feat: 支持默认搜索自动识别通配符

TODOS.md

@@ -170,6 +170,7 @@
 - [x] 实现简单现代并适配移动端的单页搜索界面
 - [x] 接入搜索排序分页详情内容变体和磁力链接复制
 - [x] 实现名称别名和文件路径的有限状态自动机正则搜索
+- [x] 在普通搜索中自动识别通配符并保留独立正则开关
 - [x] 实现种子详情文件列表后端分页并限制浏览器单页节点数量
 - [x] 保留种子详情顶部结构并使用 reka-ui 数字分页重构文件条目
 - [x] 支持用户选择并持久化文件列表每页数量

dht-search/README.md

@@ -152,9 +152,10 @@ GET /health
 GET /ready
 GET /stats
 GET /search?q=ubuntu&offset=0&limit=20
+GET /search?q=%2A.iso
 GET /search?q=&min_size=1048576&max_size=10737418240&extension=mkv
 GET /search?q=流浪地球&min_files=1&availability=active&heat=hot&sort=heat
-GET /search?q=%5ES%5Cd%7B2%7DE%5Cd%7B2%7D&regex=true
+GET /search?q=%5ES%5Cd%7B2%7DE%5Cd%7B2%7D&mode=regex
 GET /contents/{content_key}?offset=0&limit=20
 GET /torrents/{infohash}?file_offset=0&file_limit=100
 ```
@@ -163,7 +164,9 @@ GET /torrents/{infohash}?file_offset=0&file_limit=100
 
 搜索支持中文英文数字和文件名片段匹配
 
-设置 `regex=true` 后查询文本作为不区分大小写的正则表达式匹配名称 别名和文件路径 正则最长 256 字节并由 Tantivy 有限状态自动机执行
+默认搜索会自动识别不区分大小写的通配符并匹配名称 别名和文件路径 `*` 表示任意长度字符 `?` 表示一个字符 例如 `*.iso` 匹配所有以 `.iso` 结尾的已索引名称或文件路径 不含通配符时保持普通关键词和片段搜索
+
+设置 `mode=regex` 后查询文本作为不区分大小写的正则表达式匹配名称 别名和文件路径 通配符和正则最长 256 字节并由 Tantivy 有限状态自动机执行
 
 过滤参数还包括 `min_files` `max_files` `first_seen_after` `first_seen_before` `last_seen_after` `last_seen_before` `availability` 和 `heat`
 

dht-search/src/api/handlers.rs

@@ -10,7 +10,7 @@ use axum::{
 };
 use dht_search::{
     domain::{InfoHash, MetadataRejectionReason},
-    search::{SearchOptions, SearchPage, SearchSort},
+    search::{SearchMode, SearchOptions, SearchPage, SearchSort},
     storage::VerificationPriority,
 };
 
@@ -118,8 +118,10 @@ pub(crate) async fn search(
     if request.q.len() > 512 {
         return Err(ApiError::bad_request("查询文本不能超过 512 字节"));
     }
-    if request.regex && request.q.len() > 256 {
-        return Err(ApiError::bad_request("正则表达式不能超过 256 字节"));
+    if (request.mode != SearchMode::Text || request.q.contains('*') || request.q.contains('?'))
+        && request.q.len() > 256
+    {
+        return Err(ApiError::bad_request("通配符或正则表达式不能超过 256 字节"));
     }
     validate_range(request.min_size, request.max_size, "min_size", "max_size")?;
     validate_range(
@@ -141,7 +143,7 @@ pub(crate) async fn search(
         "last_seen_before",
     )?;
     let mut query = request.q;
-    let content_key = if !request.regex {
+    let content_key = if request.mode == SearchMode::Text {
         InfoHash::from_str(query.trim()).ok()
     } else {
         None
@@ -169,7 +171,7 @@ pub(crate) async fn search(
     let page = tokio::task::spawn_blocking(move || {
         state.search.search_with(SearchOptions {
             query,
-            regex: request.regex,
+            mode: request.mode,
             offset: request.offset,
             limit: request.limit,
             min_size: request.min_size,

dht-search/src/api/mod.rs

@@ -172,7 +172,7 @@ mod tests {
             .clone()
             .oneshot(
                 Request::builder()
-                    .uri("/search?q=example.%2Amovie&regex=true")
+                    .uri("/search?q=example.%2Amovie&mode=regex")
                     .body(Body::empty())
                     .unwrap(),
             )
@@ -188,7 +188,23 @@ mod tests {
             .clone()
             .oneshot(
                 Request::builder()
-                    .uri("/search?q=%5B&regex=true")
+                    .uri("/search?q=%2A.mkv")
+                    .body(Body::empty())
+                    .unwrap(),
+            )
+            .await
+            .unwrap();
+        assert_eq!(response.status(), StatusCode::OK);
+        let json: serde_json::Value =
+            serde_json::from_slice(&to_bytes(response.into_body(), usize::MAX).await.unwrap())
+                .unwrap();
+        assert_eq!(json["total"], 1);
+
+        let response = app
+            .clone()
+            .oneshot(
+                Request::builder()
+                    .uri("/search?q=%5B&mode=regex")
                     .body(Body::empty())
                     .unwrap(),
             )

dht-search/src/api/request.rs

@@ -2,7 +2,7 @@
 
 use dht_search::{
     domain::{AvailabilityStatus, HeatLevel},
-    search::SearchSort,
+    search::{SearchMode, SearchSort},
 };
 use serde::Deserialize;
 
@@ -15,7 +15,7 @@ pub(crate) struct SearchRequest {
     #[serde(default)]
     pub(crate) q: String,
     #[serde(default)]
-    pub(crate) regex: bool,
+    pub(crate) mode: SearchMode,
     #[serde(default)]
     pub(crate) offset: usize,
     #[serde(default = "default_limit")]

dht-search/src/bin/dht-benchmark/workload.rs

@@ -2,7 +2,7 @@
 
 use std::error::Error;
 
-use dht_search::search::{SearchOptions, SearchSort};
+use dht_search::search::{SearchMode, SearchOptions, SearchSort};
 
 use super::dataset::generate_record;
 
@@ -56,7 +56,7 @@ pub(crate) fn query_cases(
             name: "有限状态正则",
             options: SearchOptions {
                 query: "ubuntu.*desktop".into(),
-                regex: true,
+                mode: SearchMode::Regex,
                 limit: 20,
                 ..SearchOptions::default()
             },

dht-search/src/search/filter.rs

@@ -12,7 +12,7 @@ use unicode_normalization::UnicodeNormalization;
 use super::{
     SearchError,
     document::availability_number,
-    query::{SearchOptions, SearchSort},
+    query::{SearchMode, SearchOptions, SearchSort},
     schema::SearchFields,
 };
 
@@ -29,10 +29,14 @@ pub(crate) fn prepare(
     let query_text = options.query.trim();
     if query_text.is_empty() || query_text == "*" {
         clauses.push(Box::new(AllQuery));
-    } else if options.regex {
-        clauses.push(regex_query(query_text, fields)?);
     } else {
-        clauses.push(text_query(query_text, fields));
+        clauses.push(match options.mode {
+            SearchMode::Text if has_wildcard_syntax(query_text) => {
+                wildcard_query(query_text, fields)?
+            }
+            SearchMode::Text => text_query(query_text, fields),
+            SearchMode::Regex => regex_query(query_text, fields)?,
+        });
     }
     if let Some(content_key) = options.content_key {
         clauses.push(Box::new(TermQuery::new(
@@ -131,6 +135,50 @@ fn regex_query(pattern: &str, fields: SearchFields) -> Result<Box<dyn Query>, Se
     )?))
 }
 
+fn wildcard_query(pattern: &str, fields: SearchFields) -> Result<Box<dyn Query>, SearchError> {
+    Ok(Box::new(RegexQuery::from_pattern(
+        &wildcard_pattern(pattern),
+        fields.regex_text,
+    )?))
+}
+
+fn has_wildcard_syntax(pattern: &str) -> bool {
+    pattern.contains('*') || pattern.contains('?')
+}
+
+fn wildcard_pattern(pattern: &str) -> String {
+    let pattern = normalize_text(pattern);
+    let mut regex = String::with_capacity(pattern.len());
+    let mut escaped = false;
+    for character in pattern.chars() {
+        if escaped {
+            push_regex_literal(&mut regex, character);
+            escaped = false;
+            continue;
+        }
+        match character {
+            '\\' => escaped = true,
+            '*' => regex.push_str(".*"),
+            '?' => regex.push('.'),
+            literal => push_regex_literal(&mut regex, literal),
+        }
+    }
+    if escaped {
+        push_regex_literal(&mut regex, '\\');
+    }
+    regex
+}
+
+fn push_regex_literal(regex: &mut String, character: char) {
+    if matches!(
+        character,
+        '\\' | '.' | '+' | '*' | '?' | '(' | ')' | '|' | '[' | ']' | '{' | '}' | '^' | '$'
+    ) {
+        regex.push('\\');
+    }
+    regex.push(character);
+}
+
 fn text_query(query: &str, fields: SearchFields) -> Box<dyn Query> {
     let normalized = normalize_text(query.trim());
     if normalized.len() == 40 && normalized.bytes().all(|byte| byte.is_ascii_hexdigit()) {
@@ -220,3 +268,18 @@ fn add_range(
         .unwrap_or(Bound::Unbounded);
     clauses.push(Box::new(RangeQuery::new(lower, upper)));
 }
+
+#[cfg(test)]
+mod tests {
+    use super::{has_wildcard_syntax, wildcard_pattern};
+
+    #[test]
+    fn wildcard_conversion_is_case_insensitive_and_escapes_regex_syntax() {
+        assert_eq!(wildcard_pattern("*.ISO"), r".*\.iso");
+        assert_eq!(wildcard_pattern("file?.[ch]"), r"file.\.\[ch\]");
+        assert_eq!(wildcard_pattern(r"literal\*name"), r"literal\*name");
+        assert!(has_wildcard_syntax("*.iso"));
+        assert!(has_wildcard_syntax("file?.mkv"));
+        assert!(has_wildcard_syntax(r"literal\*name"));
+    }
+}

dht-search/src/search/indexer.rs

@@ -271,6 +271,7 @@ mod tests {
     use crate::domain::{
         AvailabilityStatus, ContentGroupBuilder, InfoHash, TorrentFile, TorrentRecord,
     };
+    use crate::search::SearchMode;
     use crate::storage::{RocksTorrentRepository, TorrentRepository};
 
     use super::*;
@@ -364,7 +365,7 @@ mod tests {
         let name = engine
             .search_with(SearchOptions {
                 query: r"ubuntu\s+linux\s+24\.0[0-9]".into(),
-                regex: true,
+                mode: SearchMode::Regex,
                 limit: 10,
                 ..SearchOptions::default()
             })
@@ -373,7 +374,7 @@ mod tests {
         let path = engine
             .search_with(SearchOptions {
                 query: r"ubuntu\.(iso|img)$".into(),
-                regex: true,
+                mode: SearchMode::Regex,
                 limit: 10,
                 ..SearchOptions::default()
             })
@@ -381,6 +382,32 @@ mod tests {
         assert_eq!(path.total, 1);
     }
 
+    #[test]
+    fn wildcard_matches_names_and_file_extensions() {
+        let directory = TempDir::new().unwrap();
+        let engine = SearchEngine::open(directory.path()).unwrap();
+        index_records(&engine, &[record()]);
+
+        for pattern in ["*.iso", "Ubuntu*", "ubuntu.?so"] {
+            let page = engine
+                .search_with(SearchOptions {
+                    query: pattern.into(),
+                    limit: 10,
+                    ..SearchOptions::default()
+                })
+                .unwrap();
+            assert_eq!(page.total, 1, "pattern {pattern}");
+        }
+        let missing = engine
+            .search_with(SearchOptions {
+                query: "*.img".into(),
+                limit: 10,
+                ..SearchOptions::default()
+            })
+            .unwrap();
+        assert_eq!(missing.total, 0);
+    }
+
     #[test]
     fn invalid_regex_is_rejected() {
         let directory = TempDir::new().unwrap();
@@ -388,7 +415,7 @@ mod tests {
         index_records(&engine, &[record()]);
         let result = engine.search_with(SearchOptions {
             query: "[".into(),
-            regex: true,
+            mode: SearchMode::Regex,
             limit: 10,
             ..SearchOptions::default()
         });

dht-search/src/search/mod.rs

@@ -7,7 +7,9 @@ mod query;
 mod schema;
 
 pub use indexer::SearchEngine;
-pub use query::{AvailabilitySummary, SearchHit, SearchOptions, SearchPage, SearchSort};
+pub use query::{
+    AvailabilitySummary, SearchHit, SearchMode, SearchOptions, SearchPage, SearchSort,
+};
 
 #[derive(Debug, thiserror::Error)]
 pub enum SearchError {

dht-search/src/search/query.rs

@@ -17,10 +17,18 @@ pub enum SearchSort {
     Discoveries,
 }
 
+#[derive(Debug, Clone, Copy, Default, PartialEq, Eq, Deserialize, Serialize)]
+#[serde(rename_all = "snake_case")]
+pub enum SearchMode {
+    #[default]
+    Text,
+    Regex,
+}
+
 #[derive(Debug, Clone, Default)]
 pub struct SearchOptions {
     pub query: String,
-    pub regex: bool,
+    pub mode: SearchMode,
     pub offset: usize,
     pub limit: usize,
     pub min_size: Option<u64>,

web/README.md

@@ -50,7 +50,7 @@ Axum 根据配置中的 `http.web_dir` 提供静态资源和单页回退 不需
 ## 已实现功能
 
 - 关键词 文件名片段和精确 infohash 搜索
-- 普通关键词和名称 文件路径正则搜索
+- 自动识别通配符的普通关键词搜索和名称 文件路径正则搜索
 - 相关度 时间 热度 大小和发现次数排序
 - 有上限的结果分页和 URL 查询恢复
 - 基于 reka-ui 的搜索结果数字分页和浏览器持久化每页数量选择

web/src/App.vue

@@ -43,6 +43,10 @@ let statsRequestActive = false
 
 const pageNumber = computed(() => Math.floor((page.value?.offset ?? 0) / limit.value) + 1)
 const paginationTotal = computed(() => Math.min(page.value?.total ?? 0, 10_000 + limit.value))
+const searchPlaceholder = computed(() => {
+  if (regexMode.value) return '输入正则表达式'
+  return '输入名称、文件名、*.iso 或 40 位 infohash'
+})
 const resultPageSizeOptions = [
   { value: '10', label: '10 条/页' },
   { value: '20', label: '20 条/页' },
@@ -58,12 +62,11 @@ const sortOptions: ReadonlyArray<{ value: SearchSort; label: string }> = [
   { value: 'size_desc', label: '大小降序' },
   { value: 'size_asc', label: '大小升序' },
 ]
-
 function updateBrowserUrl() {
   const params = new URLSearchParams()
   if (submittedQuery.value) params.set('q', submittedQuery.value)
   if (sort.value !== 'latest') params.set('sort', sort.value)
-  if (regexMode.value) params.set('regex', 'true')
+  if (regexMode.value) params.set('mode', 'regex')
   if (offset.value) params.set('offset', String(offset.value))
   history.replaceState(null, '', params.size ? `?${params}` : location.pathname)
 }
@@ -185,7 +188,7 @@ onMounted(() => {
   const params = new URLSearchParams(location.search)
   query.value = params.get('q') ?? ''
   submittedQuery.value = query.value
-  regexMode.value = params.get('regex') === 'true'
+  regexMode.value = params.get('mode') === 'regex'
   const initialSort = params.get('sort') as SearchSort | null
   if (initialSort && ['relevance', 'latest', 'oldest', 'heat', 'size_desc', 'size_asc', 'discoveries'].includes(initialSort)) sort.value = initialSort
   else if (query.value.trim()) sort.value = 'relevance'
@@ -225,7 +228,7 @@ onBeforeUnmount(() => {
           <div class="flex flex-wrap items-center gap-2 p-2">
             <Search class="ml-2.5 size-4 shrink-0 text-muted-foreground" />
             <label class="sr-only" for="search-query">搜索关键词</label>
-            <input id="search-query" v-model="query" class="h-9 min-w-48 flex-1 bg-transparent px-2 text-sm outline-none placeholder:text-muted-foreground" :maxlength="regexMode ? 256 : 512" :placeholder="regexMode ? '输入正则表达式' : '输入名称、文件名或 40 位 infohash'" type="search" />
+            <input id="search-query" v-model="query" class="h-9 min-w-48 flex-1 bg-transparent px-2 text-sm outline-none placeholder:text-muted-foreground" :maxlength="regexMode ? 256 : 512" :placeholder="searchPlaceholder" type="search" />
             <Button size="sm" type="button" :variant="regexMode ? 'secondary' : 'ghost'" :aria-pressed="regexMode" title="正则搜索" @click="toggleRegex"><span class="font-mono">.*</span><span class="hidden sm:inline">正则</span></Button>
             <AppSelect :model-value="sort" :options="sortOptions" label="结果排序" @update:model-value="changeSort" />
             <Button type="submit">搜索</Button>
@@ -235,7 +238,7 @@ onBeforeUnmount(() => {
         <div v-if="loading" class="flex min-h-72 items-center justify-center text-sm text-muted-foreground"><LoaderCircle class="mr-2 size-5 animate-spin" />正在搜索</div>
         <div v-else-if="error" class="mt-4 flex min-h-72 flex-col items-center justify-center gap-4 rounded-xl border border-dashed text-center"><Server class="size-8 text-muted-foreground" /><div><p class="font-medium">无法完成搜索</p><p class="mt-1 text-sm text-muted-foreground">{{ error }}</p></div><Button variant="outline" @click="loadSearch">重试</Button></div>
         <div v-else-if="page?.hits.length" class="mt-4 space-y-3"><SearchResultCard v-for="hit in page.hits" :key="hit.content_key" :hit="hit" @select="openDetail" /></div>
-        <div v-else class="mt-4 flex min-h-72 flex-col items-center justify-center rounded-xl border border-dashed text-center"><Search class="mb-3 size-8 text-muted-foreground" /><p class="font-medium">没有找到匹配内容</p><p class="mt-1 text-sm text-muted-foreground">尝试其他关键词或正则表达式</p></div>
+        <div v-else class="mt-4 flex min-h-72 flex-col items-center justify-center rounded-xl border border-dashed text-center"><Search class="mb-3 size-8 text-muted-foreground" /><p class="font-medium">没有找到匹配内容</p><p class="mt-1 text-sm text-muted-foreground">尝试其他关键词、通配符或正则表达式</p></div>
 
         <div v-if="page && page.total > 0" class="mt-6 flex flex-wrap items-center justify-center gap-2">
           <AppSelect :model-value="String(limit)" :options="resultPageSizeOptions" label="每页搜索结果数量" @update:model-value="changeResultPageSize" />

web/src/lib/api.ts

@@ -36,7 +36,7 @@ export function search(input: SearchInput, signal?: AbortSignal): Promise<Search
     limit: String(input.limit),
     sort: input.sort,
   })
-  if (input.regex) params.set('regex', 'true')
+  if (input.regex) params.set('mode', 'regex')
   return request<SearchPage>(`/search?${params}`, signal)
 }