hottub/archive/hentaimoon.rs

use crate::util::parse_abbreviated_number;
use crate::DbPool;
use crate::providers::Provider;
use crate::util::cache::VideoCache;
use crate::util::flaresolverr::{FlareSolverrRequest, Flaresolverr};
use crate::util::time::parse_time_to_seconds;
use crate::videos::{ServerOptions, VideoItem};
use error_chain::error_chain;
use htmlentity::entity::{ICodedDataTrait, decode};
use std::env;
use std::vec;
use wreq::{Client, Proxy};
use wreq_util::Emulation;

error_chain! {
    foreign_links {
        Io(std::io::Error);
        HttpRequest(wreq::Error);
    }
}

#[derive(Debug, Clone)]
pub struct HentaimoonProvider {
    url: String,
}
impl HentaimoonProvider {
    pub fn new() -> Self {
        HentaimoonProvider {
            url: "https://hentai-moon.com".to_string(),
        }
    }
    async fn get(
        &self,
        cache: VideoCache,
        page: u8,
        sort: &str,
    ) -> Result<Vec<VideoItem>> {
        let sort_string = match sort {
            "popular" => "/most-popular",
            "top-rated" => "/top-rated",
            _ => "/latest-updates/",
        };

        let list_str = match sort {
            "popular" => "list_videos_common_videos_list",
            "top-rated" => "list_videos_common_videos_list",
            _ => "list_videos_most_recent_videos",
        };

        let video_url = format!("{}{}?mode=async^&function=get_block^&block_id={}^&from={}", self.url, sort_string, list_str, page);
        let old_items = match cache.get(&video_url) {
            Some((time, items)) => {
                if time.elapsed().unwrap_or_default().as_secs() < 60 * 5 {
                    return Ok(items.clone());
                } else {
                    items.clone()
                }
            }
            None => {
                vec![]
            }
        };

        let proxy = Proxy::all("http://192.168.0.103:8081").unwrap();
        let client = Client::builder().cert_verification(false).emulation(Emulation::Firefox136).build()?;

        let mut response = client.get(video_url.clone())
        // .proxy(proxy.clone())
        .send().await?;
        if response.status().is_redirection(){
            response = client.get(response.headers()["Location"].to_str().unwrap())
            // .proxy(proxy.clone())
            .send().await?;
        }
        if response.status().is_success() {
            let text = response.text().await?;
            let video_items: Vec<VideoItem> = self.get_video_items_from_html(text.clone());
            if !video_items.is_empty() {
                cache.remove(&video_url);
                cache.insert(video_url.clone(), video_items.clone());
            } else {
                return Ok(old_items);
            }
            Ok(video_items)
        } else {
            let flare_url = env::var("FLARE_URL").expect("FLARE_URL not set");
            let flare = Flaresolverr::new(flare_url);
            let result = flare
                .solve(FlareSolverrRequest {
                    cmd: "request.get".to_string(),
                    url: video_url.clone(),
                    maxTimeout: 60000,
                })
                .await;
            let video_items = match result {
                Ok(res) => {
                    // println!("FlareSolverr response: {}", res);
                    self.get_video_items_from_html(res.solution.response)
                }
                Err(e) => {
                    println!("Error solving FlareSolverr: {}", e);
                    return Err("Failed to solve FlareSolverr".into());
                }
            };
            if !video_items.is_empty() {
                cache.remove(&video_url);
                cache.insert(video_url.clone(), video_items.clone());
            } else {
                return Ok(old_items);
            }
            Ok(video_items)
        }
    }
    async fn query(
        &self,
        cache: VideoCache,
        page: u8,
        query: &str,
    ) -> Result<Vec<VideoItem>> {
        let search_string = query.to_lowercase().trim().replace(" ", "-");
        let video_url = format!("{}/search/{}/?mode=async&function=get_block&block_id=list_videos_videos_list_search_result&q={}&category_ids=&sort_by=&from_videos={}&from_albums={}&", self.url, search_string, search_string, page, page);
        // Check our Video Cache. If the result is younger than 1 hour, we return it.
        let old_items = match cache.get(&video_url) {
            Some((time, items)) => {
                if time.elapsed().unwrap_or_default().as_secs() < 60 * 5 {
                    return Ok(items.clone());
                } else {
                    let _ = cache.check().await;
                    return Ok(items.clone());
                }
            }
            None => {
                vec![]
            }
        };

        let proxy = Proxy::all("http://192.168.0.103:8081").unwrap();
        let client = Client::builder().cert_verification(false).emulation(Emulation::Firefox136).build()?;

        let mut response = client.get(video_url.clone())
        // .proxy(proxy.clone())
        .send().await?;

        if response.status().is_redirection(){
            
            response = client.get(self.url.clone() + response.headers()["Location"].to_str().unwrap())
            // .proxy(proxy.clone())
            .send().await?;
        }

        if response.status().is_success() {
            let text = response.text().await?;
            let video_items: Vec<VideoItem> = self.get_video_items_from_html(text.clone());
            if !video_items.is_empty() {
                cache.remove(&video_url);
                cache.insert(video_url.clone(), video_items.clone());
            } else {
                return Ok(old_items);
            }
            Ok(video_items)
        } else {
            let flare_url = env::var("FLARE_URL").expect("FLARE_URL not set");
            let flare = Flaresolverr::new(flare_url);
            let result = flare
                .solve(FlareSolverrRequest {
                    cmd: "request.get".to_string(),
                    url: video_url.clone(),
                    maxTimeout: 60000,
                })
                .await;
            let video_items = match result {
                Ok(res) => self.get_video_items_from_html(res.solution.response),
                Err(e) => {
                    println!("Error solving FlareSolverr: {}", e);
                    return Err("Failed to solve FlareSolverr".into());
                }
            };
            if !video_items.is_empty() {
                cache.remove(&video_url);
                cache.insert(video_url.clone(), video_items.clone());
            } else {
                return Ok(old_items);
            }
            Ok(video_items)
        }
    }

    fn get_video_items_from_html(&self, html: String) -> Vec<VideoItem> {
        if html.is_empty() {
            println!("HTML is empty");
            return vec![];
        }
        let mut items: Vec<VideoItem> = Vec::new();
        let raw_videos = html.split("<div class=\"pagination\"").collect::<Vec<&str>>()[0]
            .split("<div class=\"item  \">")
            .collect::<Vec<&str>>()[1..]
            .to_vec();
        for video_segment in &raw_videos {
            // let vid = video_segment.split("\n").collect::<Vec<&str>>();
            // for (index, line) in vid.iter().enumerate() {
            //     println!("Line {}: {}", index, line);
            // }
            let video_url: String = video_segment.split("<a href=\"").collect::<Vec<&str>>()[1]
                    .split("\"")
                    .collect::<Vec<&str>>()[0].to_string();
            let mut title = video_segment.split("\" title=\"").collect::<Vec<&str>>()[1]
                .split("\"")
                .collect::<Vec<&str>>()[0]
                .to_string();
            // html decode
            title = decode(title.as_bytes()).to_string().unwrap_or(title);
            let id = video_url.split("/").collect::<Vec<&str>>()[4].to_string();
            let raw_duration = video_segment.split("<div class=\"duration\">").collect::<Vec<&str>>()[1]
                .split("<")
                .collect::<Vec<&str>>()[0]
                .to_string();
            let duration = parse_time_to_seconds(&raw_duration).unwrap_or(0) as u32;

            let thumb = video_segment.split("<img class=\"thumb ").collect::<Vec<&str>>()[1]
                .split("data-original=\"").collect::<Vec<&str>>()[1]
                .split("\"")
                .collect::<Vec<&str>>()[0]
                .to_string();

            let views_part = video_segment.split("<div class=\"views\">").collect::<Vec<&str>>()[1]
                .split("<")
                .collect::<Vec<&str>>()[0]
                .to_string();
            let views = parse_abbreviated_number(&views_part).unwrap_or(0) as u32;

            let video_item = VideoItem::new(
                id,
                title,
                video_url.to_string(),
                "hentaimoon".to_string(),
                thumb,
                duration,
            )
            .views(views)
            ;
            items.push(video_item);
        }
        return items;
    }


}

impl Provider for HentaimoonProvider {
    async fn get_videos(
        &self,
        cache: VideoCache,
        pool: DbPool,
        sort: String,
        query: Option<String>,
        page: String,
        per_page: String,
        options: ServerOptions,
    ) -> Vec<VideoItem> {
        let _ = options;
        let _ = per_page;
        let _ = pool;
        let videos: std::result::Result<Vec<VideoItem>, Error> = match query {
            Some(q) => {
                self.query(cache, page.parse::<u8>().unwrap_or(1), &q,)
                    .await
            }
            None => {
                self.get(cache, page.parse::<u8>().unwrap_or(1), &sort)
                    .await
            }
        };
        match videos {
            Ok(v) => v,
            Err(e) => {
                println!("Error fetching videos: {}", e);
                vec![]
            }
        }
    }
}