//! 文章列表查询接口。 //! //! 提供已发布文章分页、管理员全量列表、以及按标签筛选三种查询能力, //! 均通过缓存层减少重复数据库访问。 //! Dioxus server function,注册在 `/api` 路径下。 //! 仅在 `feature = "server"` 启用的服务端构建中查询数据库。 use dioxus::prelude::*; #[cfg(feature = "server")] use super::helpers::{get_current_admin_user, row_to_post_list_item}; use super::types::PostListResponse; #[cfg(feature = "server")] use crate::api::error::AppError; #[cfg(feature = "server")] use crate::db::pool::get_conn; /// 单页允许的最大文章数。 /// /// 公开的 `list_published_posts` 接口无需认证,若不对 `per_page` 设上限, /// 攻击者可传入巨大值迫使数据库扫描并实例化超大 Vec,造成内存放大与拒绝服务。 #[cfg(feature = "server")] const MAX_PER_PAGE: i32 = 50; /// 允许的最大页码。 /// /// `page` 无上限时,攻击者可用海量不同 `page` 值撑大缓存键空间(缓存污染), /// 并触发无意义的超大 `OFFSET` 扫描。10_000 对任何实际博客都足够宽裕 /// (配合 `MAX_PER_PAGE` 最多覆盖 50 万篇文章),同时把缓存键空间限制在有限范围。 #[cfg(feature = "server")] const MAX_PAGE: i32 = 10_000; /// 将分页参数钳制到安全范围:页码 1–`MAX_PAGE`,每页 1–`MAX_PER_PAGE`。 /// /// 注意:返回值必须同时用于缓存键与 SQL 查询,避免同一逻辑页落入不同缓存条目。 #[cfg(feature = "server")] fn clamp_pagination(page: i32, per_page: i32) -> (i32, i32) { (page.clamp(1, MAX_PAGE), per_page.clamp(1, MAX_PER_PAGE)) } /// 获取已发布文章分页列表。 /// /// 优先命中缓存;未命中时查询总数与分页记录,并按 published_at 降序排列。 #[server(ListPublishedPosts, "/api")] pub async fn list_published_posts( page: i32, per_page: i32, ) -> Result { // 钳制分页参数,防止无认证调用方请求超大每页数量导致内存放大 / DoS。 let (page, per_page) = clamp_pagination(page, per_page); #[cfg(feature = "server")] { let cache_key = crate::cache::CacheKey::PublishedPosts { page, per_page }; if let Some((cached_posts, cached_total)) = crate::cache::get_post_list(&cache_key).await { return Ok(PostListResponse { posts: cached_posts, total: cached_total, }); } let client = get_conn().await.map_err(AppError::db_conn)?; // 优先读取缓存中的已发布文章总数,否则查询数据库并回填缓存。 let total = if let Some(cached_total) = crate::cache::get_total_published_posts().await { cached_total } else { let count_row = client .query_one( "SELECT COUNT(*) FROM posts WHERE status = 'published' AND deleted_at IS NULL", &[], ) .await .map_err(AppError::query)?; let total: i64 = count_row.get(0); crate::cache::set_total_published_posts(total).await; total }; let offset = ((page - 1).max(0) as i64) * (per_page as i64); let limit = per_page as i64; let rows = client .query( "SELECT p.id, p.author_id, p.title, p.slug, p.summary, p.status, p.published_at, p.created_at, p.updated_at, p.cover_image, p.word_count, p.reading_time, COALESCE(array_agg(t.name) FILTER (WHERE t.name IS NOT NULL), '{}') as tags FROM posts p LEFT JOIN post_tags pt ON p.id = pt.post_id LEFT JOIN tags t ON pt.tag_id = t.id WHERE p.status = 'published' AND p.deleted_at IS NULL GROUP BY p.id ORDER BY p.published_at DESC LIMIT $1 OFFSET $2", &[&limit, &offset], ) .await .map_err(AppError::query)?; let posts: Vec<_> = rows.iter().map(row_to_post_list_item).collect(); crate::cache::set_post_list(&cache_key, posts.clone(), total).await; Ok(PostListResponse { posts, total }) } #[cfg(not(feature = "server"))] { Ok(PostListResponse { posts: Vec::new(), total: 0, }) } } /// 获取管理员视角的全部文章列表(含草稿与已发布)。 /// /// 需要 admin 权限;结果按创建时间降序,不走缓存。 #[server(ListPosts, "/api")] pub async fn list_posts( page: i32, per_page: i32, search: Option, ) -> Result { // 与公开接口保持一致的分页钳制,避免单次请求拉取过多记录。 let (page, per_page) = clamp_pagination(page, per_page); let _user = get_current_admin_user().await?; #[cfg(feature = "server")] { let client = get_conn().await.map_err(AppError::db_conn)?; // 归一化标题搜索词:trim 后为空则视为不搜索;限长 200 字符并转义 SQL // LIKE 通配符(% / _ / \),避免用户输入导致模式错配或全表误匹配。 // 与 search.rs 的全文检索不同:管理后台需覆盖草稿,且仅按标题匹配。 let title_filter: Option = search .as_deref() .map(str::trim) .filter(|s| !s.is_empty()) .map(|s| { let mut esc = String::with_capacity(s.len()); for ch in s.chars().take(200) { match ch { '\\' | '%' | '_' => { esc.push('\\'); esc.push(ch); } _ => esc.push(ch), } } esc }); let offset = ((page - 1).max(0) as i64) * (per_page as i64); let limit = per_page as i64; // 有搜索词时按 title ILIKE 子串匹配过滤(双侧 %),否则走原全量列表。 // posts 表对管理后台属低频访问、行数有限,ILIKE 全表扫可接受,靠 LIMIT 兜底。 let (total, rows) = if let Some(esc) = title_filter { let pattern = format!("%{esc}%"); let total: i64 = client .query_one( "SELECT COUNT(*) FROM posts WHERE deleted_at IS NULL AND title ILIKE $1 ESCAPE '\\'", &[&pattern], ) .await .map_err(AppError::query)? .get(0); let rows = client .query( "SELECT p.id, p.author_id, p.title, p.slug, p.summary, p.status, p.published_at, p.created_at, p.updated_at, p.cover_image, p.word_count, p.reading_time, COALESCE(array_agg(t.name) FILTER (WHERE t.name IS NOT NULL), '{}') as tags FROM posts p LEFT JOIN post_tags pt ON p.id = pt.post_id LEFT JOIN tags t ON pt.tag_id = t.id WHERE p.deleted_at IS NULL AND p.title ILIKE $3 ESCAPE '\\' GROUP BY p.id ORDER BY p.created_at DESC LIMIT $1 OFFSET $2", &[&limit, &offset, &pattern], ) .await .map_err(AppError::query)?; (total, rows) } else { let total: i64 = client .query_one("SELECT COUNT(*) FROM posts WHERE deleted_at IS NULL", &[]) .await .map_err(AppError::query)? .get(0); let rows = client .query( "SELECT p.id, p.author_id, p.title, p.slug, p.summary, p.status, p.published_at, p.created_at, p.updated_at, p.cover_image, p.word_count, p.reading_time, COALESCE(array_agg(t.name) FILTER (WHERE t.name IS NOT NULL), '{}') as tags FROM posts p LEFT JOIN post_tags pt ON p.id = pt.post_id LEFT JOIN tags t ON pt.tag_id = t.id WHERE p.deleted_at IS NULL GROUP BY p.id ORDER BY p.created_at DESC LIMIT $1 OFFSET $2", &[&limit, &offset], ) .await .map_err(AppError::query)?; (total, rows) }; let posts: Vec<_> = rows.iter().map(row_to_post_list_item).collect(); Ok(PostListResponse { posts, total }) } #[cfg(not(feature = "server"))] { Ok(PostListResponse { posts: Vec::new(), total: 0, }) } } /// 获取回收站中已软删除的文章列表。 /// /// 需要 admin 权限;按删除时间降序,不走缓存。 #[server(ListDeletedPosts, "/api")] pub async fn list_deleted_posts( page: i32, per_page: i32, ) -> Result { // 与 list_posts 一致的分页钳制。 let (page, per_page) = clamp_pagination(page, per_page); let _user = get_current_admin_user().await?; #[cfg(feature = "server")] { let client = get_conn().await.map_err(AppError::db_conn)?; let count_row = client .query_one( "SELECT COUNT(*) FROM posts WHERE deleted_at IS NOT NULL", &[], ) .await .map_err(AppError::query)?; let total: i64 = count_row.get(0); let offset = ((page - 1).max(0) as i64) * (per_page as i64); let limit = per_page as i64; let rows = client .query( "SELECT p.id, p.author_id, p.title, p.slug, p.summary, p.status, p.published_at, p.created_at, p.updated_at, p.cover_image, p.deleted_at, p.word_count, p.reading_time, COALESCE(array_agg(t.name) FILTER (WHERE t.name IS NOT NULL), '{}') as tags FROM posts p LEFT JOIN post_tags pt ON p.id = pt.post_id LEFT JOIN tags t ON pt.tag_id = t.id WHERE p.deleted_at IS NOT NULL GROUP BY p.id ORDER BY p.deleted_at DESC LIMIT $1 OFFSET $2", &[&limit, &offset], ) .await .map_err(AppError::query)?; let posts: Vec<_> = rows.iter().map(row_to_post_list_item).collect(); Ok(PostListResponse { posts, total }) } #[cfg(not(feature = "server"))] { Ok(PostListResponse { posts: Vec::new(), total: 0, }) } } /// 获取指定标签下的已发布文章列表。 /// /// 分页参数为可选: /// - `page` 与 `per_page` 均为 `None` 时返回该标签下全部已发布文章(上限 200), /// 用于无分页 UI 的标签详情页。 /// - 两者均提供时走标准分页(经 `clamp_pagination` 钳制)。 /// 结果缓存于按标签的分页键空间。 #[server(GetPostsByTag, "/api")] pub async fn get_posts_by_tag( tag_name: String, page: Option, per_page: Option, ) -> Result { #[cfg(feature = "server")] { // 仅当两个分页参数都提供时才走分页路径;任一为 None 视为不分页。 let (page, per_page) = match (page, per_page) { (Some(p), Some(pp)) => (Some(p), Some(pp)), _ => (None, None), }; let client = get_conn().await.map_err(AppError::db_conn)?; if let (Some(page), Some(per_page)) = (page, per_page) { // 分页路径:钳制参数,走分页缓存键。 let (page, per_page) = clamp_pagination(page, per_page); let cache_key = crate::cache::CacheKey::PostsByTagPage { tag: tag_name.clone(), page, per_page, }; if let Some((cached_posts, cached_total)) = crate::cache::get_posts_by_tag_paged(&cache_key).await { return Ok(PostListResponse { posts: cached_posts, total: cached_total, }); } // 标签下已发布文章总数。 let total: i64 = client .query_one( "SELECT COUNT(*) FROM posts p JOIN post_tags pt ON p.id = pt.post_id JOIN tags t ON pt.tag_id = t.id WHERE t.name = $1 AND p.status = 'published' AND p.deleted_at IS NULL", &[&tag_name], ) .await .map_err(AppError::query)? .get(0); let offset = ((page - 1).max(0) as i64) * (per_page as i64); let limit = per_page as i64; let rows = client .query( "SELECT p.id, p.author_id, p.title, p.slug, p.summary, p.status, p.published_at, p.created_at, p.updated_at, p.cover_image, p.word_count, p.reading_time, COALESCE(array_agg(t2.name) FILTER (WHERE t2.name IS NOT NULL), '{}') as tags FROM posts p JOIN post_tags pt ON p.id = pt.post_id JOIN tags t ON pt.tag_id = t.id LEFT JOIN post_tags pt2 ON p.id = pt2.post_id LEFT JOIN tags t2 ON pt2.tag_id = t2.id WHERE t.name = $1 AND p.status = 'published' AND p.deleted_at IS NULL GROUP BY p.id ORDER BY p.published_at DESC LIMIT $2 OFFSET $3", &[&tag_name, &limit, &offset], ) .await .map_err(AppError::query)?; let posts: Vec<_> = rows.iter().map(row_to_post_list_item).collect(); crate::cache::set_posts_by_tag_paged(&cache_key, posts.clone(), total).await; Ok(PostListResponse { posts, total }) } else { // 不分页路径:返回全部(上限 200),用于无翻页 UI 的标签详情页。 if let Some((cached_posts, cached_total)) = crate::cache::get_posts_by_tag(&tag_name).await { return Ok(PostListResponse { posts: cached_posts, total: cached_total, }); } // 真实总数(即使被 LIMIT 截断也返回完整计数)。 let total: i64 = client .query_one( "SELECT COUNT(*) FROM posts p JOIN post_tags pt ON p.id = pt.post_id JOIN tags t ON pt.tag_id = t.id WHERE t.name = $1 AND p.status = 'published' AND p.deleted_at IS NULL", &[&tag_name], ) .await .map_err(AppError::query)? .get(0); let rows = client .query( "SELECT p.id, p.author_id, p.title, p.slug, p.summary, p.status, p.published_at, p.created_at, p.updated_at, p.cover_image, p.word_count, p.reading_time, COALESCE(array_agg(t2.name) FILTER (WHERE t2.name IS NOT NULL), '{}') as tags FROM posts p JOIN post_tags pt ON p.id = pt.post_id JOIN tags t ON pt.tag_id = t.id LEFT JOIN post_tags pt2 ON p.id = pt2.post_id LEFT JOIN tags t2 ON pt2.tag_id = t2.id WHERE t.name = $1 AND p.status = 'published' AND p.deleted_at IS NULL GROUP BY p.id ORDER BY p.published_at DESC LIMIT 200", &[&tag_name], ) .await .map_err(AppError::query)?; let posts: Vec<_> = rows.iter().map(row_to_post_list_item).collect(); // total 为真实 COUNT(*),不再用 posts.len()。 crate::cache::set_posts_by_tag(&tag_name, posts.clone(), total).await; Ok(PostListResponse { posts, total }) } } #[cfg(not(feature = "server"))] { Ok(PostListResponse { posts: Vec::new(), total: 0, }) } } #[cfg(test)] mod tests { use super::*; #[test] fn clamp_pagination_keeps_valid_values() { assert_eq!(clamp_pagination(1, 10), (1, 10)); assert_eq!(clamp_pagination(3, 20), (3, 20)); } #[test] fn clamp_pagination_clamps_oversized_per_page() { // 攻击者传入超大 per_page 必须被压回上限,避免内存放大 / DoS。 assert_eq!(clamp_pagination(1, 1_000_000_000), (1, MAX_PER_PAGE)); assert_eq!(clamp_pagination(2, 51), (2, MAX_PER_PAGE)); } #[test] fn clamp_pagination_clamps_non_positive() { assert_eq!(clamp_pagination(0, 10), (1, 10)); assert_eq!(clamp_pagination(-5, 10), (1, 10)); assert_eq!(clamp_pagination(1, 0), (1, 1)); assert_eq!(clamp_pagination(1, -100), (1, 1)); } #[test] fn clamp_pagination_clamps_oversized_page() { // 巨大 page 必须被压回上限,避免无界 OFFSET 扫描与缓存键扇出。 assert_eq!(clamp_pagination(i32::MAX, 10), (MAX_PAGE, 10)); assert_eq!(clamp_pagination(MAX_PAGE + 1, 10), (MAX_PAGE, 10)); } #[test] fn clamp_pagination_max_page_boundary() { assert_eq!(clamp_pagination(MAX_PAGE, 10), (MAX_PAGE, 10)); assert_eq!(clamp_pagination(MAX_PAGE - 1, 10), (MAX_PAGE - 1, 10)); } #[test] fn clamp_pagination_max_per_page_boundary() { assert_eq!(clamp_pagination(1, MAX_PER_PAGE), (1, MAX_PER_PAGE)); assert_eq!(clamp_pagination(1, MAX_PER_PAGE - 1), (1, MAX_PER_PAGE - 1)); } }