Files
hottub/src/providers/hentaimama.rs
2026-09-07 19:31:34 +00:00

669 lines
24 KiB
Rust

use crate::DbPool;
use crate::api::ClientVersion;
use crate::providers::{Provider, report_provider_error, report_provider_error_background, requester_or_default};
use crate::status::*;
use crate::util::cache::VideoCache;
use crate::util::requester::Requester;
use crate::videos::{ServerOptions, VideoFormat, VideoItem};
use async_trait::async_trait;
use base64::{Engine, engine::general_purpose::STANDARD};
use chrono::DateTime;
use error_chain::error_chain;
use futures::stream::{self, StreamExt};
use htmlentity::entity::{ICodedDataTrait, decode};
use percent_encoding::{NON_ALPHANUMERIC, utf8_percent_encode};
use regex::Regex;
use scraper::{ElementRef, Html, Selector};
use std::collections::HashSet;
use std::sync::{Arc, RwLock};
use std::thread;
use wreq::Version;
pub const CHANNEL_METADATA: crate::providers::ProviderChannelMetadata =
crate::providers::ProviderChannelMetadata {
group_id: "hentai-animation",
tags: &["hentai", "anime", "episodes"],
};
error_chain! {
foreign_links {
Io(std::io::Error);
}
errors {
Parse(msg: String) {
description("parse error")
display("parse error: {}", msg)
}
}
}
const CHANNEL_ID: &str = "hentaimama";
const BASE_URL: &str = "https://hentaimama.io";
// Static-file mirror the site's own "rtmp" player option points at. Serves
// direct 200s with Accept-Ranges and no Referer/token needed, so the format
// URL can be built locally from the AJAX response with no extra fetch.
const MEDIA_HOST: &str = "https://gdvid.info/";
const USER_AGENT: &str =
"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/146.0.0.0 Safari/537.36";
#[derive(Debug, Clone)]
enum Target {
Latest,
Genre(String),
Studio(String),
Search(String),
}
#[derive(Debug, Clone)]
struct EpisodeCard {
id: String,
title: String,
url: String,
thumb: String,
rating: Option<f32>,
}
#[derive(Debug, Clone)]
pub struct HentaimamaProvider {
genres: Arc<RwLock<Vec<FilterOption>>>,
}
impl HentaimamaProvider {
pub fn new() -> Self {
let provider = Self {
genres: Arc::new(RwLock::new(vec![])),
};
provider.spawn_initial_load();
provider
}
fn spawn_initial_load(&self) {
let genres = Arc::clone(&self.genres);
thread::spawn(move || {
let runtime = match tokio::runtime::Builder::new_current_thread().enable_all().build() {
Ok(runtime) => runtime,
Err(e) => {
report_provider_error_background(
CHANNEL_ID,
"spawn_initial_load.runtime_build",
&e.to_string(),
);
return;
}
};
runtime.block_on(async move {
if let Err(e) = Self::load_genres(genres).await {
report_provider_error_background(CHANNEL_ID, "load_genres", &e.to_string());
}
});
});
}
async fn load_genres(genres: Arc<RwLock<Vec<FilterOption>>>) -> Result<()> {
let mut requester = Requester::new();
let html = requester
.get_with_headers(
&format!("{BASE_URL}/genres-filter/"),
Self::html_headers(BASE_URL),
Some(Version::HTTP_2),
)
.await
.map_err(|e| Error::from(format!("genres fetch failed: {e}")))?;
let document = Html::parse_document(&html);
let selector = Self::selector("a.genreitem")?;
let mut options = Vec::new();
let mut seen = HashSet::new();
for element in document.select(&selector) {
let Some(href) = element.value().attr("href") else {
continue;
};
let slug = href.trim_end_matches('/').rsplit('/').next().unwrap_or("").to_string();
if slug.is_empty() || !seen.insert(slug.clone()) {
continue;
}
let title = Self::decode_entities(&element.text().collect::<String>());
if title.is_empty() {
continue;
}
options.push(FilterOption { id: slug, title });
}
if !options.is_empty() {
if let Ok(mut guard) = genres.write() {
*guard = options;
}
}
Ok(())
}
fn build_channel(&self, _clientversion: ClientVersion) -> Channel {
let genres = self.genres.read().map(|g| g.clone()).unwrap_or_default();
Channel {
id: CHANNEL_ID.to_string(),
name: "Hentaimama".to_string(),
description: "Watch hentai episodes online free in HD.".to_string(),
premium: false,
favicon: "https://www.google.com/s2/favicons?sz=64&domain=hentaimama.io".to_string(),
status: "active".to_string(),
categories: genres.iter().map(|g| g.title.clone()).collect(),
options: vec![ChannelOption {
id: "categories".to_string(),
title: "Genres".to_string(),
description: "Filter by genre".to_string(),
systemImage: "tag.fill".to_string(),
colorName: "green".to_string(),
options: genres,
multiSelect: false,
}],
nsfw: true,
cacheDuration: Some(1800),
ytdlpCommand: Some("yt-dlp".to_string()),
}
}
fn selector(value: &str) -> Result<Selector> {
Selector::parse(value).map_err(|e| Error::from(format!("selector `{value}` parse failed: {e}")))
}
fn decode_entities(text: &str) -> String {
decode(text.as_bytes())
.to_string()
.unwrap_or_else(|_| text.to_string())
.split_whitespace()
.collect::<Vec<_>>()
.join(" ")
.trim()
.to_string()
}
fn html_headers(referer: &str) -> Vec<(String, String)> {
vec![
("Referer".to_string(), referer.to_string()),
("User-Agent".to_string(), USER_AGENT.to_string()),
]
}
async fn fetch_html(requester: &mut Requester, url: &str, referer: &str) -> Result<String> {
requester
.get_with_headers(url, Self::html_headers(referer), Some(Version::HTTP_2))
.await
.map_err(|e| Error::from(format!("request failed for {url}: {e}")))
}
fn slugify(value: &str) -> String {
value
.trim()
.to_lowercase()
.chars()
.map(|c| if c.is_ascii_alphanumeric() { c } else { '-' })
.collect::<String>()
.split('-')
.filter(|s| !s.is_empty())
.collect::<Vec<_>>()
.join("-")
}
fn find_genre_slug(&self, value: &str) -> Option<String> {
let normalized = value.trim().to_lowercase();
let genres = self.genres.read().ok()?;
genres
.iter()
.find(|g| g.id.eq_ignore_ascii_case(value.trim()) || g.title.to_lowercase() == normalized)
.map(|g| g.id.clone())
}
fn resolve_target(&self, query: Option<&str>, categories: Option<&str>) -> Target {
if let Some(value) = categories {
if let Some(slug) = self.find_genre_slug(value) {
return Target::Genre(slug);
}
let slug = Self::slugify(value);
if !slug.is_empty() {
return Target::Genre(slug);
}
}
if let Some(q) = query {
let q = q.trim();
if let Some(rest) = q
.strip_prefix("genre:")
.or_else(|| q.strip_prefix("cat:"))
.or_else(|| q.strip_prefix("category:"))
{
let slug = self.find_genre_slug(rest).unwrap_or_else(|| Self::slugify(rest));
return Target::Genre(slug);
}
if let Some(rest) = q.strip_prefix("studio:").or_else(|| q.strip_prefix("uploader:")) {
return Target::Studio(Self::slugify(rest));
}
if let Some(slug) = self.find_genre_slug(q) {
return Target::Genre(slug);
}
if !q.is_empty() {
return Target::Search(q.to_string());
}
}
Target::Latest
}
fn build_list_url(target: &Target, page: u32) -> String {
match target {
Target::Latest => {
if page <= 1 {
format!("{BASE_URL}/episodes/")
} else {
format!("{BASE_URL}/episodes/page/{page}/")
}
}
Target::Genre(slug) => {
if page <= 1 {
format!("{BASE_URL}/genre/{slug}/")
} else {
format!("{BASE_URL}/genre/{slug}/page/{page}/")
}
}
Target::Studio(slug) => {
if page <= 1 {
format!("{BASE_URL}/studio/{slug}/")
} else {
format!("{BASE_URL}/studio/{slug}/page/{page}/")
}
}
Target::Search(query) => {
let encoded = utf8_percent_encode(query, NON_ALPHANUMERIC).to_string();
if page <= 1 {
format!("{BASE_URL}/?s={encoded}")
} else {
format!("{BASE_URL}/page/{page}/?s={encoded}")
}
}
}
}
fn parse_episode_cards(html: &str) -> Result<Vec<EpisodeCard>> {
let document = Html::parse_document(html);
let card_selector = Self::selector("article.se.episodes")?;
let link_selector = Self::selector("div.season_m a")?;
let img_selector = Self::selector("img")?;
let rating_selector = Self::selector("div.rating")?;
let rating_re = Regex::new(r"([0-9]+(?:\.[0-9]+)?)").unwrap();
let mut cards = Vec::new();
for card in document.select(&card_selector) {
let Some(id) = card.value().attr("rel").filter(|s| !s.is_empty()) else {
continue;
};
let Some(link) = card.select(&link_selector).next() else {
continue;
};
let Some(url) = link.value().attr("href") else {
continue;
};
let title = Self::decode_entities(&link.text().collect::<String>());
if title.is_empty() {
continue;
}
let thumb = card
.select(&img_selector)
.next()
.and_then(|img| img.value().attr("src"))
.unwrap_or("")
.to_string();
let rating = card
.select(&rating_selector)
.next()
.map(|r| r.text().collect::<String>())
.and_then(|text| rating_re.captures(&text).and_then(|c| c[1].parse::<f32>().ok()));
cards.push(EpisodeCard {
id: id.to_string(),
title,
url: url.to_string(),
thumb,
rating,
});
}
Ok(cards)
}
fn parse_series_card_urls(html: &str) -> Result<Vec<String>> {
let document = Html::parse_document(html);
let selector = Self::selector("a.sc-poster")?;
Ok(document
.select(&selector)
.filter_map(|e| e.value().attr("href").map(|s| s.to_string()))
.collect())
}
fn last_episode_url(html: &str) -> Result<Option<String>> {
let document = Html::parse_document(html);
let selector = Self::selector("a.dt-se-item")?;
Ok(document
.select(&selector)
.filter_map(|e: ElementRef| e.value().attr("href").map(|s| s.to_string()))
.last())
}
fn post_id_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"data-post-id="(\d+)""#).expect("valid regex"))
}
fn title_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"<meta itemprop="name" content="([^"]+)">"#).expect("valid regex"))
}
fn thumb_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"<meta property="og:image" content="([^"]+)""#).expect("valid regex"))
}
fn rating_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"data-rating="([0-9.]+)""#).expect("valid regex"))
}
fn published_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| {
Regex::new(r#"<meta property="article:published_time" content="([^"]+)""#).expect("valid regex")
})
}
fn tag_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| {
Regex::new(r#"href="https://hentaimama\.io/(?:genre|studio)/[a-z0-9-]+/"[^>]*rel="tag">([^<]+)</a>"#)
.expect("valid regex")
})
}
fn rtmp_p_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"dt_embed=rtmp[^"]*?[?&]p=([A-Za-z0-9+/=]+)"#).expect("valid regex"))
}
fn iframe_src_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"src="([^"]+)""#).expect("valid regex"))
}
fn jwplayer_file_regex() -> &'static Regex {
static RE: std::sync::OnceLock<Regex> = std::sync::OnceLock::new();
RE.get_or_init(|| Regex::new(r#"file:\s*"([^"]+)""#).expect("valid regex"))
}
fn pad_base64(value: &str) -> String {
let padding = (4 - value.len() % 4) % 4;
format!("{value}{}", "=".repeat(padding))
}
/// The site's "rtmp" player option embeds the raw storage path as a base64
/// query param; the direct mirror URL is just that path appended to
/// `MEDIA_HOST`, so no extra request is needed to resolve it.
fn format_from_p_param(iframe_html: &str) -> Option<VideoFormat> {
let captures = Self::rtmp_p_regex().captures(iframe_html)?;
let decoded = STANDARD.decode(Self::pad_base64(&captures[1])).ok()?;
let path = String::from_utf8(decoded).ok()?;
let path = path.trim().trim_start_matches('/');
if path.is_empty() {
return None;
}
Some(VideoFormat::new(
format!("{MEDIA_HOST}{path}"),
"auto".to_string(),
"mp4".to_string(),
))
}
fn format_from_jwplayer(embed_html: &str) -> Option<VideoFormat> {
let src = &Self::jwplayer_file_regex().captures(embed_html)?[1];
let container = if src.to_lowercase().contains(".m3u8") { "m3u8" } else { "mp4" };
Some(VideoFormat::new(src.to_string(), "auto".to_string(), container.to_string()))
}
async fn fetch_player_iframe(
requester: &mut Requester,
referer: &str,
post_id: &str,
slot: u8,
) -> Option<String> {
let body = format!("action=get_player_contents&a={post_id}&i={slot}");
let headers = vec![
("Content-Type", "application/x-www-form-urlencoded"),
("X-Requested-With", "XMLHttpRequest"),
("Referer", referer),
("User-Agent", USER_AGENT),
];
let response = requester
.post(&format!("{BASE_URL}/wp-admin/admin-ajax.php"), &body, headers)
.await
.ok()?;
let text = response.text().await.ok()?;
let slots: Vec<String> = serde_json::from_str(&text).ok()?;
slots.into_iter().find(|s| !s.trim().is_empty())
}
async fn resolve_format(requester: &mut Requester, referer: &str, post_id: &str) -> Option<VideoFormat> {
let iframe_html = Self::fetch_player_iframe(requester, referer, post_id, 1).await?;
if let Some(format) = Self::format_from_p_param(&iframe_html) {
return Some(format);
}
// Fallback for embed shapes we haven't seen: follow the iframe itself and
// pull the jwplayer source out of its markup.
let src = &Self::iframe_src_regex().captures(&iframe_html)?[1];
let embed_html = Self::fetch_html(requester, src, referer).await.ok()?;
Self::format_from_jwplayer(&embed_html)
}
async fn build_item_from_episode_card(requester: &mut Requester, entry: &EpisodeCard) -> Option<VideoItem> {
let format = Self::resolve_format(requester, &entry.url, &entry.id).await?;
let mut item = VideoItem::new(
entry.id.clone(),
entry.title.clone(),
entry.url.clone(),
CHANNEL_ID.to_string(),
entry.thumb.clone(),
0,
)
.formats(vec![format])
.aspect_ratio(16.0 / 9.0);
if let Some(rating) = entry.rating {
item = item.rating(rating * 10.0);
}
Some(item)
}
/// Search/genre/studio archives only expose series (show) cards, so a match
/// is represented by its most recent episode - the same shape the latest
/// feed already returns, just reached through one extra hop.
async fn resolve_series_to_item(requester: &mut Requester, series_url: &str) -> Option<VideoItem> {
let html = Self::fetch_html(requester, series_url, BASE_URL).await.ok()?;
let episode_url = Self::last_episode_url(&html).ok().flatten()?;
Self::resolve_episode_url(requester, &episode_url).await
}
async fn resolve_episode_url(requester: &mut Requester, episode_url: &str) -> Option<VideoItem> {
let html = Self::fetch_html(requester, episode_url, BASE_URL).await.ok()?;
let post_id = Self::post_id_regex().captures(&html)?[1].to_string();
let title = Self::decode_entities(&Self::title_regex().captures(&html)?[1]);
if title.is_empty() {
return None;
}
let thumb = Self::thumb_regex()
.captures(&html)
.map(|c| c[1].to_string())
.unwrap_or_default();
let rating = Self::rating_regex().captures(&html).and_then(|c| c[1].parse::<f32>().ok());
let uploaded_at = Self::published_regex()
.captures(&html)
.and_then(|c| DateTime::parse_from_rfc3339(&c[1]).ok())
.map(|dt| dt.timestamp().max(0) as u64);
let mut seen = HashSet::new();
let tags: Vec<String> = Self::tag_regex()
.captures_iter(&html)
.map(|c| Self::decode_entities(&c[1]))
.filter(|t| !t.is_empty() && seen.insert(t.clone()))
.collect();
let format = Self::resolve_format(requester, episode_url, &post_id).await?;
let mut item = VideoItem::new(
post_id,
title,
episode_url.to_string(),
CHANNEL_ID.to_string(),
thumb,
0,
)
.formats(vec![format])
.tags(tags)
.aspect_ratio(16.0 / 9.0);
if let Some(rating) = rating {
item = item.rating(rating * 10.0);
}
if let Some(uploaded_at) = uploaded_at {
item = item.uploaded_at(uploaded_at);
}
Some(item)
}
async fn get(
&self,
cache: VideoCache,
page: u32,
per_page: usize,
query: Option<&str>,
options: ServerOptions,
) -> Result<Vec<VideoItem>> {
let target = self.resolve_target(query, options.categories.as_deref());
let list_url = Self::build_list_url(&target, page);
if let Some((time, items)) = cache.get(&list_url) {
if time.elapsed().unwrap_or_default().as_secs() < 300 {
return Ok(items);
}
}
let mut requester = requester_or_default(&options, CHANNEL_ID, "get_videos");
let html = Self::fetch_html(&mut requester, &list_url, BASE_URL).await?;
let items = match target {
Target::Latest => {
let cards = Self::parse_episode_cards(&html)?;
stream::iter(cards.into_iter().take(per_page.max(1)).map(|entry| {
let mut req = requester.clone();
async move { Self::build_item_from_episode_card(&mut req, &entry).await }
}))
.buffer_unordered(6)
.filter_map(|item| async move { item })
.collect::<Vec<_>>()
.await
}
_ => {
let urls = Self::parse_series_card_urls(&html)?;
stream::iter(urls.into_iter().take(per_page.max(1)).map(|url| {
let mut req = requester.clone();
async move { Self::resolve_series_to_item(&mut req, &url).await }
}))
.buffer_unordered(4)
.filter_map(|item| async move { item })
.collect::<Vec<_>>()
.await
}
};
if !items.is_empty() {
cache.insert(list_url, items.clone());
}
Ok(items)
}
}
#[async_trait]
impl Provider for HentaimamaProvider {
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::<u32>().unwrap_or(1).max(1);
let per_page = per_page.parse::<usize>().unwrap_or(24);
let query_ref = query.as_deref().filter(|q| !q.trim().is_empty());
match self.get(cache, page, per_page, query_ref, options).await {
Ok(items) => items,
Err(e) => {
report_provider_error(CHANNEL_ID, "get_videos", &e.to_string()).await;
vec![]
}
}
}
fn get_channel(&self, clientversion: ClientVersion) -> Option<Channel> {
Some(self.build_channel(clientversion))
}
}
#[cfg(test)]
mod tests {
use super::*;
fn provider() -> HentaimamaProvider {
HentaimamaProvider {
genres: Arc::new(RwLock::new(vec![FilterOption {
id: "maid".to_string(),
title: "Maid".to_string(),
}])),
}
}
#[test]
fn builds_latest_urls() {
assert_eq!(
HentaimamaProvider::build_list_url(&Target::Latest, 1),
"https://hentaimama.io/episodes/"
);
assert_eq!(
HentaimamaProvider::build_list_url(&Target::Latest, 2),
"https://hentaimama.io/episodes/page/2/"
);
}
#[test]
fn builds_genre_and_search_urls() {
assert_eq!(
HentaimamaProvider::build_list_url(&Target::Genre("maid".to_string()), 1),
"https://hentaimama.io/genre/maid/"
);
assert_eq!(
HentaimamaProvider::build_list_url(&Target::Search("school girl".to_string()), 2),
"https://hentaimama.io/page/2/?s=school%20girl"
);
}
#[test]
fn routes_query_shortcuts() {
let provider = provider();
assert!(matches!(provider.resolve_target(Some("maid"), None), Target::Genre(s) if s == "maid"));
assert!(matches!(provider.resolve_target(Some("genre:blowjob"), None), Target::Genre(s) if s == "blowjob"));
assert!(matches!(provider.resolve_target(Some("studio:majin"), None), Target::Studio(s) if s == "majin"));
assert!(matches!(provider.resolve_target(Some("random keyword"), None), Target::Search(s) if s == "random keyword"));
assert!(matches!(provider.resolve_target(None, None), Target::Latest));
}
#[test]
fn decodes_gdvid_path_from_p_param() {
let iframe = r#"<iframe src="https://hentaimama.io/?dt_embed=rtmp&#038;p=UC9wdXJlLXgtaG9saWMtanVua2V0c3Utb3RvbWUtdG8ta29uaW4ta2Fua2VpLXRoZS1hbmltYXRpb24tMi5tcDQ&#038;ep=17001"></iframe>"#;
let format = HentaimamaProvider::format_from_p_param(iframe).unwrap();
assert_eq!(
format.url,
"https://gdvid.info/P/pure-x-holic-junketsu-otome-to-konin-kankei-the-animation-2.mp4"
);
}
}