feat: classic episodes scraper

apps/recorder/Cargo.toml

@@ -2,8 +2,20 @@
 name = "recorder"
 version = "0.1.0"
 edition = "2024"
-
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
+
+[features]
+default = ["jxl"]
+playground = ["dep:inquire", "dep:color-eyre", "dep:polars"]
+testcontainers = [
+    "dep:testcontainers",
+    "dep:testcontainers-modules",
+    "dep:testcontainers-ext",
+    "downloader/testcontainers",
+    "testcontainers-modules/postgres",
+]
+jxl = ["dep:jpegxl-rs", "dep:jpegxl-sys"]
+
 [lib]
 name = "recorder"
 path = "src/lib.rs"
@@ -13,17 +25,25 @@ name = "recorder_cli"
 path = "src/bin/main.rs"
 required-features = []
 
-[features]
+[[example]]
-default = ["jxl"]
+name = "mikan_collect_classic_eps"
-playground = ["dep:inquire", "dep:color-eyre"]
+path = "examples/mikan_collect_classic_eps.rs"
-testcontainers = [
+required-features = ["playground"]
-    "dep:testcontainers",
+
-    "dep:testcontainers-modules",
+[[example]]
-    "dep:testcontainers-ext",
+name = "mikan_doppel_season_subscription"
-    "downloader/testcontainers",
+path = "examples/mikan_doppel_season_subscription.rs"
-    "testcontainers-modules/postgres",
+required-features = ["playground"]
-]
+
-jxl = ["dep:jpegxl-rs", "dep:jpegxl-sys"]
+[[example]]
+name = "mikan_doppel_subscriber_subscription"
+path = "examples/mikan_doppel_subscriber_subscription.rs"
+required-features = ["playground"]
+
+[[example]]
+name = "playground"
+path = "examples/playground.rs"
+required-features = ["playground"]
 
 [dependencies]
 downloader = { workspace = true }
@@ -93,7 +113,7 @@ fancy-regex = "0.14"
 lightningcss = "1.0.0-alpha.66"
 html-escape = "0.2.13"
 opendal = { version = "0.53", features = ["default", "services-fs"] }
-scraper = "0.23"
+scraper = "0.23.1"
 async-graphql = { version = "7", features = ["dynamic-schema"] }
 async-graphql-axum = "7"
 seaography = { version = "1.1", features = [
@@ -134,11 +154,11 @@ icu = "2.0.0"
 tracing-tree = "0.4.0"
 num_cpus = "1.17.0"
 headers-accept = "0.1.4"
+polars = { version = "0.49.1", features = ["parquet"], optional = true }
 
 [dev-dependencies]
 inquire = { workspace = true }
 color-eyre = { workspace = true }
-
 serial_test = "3"
 insta = { version = "1", features = ["redactions", "toml", "filters"] }
 rstest = "0.25"

apps/recorder/examples/mikan_collect_classic_eps.rs

@@ -0,0 +1,443 @@
+use std::collections::HashSet;
+
+use chrono::{DateTime, Duration, FixedOffset, NaiveDate, NaiveTime, TimeZone, Utc};
+use fetch::{HttpClientConfig, fetch_html};
+use lazy_static::lazy_static;
+use nom::{
+    IResult, Parser,
+    branch::alt,
+    bytes::complete::{tag, take, take_till1},
+    character::complete::space1,
+    combinator::map,
+};
+use recorder::{
+    errors::{RecorderError, RecorderResult},
+    extract::{
+        html::extract_inner_text_from_element_ref,
+        mikan::{MikanClient, MikanConfig, MikanEpisodeHash, MikanFansubHash},
+    },
+};
+use regex::Regex;
+use scraper::{ElementRef, Html, Selector};
+use snafu::FromString;
+use url::Url;
+
+lazy_static! {
+    static ref TEST_FOLDER: std::path::PathBuf =
+        if cfg!(any(test, debug_assertions, feature = "playground")) {
+            std::path::PathBuf::from(format!(
+                "{}/tests/resources/mikan/classic_episodes",
+                env!("CARGO_MANIFEST_DIR")
+            ))
+        } else {
+            std::path::PathBuf::from("tests/resources/mikan/classic_episodes")
+        };
+}
+
+lazy_static! {
+    static ref TOTAL_PAGE_REGEX: Regex =
+        Regex::new(r#"\$\(\'\.classic-view-pagination2\'\)\.bootpag\(\{\s*total:\s*(\d+)"#)
+            .unwrap();
+}
+
+pub struct MikanClassicEpisodeTableRow {
+    pub id: i32,
+    pub publish_at: DateTime<Utc>,
+    pub mikan_fansub_id: Option<String>,
+    pub fansub_name: Option<String>,
+    pub mikan_episode_id: String,
+    pub original_name: String,
+    pub magnet_link: Option<String>,
+    pub file_size: Option<String>,
+    pub torrent_link: Option<String>,
+}
+
+impl MikanClassicEpisodeTableRow {
+    fn timezone() -> FixedOffset {
+        FixedOffset::east_opt(8 * 3600).unwrap()
+    }
+
+    fn fixed_date_parser(input: &str) -> IResult<&str, NaiveDate> {
+        alt((
+            map(tag("今天"), move |_| {
+                Utc::now().with_timezone(&Self::timezone()).date_naive()
+            }),
+            map(tag("昨天"), move |_| {
+                Utc::now().with_timezone(&Self::timezone()).date_naive() - Duration::days(1)
+            }),
+        ))
+        .parse(input)
+    }
+
+    fn formatted_date_parser(input: &str) -> IResult<&str, NaiveDate> {
+        let (remain, date_str) = take_till1(|c: char| c.is_whitespace()).parse(input)?;
+        let date = NaiveDate::parse_from_str(date_str, "%Y/%m/%d").map_err(|_| {
+            nom::Err::Error(nom::error::Error::new(input, nom::error::ErrorKind::Verify))
+        })?;
+        Ok((remain, date))
+    }
+
+    fn date_parser(input: &str) -> IResult<&str, NaiveDate> {
+        alt((Self::fixed_date_parser, Self::formatted_date_parser)).parse(input)
+    }
+
+    fn time_parser(input: &str) -> IResult<&str, NaiveTime> {
+        let (remain, time_str) = take(5usize).parse(input)?;
+        let time = NaiveTime::parse_from_str(time_str, "%H:%M").map_err(|_| {
+            nom::Err::Error(nom::error::Error::new(input, nom::error::ErrorKind::Verify))
+        })?;
+        Ok((remain, time))
+    }
+
+    fn extract_publish_at(text: &str) -> Option<DateTime<Utc>> {
+        let (_, (date, _, time)) = (Self::date_parser, space1, Self::time_parser)
+            .parse(text)
+            .ok()?;
+        let local_dt = Self::timezone()
+            .from_local_datetime(&date.and_time(time))
+            .single()?;
+        Some(local_dt.with_timezone(&Utc))
+    }
+
+    pub fn from_element_ref(
+        row: ElementRef<'_>,
+        rev_id: i32,
+        idx: i32,
+        mikan_base_url: &Url,
+    ) -> RecorderResult<Self> {
+        let publish_at_selector = &Selector::parse("td:nth-of-type(1)").unwrap();
+        let fansub_selector = &Selector::parse("td:nth-of-type(2) > a").unwrap();
+        let original_name_selector =
+            &Selector::parse("td:nth-of-type(3) > a:nth-of-type(1)").unwrap();
+        let magnet_link_selector =
+            &Selector::parse("td:nth-of-type(3) > a:nth-of-type(2)").unwrap();
+        let file_size_selector = &Selector::parse("td:nth-of-type(4)").unwrap();
+        let torrent_link_selector = &Selector::parse("td:nth-of-type(5) > a").unwrap();
+
+        let publish_at = row
+            .select(publish_at_selector)
+            .next()
+            .map(extract_inner_text_from_element_ref)
+            .and_then(|e| Self::extract_publish_at(&e));
+
+        let (mikan_fansub_hash, fansub_name) = row
+            .select(fansub_selector)
+            .next()
+            .and_then(|e| {
+                e.attr("href")
+                    .and_then(|s| mikan_base_url.join(s).ok())
+                    .and_then(|u| MikanFansubHash::from_homepage_url(&u))
+                    .map(|h| (h, extract_inner_text_from_element_ref(e)))
+            })
+            .unzip();
+
+        let (mikan_episode_hash, original_name) = row
+            .select(original_name_selector)
+            .next()
+            .and_then(|el| {
+                el.attr("href")
+                    .and_then(|s| mikan_base_url.join(s).ok())
+                    .and_then(|u| MikanEpisodeHash::from_homepage_url(&u))
+                    .map(|h| (h, extract_inner_text_from_element_ref(el)))
+            })
+            .unzip();
+
+        let magnet_link = row
+            .select(magnet_link_selector)
+            .next()
+            .and_then(|el| el.attr("data-clipboard-text"));
+
+        let file_size = row
+            .select(file_size_selector)
+            .next()
+            .map(extract_inner_text_from_element_ref);
+
+        let torrent_link = row
+            .select(torrent_link_selector)
+            .next()
+            .and_then(|el| el.attr("href"));
+
+        if let (Some(mikan_episode_hash), Some(original_name), Some(publish_at)) = (
+            mikan_episode_hash.as_ref(),
+            original_name.as_ref(),
+            publish_at.as_ref(),
+        ) {
+            Ok(Self {
+                id: rev_id * 1000 + idx,
+                publish_at: *publish_at,
+                mikan_fansub_id: mikan_fansub_hash.map(|h| h.mikan_fansub_id.clone()),
+                fansub_name,
+                mikan_episode_id: mikan_episode_hash.mikan_episode_id.clone(),
+                original_name: original_name.clone(),
+                magnet_link: magnet_link.map(|s| s.to_string()),
+                file_size: file_size.map(|s| s.to_string()),
+                torrent_link: torrent_link.map(|s| s.to_string()),
+            })
+        } else {
+            let mut missing_fields = vec![];
+            if mikan_episode_hash.is_none() {
+                missing_fields.push("mikan_episode_id");
+            }
+            if original_name.is_none() {
+                missing_fields.push("original_name");
+            }
+            if publish_at.is_none() {
+                missing_fields.push("publish_at");
+            }
+            Err(RecorderError::without_source(format!(
+                "Failed to parse episode table row, missing fields: {missing_fields:?}, row \
+                 index: {idx}"
+            )))
+        }
+    }
+}
+
+pub struct MikanClassicEpisodeTablePage {
+    pub page: i32,
+    pub total: i32,
+    pub html: String,
+    pub rows: Vec<MikanClassicEpisodeTableRow>,
+}
+
+impl MikanClassicEpisodeTablePage {
+    pub fn from_html(
+        html: String,
+        mikan_base_url: &Url,
+        page: i32,
+        updated_info: Option<(i32, i32)>,
+    ) -> RecorderResult<Self> {
+        let tr_selector = &Selector::parse("tbody tr").unwrap();
+        let doc = Html::parse_document(&html);
+        if let Some(mut total) = TOTAL_PAGE_REGEX
+            .captures(&html)
+            .and_then(|c| c.get(1))
+            .and_then(|s| s.as_str().parse::<i32>().ok())
+        {
+            if let Some((_, update_total)) = updated_info {
+                total = update_total;
+            }
+
+            let rev_id = total - page;
+            let rows = doc
+                .select(tr_selector)
+                .rev()
+                .enumerate()
+                .map(|(idx, tr)| {
+                    MikanClassicEpisodeTableRow::from_element_ref(
+                        tr,
+                        rev_id,
+                        idx as i32,
+                        mikan_base_url,
+                    )
+                })
+                .collect::<RecorderResult<Vec<_>>>()?;
+            Ok(Self {
+                page,
+                total,
+                html,
+                rows,
+            })
+        } else {
+            Err(RecorderError::without_source(
+                "Failed to parse pagination meta and rows".into(),
+            ))
+        }
+    }
+
+    pub fn save_to_files(&self) -> RecorderResult<()> {
+        use polars::prelude::*;
+
+        let rev_id = self.total - self.page;
+        let parquet_path = TEST_FOLDER.join(format!("parquet/rev_{rev_id}.parquet"));
+        let csv_path = TEST_FOLDER.join(format!("csv/rev_{rev_id}.csv"));
+        let html_path = TEST_FOLDER.join(format!("html/rev_{rev_id}.html"));
+
+        std::fs::write(html_path, self.html.clone())?;
+
+        let mut publish_at_vec = Vec::new();
+        let mut mikan_fansub_id_vec = Vec::new();
+        let mut fansub_name_vec = Vec::new();
+        let mut mikan_episode_id_vec = Vec::new();
+        let mut original_name_vec = Vec::new();
+        let mut magnet_link_vec = Vec::new();
+        let mut file_size_vec = Vec::new();
+        let mut torrent_link_vec = Vec::new();
+
+        for row in &self.rows {
+            publish_at_vec.push(row.publish_at.to_rfc3339());
+            mikan_fansub_id_vec.push(row.mikan_fansub_id.clone());
+            fansub_name_vec.push(row.fansub_name.clone());
+            mikan_episode_id_vec.push(row.mikan_episode_id.clone());
+            original_name_vec.push(row.original_name.clone());
+            magnet_link_vec.push(row.magnet_link.clone());
+            file_size_vec.push(row.file_size.clone());
+            torrent_link_vec.push(row.torrent_link.clone());
+        }
+
+        let df = df! [
+            "publish_at_timestamp" => publish_at_vec,
+            "mikan_fansub_id" => mikan_fansub_id_vec,
+            "fansub_name" => fansub_name_vec,
+            "mikan_episode_id" => mikan_episode_id_vec,
+            "original_name" => original_name_vec,
+            "magnet_link" => magnet_link_vec,
+            "file_size" => file_size_vec,
+            "torrent_link" => torrent_link_vec,
+        ]
+        .map_err(|e| {
+            let message = format!("Failed to create DataFrame: {e}");
+            RecorderError::with_source(Box::new(e), message)
+        })?;
+
+        let mut parquet_file = std::fs::File::create(&parquet_path)?;
+
+        ParquetWriter::new(&mut parquet_file)
+            .finish(&mut df.clone())
+            .map_err(|e| {
+                let message = format!("Failed to write parquet file: {e}");
+                RecorderError::with_source(Box::new(e), message)
+            })?;
+
+        let mut csv_file = std::fs::File::create(&csv_path)?;
+
+        CsvWriter::new(&mut csv_file)
+            .include_header(true)
+            .with_quote_style(QuoteStyle::Always)
+            .finish(&mut df.clone())
+            .map_err(|e| {
+                let message = format!("Failed to write csv file: {e}");
+                RecorderError::with_source(Box::new(e), message)
+            })?;
+
+        println!(
+            "[{}/{}] Saved {} rows to rev_{}.{{parquet,html,csv}}",
+            self.page,
+            self.total,
+            self.rows.len(),
+            rev_id
+        );
+
+        Ok(())
+    }
+
+    pub fn waiting_rev_ids(total: i32) -> RecorderResult<Vec<i32>> {
+        let dir = TEST_FOLDER.join("csv");
+
+        let files = std::fs::read_dir(dir)?;
+
+        let rev_ids = files
+            .filter_map(|f| f.ok())
+            .filter_map(|f| {
+                f.path().file_stem().and_then(|s| {
+                    s.to_str().and_then(|s| {
+                        if s.starts_with("rev_") {
+                            s.replace("rev_", "").parse::<i32>().ok()
+                        } else {
+                            None
+                        }
+                    })
+                })
+            })
+            .collect::<HashSet<_>>();
+
+        Ok((0..total)
+            .filter(|rev_id| !rev_ids.contains(rev_id))
+            .collect::<Vec<_>>())
+    }
+}
+
+async fn scrape_mikan_classic_episode_table_page(
+    mikan_client: &MikanClient,
+    page: i32,
+    updated_info: Option<(i32, i32)>,
+) -> RecorderResult<MikanClassicEpisodeTablePage> {
+    let mikan_base_url = mikan_client.base_url();
+    let url = mikan_base_url.join(&format!("/Home/Classic/{page}"))?;
+
+    if let Some((rev_id, update_total)) = updated_info.as_ref() {
+        let html_path = TEST_FOLDER.join(format!("html/rev_{rev_id}.html"));
+        if html_path.exists() {
+            let html = std::fs::read_to_string(&html_path)?;
+            println!("[{page}/{update_total}] html exists, skipping fetch");
+            return MikanClassicEpisodeTablePage::from_html(
+                html,
+                mikan_base_url,
+                page,
+                updated_info,
+            );
+        }
+    }
+
+    let total = if let Some((_, update_total)) = updated_info.as_ref() {
+        update_total.to_string()
+    } else {
+        "Unknown".to_string()
+    };
+
+    println!("[{page}/{total}] fetching html...");
+
+    let html = fetch_html(mikan_client, url).await?;
+
+    println!("[{page}/{total}] fetched html done");
+
+    std::fs::write(TEST_FOLDER.join("html/temp.html"), html.clone())?;
+
+    MikanClassicEpisodeTablePage::from_html(html, mikan_base_url, page, updated_info)
+}
+
+async fn scrape_mikan_classic_episode_table_page_from_rev_id(
+    mikan_client: &MikanClient,
+    total: i32,
+    rev_idx: i32,
+) -> RecorderResult<MikanClassicEpisodeTablePage> {
+    let page = total - rev_idx;
+
+    scrape_mikan_classic_episode_table_page(mikan_client, page, Some((rev_idx, total))).await
+}
+
+#[tokio::main]
+async fn main() -> RecorderResult<()> {
+    std::fs::create_dir_all(TEST_FOLDER.join("html"))?;
+    std::fs::create_dir_all(TEST_FOLDER.join("parquet"))?;
+    std::fs::create_dir_all(TEST_FOLDER.join("csv"))?;
+
+    let mikan_scrape_client = MikanClient::from_config(MikanConfig {
+        http_client: HttpClientConfig {
+            exponential_backoff_max_retries: Some(3),
+            leaky_bucket_max_tokens: Some(2),
+            leaky_bucket_initial_tokens: Some(0),
+            leaky_bucket_refill_tokens: Some(1),
+            leaky_bucket_refill_interval: Some(std::time::Duration::from_millis(1000)),
+            user_agent: Some(
+                "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) \
+                 Chrome/136.0.0.0 Safari/537.36 Edg/136.0.0.0"
+                    .to_string(),
+            ),
+            ..Default::default()
+        },
+        base_url: Url::parse("https://mikanani.me")?,
+    })
+    .await?;
+
+    let first_page_and_pagination_info =
+        scrape_mikan_classic_episode_table_page(&mikan_scrape_client, 1, None).await?;
+
+    let total_page = first_page_and_pagination_info.total;
+
+    first_page_and_pagination_info.save_to_files()?;
+
+    let next_rev_ids = MikanClassicEpisodeTablePage::waiting_rev_ids(total_page)?;
+
+    for todo_rev_id in next_rev_ids {
+        let page = scrape_mikan_classic_episode_table_page_from_rev_id(
+            &mikan_scrape_client,
+            total_page,
+            todo_rev_id,
+        )
+        .await?;
+
+        page.save_to_files()?;
+    }
+
+    Ok(())
+}

apps/recorder/src/extract/mikan/constants.rs

@@ -12,6 +12,7 @@ pub const MIKAN_BANGUMI_POSTER_PATH: &str = "/images/Bangumi";
 pub const MIKAN_EPISODE_TORRENT_PATH: &str = "/Download";
 pub const MIKAN_SUBSCRIBER_SUBSCRIPTION_RSS_PATH: &str = "/RSS/MyBangumi";
 pub const MIKAN_BANGUMI_RSS_PATH: &str = "/RSS/Bangumi";
+pub const MIKAN_FANSUB_HOMEPAGE_PATH: &str = "/Home/PublishGroup";
 pub const MIKAN_BANGUMI_ID_QUERY_KEY: &str = "bangumiId";
 pub const MIKAN_FANSUB_ID_QUERY_KEY: &str = "subgroupid";
 pub const MIKAN_SUBSCRIBER_SUBSCRIPTION_TOKEN_QUERY_KEY: &str = "token";

apps/recorder/src/extract/mikan/mod.rs

@@ -11,10 +11,11 @@ pub use constants::{
     MIKAN_ACCOUNT_MANAGE_PAGE_PATH, MIKAN_BANGUMI_EXPAND_SUBSCRIBED_PAGE_PATH,
     MIKAN_BANGUMI_HOMEPAGE_PATH, MIKAN_BANGUMI_ID_QUERY_KEY, MIKAN_BANGUMI_POSTER_PATH,
     MIKAN_BANGUMI_RSS_PATH, MIKAN_EPISODE_HOMEPAGE_PATH, MIKAN_EPISODE_TORRENT_PATH,
-    MIKAN_FANSUB_ID_QUERY_KEY, MIKAN_LOGIN_PAGE_PATH, MIKAN_LOGIN_PAGE_SEARCH,
+    MIKAN_FANSUB_HOMEPAGE_PATH, MIKAN_FANSUB_ID_QUERY_KEY, MIKAN_LOGIN_PAGE_PATH,
-    MIKAN_POSTER_BUCKET_KEY, MIKAN_SEASON_FLOW_PAGE_PATH, MIKAN_SEASON_STR_QUERY_KEY,
+    MIKAN_LOGIN_PAGE_SEARCH, MIKAN_POSTER_BUCKET_KEY, MIKAN_SEASON_FLOW_PAGE_PATH,
-    MIKAN_SUBSCRIBER_SUBSCRIPTION_RSS_PATH, MIKAN_SUBSCRIBER_SUBSCRIPTION_TOKEN_QUERY_KEY,
+    MIKAN_SEASON_STR_QUERY_KEY, MIKAN_SUBSCRIBER_SUBSCRIPTION_RSS_PATH,
-    MIKAN_UNKNOWN_FANSUB_ID, MIKAN_UNKNOWN_FANSUB_NAME, MIKAN_YEAR_QUERY_KEY,
+    MIKAN_SUBSCRIBER_SUBSCRIPTION_TOKEN_QUERY_KEY, MIKAN_UNKNOWN_FANSUB_ID,
+    MIKAN_UNKNOWN_FANSUB_NAME, MIKAN_YEAR_QUERY_KEY,
 };
 pub use credential::MikanCredentialForm;
 pub use subscription::{
@@ -22,11 +23,12 @@ pub use subscription::{
 };
 pub use web::{
     MikanBangumiHash, MikanBangumiIndexHash, MikanBangumiIndexMeta, MikanBangumiMeta,
-    MikanBangumiPosterMeta, MikanEpisodeHash, MikanEpisodeMeta, MikanRssEpisodeItem,
+    MikanBangumiPosterMeta, MikanEpisodeHash, MikanEpisodeMeta, MikanFansubHash,
-    MikanSeasonFlowUrlMeta, MikanSeasonStr, MikanSubscriberSubscriptionRssUrlMeta,
+    MikanRssEpisodeItem, MikanSeasonFlowUrlMeta, MikanSeasonStr,
-    build_mikan_bangumi_expand_subscribed_url, build_mikan_bangumi_homepage_url,
+    MikanSubscriberSubscriptionRssUrlMeta, build_mikan_bangumi_expand_subscribed_url,
-    build_mikan_bangumi_subscription_rss_url, build_mikan_episode_homepage_url,
+    build_mikan_bangumi_homepage_url, build_mikan_bangumi_subscription_rss_url,
-    build_mikan_season_flow_url, build_mikan_subscriber_subscription_rss_url,
+    build_mikan_episode_homepage_url, build_mikan_season_flow_url,
+    build_mikan_subscriber_subscription_rss_url,
     extract_mikan_bangumi_index_meta_list_from_season_flow_fragment,
     extract_mikan_bangumi_meta_from_expand_subscribed_fragment,
     extract_mikan_episode_meta_from_episode_homepage_html,

apps/recorder/src/extract/mikan/web.rs

@@ -22,8 +22,8 @@ use crate::{
         mikan::{
             MIKAN_BANGUMI_EXPAND_SUBSCRIBED_PAGE_PATH, MIKAN_BANGUMI_HOMEPAGE_PATH,
             MIKAN_BANGUMI_ID_QUERY_KEY, MIKAN_BANGUMI_POSTER_PATH, MIKAN_BANGUMI_RSS_PATH,
-            MIKAN_EPISODE_HOMEPAGE_PATH, MIKAN_FANSUB_ID_QUERY_KEY, MIKAN_POSTER_BUCKET_KEY,
+            MIKAN_EPISODE_HOMEPAGE_PATH, MIKAN_FANSUB_HOMEPAGE_PATH, MIKAN_FANSUB_ID_QUERY_KEY,
-            MIKAN_SEASON_FLOW_PAGE_PATH, MIKAN_SEASON_STR_QUERY_KEY,
+            MIKAN_POSTER_BUCKET_KEY, MIKAN_SEASON_FLOW_PAGE_PATH, MIKAN_SEASON_STR_QUERY_KEY,
             MIKAN_SUBSCRIBER_SUBSCRIPTION_RSS_PATH, MIKAN_SUBSCRIBER_SUBSCRIPTION_TOKEN_QUERY_KEY,
             MIKAN_YEAR_QUERY_KEY, MikanClient,
         },
@@ -205,6 +205,32 @@ impl MikanBangumiMeta {
     }
 }
 
+#[derive(Clone, Debug, PartialEq, Eq, Hash)]
+pub struct MikanFansubHash {
+    pub mikan_fansub_id: String,
+}
+
+impl MikanFansubHash {
+    pub fn from_homepage_url(url: &Url) -> Option<Self> {
+        let path = url.path();
+        if path.starts_with(MIKAN_FANSUB_HOMEPAGE_PATH) {
+            let mikan_fansub_id = path.replace(&format!("{MIKAN_FANSUB_HOMEPAGE_PATH}/"), "");
+            Some(Self { mikan_fansub_id })
+        } else {
+            None
+        }
+    }
+
+    pub fn build_homepage_url(self, mikan_base_url: Url) -> Url {
+        let mut url = mikan_base_url;
+        url.set_path(&format!(
+            "{MIKAN_FANSUB_HOMEPAGE_PATH}/{}",
+            self.mikan_fansub_id
+        ));
+        url
+    }
+}
+
 #[derive(Clone, Debug, PartialEq)]
 pub struct MikanEpisodeMeta {
     pub homepage: Url,

apps/recorder/src/models/bangumi.rs

@@ -152,7 +152,10 @@ impl ActiveModel {
             season_raw: ActiveValue::Set(season_raw),
             fansub: ActiveValue::Set(Some(meta.fansub)),
             poster_link: ActiveValue::Set(poster_link),
-            origin_poster_link: ActiveValue::Set(meta.origin_poster_src.map(|src| src.to_string())),
+            origin_poster_link: ActiveValue::Set(
+                meta.origin_poster_src
+                    .map(|src| src[url::Position::BeforePath..].to_string()),
+            ),
             homepage: ActiveValue::Set(Some(meta.homepage.to_string())),
             rss_link: ActiveValue::Set(Some(rss_url.to_string())),
             ..Default::default()