use crate::DbPool;
use crate::api::ClientVersion;
use crate::providers::Provider;
use crate::status::*;
use crate::util::cache::VideoCache;
use crate::util::parse_abbreviated_number;
use crate::util::time::parse_time_to_seconds;
use crate::videos::{ServerOptions, VideoFormat, VideoItem};

use async_trait::async_trait;
use error_chain::error_chain;
use htmlentity::entity::{ICodedDataTrait, decode};
use regex::Regex;
use std::net::IpAddr;
use std::vec;
use titlecase::Titlecase;
use url::Url;
use wreq::Version;

pub const CHANNEL_METADATA: crate::providers::ProviderChannelMetadata =
    crate::providers::ProviderChannelMetadata {
        group_id: "mainstream-tube",
        tags: &["search", "mixed", "user-upload"],
    };

error_chain! {
    foreign_links {
        Io(std::io::Error);
        HttpRequest(wreq::Error);
    }
    errors {
        Parse(msg: String)
    }
}

#[derive(Debug, Clone)]
pub struct NoodlemagazineProvider {
    url: String,
}

impl NoodlemagazineProvider {
    pub fn new() -> Self {
        Self {
            url: "https://noodlemagazine.com".to_string(),
        }
    }

    fn build_channel(&self, _clientversion: ClientVersion) -> Channel {
        Channel {
            id: "noodlemagazine".into(),
            name: "Noodlemagazine".into(),
            description: "The Best Search Engine of HD Videos".into(),
            premium: false,
            favicon: "https://www.google.com/s2/favicons?sz=64&domain=noodlemagazine.com".into(),
            status: "active".into(),
            categories: vec![],
            options: vec![
                ChannelOption {
                    id: "category".into(),
                    title: "Popular Period".into(),
                    description: "Pick which popular feed to browse.".into(),
                    systemImage: "clock".into(),
                    colorName: "blue".into(),
                    options: vec![
                        FilterOption {
                            id: "recent".into(),
                            title: "Recent".into(),
                        },
                        FilterOption {
                            id: "week".into(),
                            title: "This Week".into(),
                        },
                        FilterOption {
                            id: "month".into(),
                            title: "This Month".into(),
                        },
                    ],
                    multiSelect: false,
                },
                ChannelOption {
                    id: "sort".into(),
                    title: "Sort By".into(),
                    description: "Sort popular feed results.".into(),
                    systemImage: "arrow.up.arrow.down".into(),
                    colorName: "orange".into(),
                    options: vec![
                        FilterOption {
                            id: "views".into(),
                            title: "Views".into(),
                        },
                        FilterOption {
                            id: "date".into(),
                            title: "Newest".into(),
                        },
                        FilterOption {
                            id: "duration".into(),
                            title: "Duration".into(),
                        },
                    ],
                    multiSelect: false,
                },
                ChannelOption {
                    id: "filter".into(),
                    title: "Order".into(),
                    description: "Ascending or descending order.".into(),
                    systemImage: "list.number".into(),
                    colorName: "green".into(),
                    options: vec![
                        FilterOption {
                            id: "desc".into(),
                            title: "Descending".into(),
                        },
                        FilterOption {
                            id: "asc".into(),
                            title: "Ascending".into(),
                        },
                    ],
                    multiSelect: false,
                },
            ],
            nsfw: true,
            cacheDuration: Some(1800),
        }
    }

    fn resolve_popular_period(options: &ServerOptions) -> &'static str {
        match options.category.as_deref() {
            Some("week") => "week",
            Some("month") => "month",
            // The upstream site does not expose a valid /popular/all route.
            // Keep "all" as a backward-compatible alias for stale clients.
            Some("all") => "recent",
            _ => "recent",
        }
    }

    fn resolve_sort_by(sort: &str, options: &ServerOptions) -> &'static str {
        match options.sort.as_deref().unwrap_or(sort) {
            "date" | "new" | "latest" => "date",
            "duration" | "length" => "duration",
            _ => "views",
        }
    }

    fn resolve_sort_order(options: &ServerOptions) -> &'static str {
        match options.filter.as_deref() {
            Some("asc") => "asc",
            _ => "desc",
        }
    }

    fn mirror_url(url: &str) -> String {
        let stripped = url
            .strip_prefix("https://")
            .or_else(|| url.strip_prefix("http://"))
            .unwrap_or(url);
        format!("https://r.jina.ai/http://{stripped}")
    }

    fn looks_like_bot_challenge_or_block(html: &str) -> bool {
        let lower = html.to_ascii_lowercase();
        lower.contains("just a moment")
            || lower.contains("cf-browser-verification")
            || lower.contains("cf-chl")
            || lower.contains("access restricted")
            || lower.contains("cloudflare")
    }

    fn parse_markdown_listing_items(
        &self,
        markdown: &str,
        options: &ServerOptions,
    ) -> Vec<VideoItem> {
        let Some(regex) = Regex::new(
            r#"(?is)\[\!\[Image\s+\d+:\s*(?P<title>.*?)\]\((?P<thumb>https?://[^)\s]+)\)(?P<meta>.*?)\]\((?P<url>https?://noodlemagazine\.com/watch/[^)\s]+)\)"#,
        )
        .ok() else {
            return vec![];
        };
        let Some(duration_regex) = Regex::new(r"(?P<duration>\d{1,2}:\d{2}(?::\d{2})?)").ok() else {
            return vec![];
        };
        let Some(views_regex) = Regex::new(r"(?P<views>[0-9]+(?:\.[0-9]+)?[KMB]?)\s+\d{1,2}:\d{2}(?::\d{2})?").ok() else {
            return vec![];
        };

        regex
            .captures_iter(markdown)
            .filter_map(|caps| {
                let title_raw = caps.name("title")?.as_str().trim();
                let thumb = caps.name("thumb")?.as_str().trim();
                let video_url = caps.name("url")?.as_str().trim();
                let meta = caps.name("meta").map(|m| m.as_str()).unwrap_or("");

                let parsed_url = Url::parse(video_url).ok()?;
                let id = parsed_url
                    .path_segments()
                    .and_then(|mut segs| segs.next_back())
                    .filter(|value| !value.is_empty())
                    .map(|value| value.to_string())?;

                let duration = duration_regex
                    .captures(meta)
                    .and_then(|m| m.name("duration").map(|v| v.as_str()))
                    .and_then(|v| parse_time_to_seconds(v))
                    .unwrap_or(0) as u32;

                let views = views_regex
                    .captures(meta)
                    .and_then(|m| m.name("views").map(|v| v.as_str()))
                    .and_then(|v| parse_abbreviated_number(v.trim()))
                    .unwrap_or(0);

                let title = decode(title_raw.as_bytes())
                    .to_string()
                    .unwrap_or_else(|_| title_raw.to_string())
                    .titlecase();
                let proxy_url = self.proxy_url(options, video_url);
                let proxied_thumb = self.proxied_thumb(options, thumb);

                Some(
                    VideoItem::new(
                        id,
                        title,
                        proxy_url.clone(),
                        "noodlemagazine".into(),
                        proxied_thumb,
                        duration,
                    )
                    .views(views)
                    .formats(vec![
                        VideoFormat::new(proxy_url, "auto".into(), "video/mp4".into())
                            .format_id("auto".into())
                            .format_note("proxied".into())
                            .http_header("Referer".into(), video_url.to_string()),
                    ]),
                )
            })
            .collect()
    }

    async fn fetch_listing_items(
        &self,
        requester: &mut crate::util::requester::Requester,
        page_url: &str,
        options: &ServerOptions,
    ) -> Vec<VideoItem> {
        let html = requester
            .get(page_url, Some(Version::HTTP_2))
            .await
            .unwrap_or_default();
        let mut items = self.get_video_items_from_html(html.clone(), options);
        if !items.is_empty() {
            return items;
        }

        if !Self::looks_like_bot_challenge_or_block(&html) {
            return items;
        }

        let mirror = requester
            .get(&Self::mirror_url(page_url), Some(Version::HTTP_11))
            .await
            .unwrap_or_default();
        items = self.parse_markdown_listing_items(&mirror, options);
        items
    }

    async fn get(
        &self,
        cache: VideoCache,
        page: u8,
        sort: &str,
        options: ServerOptions,
    ) -> Result<Vec<VideoItem>> {
        let period = Self::resolve_popular_period(&options);
        let sort_by = Self::resolve_sort_by(sort, &options);
        let sort_order = Self::resolve_sort_order(&options);
        let video_url = format!(
            "{}/popular/{period}?sort_by={sort_by}&sort_order={sort_order}&p={}",
            self.url,
            page.saturating_sub(1)
        );

        let old_items = match cache.get(&video_url) {
            Some((t, i)) if t.elapsed().unwrap_or_default().as_secs() < 300 => return Ok(i.clone()),
            Some((_, i)) => i.clone(),
            None => vec![],
        };

        let mut requester = match options.requester.clone() {
            Some(r) => r,
            None => return Ok(old_items),
        };

        let items = self
            .fetch_listing_items(&mut requester, &video_url, &options)
            .await;

        if items.is_empty() {
            Ok(old_items)
        } else {
            cache.remove(&video_url);
            cache.insert(video_url, items.clone());
            Ok(items)
        }
    }

    async fn query(
        &self,
        cache: VideoCache,
        page: u8,
        query: &str,
        options: ServerOptions,
    ) -> Result<Vec<VideoItem>> {
        let q = query.trim().replace(' ', "%20");
        let video_url = format!("{}/video/{}?p={}", self.url, q, page.saturating_sub(1));

        let old_items = match cache.get(&video_url) {
            Some((t, i)) if t.elapsed().unwrap_or_default().as_secs() < 300 => return Ok(i.clone()),
            Some((_, i)) => i.clone(),
            None => vec![],
        };

        let mut requester = match options.requester.clone() {
            Some(r) => r,
            None => return Ok(old_items),
        };

        let items = self
            .fetch_listing_items(&mut requester, &video_url, &options)
            .await;

        if items.is_empty() {
            Ok(old_items)
        } else {
            cache.remove(&video_url);
            cache.insert(video_url, items.clone());
            Ok(items)
        }
    }

    fn get_video_items_from_html(&self, html: String, options: &ServerOptions) -> Vec<VideoItem> {
        if html.is_empty() || html.contains("404 Not Found") {
            return vec![];
        }

        let section = match html.split(">Show more</div>").next() {
            Some(s) => s,
            None => return vec![],
        };

        let list = match section
            .split("<div class=\"list_videos\" id=\"list_videos\">")
            .nth(1)
        {
            Some(l) => l,
            None => return vec![],
        };

        list.split("<div class=\"item")
            .skip(1)
            .filter_map(|segment| {
                self.get_video_item(segment.to_string(), options).ok()
            })
            .collect()
    }

    fn proxy_url(&self, options: &ServerOptions, video_url: &str) -> String {
        crate::providers::build_proxy_url(
            options,
            "noodlemagazine",
            &crate::providers::strip_url_scheme(video_url),
        )
    }

    fn normalize_thumb_url(&self, thumb: &str) -> String {
        let thumb = thumb.trim();
        if thumb.is_empty() {
            return String::new();
        }

        if thumb.starts_with("http://") || thumb.starts_with("https://") {
            return thumb.to_string();
        }

        if thumb.starts_with("//") {
            return format!("https:{thumb}");
        }

        if thumb.starts_with('/') {
            return format!("{}{}", self.url, thumb);
        }

        format!("{}/{}", self.url.trim_end_matches('/'), thumb.trim_start_matches('/'))
    }

    fn has_allowed_image_extension(path: &str) -> bool {
        let path = path.to_ascii_lowercase();
        [".jpg", ".jpeg", ".png", ".webp", ".avif", ".gif"]
            .iter()
            .any(|ext| path.ends_with(ext))
    }

    fn is_known_preview_host(host: &str) -> bool {
        let host = host.to_ascii_lowercase();
        host.ends_with("pvvstream.pro")
            || host.ends_with("okcdn.ru")
            || host.ends_with("vkuserphoto.ru")
            || host.ends_with("noodlemagazine.com")
    }

    fn has_preview_signature(url: &Url) -> bool {
        let path = url.path().to_ascii_lowercase();
        let query = url.query().unwrap_or("").to_ascii_lowercase();
        path.contains("/preview/")
            || path.contains("/poster/")
            || path.contains("getvideopreview")
            || query.contains("type=video_thumb")
            || query.contains("keep_aspect_ratio=")
    }

    fn is_disallowed_thumb_host(host: &str) -> bool {
        if host.eq_ignore_ascii_case("localhost") {
            return true;
        }

        match host.parse::<IpAddr>() {
            Ok(IpAddr::V4(ip)) => {
                ip.is_private()
                    || ip.is_loopback()
                    || ip.is_link_local()
                    || ip.is_broadcast()
                    || ip.is_documentation()
                    || ip.is_unspecified()
            }
            Ok(IpAddr::V6(ip)) => {
                ip.is_loopback()
                    || ip.is_unspecified()
                    || ip.is_multicast()
                    || ip.is_unique_local()
                    || ip.is_unicast_link_local()
            }
            Err(_) => false,
        }
    }

    fn is_allowed_thumb_url(&self, url: &str) -> bool {
        let Some(url) = Url::parse(url).ok() else {
            return false;
        };
        if url.scheme() != "https" {
            return false;
        }
        let Some(host) = url.host_str() else {
            return false;
        };

        if Self::is_disallowed_thumb_host(host) {
            return false;
        }

        if Self::has_allowed_image_extension(url.path()) {
            return true;
        }

        Self::is_known_preview_host(host) && Self::has_preview_signature(&url)
    }

    fn proxied_thumb(&self, _options: &ServerOptions, thumb: &str) -> String {
        let normalized = self.normalize_thumb_url(thumb);
        if normalized.is_empty() || !self.is_allowed_thumb_url(&normalized) {
            return String::new();
        }
        let Some(url) = Url::parse(&normalized).ok() else {
            return String::new();
        };
        if url
            .host_str()
            .is_some_and(|host| host.eq_ignore_ascii_case("img.pvvstream.pro"))
        {
            return crate::providers::build_proxy_url(
                _options,
                "noodlemagazine-thumb",
                &crate::providers::strip_url_scheme(&normalized),
            );
        }
        normalized
    }

    fn get_video_item(&self, video_segment: String, options: &ServerOptions) -> Result<VideoItem> {
        let href = video_segment
            .split("<a href=\"")
            .nth(1)
            .and_then(|s| s.split('"').next())
            .ok_or_else(|| Error::from("missing href"))?;

        let video_url = format!("{}{}", self.url, href);

        let mut title = video_segment
            .split("<div class=\"title\">")
            .nth(1)
            .and_then(|s| s.split('<').next())
            .unwrap_or("")
            .trim()
            .to_string();

        title = decode(title.as_bytes())
            .to_string()
            .unwrap_or(title)
            .titlecase();

        let id = video_url
            .split('/')
            .nth(4)
            .and_then(|s| s.split('.').next())
            .ok_or_else(|| Error::from("missing id"))?
            .to_string();

        let thumb = Regex::new(
            r#"(?i)(?:data-src|data-original|data-webp|src|poster)\s*=\s*"(?P<url>[^"]+)""#,
        )
        .ok()
        .and_then(|regex| {
            regex
                .captures_iter(&video_segment)
                .filter_map(|captures| captures.name("url").map(|value| value.as_str().to_string()))
                .find(|candidate| !candidate.starts_with("data:image/"))
        })
        .unwrap_or_default();

        let raw_duration = video_segment
            .split("#clock-o\"></use></svg>")
            .nth(1)
            .and_then(|s| s.split('<').next())
            .unwrap_or("0:00");

        let duration = parse_time_to_seconds(raw_duration).unwrap_or(0) as u32;

        let views = video_segment
            .split("#eye\"></use></svg>")
            .nth(1)
            .and_then(|s| s.split('<').next())
            .and_then(|v| parse_abbreviated_number(v.trim()))
            .unwrap_or(0);
        let proxy_url = self.proxy_url(options, &video_url);
        let proxied_thumb = self.proxied_thumb(options, &thumb);

        Ok(VideoItem::new(
            id,
            title,
            proxy_url.clone(),
            "noodlemagazine".into(),
            proxied_thumb,
            duration,
        )
        .views(views)
        .formats(vec![
            VideoFormat::new(proxy_url, "auto".into(), "video/mp4".into())
                .format_id("auto".into())
                .format_note("proxied".into())
                .http_header("Referer".into(), video_url),
        ]))
    }
}

#[async_trait]
impl Provider for NoodlemagazineProvider {
    async fn get_videos(
        &self,
        cache: VideoCache,
        pool: DbPool,
        sort: String,
        query: Option<String>,
        page: String,
        per_page: String,
        options: ServerOptions,
    ) -> Vec<VideoItem> {
        let _ = pool;
        let _ = per_page;

        let page = page.parse::<u8>().unwrap_or(1);

        let res = match query {
            Some(q) => self.query(cache, page, &q, options).await,
            None => self.get(cache, page, &sort, options).await,
        };

        res.unwrap_or_else(|e| {
            eprintln!("Noodlemagazine error: {e}");
            vec![]
        })
    }

    fn get_channel(&self, clientversion: ClientVersion) -> Option<Channel> {
        Some(self.build_channel(clientversion))
    }
}

#[cfg(test)]
mod tests {
    use super::NoodlemagazineProvider;
    use crate::videos::ServerOptions;

    fn options() -> ServerOptions {
        ServerOptions {
            featured: None,
            category: None,
            sites: None,
            filter: None,
            language: None,
            public_url_base: Some("https://example.com".to_string()),
            requester: None,
            network: None,
            stars: None,
            categories: None,
            duration: None,
            sort: None,
            sexuality: None,
        }
    }

    #[test]
    fn rewrites_video_pages_to_hottub_proxy() {
        let provider = NoodlemagazineProvider::new();
        let options = options();

        assert_eq!(
            provider.proxy_url(&options, "https://noodlemagazine.com/watch/-123_456"),
            "https://example.com/proxy/noodlemagazine/noodlemagazine.com/watch/-123_456"
        );
    }

    #[test]
    fn parses_listing_without_detail_page_requests() {
        let provider = NoodlemagazineProvider::new();
        let options = options();
        let html = r#"
            <div class="list_videos" id="list_videos">
                <div class="item">
                    <a href="/watch/-123_456">
                        <img data-src="https://noodlemagazine.com/thumbs/test.jpg" />
                    </a>
                    <div class="title">sample &amp; title</div>
                    <svg><use></use></svg>#clock-o"></use></svg>12:34<
                    <svg><use></use></svg>#eye"></use></svg>1.2K<
                </div>
            >Show more</div>
        "#;

        let items = provider.get_video_items_from_html(html.to_string(), &options);

        assert_eq!(items.len(), 1);
        assert_eq!(
            items[0].url,
            "https://example.com/proxy/noodlemagazine/noodlemagazine.com/watch/-123_456"
        );
        assert_eq!(
            items[0].thumb,
            "https://noodlemagazine.com/thumbs/test.jpg"
        );
        assert_eq!(items[0].formats.as_ref().map(|f| f.len()), Some(1));
    }

    #[test]
    fn keeps_https_cdn_thumbs_but_drops_non_images() {
        let provider = NoodlemagazineProvider::new();
        let options = options();
        let html = r#"
            <div class="list_videos" id="list_videos">
                <div class="item">
                    <a href="/watch/-123_456">
                        <img data-src="https://cdn.example/thumb.jpg" />
                    </a>
                    <div class="title">sample</div>
                    <svg><use></use></svg>#clock-o"></use></svg>12:34<
                    <svg><use></use></svg>#eye"></use></svg>1.2K<
                </div>
                <div class="item">
                    <a href="/watch/-555_666">
                        <img data-src="https://noodlemagazine.com/watch/not-an-image" />
                    </a>
                    <div class="title">sample 2</div>
                    <svg><use></use></svg>#clock-o"></use></svg>00:42<
                    <svg><use></use></svg>#eye"></use></svg>123<
                </div>
            >Show more</div>
        "#;

        let items = provider.get_video_items_from_html(html.to_string(), &options);

        assert_eq!(items.len(), 2);
        assert_eq!(
            items[0].thumb,
            "https://cdn.example/thumb.jpg"
        );
        assert!(items[1].thumb.is_empty());
    }

    #[test]
    fn keeps_preview_urls_without_file_extension() {
        let provider = NoodlemagazineProvider::new();
        let options = options();
        let html = r#"
            <div class="list_videos" id="list_videos">
                <div class="item">
                    <a href="/watch/-111_222">
                        <img data-src="https://img.pvvstream.pro/preview/abc/-111_222/240/iv.okcdn.ru/getVideoPreview?id=1&type=39&fn=vid_l" />
                    </a>
                    <div class="title">sample</div>
                    <svg><use></use></svg>#clock-o"></use></svg>12:34<
                    <svg><use></use></svg>#eye"></use></svg>1.2K<
                </div>
            >Show more</div>
        "#;

        let items = provider.get_video_items_from_html(html.to_string(), &options);
        assert_eq!(items.len(), 1);
        assert_eq!(
            items[0].thumb,
            "https://example.com/proxy/noodlemagazine-thumb/img.pvvstream.pro/preview/abc/-111_222/240/iv.okcdn.ru/getVideoPreview?id=1&type=39&fn=vid_l"
        );
    }

    #[test]
    fn parses_item_variants_and_alternate_thumb_attributes() {
        let provider = NoodlemagazineProvider::new();
        let options = options();
        let html = r#"
            <div class="list_videos" id="list_videos">
                <div class="item has-video" data-id="123">
                    <a href="/watch/-333_444">
                        <img data-original="https://cdn2.pvvstream.pro/videos/-333/444/preview_320.jpg" />
                    </a>
                    <div class="title">sample alt</div>
                    <svg><use></use></svg>#clock-o"></use></svg>00:42<
                    <svg><use></use></svg>#eye"></use></svg>123<
                </div>
            >Show more</div>
        "#;

        let items = provider.get_video_items_from_html(html.to_string(), &options);
        assert_eq!(items.len(), 1);
        assert_eq!(
            items[0].thumb,
            "https://cdn2.pvvstream.pro/videos/-333/444/preview_320.jpg"
        );
    }

    #[test]
    fn resolves_popular_filters_for_usability_options() {
        let mut options = options();
        options.category = Some("month".to_string());
        options.sort = Some("date".to_string());
        options.filter = Some("asc".to_string());

        assert_eq!(NoodlemagazineProvider::resolve_popular_period(&options), "month");
        assert_eq!(NoodlemagazineProvider::resolve_sort_by("views", &options), "date");
        assert_eq!(NoodlemagazineProvider::resolve_sort_order(&options), "asc");
    }

    #[test]
    fn maps_legacy_all_time_period_to_recent_feed() {
        let mut options = options();
        options.category = Some("all".to_string());
        options.sort = Some("views".to_string());
        options.filter = Some("desc".to_string());

        assert_eq!(NoodlemagazineProvider::resolve_popular_period(&options), "recent");
    }
}