|
|
entry: &ListingEntry,
|
|
requester: &mut Requester,
|
|
) -> Result<(VideoItem, bool)> {
|
|
let html = Self::get_with_retry(requester, &entry.url, 3)
|
|
.await
|
|
.map_err(|e| Error::from(format!("Failed to fetch series page: {e}")))?;
|
|
|
|
let genres = Self::parse_genres(&html);
|
|
for genre in &genres {
|
|
Self::push_unique(&self.categories, genre.clone());
|
|
}
|
|
let tags: Vec<String> = genres.into_iter().map(|g| g.title).collect();
|
|
|
|
let episodes = Self::parse_episode_paths(&html, &entry.slug)?;
|
|
let expected = episodes.len();
|
|
let formats: Vec<(VideoFormat, u32)> = if episodes.is_empty() {
|
|
// A handful of one-shot entries render the player straight onto the
|
|
// series page instead of linking out to an episode.
|
|
vec![Self::format_from_player_html(&html, "Episode 1").ok_or_else(|| {
|
|
ErrorKind::Parse(format!("no episodes and no player on {}", entry.url))
|
|
})?]
|
|
} else {
|
|
stream::iter(episodes.into_iter().map(|(number, url)| {
|
|
let mut req = requester.clone();
|
|
async move {
|
|
let label = if number == u32::MAX {
|
|
"Episode".to_string()
|
|
} else {
|
|
format!("Episode {number}")
|
|
};
|
|
match Self::get_with_retry(&mut req, &url, 3).await {
|
|
Ok(html) => Self::format_from_player_html(&html, &label),
|
|
Err(e) => {
|
|
eprintln!("Hentai Haven Provider: Failed to fetch episode {url}: {e}");
|
|
None
|
|
}
|
|
}
|
|
}
|
|
}))
|
|
.buffered(2)
|
|
.filter_map(|format| async move { format })
|
|
.collect::<Vec<_>>()
|
|
.await
|
|
};
|
|
|
|
if formats.is_empty() {
|
|
return Err(Error::from(format!(
|
|
"No formats found for series: {}",
|
|
entry.url
|
|
)));
|
|
}
|
|
|
|
// A multi-episode series plays its first episode by default, so that is
|
|
// the duration the client should show.
|
|
let format_count = formats.len();
|
|
let duration = formats.first().map(|(_, secs)| *secs).unwrap_or(0);
|
|
let mut title = entry.title.clone();
|
|
if format_count > 1 {
|
|
title = format!("{} ({} Episodes)", title, format_count);
|
|
}
|
|
|
|
let mut item = VideoItem::new(
|
|
entry.id.to_string(),
|
|
title,
|
|
entry.url.clone(),
|
|
"hentaihaven".into(),
|
|
entry.thumb.clone(),
|
|
duration,
|
|
)
|
|
.formats(formats.into_iter().map(|(format, _)| format).collect())
|
|
.tags(tags)
|
|
.views(entry.views)
|
|
.aspect_ratio(0.667);
|
|
|
|
if entry.rating > 0.0 {
|
|
item = item.rating(entry.rating);
|
|
}
|
|
if entry.uploaded_at > 0 {
|
|
item = item.uploaded_at(entry.uploaded_at);
|
|
}
|
|
|
|
// The single-player fallback has nothing to compare against, so it only
|
|
// gets here when it did produce a format — treat that as complete too.
|
|
let complete = expected == 0 || format_count == expected;
|
|
Ok((item, complete))
|
|
}
|
|
|
|
/// Episode pages publish the stream in their `VideoObject` JSON-LD block, so
|
|
/// no player handshake is needed any more. The page also carries an
|
|
/// `ImageObject` whose `contentUrl` is the poster and which is emitted
|
|
/// *before* the video, so the playlist has to be picked out by extension
|
|
/// rather than by taking the first match.
|
|
fn format_from_player_html(html: &str, label: &str) -> Option<(VideoFormat, u32)> {
|
|
let src = Self::content_url_regex()
|
|
.captures_iter(html)
|
|
.map(|c| c[1].to_string())
|
|
.find(|url| Self::is_stream_url(url))?;
|
|
let duration = Self::duration_regex()
|
|
.captures(html)
|
|
.map(|c| Self::parse_iso8601_duration(&c[1]))
|
|
.unwrap_or(0);
|
|
let container = if src.to_lowercase().contains(".mp4") {
|
|
"mp4"
|
|
} else {
|
|
"m3u8"
|
|
};
|
|
let format = VideoFormat::new(src, "auto".to_string(), container.to_string())
|
|
.format_id(label.to_string())
|
|
.format_note(label.to_string());
|
|
Some((format, duration))
|
|
}
|
|
|
|
/// Distinguishes the `VideoObject` stream from the poster art that shares
|
|
/// the `contentUrl` key.
|
|
fn is_stream_url(url: &str) -> bool {
|
|
let path = url.split(['?', '#']).next().unwrap_or(url).to_lowercase();
|
|
path.ends_with(".m3u8") || path.ends_with(".mp4")
|
|
}
|
|
|
|
async fn get_with_retry(
|
|
requester: &mut Requester,
|
|
url: &str,
|
|
attempts: u32,
|
|
) -> std::result::Result<String, String> {
|
|
let mut last_err = String::new();
|
|
for attempt in 0..attempts {
|
|
if attempt > 0 {
|
|
let backoff_ms = 500u64 * (1u64 << (attempt - 1).min(3));
|
|
tokio::time::sleep(std::time::Duration::from_millis(backoff_ms)).await;
|
|
}
|
|
match requester.get(url, Some(Version::HTTP_2)).await {
|
|
Ok(text) => return Ok(text),
|
|
Err(e) => last_err = e.to_string(),
|
|
}
|
|
}
|
|
Err(last_err)
|
|
}
|
|
}
|
|
|
|
#[async_trait]
|
|
impl Provider for HentaihavenProvider {
|
|
async fn get_videos(
|
|
&self,
|
|
cache: VideoCache,
|
|
pool: DbPool,
|
|
sort: String,
|
|
query: Option<String>,
|
|
page: String,
|
|
per_page: String,
|
|
options: ServerOptions,
|
|
) -> Vec<VideoItem> {
|
|
let page = page.parse::<u8>().unwrap_or(1);
|
|
let per_page = per_page.parse::<usize>().unwrap_or(20);
|
|
|
|
let res = self
|
|
.list(
|
|
cache,
|
|
page,
|
|
per_page,
|
|
&sort,
|
|
query.as_deref(),
|
|
options,
|
|
pool,
|
|
)
|
|
.await;
|
|
|
|
res.unwrap_or_else(|e| {
|
|
eprintln!("hentai haven error: {e}");
|
|
vec![]
|
|
})
|
|
}
|
|
|
|
fn get_channel(&self, v: ClientVersion) -> Option<Channel> {
|
|
Some(self.build_channel(v))
|
|
}
|
|
}
|