fix: resolve 10 code review issues

High priority:
- Fix concurrent race condition for view_count/like_count (atomic update)
- Add route request ID tracking to prevent race conditions
- Filter get_joke by status=approved (no pending content leak)
- Add error feedback for like button

Performance:
- Optimize random joke query (avoid full table sort)
- Limit page_size max to 100 (DoS prevention)

Medium:
- Add localStorage quota error handling
- Handle empty AI response gracefully
- Fix generate content title extraction

Low:
- Add rejected_jokes to stats API
- Update dashboard to show rejected count

.claude/settings.local.json

@@ -6,7 +6,15 @@
       "Skill(subagent-driven-development)",
       "Bash(git add *)",
       "Bash(git commit *)",
-      "Bash(python -c \"from app.models import Joke, JokeType, JokeCrowd, AdminUser; print\\('All models imported successfully'\\)\")"
+      "Bash(python -c \"from app.models import Joke, JokeType, JokeCrowd, AdminUser; print\\('All models imported successfully'\\)\")",
+      "Bash(curl -s -o /dev/null -w \"%{http_code}\" http://localhost:8001/health)",
+      "Bash(curl -s http://localhost:8001/api/auth/login -H \"Content-Type: application/json\" -d \"{\\\\\"username\\\\\":\\\\\"admin\\\\\",\\\\\"password\\\\\":\\\\\"admin123\\\\\"}\")",
+      "Bash(python -c \"import sys,json;d=json.load\\(sys.stdin\\);print\\(f'Jokes: {d[\\\\\"total\\\\\"]} total'\\)\")",
+      "Bash(npm run *)",
+      "Bash(curl -s -o /dev/null -w \"%{http_code}\" http://localhost:3000)",
+      "Bash(python -m uvicorn main:app --reload --port 8001)",
+      "Bash(python -c \"from app.schemas.joke import GenerateRequest, GenerateResponse; print\\('OK'\\)\")",
+      "Bash(python -c \"from main import app; print\\('API imports OK'\\)\")"
     ]
   }
 }

__joke.json

@@ -0,0 +1 @@
+{"title":"程序员的幽默","content":"程序员去相亲，女方问：你有什么优点？程序员答：我bug少。女方：...","type_id":2,"crowd_id":2,"id":1,"status":"approved","view_count":2,"like_count":0,"created_at":"2026-05-21T14:05:38","updated_at":"2026-05-22T10:40:03","type_name":"谐音梗","crowd_name":"职场"}

__login.json

@@ -0,0 +1 @@
+{"access_token":"eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxIiwidXNlcm5hbWUiOiJhZG1pbiIsImV4cCI6MTc4MDM1NDE1OX0.jqZHtndPbRhPYY7ZrkRQGqLzAMtfbejzkhRCq7fwqGY","token_type":"bearer"}

admin/node_modules/.vite/deps/_metadata.json

@@ -0,0 +1,49 @@
+{
+  "hash": "529e4e6a",
+  "configHash": "47bdc112",
+  "lockfileHash": "81bd2f5f",
+  "browserHash": "e99f2664",
+  "optimized": {
+    "axios": {
+      "src": "../../axios/index.js",
+      "file": "axios.js",
+      "fileHash": "64a84452",
+      "needsInterop": false
+    },
+    "element-plus": {
+      "src": "../../element-plus/es/index.mjs",
+      "file": "element-plus.js",
+      "fileHash": "d8579f02",
+      "needsInterop": false
+    },
+    "pinia": {
+      "src": "../../pinia/dist/pinia.mjs",
+      "file": "pinia.js",
+      "fileHash": "cc8debcd",
+      "needsInterop": false
+    },
+    "vue": {
+      "src": "../../vue/dist/vue.runtime.esm-bundler.js",
+      "file": "vue.js",
+      "fileHash": "7c6ca729",
+      "needsInterop": false
+    },
+    "vue-router": {
+      "src": "../../vue-router/dist/vue-router.mjs",
+      "file": "vue-router.js",
+      "fileHash": "bb6a478b",
+      "needsInterop": false
+    }
+  },
+  "chunks": {
+    "chunk-YFT6OQ5R": {
+      "file": "chunk-YFT6OQ5R.js"
+    },
+    "chunk-UC4PQXLF": {
+      "file": "chunk-UC4PQXLF.js"
+    },
+    "chunk-G3PMV62Z": {
+      "file": "chunk-G3PMV62Z.js"
+    }
+  }
+}

admin/node_modules/.vite/deps/chunk-G3PMV62Z.js

@@ -0,0 +1,36 @@
+var __create = Object.create;
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __commonJS = (cb, mod) => function __require() {
+  return mod || (0, cb[__getOwnPropNames(cb)[0]])((mod = { exports: {} }).exports, mod), mod.exports;
+};
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
+  // If the importer is in node compatibility mode or this is not an ESM
+  // file that has been converted to a CommonJS file using a Babel-
+  // compatible transform (i.e. "__esModule" has not been set), then set
+  // "default" to the CommonJS "module.exports" for node compatibility.
+  isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
+  mod
+));
+
+export {
+  __commonJS,
+  __export,
+  __toESM
+};
+//# sourceMappingURL=chunk-G3PMV62Z.js.map

admin/node_modules/.vite/deps/chunk-G3PMV62Z.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": [],
+  "sourcesContent": [],
+  "mappings": "",
+  "names": []
+}

admin/node_modules/.vite/deps/chunk-YFT6OQ5R.js

@@ -0,0 +1,162 @@
+// node_modules/@vue/devtools-api/lib/esm/env.js
+function getDevtoolsGlobalHook() {
+  return getTarget().__VUE_DEVTOOLS_GLOBAL_HOOK__;
+}
+function getTarget() {
+  return typeof navigator !== "undefined" && typeof window !== "undefined" ? window : typeof globalThis !== "undefined" ? globalThis : {};
+}
+var isProxyAvailable = typeof Proxy === "function";
+
+// node_modules/@vue/devtools-api/lib/esm/const.js
+var HOOK_SETUP = "devtools-plugin:setup";
+var HOOK_PLUGIN_SETTINGS_SET = "plugin:settings:set";
+
+// node_modules/@vue/devtools-api/lib/esm/time.js
+var supported;
+var perf;
+function isPerformanceSupported() {
+  var _a;
+  if (supported !== void 0) {
+    return supported;
+  }
+  if (typeof window !== "undefined" && window.performance) {
+    supported = true;
+    perf = window.performance;
+  } else if (typeof globalThis !== "undefined" && ((_a = globalThis.perf_hooks) === null || _a === void 0 ? void 0 : _a.performance)) {
+    supported = true;
+    perf = globalThis.perf_hooks.performance;
+  } else {
+    supported = false;
+  }
+  return supported;
+}
+function now() {
+  return isPerformanceSupported() ? perf.now() : Date.now();
+}
+
+// node_modules/@vue/devtools-api/lib/esm/proxy.js
+var ApiProxy = class {
+  constructor(plugin, hook) {
+    this.target = null;
+    this.targetQueue = [];
+    this.onQueue = [];
+    this.plugin = plugin;
+    this.hook = hook;
+    const defaultSettings = {};
+    if (plugin.settings) {
+      for (const id in plugin.settings) {
+        const item = plugin.settings[id];
+        defaultSettings[id] = item.defaultValue;
+      }
+    }
+    const localSettingsSaveId = `__vue-devtools-plugin-settings__${plugin.id}`;
+    let currentSettings = Object.assign({}, defaultSettings);
+    try {
+      const raw = localStorage.getItem(localSettingsSaveId);
+      const data = JSON.parse(raw);
+      Object.assign(currentSettings, data);
+    } catch (e) {
+    }
+    this.fallbacks = {
+      getSettings() {
+        return currentSettings;
+      },
+      setSettings(value) {
+        try {
+          localStorage.setItem(localSettingsSaveId, JSON.stringify(value));
+        } catch (e) {
+        }
+        currentSettings = value;
+      },
+      now() {
+        return now();
+      }
+    };
+    if (hook) {
+      hook.on(HOOK_PLUGIN_SETTINGS_SET, (pluginId, value) => {
+        if (pluginId === this.plugin.id) {
+          this.fallbacks.setSettings(value);
+        }
+      });
+    }
+    this.proxiedOn = new Proxy({}, {
+      get: (_target, prop) => {
+        if (this.target) {
+          return this.target.on[prop];
+        } else {
+          return (...args) => {
+            this.onQueue.push({
+              method: prop,
+              args
+            });
+          };
+        }
+      }
+    });
+    this.proxiedTarget = new Proxy({}, {
+      get: (_target, prop) => {
+        if (this.target) {
+          return this.target[prop];
+        } else if (prop === "on") {
+          return this.proxiedOn;
+        } else if (Object.keys(this.fallbacks).includes(prop)) {
+          return (...args) => {
+            this.targetQueue.push({
+              method: prop,
+              args,
+              resolve: () => {
+              }
+            });
+            return this.fallbacks[prop](...args);
+          };
+        } else {
+          return (...args) => {
+            return new Promise((resolve) => {
+              this.targetQueue.push({
+                method: prop,
+                args,
+                resolve
+              });
+            });
+          };
+        }
+      }
+    });
+  }
+  async setRealTarget(target) {
+    this.target = target;
+    for (const item of this.onQueue) {
+      this.target.on[item.method](...item.args);
+    }
+    for (const item of this.targetQueue) {
+      item.resolve(await this.target[item.method](...item.args));
+    }
+  }
+};
+
+// node_modules/@vue/devtools-api/lib/esm/index.js
+function setupDevtoolsPlugin(pluginDescriptor, setupFn) {
+  const descriptor = pluginDescriptor;
+  const target = getTarget();
+  const hook = getDevtoolsGlobalHook();
+  const enableProxy = isProxyAvailable && descriptor.enableEarlyProxy;
+  if (hook && (target.__VUE_DEVTOOLS_PLUGIN_API_AVAILABLE__ || !enableProxy)) {
+    hook.emit(HOOK_SETUP, pluginDescriptor, setupFn);
+  } else {
+    const proxy = enableProxy ? new ApiProxy(descriptor, hook) : null;
+    const list = target.__VUE_DEVTOOLS_PLUGINS__ = target.__VUE_DEVTOOLS_PLUGINS__ || [];
+    list.push({
+      pluginDescriptor: descriptor,
+      setupFn,
+      proxy
+    });
+    if (proxy) {
+      setupFn(proxy.proxiedTarget);
+    }
+  }
+}
+
+export {
+  setupDevtoolsPlugin
+};
+//# sourceMappingURL=chunk-YFT6OQ5R.js.map

admin/node_modules/.vite/deps/package.json

@@ -0,0 +1,3 @@
+{
+  "type": "module"
+}

admin/node_modules/.vite/deps/vue.js

@@ -0,0 +1,348 @@
+import {
+  BaseTransition,
+  BaseTransitionPropsValidators,
+  Comment,
+  DeprecationTypes,
+  EffectScope,
+  ErrorCodes,
+  ErrorTypeStrings,
+  Fragment,
+  KeepAlive,
+  ReactiveEffect,
+  Static,
+  Suspense,
+  Teleport,
+  Text,
+  TrackOpTypes,
+  Transition,
+  TransitionGroup,
+  TriggerOpTypes,
+  VueElement,
+  assertNumber,
+  callWithAsyncErrorHandling,
+  callWithErrorHandling,
+  camelize,
+  capitalize,
+  cloneVNode,
+  compatUtils,
+  compile,
+  computed,
+  createApp,
+  createBaseVNode,
+  createBlock,
+  createCommentVNode,
+  createElementBlock,
+  createHydrationRenderer,
+  createPropsRestProxy,
+  createRenderer,
+  createSSRApp,
+  createSlots,
+  createStaticVNode,
+  createTextVNode,
+  createVNode,
+  customRef,
+  defineAsyncComponent,
+  defineComponent,
+  defineCustomElement,
+  defineEmits,
+  defineExpose,
+  defineModel,
+  defineOptions,
+  defineProps,
+  defineSSRCustomElement,
+  defineSlots,
+  devtools,
+  effect,
+  effectScope,
+  getCurrentInstance,
+  getCurrentScope,
+  getCurrentWatcher,
+  getTransitionRawChildren,
+  guardReactiveProps,
+  h,
+  handleError,
+  hasInjectionContext,
+  hydrate,
+  hydrateOnIdle,
+  hydrateOnInteraction,
+  hydrateOnMediaQuery,
+  hydrateOnVisible,
+  initCustomFormatter,
+  initDirectivesForSSR,
+  inject,
+  isMemoSame,
+  isProxy,
+  isReactive,
+  isReadonly,
+  isRef,
+  isRuntimeOnly,
+  isShallow,
+  isVNode,
+  markRaw,
+  mergeDefaults,
+  mergeModels,
+  mergeProps,
+  nextTick,
+  nodeOps,
+  normalizeClass,
+  normalizeProps,
+  normalizeStyle,
+  onActivated,
+  onBeforeMount,
+  onBeforeUnmount,
+  onBeforeUpdate,
+  onDeactivated,
+  onErrorCaptured,
+  onMounted,
+  onRenderTracked,
+  onRenderTriggered,
+  onScopeDispose,
+  onServerPrefetch,
+  onUnmounted,
+  onUpdated,
+  onWatcherCleanup,
+  openBlock,
+  patchProp,
+  popScopeId,
+  provide,
+  proxyRefs,
+  pushScopeId,
+  queuePostFlushCb,
+  reactive,
+  readonly,
+  ref,
+  registerRuntimeCompiler,
+  render,
+  renderList,
+  renderSlot,
+  resolveComponent,
+  resolveDirective,
+  resolveDynamicComponent,
+  resolveFilter,
+  resolveTransitionHooks,
+  setBlockTracking,
+  setDevtoolsHook,
+  setTransitionHooks,
+  shallowReactive,
+  shallowReadonly,
+  shallowRef,
+  ssrContextKey,
+  ssrUtils,
+  stop,
+  toDisplayString,
+  toHandlerKey,
+  toHandlers,
+  toRaw,
+  toRef,
+  toRefs,
+  toValue,
+  transformVNodeArgs,
+  triggerRef,
+  unref,
+  useAttrs,
+  useCssModule,
+  useCssVars,
+  useHost,
+  useId,
+  useModel,
+  useSSRContext,
+  useShadowRoot,
+  useSlots,
+  useTemplateRef,
+  useTransitionState,
+  vModelCheckbox,
+  vModelDynamic,
+  vModelRadio,
+  vModelSelect,
+  vModelText,
+  vShow,
+  version,
+  warn,
+  watch,
+  watchEffect,
+  watchPostEffect,
+  watchSyncEffect,
+  withAsyncContext,
+  withCtx,
+  withDefaults,
+  withDirectives,
+  withKeys,
+  withMemo,
+  withModifiers,
+  withScopeId
+} from "./chunk-UC4PQXLF.js";
+import "./chunk-G3PMV62Z.js";
+export {
+  BaseTransition,
+  BaseTransitionPropsValidators,
+  Comment,
+  DeprecationTypes,
+  EffectScope,
+  ErrorCodes,
+  ErrorTypeStrings,
+  Fragment,
+  KeepAlive,
+  ReactiveEffect,
+  Static,
+  Suspense,
+  Teleport,
+  Text,
+  TrackOpTypes,
+  Transition,
+  TransitionGroup,
+  TriggerOpTypes,
+  VueElement,
+  assertNumber,
+  callWithAsyncErrorHandling,
+  callWithErrorHandling,
+  camelize,
+  capitalize,
+  cloneVNode,
+  compatUtils,
+  compile,
+  computed,
+  createApp,
+  createBlock,
+  createCommentVNode,
+  createElementBlock,
+  createBaseVNode as createElementVNode,
+  createHydrationRenderer,
+  createPropsRestProxy,
+  createRenderer,
+  createSSRApp,
+  createSlots,
+  createStaticVNode,
+  createTextVNode,
+  createVNode,
+  customRef,
+  defineAsyncComponent,
+  defineComponent,
+  defineCustomElement,
+  defineEmits,
+  defineExpose,
+  defineModel,
+  defineOptions,
+  defineProps,
+  defineSSRCustomElement,
+  defineSlots,
+  devtools,
+  effect,
+  effectScope,
+  getCurrentInstance,
+  getCurrentScope,
+  getCurrentWatcher,
+  getTransitionRawChildren,
+  guardReactiveProps,
+  h,
+  handleError,
+  hasInjectionContext,
+  hydrate,
+  hydrateOnIdle,
+  hydrateOnInteraction,
+  hydrateOnMediaQuery,
+  hydrateOnVisible,
+  initCustomFormatter,
+  initDirectivesForSSR,
+  inject,
+  isMemoSame,
+  isProxy,
+  isReactive,
+  isReadonly,
+  isRef,
+  isRuntimeOnly,
+  isShallow,
+  isVNode,
+  markRaw,
+  mergeDefaults,
+  mergeModels,
+  mergeProps,
+  nextTick,
+  nodeOps,
+  normalizeClass,
+  normalizeProps,
+  normalizeStyle,
+  onActivated,
+  onBeforeMount,
+  onBeforeUnmount,
+  onBeforeUpdate,
+  onDeactivated,
+  onErrorCaptured,
+  onMounted,
+  onRenderTracked,
+  onRenderTriggered,
+  onScopeDispose,
+  onServerPrefetch,
+  onUnmounted,
+  onUpdated,
+  onWatcherCleanup,
+  openBlock,
+  patchProp,
+  popScopeId,
+  provide,
+  proxyRefs,
+  pushScopeId,
+  queuePostFlushCb,
+  reactive,
+  readonly,
+  ref,
+  registerRuntimeCompiler,
+  render,
+  renderList,
+  renderSlot,
+  resolveComponent,
+  resolveDirective,
+  resolveDynamicComponent,
+  resolveFilter,
+  resolveTransitionHooks,
+  setBlockTracking,
+  setDevtoolsHook,
+  setTransitionHooks,
+  shallowReactive,
+  shallowReadonly,
+  shallowRef,
+  ssrContextKey,
+  ssrUtils,
+  stop,
+  toDisplayString,
+  toHandlerKey,
+  toHandlers,
+  toRaw,
+  toRef,
+  toRefs,
+  toValue,
+  transformVNodeArgs,
+  triggerRef,
+  unref,
+  useAttrs,
+  useCssModule,
+  useCssVars,
+  useHost,
+  useId,
+  useModel,
+  useSSRContext,
+  useShadowRoot,
+  useSlots,
+  useTemplateRef,
+  useTransitionState,
+  vModelCheckbox,
+  vModelDynamic,
+  vModelRadio,
+  vModelSelect,
+  vModelText,
+  vShow,
+  version,
+  warn,
+  watch,
+  watchEffect,
+  watchPostEffect,
+  watchSyncEffect,
+  withAsyncContext,
+  withCtx,
+  withDefaults,
+  withDirectives,
+  withKeys,
+  withMemo,
+  withModifiers,
+  withScopeId
+};
+//# sourceMappingURL=vue.js.map

admin/node_modules/.vite/deps/vue.js.map

@@ -0,0 +1,7 @@
+{
+  "version": 3,
+  "sources": [],
+  "sourcesContent": [],
+  "mappings": "",
+  "names": []
+}

admin/src/api/category.js

@@ -1,7 +1,7 @@
 import request from './request'
 
-export const getTypes = () => request.get('/types')
-export const getCrowds = () => request.get('/crowds')
+export const getTypes = () => request.get('/categories/types')
+export const getCrowds = () => request.get('/categories/crowds')
 export const createType = data => request.post('/admin/types', data)
 export const updateType = (id, data) => request.put(`/admin/types/${id}`, data)
 export const deleteType = id => request.delete(`/admin/types/${id}`)

admin/src/api/feedback.js

@@ -0,0 +1,4 @@
+import request from './request'
+
+export const getFeedbacks = () => request.get('/admin/feedbacks')
+export const deleteFeedback = id => request.delete('/admin/feedbacks/' + id)

admin/src/api/link.js

@@ -0,0 +1,7 @@
+import request from './request'
+
+export const getLinks = () => request.get('/admin/links')
+export const getLink = id => request.get('/admin/links/' + id)
+export const createLink = data => request.post('/admin/links', data)
+export const updateLink = (id, data) => request.put('/admin/links/' + id, data)
+export const deleteLink = id => request.delete('/admin/links/' + id)

admin/src/api/request.js

@@ -13,15 +13,15 @@ request.interceptors.request.use(config => {
   return config
 })
 
-request.interceptors.response.use({
-  successHandler: res => res.data,
-  errorHandler: err => {
+request.interceptors.response.use(
+  res => res.data,
+  err => {
     if (err.response?.status === 401) {
       localStorage.removeItem('token')
       window.location.href = '/login'
     }
     return Promise.reject(err)
   }
-})
+)
 
 export default request

admin/src/api/setting.js

@@ -0,0 +1,8 @@
+import request from './request'
+
+export const getSettings = () => request.get('/admin/settings')
+export const getActiveSetting = () => request.get('/admin/settings/active')
+export const createSetting = data => request.post('/admin/settings', data)
+export const updateSetting = (id, data) => request.put(`/admin/settings/${id}`, data)
+export const toggleSetting = id => request.put(`/admin/settings/${id}/toggle`)
+export const deleteSetting = id => request.delete(`/admin/settings/${id}`)

admin/src/router/index.js

@@ -16,6 +16,10 @@ const routes = [
       { path: 'jokes/edit/:id?', name: 'JokeEdit', component: () => import('@/views/joke/Edit.vue') },
       { path: 'category/type', name: 'CategoryType', component: () => import('@/views/category/Type.vue') },
       { path: 'category/crowd', name: 'CategoryCrowd', component: () => import('@/views/category/Crowd.vue') },
+      { path: 'settings', name: 'Settings', component: () => import('@/views/setting/Index.vue') },
+      { path: 'links', name: 'LinkList', component: () => import('@/views/links/List.vue') },
+      { path: 'links/edit/:id?', name: 'LinkEdit', component: () => import('@/views/links/Edit.vue') },
+      { path: 'feedbacks', name: 'FeedbackList', component: () => import('@/views/feedback/List.vue') },
     ]
   }
 ]

admin/src/stores/joke.js

@@ -11,7 +11,10 @@ export const useJokeStore = defineStore('joke', {
     async fetchJokes(params) {
       this.loading = true
       try {
-        const res = await getJokes(params)
+        const cleanParams = Object.fromEntries(
+          Object.entries(params).filter(([_, v]) => v != null && v !== '')
+        )
+        const res = await getJokes(cleanParams)
         this.jokes = res.items
         this.total = res.total
       } finally {

admin/src/views/dashboard/index.vue

@@ -10,14 +10,6 @@
           </div>
         </el-card>
       </el-col>
-      <el-col :span="6">
-        <el-card>
-          <div class="stat-card">
-            <div class="stat-value">{{ stats.pending }}</div>
-            <div class="stat-label">待审核</div>
-          </div>
-        </el-card>
-      </el-col>
       <el-col :span="6">
         <el-card>
           <div class="stat-card">
@@ -26,6 +18,24 @@
           </div>
         </el-card>
       </el-col>
+      <el-col :span="6">
+        <el-card>
+          <div class="stat-card">
+            <div class="stat-value">{{ stats.pending }}</div>
+            <div class="stat-label">待审核</div>
+          </div>
+        </el-card>
+      </el-col>
+      <el-col :span="6">
+        <el-card>
+          <div class="stat-card">
+            <div class="stat-value">{{ stats.rejected }}</div>
+            <div class="stat-label">已拒绝</div>
+          </div>
+        </el-card>
+      </el-col>
+    </el-row>
+    <el-row :gutter="20" style="margin-top: 20px">
       <el-col :span="6">
         <el-card>
           <div class="stat-card">
@@ -34,6 +44,14 @@
           </div>
         </el-card>
       </el-col>
+      <el-col :span="6">
+        <el-card>
+          <div class="stat-card">
+            <div class="stat-value">{{ stats.likes }}</div>
+            <div class="stat-label">总点赞</div>
+          </div>
+        </el-card>
+      </el-col>
     </el-row>
   </div>
 </template>
@@ -42,11 +60,22 @@
 import { ref, onMounted } from 'vue'
 import { getStats } from '@/api/joke'
 
-const stats = ref({ total: 0, pending: 0, approved: 0, views: 0 })
+const stats = ref({ total: 0, pending: 0, approved: 0, rejected: 0, views: 0, likes: 0 })
 
 onMounted(async () => {
+  try {
     const res = await getStats()
-  stats.value = res
+    stats.value = {
+      total: res.total_jokes || 0,
+      pending: res.pending_jokes || 0,
+      approved: res.approved_jokes || 0,
+      rejected: res.rejected_jokes || 0,
+      views: res.total_views || 0,
+      likes: res.total_likes || 0,
+    }
+  } catch (e) {
+    console.error('获取统计数据失败', e)
+  }
 })
 </script>
 

admin/src/views/feedback/List.vue

@@ -0,0 +1,54 @@
+<template>
+  <div class="feedback-list">
+    <div class="header">
+      <h2>反馈管理</h2>
+    </div>
+    <el-table :data="feedbacks" v-loading="loading" style="margin-top: 20px">
+      <el-table-column prop="id" label="ID" width="60" />
+      <el-table-column prop="name" label="名称" width="120" />
+      <el-table-column prop="email" label="邮箱" width="200" />
+      <el-table-column prop="content" label="内容" min-width="300" show-overflow-tooltip />
+      <el-table-column prop="created_at" label="提交时间" width="180" />
+      <el-table-column label="操作" width="120">
+        <template #default="{ row }">
+          <el-button size="small" type="danger" @click="handleDelete(row.id)">删除</el-button>
+        </template>
+      </el-table-column>
+    </el-table>
+  </div>
+</template>
+
+<script setup>
+import { ref, onMounted } from 'vue'
+import { getFeedbacks, deleteFeedback } from '@/api/feedback'
+import { ElMessage, ElMessageBox } from 'element-plus'
+
+const feedbacks = ref([])
+const loading = ref(false)
+
+const loadFeedbacks = async () => {
+  loading.value = true
+  try {
+    feedbacks.value = await getFeedbacks()
+  } finally {
+    loading.value = false
+  }
+}
+
+const handleDelete = async (id) => {
+  await ElMessageBox.confirm('确定删除?')
+  await deleteFeedback(id)
+  ElMessage.success('删除成功')
+  loadFeedbacks()
+}
+
+onMounted(() => loadFeedbacks())
+</script>
+
+<style scoped>
+.header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+}
+</style>

admin/src/views/joke/Edit.vue

@@ -8,13 +8,13 @@
       <el-form-item label="内容" prop="content">
         <el-input v-model="form.content" type="textarea" :rows="6" />
       </el-form-item>
-      <el-form-item label="类型" prop="type_id">
-        <el-select v-model="form.type_id" placeholder="选择类型">
+      <el-form-item label="类型" prop="type_ids">
+        <el-select v-model="form.type_ids" multiple placeholder="选择类型（可多选）">
           <el-option v-for="t in types" :key="t.id" :label="t.name" :value="t.id" />
         </el-select>
       </el-form-item>
-      <el-form-item label="人群" prop="crowd_id">
-        <el-select v-model="form.crowd_id" placeholder="选择人群">
+      <el-form-item label="人群" prop="crowd_ids">
+        <el-select v-model="form.crowd_ids" multiple placeholder="选择人群（可多选）">
           <el-option v-for="c in crowds" :key="c.id" :label="c.name" :value="c.id" />
         </el-select>
       </el-form-item>
@@ -46,7 +46,7 @@ const formRef = ref()
 const isEdit = ref(false)
 
 const form = reactive({
-  title: '', content: '', type_id: null, crowd_id: null, status: 'pending'
+  title: '', content: '', type_ids: [], crowd_ids: [], status: 'pending'
 })
 
 const types = ref([])

admin/src/views/joke/List.vue

@@ -7,8 +7,16 @@
     <el-table :data="jokeStore.jokes" v-loading="jokeStore.loading" style="margin-top: 20px">
       <el-table-column prop="id" label="ID" width="60" />
       <el-table-column prop="title" label="标题" />
-      <el-table-column prop="type_name" label="类型" width="100" />
-      <el-table-column prop="crowd_name" label="人群" width="100" />
+      <el-table-column label="类型" width="140">
+        <template #default="{ row }">
+          <el-tag v-for="name in (row.type_names || [])" :key="name" size="small" style="margin:1px">{{ name }}</el-tag>
+        </template>
+      </el-table-column>
+      <el-table-column label="人群" width="140">
+        <template #default="{ row }">
+          <el-tag v-for="name in (row.crowd_names || [])" :key="name" size="small" type="success" style="margin:1px">{{ name }}</el-tag>
+        </template>
+      </el-table-column>
       <el-table-column prop="status" label="状态" width="100">
         <template #default="{ row }">
           <el-tag :type="row.status === 'approved' ? 'success' : row.status === 'pending' ? 'warning' : 'danger'">

admin/src/views/links/Edit.vue

@@ -0,0 +1,60 @@
+<template>
+  <div class="link-edit">
+    <h2>{{ isEdit ? '编辑友链' : '添加友链' }}</h2>
+    <el-form :model="form" ref="formRef" label-width="120px" style="margin-top: 20px; max-width: 600px">
+      <el-form-item label="名称" prop="name">
+        <el-input v-model="form.name" />
+      </el-form-item>
+      <el-form-item label="URL" prop="url">
+        <el-input v-model="form.url" placeholder="https://" />
+      </el-form-item>
+      <el-form-item label="描述" prop="description">
+        <el-input v-model="form.description" type="textarea" :rows="3" />
+      </el-form-item>
+      <el-form-item label="排序" prop="sort_order">
+        <el-input-number v-model="form.sort_order" :min="0" />
+      </el-form-item>
+      <el-form-item>
+        <el-button type="primary" @click="handleSave">保存</el-button>
+        <el-button @click="$router.back()">取消</el-button>
+      </el-form-item>
+    </el-form>
+  </div>
+</template>
+
+<script setup>
+import { ref, reactive, onMounted } from 'vue'
+import { useRoute, useRouter } from 'vue-router'
+import { getLink, createLink, updateLink } from '@/api/link'
+import { ElMessage } from 'element-plus'
+
+const route = useRoute()
+const router = useRouter()
+const formRef = ref()
+const isEdit = ref(false)
+
+const form = reactive({
+  name: '', url: '', description: '', sort_order: 0
+})
+
+const loadData = async () => {
+  if (route.params.id) {
+    isEdit.value = true
+    const res = await getLink(route.params.id)
+    Object.assign(form, res)
+  }
+}
+
+const handleSave = async () => {
+  if (isEdit.value) {
+    await updateLink(route.params.id, form)
+    ElMessage.success('更新成功')
+  } else {
+    await createLink(form)
+    ElMessage.success('创建成功')
+  }
+  router.push('/links')
+}
+
+onMounted(() => loadData())
+</script>

admin/src/views/links/List.vue

@@ -0,0 +1,61 @@
+<template>
+  <div class="link-list">
+    <div class="header">
+      <h2>友链管理</h2>
+      <el-button type="primary" @click="$router.push('/links/edit')">添加友链</el-button>
+    </div>
+    <el-table :data="links" v-loading="loading" style="margin-top: 20px">
+      <el-table-column prop="id" label="ID" width="60" />
+      <el-table-column prop="name" label="名称" />
+      <el-table-column prop="url" label="URL" min-width="200">
+        <template #default="{ row }">
+          <a :href="row.url" target="_blank">{{ row.url }}</a>
+        </template>
+      </el-table-column>
+      <el-table-column prop="description" label="描述" min-width="200" />
+      <el-table-column prop="sort_order" label="排序" width="80" />
+      <el-table-column prop="created_at" label="创建时间" width="180" />
+      <el-table-column label="操作" width="200">
+        <template #default="{ row }">
+          <el-button size="small" @click="$router.push(`/links/edit/${row.id}`)">编辑</el-button>
+          <el-button size="small" type="danger" @click="handleDelete(row.id)">删除</el-button>
+        </template>
+      </el-table-column>
+    </el-table>
+  </div>
+</template>
+
+<script setup>
+import { ref, onMounted } from 'vue'
+import { getLinks, deleteLink } from '@/api/link'
+import { ElMessage, ElMessageBox } from 'element-plus'
+
+const links = ref([])
+const loading = ref(false)
+
+const loadLinks = async () => {
+  loading.value = true
+  try {
+    links.value = await getLinks()
+  } finally {
+    loading.value = false
+  }
+}
+
+const handleDelete = async (id) => {
+  await ElMessageBox.confirm('确定删除?')
+  await deleteLink(id)
+  ElMessage.success('删除成功')
+  loadLinks()
+}
+
+onMounted(() => loadLinks())
+</script>
+
+<style scoped>
+.header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+}
+</style>

admin/src/views/setting/Index.vue

@@ -0,0 +1,198 @@
+<template>
+  <div class="setting-page">
+    <h2>AI 配置</h2>
+
+    <!-- AI 配置表单 -->
+    <el-card style="margin-top: 20px">
+      <template #header>
+        <span>AI 配置</span>
+      </template>
+      <el-form :model="form" label-width="120px">
+        <el-form-item label="提供商">
+          <el-select v-model="form.provider" style="width: 200px">
+            <el-option label="NVIDIA NIM" value="nvidia" />
+            <el-option label="OpenAI" value="openai" />
+            <el-option label="Anthropic" value="anthropic" />
+          </el-select>
+        </el-form-item>
+        <el-form-item label="API 地址">
+          <el-input v-model="form.api_base" placeholder="https://integrate.api.nvidia.com/v1" style="width: 400px" />
+        </el-form-item>
+        <el-form-item label="API 密钥">
+          <el-input v-model="form.api_key" type="password" show-password style="width: 400px" placeholder="nvapi-..." />
+        </el-form-item>
+        <el-form-item label="模型名称">
+          <el-input v-model="form.model_name" style="width: 300px" placeholder="nvidia/llama-3.1-nemotron-70b-instruct" />
+        </el-form-item>
+        <el-form-item label="温度">
+          <el-input-number v-model="form.temperature" :min="0" :max="2" :step="0.1" />
+        </el-form-item>
+        <el-form-item label="最大 Token">
+          <el-input-number v-model="form.max_tokens" :min="1" :max="8192" />
+        </el-form-item>
+      </el-form>
+    </el-card>
+
+    <!-- 爬虫配置 -->
+    <el-card style="margin-top: 20px">
+      <template #header>
+        <span>爬虫配置</span>
+      </template>
+      <el-form :model="form" label-width="120px">
+        <el-form-item label="启用爬虫">
+          <el-switch v-model="form.crawl_enabled" />
+        </el-form-item>
+        <el-form-item label="搜索关键词">
+          <el-input
+            v-model="form.crawl_keywords"
+            type="textarea"
+            :rows="2"
+            style="width: 400px"
+            placeholder="多个关键词用逗号分隔，如：冷笑话,段子,谐音梗"
+          />
+        </el-form-item>
+        <el-form-item label="每轮页数">
+          <el-input-number v-model="form.max_pages_per_run" :min="1" :max="20" />
+        </el-form-item>
+      </el-form>
+    </el-card>
+
+    <div style="margin-top: 20px">
+      <el-button type="primary" @click="handleSave" :loading="saving">保存配置</el-button>
+    </div>
+
+    <!-- 历史配置 -->
+    <el-card style="margin-top: 20px">
+      <template #header>
+        <span>历史配置</span>
+      </template>
+      <el-table :data="allSettings" style="width: 100%">
+        <el-table-column prop="id" label="ID" width="60" />
+        <el-table-column prop="provider" label="提供商" width="100" />
+        <el-table-column prop="model_name" label="模型" />
+        <el-table-column prop="api_base" label="API 地址" show-overflow-tooltip />
+        <el-table-column prop="is_active" label="激活" width="80">
+          <template #default="{ row }">
+            <el-tag :type="row.is_active ? 'success' : 'info'" size="small">
+              {{ row.is_active ? '是' : '否' }}
+            </el-tag>
+          </template>
+        </el-table-column>
+        <el-table-column prop="created_at" label="创建时间" width="180" />
+        <el-table-column label="操作" width="200">
+          <template #default="{ row }">
+            <el-button size="small" @click="handleEdit(row)">编辑</el-button>
+            <el-button size="small" type="warning" @click="handleToggle(row)" v-if="!row.is_active">激活</el-button>
+            <el-button size="small" type="danger" @click="handleDelete(row.id)">删除</el-button>
+          </template>
+        </el-table-column>
+      </el-table>
+    </el-card>
+  </div>
+</template>
+
+<script setup>
+import { ref, onMounted } from 'vue'
+import { ElMessage, ElMessageBox } from 'element-plus'
+import { getSettings, getActiveSetting, createSetting, updateSetting, toggleSetting, deleteSetting } from '@/api/setting'
+
+const form = ref({
+  provider: 'nvidia',
+  api_base: 'https://integrate.api.nvidia.com/v1',
+  api_key: '',
+  model_name: 'nvidia/llama-3.1-nemotron-70b-instruct',
+  temperature: 0.7,
+  max_tokens: 2048,
+  crawl_enabled: false,
+  crawl_keywords: '冷笑话,段子,谐音梗',
+  max_pages_per_run: 3,
+})
+
+const allSettings = ref([])
+const saving = ref(false)
+const editingId = ref(null)
+
+const loadData = async () => {
+  try {
+    const active = await getActiveSetting()
+    editingId.value = active.id
+    form.value = {
+      provider: active.provider,
+      api_base: active.api_base,
+      api_key: active.api_key,
+      model_name: active.model_name,
+      temperature: active.temperature,
+      max_tokens: active.max_tokens,
+      crawl_enabled: active.crawl_enabled,
+      crawl_keywords: active.crawl_keywords || '',
+      max_pages_per_run: active.max_pages_per_run,
+    }
+  } catch (e) {
+    // 没有激活配置，使用默认值
+  }
+
+  try {
+    allSettings.value = await getSettings()
+  } catch (e) {
+    console.error(e)
+  }
+}
+
+const handleSave = async () => {
+  saving.value = true
+  try {
+    if (editingId.value) {
+      await updateSetting(editingId.value, form.value)
+      ElMessage.success('保存成功')
+    } else {
+      const created = await createSetting(form.value)
+      editingId.value = created.id
+      await toggleSetting(created.id)  // 设为激活
+      ElMessage.success('创建并激活成功')
+    }
+    await loadData()
+  } catch (e) {
+    ElMessage.error('保存失败: ' + (e.message || ''))
+  } finally {
+    saving.value = false
+  }
+}
+
+const handleEdit = (row) => {
+  editingId.value = row.id
+  form.value = {
+    provider: row.provider,
+    api_base: row.api_base,
+    api_key: row.api_key,
+    model_name: row.model_name,
+    temperature: row.temperature,
+    max_tokens: row.max_tokens,
+    crawl_enabled: row.crawl_enabled,
+    crawl_keywords: row.crawl_keywords || '',
+    max_pages_per_run: row.max_pages_per_run,
+  }
+}
+
+const handleToggle = async (row) => {
+  await ElMessageBox.confirm('确定激活此配置？')
+  await toggleSetting(row.id)
+  ElMessage.success('激活成功')
+  await loadData()
+}
+
+const handleDelete = async (id) => {
+  await ElMessageBox.confirm('确定删除此配置？')
+  await deleteSetting(id)
+  ElMessage.success('删除成功')
+  if (editingId.value === id) {
+    editingId.value = null
+  }
+  await loadData()
+}
+
+onMounted(() => loadData())
+</script>
+
+<style scoped>
+.setting-page { max-width: 900px }
+</style>

admin/vite.config.js

@@ -13,7 +13,7 @@ export default defineConfig({
     port: 3001,
     proxy: {
       '/api': {
-        target: 'http://localhost:8000',
+        target: 'http://localhost:8001',
         changeOrigin: true
       }
     }

api/API操作文档.md

@@ -0,0 +1,125 @@
+# API 操作文档 - 笑话管理系统
+
+## 基础信息
+
+- **API 地址**: `http://<服务器IP>:8001`
+- **默认管理员**: `admin` / `admin123`
+- **认证方式**: JWT Bearer Token
+
+---
+
+## 一、登录获取 Token
+
+```bash
+curl -X POST http://<服务器IP>:8001/api/auth/login \
+  -H "Content-Type: application/json" \
+  -d '{"username": "admin", "password": "admin123"}'
+```
+
+**返回示例**：
+```json
+{
+  "access_token": "eyJhbGciOiJIUzI1NiIs...",
+  "token_type": "bearer"
+}
+```
+
+---
+
+## 二、提交笑话
+
+```bash
+curl -X POST http://<服务器IP>:8001/api/admin/jokes \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer <TOKEN>" \
+  -d '{
+    "title": "笑话标题",
+    "content": "笑话内容正文",
+    "type_id": 1,
+    "crowd_id": 2
+  }'
+```
+
+| 字段 | 类型 | 必填 | 说明 |
+|------|------|------|------|
+| `title` | string | 是 | 笑话标题 |
+| `content` | string | 是 | 笑话正文 |
+| `type_id` | int | 否 | 类型 ID（查分类列表获取） |
+| `crowd_id` | int | 否 | 人群 ID（查分类列表获取） |
+
+提交后 status 默认为 `pending`（待审核）。
+
+---
+
+## 三、查询分类 ID
+
+```bash
+# 查看所有类型
+curl http://<服务器IP>:8001/api/categories/types
+
+# 查看所有人群
+curl http://<服务器IP>:8001/api/categories/crowds
+```
+
+---
+
+## 四、完整脚本示例（批量提交）
+
+```bash
+#!/bin/bash
+SERVER="http://192.168.1.10:8001"
+
+# 登录获取 Token
+TOKEN=$(curl -s -X POST "$SERVER/api/auth/login" \
+  -H "Content-Type: application/json" \
+  -d '{"username":"admin","password":"admin123"}' \
+  | python -c "import sys,json; print(json.load(sys.stdin)['access_token'])")
+
+echo "Token: $TOKEN"
+
+# 批量提交笑话
+jokes=(
+  '{"title":"笑话1","content":"内容1","type_id":1,"crowd_id":2}'
+  '{"title":"笑话2","content":"内容2","type_id":1,"crowd_id":null}'
+)
+
+for joke in "${jokes[@]}"; do
+  curl -X POST "$SERVER/api/admin/jokes" \
+    -H "Content-Type: application/json" \
+    -H "Authorization: Bearer $TOKEN" \
+    -d "$joke"
+  echo ""
+done
+```
+
+---
+
+## 五、其他管理接口
+
+### 查询笑话列表
+```bash
+curl -H "Authorization: Bearer <TOKEN>" \
+  "http://<服务器IP>:8001/api/admin/jokes?page=1&page_size=20&status=pending"
+```
+
+### 审核通过
+```bash
+curl -X PUT http://<服务器IP>:8001/api/admin/jokes/batch-approve \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer <TOKEN>" \
+  -d '[1, 2, 3]'
+```
+
+### 删除笑话
+```bash
+curl -X DELETE http://<服务器IP>:8001/api/admin/jokes/1 \
+  -H "Authorization: Bearer <TOKEN>"
+```
+
+---
+
+## 六、注意事项
+
+1. 将 `<服务器IP>` 替换为实际的服务器 IP 地址
+2. 确保服务器防火墙已放行 **8001** 端口
+3. Token 有过期时间，过期后需重新登录获取

api/app/models/__init__.py

@@ -1,5 +1,8 @@
 from app.models.joke import Joke
 from app.models.category import JokeType, JokeCrowd
 from app.models.user import AdminUser
+from app.models.setting import AiSetting
+from app.models.link import Link
+from app.models.feedback import Feedback
 
-__all__ = ["Joke", "JokeType", "JokeCrowd", "AdminUser"]
+__all__ = ["Joke", "JokeType", "JokeCrowd", "AdminUser", "AiSetting", "Link", "Feedback"]

api/app/models/feedback.py

@@ -0,0 +1,13 @@
+from sqlalchemy import Column, Integer, String, Text, DateTime
+from sqlalchemy.sql import func
+
+from app.database import Base
+
+
+class Feedback(Base):
+    __tablename__ = "feedbacks"
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    name = Column(String(50), nullable=True)
+    email = Column(String(100), nullable=True)
+    content = Column(Text, nullable=False)
+    created_at = Column(DateTime, default=func.now())

api/app/models/joke.py

@@ -12,6 +12,9 @@ class Joke(Base):
     id = Column(Integer, primary_key=True, autoincrement=True)
     title = Column(String(200), nullable=False)
     content = Column(Text, nullable=False)
+    polished_content = Column(Text, nullable=True)
+    type_ids = Column(Text, nullable=True)
+    crowd_ids = Column(Text, nullable=True)
     type_id = Column(Integer, ForeignKey("joke_types.id"), nullable=True)
     crowd_id = Column(Integer, ForeignKey("joke_crowds.id"), nullable=True)
     status = Column(String(20), default="pending")

api/app/models/link.py

@@ -0,0 +1,14 @@
+from sqlalchemy import Column, Integer, String, DateTime
+from sqlalchemy.sql import func
+
+from app.database import Base
+
+
+class Link(Base):
+    __tablename__ = "links"
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    name = Column(String(100), nullable=False)
+    url = Column(String(500), nullable=False)
+    description = Column(String(200), nullable=True)
+    sort_order = Column(Integer, default=0)
+    created_at = Column(DateTime, default=func.now())

api/app/models/setting.py

@@ -0,0 +1,23 @@
+from sqlalchemy import Boolean, Column, DateTime, Float, Integer, String, Text, func
+
+from app.database import Base
+
+
+class AiSetting(Base):
+    __tablename__ = "ai_settings"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    provider = Column(String(50), nullable=False, default="nvidia")
+    api_base = Column(String(500), nullable=False, default="https://integrate.api.nvidia.com/v1")
+    api_key = Column(String(500), nullable=False, default="")
+    model_name = Column(String(100), nullable=False, default="nvidia/llama-3.1-nemotron-70b-instruct")
+    temperature = Column(Float, nullable=False, default=0.7)
+    max_tokens = Column(Integer, nullable=False, default=2048)
+
+    crawl_enabled = Column(Boolean, nullable=False, default=False)
+    crawl_keywords = Column(Text, nullable=True, default="")
+    max_pages_per_run = Column(Integer, nullable=False, default=3)
+
+    is_active = Column(Boolean, nullable=False, default=True)
+    created_at = Column(DateTime, default=func.now())
+    updated_at = Column(DateTime, default=func.now(), onupdate=func.now())

api/app/routers/__init__.py.bak

@@ -0,0 +1,7 @@
+from .jokes import router as jokes_router
+from .categories import router as categories_router
+from .auth import router as auth_router
+from .admin import router as admin_router
+from .settings import router as settings_router
+
+__all__ = ["jokes_router", "categories_router", "auth_router", "admin_router", "settings_router"]

api/app/routers/admin.py

@@ -1,3 +1,5 @@
+import json
+
 from fastapi import APIRouter, Depends, HTTPException, Header
 from jose import JWTError, jwt
 from sqlalchemy import func
@@ -8,8 +10,12 @@ from app.database import get_db
 from app.models.joke import Joke
 from app.models.category import JokeCrowd, JokeType
 from app.models.user import AdminUser
+from app.models.link import Link
+from app.models.feedback import Feedback
 from app.schemas.category import JokeCrowdCreate, JokeCrowdResponse, JokeTypeCreate, JokeTypeResponse
 from app.schemas.joke import JokeCreate, JokeResponse, JokeUpdate, PaginatedJokeResponse
+from app.schemas.link import LinkCreate, LinkResponse
+from app.schemas.feedback import FeedbackResponse
 
 router = APIRouter(prefix="/admin", tags=["管理后台"])
 
@@ -39,21 +45,48 @@ def get_current_admin_user(
     return user
 
 
-def joke_to_response(joke: Joke) -> JokeResponse:
+def _parse_ids(raw) -> list[int]:
+    """解析 DB 中的 JSON 列表字符串为 Python list"""
+    if not raw:
+        return []
+    if isinstance(raw, list):
+        return raw
+    try:
+        return json.loads(raw)
+    except Exception:
+        return []
+
+
+def _get_type_names(db: Session, ids: list[int]) -> list[str]:
+    rows = db.query(JokeType).filter(JokeType.id.in_(ids)).all()
+    return [r.name for r in rows]
+
+
+def _get_crowd_names(db: Session, ids: list[int]) -> list[str]:
+    rows = db.query(JokeCrowd).filter(JokeCrowd.id.in_(ids)).all()
+    return [r.name for r in rows]
+
+
+def joke_to_response(joke: Joke, db: Session = None) -> JokeResponse:
     """Convert Joke model to JokeResponse schema."""
+    ids = _parse_ids(joke.type_ids)
+    crowd_ids = _parse_ids(joke.crowd_ids)
+    type_names = _get_type_names(db, ids) if db else []
+    crowd_names = _get_crowd_names(db, crowd_ids) if db else []
     return JokeResponse(
         id=joke.id,
         title=joke.title,
         content=joke.content,
-        type_id=joke.type_id,
-        crowd_id=joke.crowd_id,
+        polished_content=joke.polished_content,
+        type_ids=ids,
+        crowd_ids=crowd_ids,
         status=joke.status,
         view_count=joke.view_count,
         like_count=joke.like_count,
         created_at=joke.created_at,
         updated_at=joke.updated_at,
-        type_name=joke.type.name if joke.type else None,
-        crowd_name=joke.crowd.name if joke.crowd else None,
+        type_names=type_names,
+        crowd_names=crowd_names,
     )
 
 
@@ -67,6 +100,8 @@ def admin_list_jokes(
     current_user: AdminUser = Depends(get_current_admin_user),
 ):
     """获取所有笑话（支持状态筛选）"""
+    # 限制 page_size 防止 DoS
+    page_size = max(1, min(page_size, 100))
     query = db.query(Joke)
     if status:
         query = query.filter(Joke.status == status)
@@ -74,7 +109,7 @@ def admin_list_jokes(
     offset = (page - 1) * page_size
     jokes = query.order_by(Joke.created_at.desc()).offset(offset).limit(page_size).all()
     return PaginatedJokeResponse(
-        items=[joke_to_response(j) for j in jokes],
+        items=[joke_to_response(j, db) for j in jokes],
         total=total,
         page=page,
         page_size=page_size,
@@ -88,11 +123,29 @@ def admin_create_joke(
     current_user: AdminUser = Depends(get_current_admin_user),
 ):
     """创建笑话"""
-    db_joke = Joke(**joke.model_dump())
+    data = joke.model_dump()
+    if data.get("type_ids") is not None:
+        data["type_ids"] = json.dumps(data["type_ids"])
+    if data.get("crowd_ids") is not None:
+        data["crowd_ids"] = json.dumps(data["crowd_ids"])
+    db_joke = Joke(**data)
     db.add(db_joke)
     db.commit()
     db.refresh(db_joke)
-    return joke_to_response(db_joke)
+    return joke_to_response(db_joke, db)
+
+
+@router.get("/jokes/{joke_id}", response_model=JokeResponse)
+def admin_get_joke(
+    joke_id: int,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """获取单个笑话"""
+    joke = db.query(Joke).filter(Joke.id == joke_id).first()
+    if not joke:
+        raise HTTPException(status_code=404, detail="笑话不存在")
+    return joke_to_response(joke, db)
 
 
 @router.put("/jokes/{joke_id}", response_model=JokeResponse)
@@ -108,10 +161,12 @@ def admin_update_joke(
         raise HTTPException(status_code=404, detail="笑话不存在")
     update_data = joke.model_dump(exclude_unset=True)
     for key, value in update_data.items():
+        if key in ("type_ids", "crowd_ids") and value is not None:
+            value = json.dumps(value)
         setattr(db_joke, key, value)
     db.commit()
     db.refresh(db_joke)
-    return joke_to_response(db_joke)
+    return joke_to_response(db_joke, db)
 
 
 @router.delete("/jokes/{joke_id}")
@@ -152,12 +207,14 @@ def admin_stats(
     total_jokes = db.query(Joke).count()
     approved_jokes = db.query(Joke).filter(Joke.status == "approved").count()
     pending_jokes = db.query(Joke).filter(Joke.status == "pending").count()
+    rejected_jokes = db.query(Joke).filter(Joke.status == "rejected").count()
     total_views = db.query(Joke).with_entities(func.sum(Joke.view_count)).scalar() or 0
     total_likes = db.query(Joke).with_entities(func.sum(Joke.like_count)).scalar() or 0
     return {
         "total_jokes": total_jokes,
         "approved_jokes": approved_jokes,
         "pending_jokes": pending_jokes,
+        "rejected_jokes": rejected_jokes,
         "total_views": total_views,
         "total_likes": total_likes,
     }
@@ -209,8 +266,11 @@ def admin_delete_type(
     db_type = db.query(JokeType).filter(JokeType.id == type_id).first()
     if not db_type:
         raise HTTPException(status_code=404, detail="类型不存在")
-    # Check if there are jokes using this type
-    joke_count = db.query(Joke).filter(Joke.type_id == type_id).count()
+    # Check both old single-field and new array-field associations
+    old_count = db.query(Joke).filter(Joke.type_id == type_id).count()
+    all_jokes = db.query(Joke.type_ids).filter(Joke.type_ids.isnot(None)).all()
+    new_count = sum(1 for (raw,) in all_jokes if _parse_ids(raw) and type_id in _parse_ids(raw))
+    joke_count = old_count + new_count
     if joke_count > 0:
         raise HTTPException(status_code=400, detail=f"有 {joke_count} 条笑话使用此类型，无法删除")
     db.delete(db_type)
@@ -263,10 +323,101 @@ def admin_delete_crowd(
     db_crowd = db.query(JokeCrowd).filter(JokeCrowd.id == crowd_id).first()
     if not db_crowd:
         raise HTTPException(status_code=404, detail="人群分类不存在")
-    # Check if there are jokes using this crowd
-    joke_count = db.query(Joke).filter(Joke.crowd_id == crowd_id).count()
+    # Check both old single-field and new array-field associations
+    old_count = db.query(Joke).filter(Joke.crowd_id == crowd_id).count()
+    all_jokes = db.query(Joke.crowd_ids).filter(Joke.crowd_ids.isnot(None)).all()
+    new_count = sum(1 for (raw,) in all_jokes if _parse_ids(raw) and crowd_id in _parse_ids(raw))
+    joke_count = old_count + new_count
     if joke_count > 0:
         raise HTTPException(status_code=400, detail=f"有 {joke_count} 条笑话使用此人群，无法删除")
     db.delete(db_crowd)
     db.commit()
     return {"message": "删除成功"}
+
+
+# ============================================================
+# 友情链接管理
+# ============================================================
+
+@router.get("/links", response_model=list[LinkResponse])
+def admin_list_links(
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """获取所有友情链接"""
+    return db.query(Link).order_by(Link.sort_order, Link.id).all()
+
+
+@router.post("/links", response_model=LinkResponse)
+def admin_create_link(
+    link: LinkCreate,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """创建友情链接"""
+    db_link = Link(**link.model_dump())
+    db.add(db_link)
+    db.commit()
+    db.refresh(db_link)
+    return db_link
+
+
+@router.put("/links/{link_id}", response_model=LinkResponse)
+def admin_update_link(
+    link_id: int,
+    link: LinkCreate,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """更新友情链接"""
+    db_link = db.query(Link).filter(Link.id == link_id).first()
+    if not db_link:
+        raise HTTPException(status_code=404, detail="链接不存在")
+    for key, value in link.model_dump().items():
+        setattr(db_link, key, value)
+    db.commit()
+    db.refresh(db_link)
+    return db_link
+
+
+@router.delete("/links/{link_id}")
+def admin_delete_link(
+    link_id: int,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """删除友情链接"""
+    db_link = db.query(Link).filter(Link.id == link_id).first()
+    if not db_link:
+        raise HTTPException(status_code=404, detail="链接不存在")
+    db.delete(db_link)
+    db.commit()
+    return {"message": "删除成功"}
+
+
+# ============================================================
+# 反馈建议管理
+# ============================================================
+
+@router.get("/feedbacks", response_model=list[FeedbackResponse])
+def admin_list_feedbacks(
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """获取所有反馈建议"""
+    return db.query(Feedback).order_by(Feedback.created_at.desc()).all()
+
+
+@router.delete("/feedbacks/{feedback_id}")
+def admin_delete_feedback(
+    feedback_id: int,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """删除反馈"""
+    db_feedback = db.query(Feedback).filter(Feedback.id == feedback_id).first()
+    if not db_feedback:
+        raise HTTPException(status_code=404, detail="反馈不存在")
+    db.delete(db_feedback)
+    db.commit()
+    return {"message": "删除成功"}

api/app/routers/feedback.py

@@ -0,0 +1,18 @@
+from fastapi import APIRouter, Depends
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.feedback import Feedback
+from app.schemas.feedback import FeedbackCreate, FeedbackResponse
+
+router = APIRouter(tags=["反馈建议"])
+
+
+@router.post("/feedback", response_model=FeedbackResponse)
+def create_feedback(feedback: FeedbackCreate, db: Session = Depends(get_db)):
+    """公开：提交反馈建议"""
+    db_feedback = Feedback(**feedback.model_dump())
+    db.add(db_feedback)
+    db.commit()
+    db.refresh(db_feedback)
+    return db_feedback

api/app/routers/generate.py

@@ -76,8 +76,12 @@ def generate_joke(
         raise HTTPException(status_code=500, detail=f"生成失败: {str(e)}")
 
 
-def _parse_response(raw: str) -> GenerateResponse:
+def _parse_response(raw: str | None) -> GenerateResponse:
     """解析 AI 返回内容，提取标题和内容"""
+    # 防御：处理空或 None 输入
+    if not raw or not raw.strip():
+        raise ValueError("AI 返回内容为空")
+
     title = ""
     content = raw
 
@@ -86,16 +90,26 @@ def _parse_response(raw: str) -> GenerateResponse:
         line = line.strip()
         if line.startswith("标题：") or line.startswith("标题:"):
             title = line.split("：", 1)[-1].split(":", 1)[-1].strip()
-            content = content.replace(line, "").strip()
+            # 只替换这一行，不要 replace 全局
+            lines = content.split("\n")
+            for i, l in enumerate(lines):
+                if l.strip() == line:
+                    lines[i] = ""
+                    break
+            content = "\n".join(lines).strip()
             break
 
-    # 如果没有提取到标题，取第一行或前20字
+    # 如果没有提取到标题，取第一行
     if not title:
         first_line = raw.split("\n")[0].strip()
         if first_line.startswith("标题"):
             first_line = first_line.split("：", 1)[-1].split(":", 1)[-1].strip()
         title = first_line[:30] if len(first_line) > 30 else first_line
 
+    # 防御：content 不能为空
+    if not content.strip():
+        content = "（内容生成失败，请重新生成）"
+
     return GenerateResponse(
         title=title or "生成的笑话",
         content=content.strip(),

api/app/routers/jokes.py

@@ -1,3 +1,4 @@
+import json
 import random
 
 from fastapi import APIRouter, Depends, HTTPException, Query
@@ -6,26 +7,56 @@ from sqlalchemy.orm import Session
 
 from app.database import get_db
 from app.models.joke import Joke
+from app.models.category import JokeCrowd, JokeType
 from app.schemas.joke import JokeResponse, PaginatedJokeResponse
 
 router = APIRouter(prefix="/jokes", tags=["笑话"])
 
 
-def joke_to_response(joke: Joke) -> JokeResponse:
+def _parse_ids(raw) -> list[int]:
+    if not raw:
+        return []
+    if isinstance(raw, list):
+        return raw
+    try:
+        return json.loads(raw)
+    except Exception:
+        return []
+
+
+def _get_type_names(db: Session, ids: list[int]) -> list[str]:
+    if not ids:
+        return []
+    rows = db.query(JokeType).filter(JokeType.id.in_(ids)).all()
+    return [r.name for r in rows]
+
+
+def _get_crowd_names(db: Session, ids: list[int]) -> list[str]:
+    if not ids:
+        return []
+    rows = db.query(JokeCrowd).filter(JokeCrowd.id.in_(ids)).all()
+    return [r.name for r in rows]
+
+
+def joke_to_response(joke: Joke, db: Session = None) -> JokeResponse:
     """Convert Joke model to JokeResponse schema."""
+    ids = _parse_ids(joke.type_ids)
+    crowd_ids = _parse_ids(joke.crowd_ids)
+    type_names = _get_type_names(db, ids) if db else []
+    crowd_names = _get_crowd_names(db, crowd_ids) if db else []
     return JokeResponse(
         id=joke.id,
         title=joke.title,
         content=joke.content,
-        type_id=joke.type_id,
-        crowd_id=joke.crowd_id,
+        type_ids=ids,
+        crowd_ids=crowd_ids,
         status=joke.status,
         view_count=joke.view_count,
         like_count=joke.like_count,
         created_at=joke.created_at,
         updated_at=joke.updated_at,
-        type_name=joke.type.name if joke.type else None,
-        crowd_name=joke.crowd.name if joke.crowd else None,
+        type_names=type_names,
+        crowd_names=crowd_names,
     )
 
 
@@ -33,24 +64,46 @@ def joke_to_response(joke: Joke) -> JokeResponse:
 def list_jokes(
     page: int = 1,
     page_size: int = 20,
-    type_id: int | None = None,
-    crowd_id: int | None = None,
+    type_ids: str | None = Query(None, description="逗号分隔的类型 ID，如 1,3,5"),
+    crowd_ids: str | None = Query(None, description="逗号分隔的人群 ID，如 2,4"),
     db: Session = Depends(get_db),
 ):
     """获取笑话列表（仅返回已审核通过的笑话）"""
     query = db.query(Joke).filter(Joke.status == "approved")
 
-    if type_id is not None:
-        query = query.filter(Joke.type_id == type_id)
-    if crowd_id is not None:
-        query = query.filter(Joke.crowd_id == crowd_id)
+    # 支持多选过滤：逗号分隔的 ID
+    if type_ids:
+        filter_set = {int(x.strip()) for x in type_ids.split(",") if x.strip().isdigit()}
+        if filter_set:
+            # 兼容旧单值字段 type_id
+            old_filter = Joke.type_id.in_(filter_set)
+            # 兼容新数组字段 type_ids（JSON 包含任一）
+            new_filter = [
+                Joke.type_ids.contains(str(tid)) for tid in filter_set
+            ]
+            combined = old_filter
+            for nf in new_filter:
+                combined = combined | nf
+            query = query.filter(combined)
+
+    if crowd_ids:
+        filter_set = {int(x.strip()) for x in crowd_ids.split(",") if x.strip().isdigit()}
+        if filter_set:
+            old_filter = Joke.crowd_id.in_(filter_set)
+            new_filter = [
+                Joke.crowd_ids.contains(str(cid)) for cid in filter_set
+            ]
+            combined = old_filter
+            for nf in new_filter:
+                combined = combined | nf
+            query = query.filter(combined)
 
     total = query.count()
     offset = (page - 1) * page_size
     jokes = query.order_by(Joke.created_at.desc()).offset(offset).limit(page_size).all()
 
     return PaginatedJokeResponse(
-        items=[joke_to_response(j) for j in jokes],
+        items=[joke_to_response(j, db) for j in jokes],
         total=total,
         page=page,
         page_size=page_size,
@@ -60,19 +113,59 @@ def list_jokes(
 @router.get("/{joke_id}", response_model=JokeResponse)
 def get_joke(joke_id: int, db: Session = Depends(get_db)):
     """获取单条笑话详情"""
-    joke = db.query(Joke).filter(Joke.id == joke_id).first()
+    # 只返回已审核通过的笑话
+    joke = db.query(Joke).filter(
+        Joke.id == joke_id,
+        Joke.status == "approved"
+    ).first()
     if not joke:
         raise HTTPException(status_code=404, detail="笑话不存在")
-    # 增加浏览次数
-    joke.view_count += 1
+    # 使用原子更新避免并发竞态
+    db.query(Joke).filter(Joke.id == joke_id).update({Joke.view_count: Joke.view_count + 1})
     db.commit()
-    return joke_to_response(joke)
+    # 重新查询获取更新后的数据
+    db.refresh(joke)
+    return joke_to_response(joke, db)
 
 
 @router.get("/random", response_model=JokeResponse)
 def get_random_joke(db: Session = Depends(get_db)):
     """随机获取一条已审核通过的笑话"""
+    # 使用效率更高的方式：随机 ID 取模
+    max_id = db.query(func.max(Joke.id)).filter(Joke.status == "approved").scalar()
+    if not max_id:
+        raise HTTPException(status_code=404, detail="暂无笑话")
+
+    # 尝试最多 10 次找到有效笑话
+    for _ in range(10):
+        random_id = random.randint(1, max_id)
+        joke = db.query(Joke).filter(
+            Joke.id >= random_id,
+            Joke.status == "approved"
+        ).first()
+        if joke:
+            return joke_to_response(joke, db)
+
+    # 兜底：全表随机
     joke = db.query(Joke).filter(Joke.status == "approved").order_by(func.random()).first()
     if not joke:
         raise HTTPException(status_code=404, detail="暂无笑话")
-    return joke_to_response(joke)
+    return joke_to_response(joke, db)
+
+
+@router.post("/{joke_id}/like")
+def like_joke(joke_id: int, db: Session = Depends(get_db)):
+    """为笑话点赞（仅允许已审核通过的笑话）"""
+    # 先检查笑话是否存在且已审核
+    joke = db.query(Joke).filter(
+        Joke.id == joke_id,
+        Joke.status == "approved"
+    ).first()
+    if not joke:
+        raise HTTPException(status_code=404, detail="笑话不存在")
+    # 使用原子更新避免并发竞态
+    db.query(Joke).filter(Joke.id == joke_id).update({Joke.like_count: Joke.like_count + 1})
+    db.commit()
+    # 获取更新后的值
+    db.refresh(joke)
+    return {"message": "点赞成功", "like_count": joke.like_count}

api/app/routers/links.py

@@ -0,0 +1,14 @@
+from fastapi import APIRouter, Depends
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.link import Link
+from app.schemas.link import LinkCreate, LinkResponse
+
+router = APIRouter(tags=["友情链接"])
+
+
+@router.get("/links", response_model=list[LinkResponse])
+def list_links(db: Session = Depends(get_db)):
+    """公开：获取所有友情链接"""
+    return db.query(Link).order_by(Link.sort_order, Link.id).all()

api/app/routers/settings.py

@@ -0,0 +1,96 @@
+from fastapi import APIRouter, Depends, HTTPException
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.setting import AiSetting
+from app.schemas.setting import AiSettingCreate
+from app.routers.admin import get_current_admin_user
+from app.models.user import AdminUser
+
+router = APIRouter(prefix="/admin/settings", tags=["AI设置"])
+
+
+@router.get("", response_model=list)
+def list_settings(
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """列出所有 AI 配置"""
+    return db.query(AiSetting).order_by(AiSetting.created_at.desc()).all()
+
+
+@router.get("/active")
+def get_active_setting(
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """获取当前激活的 AI 配置（爬虫调用，无需用户认证，token 校验仍保留）"""
+    setting = db.query(AiSetting).filter(AiSetting.is_active == True).first()
+    if not setting:
+        raise HTTPException(status_code=404, detail="未找到激活的 AI 配置")
+    return setting
+
+
+@router.post("")
+def create_setting(
+    setting: AiSettingCreate,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """新建 AI 配置"""
+    db_setting = AiSetting(**setting.model_dump())
+    db.add(db_setting)
+    db.commit()
+    db.refresh(db_setting)
+    return db_setting
+
+
+@router.put("/{setting_id}")
+def update_setting(
+    setting_id: int,
+    setting: AiSettingCreate,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """更新 AI 配置"""
+    db_setting = db.query(AiSetting).filter(AiSetting.id == setting_id).first()
+    if not db_setting:
+        raise HTTPException(status_code=404, detail="配置不存在")
+    for key, value in setting.model_dump().items():
+        setattr(db_setting, key, value)
+    db.commit()
+    db.refresh(db_setting)
+    return db_setting
+
+
+@router.put("/{setting_id}/toggle")
+def toggle_setting(
+    setting_id: int,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """切换激活状态（只能有一个活跃）"""
+    db_setting = db.query(AiSetting).filter(AiSetting.id == setting_id).first()
+    if not db_setting:
+        raise HTTPException(status_code=404, detail="配置不存在")
+
+    # 先全部设为非活跃
+    db.query(AiSetting).update({AiSetting.is_active: False})
+    db_setting.is_active = True
+    db.commit()
+    return {"message": "已激活", "id": setting_id}
+
+
+@router.delete("/{setting_id}")
+def delete_setting(
+    setting_id: int,
+    db: Session = Depends(get_db),
+    current_user: AdminUser = Depends(get_current_admin_user),
+):
+    """删除 AI 配置"""
+    db_setting = db.query(AiSetting).filter(AiSetting.id == setting_id).first()
+    if not db_setting:
+        raise HTTPException(status_code=404, detail="配置不存在")
+    db.delete(db_setting)
+    db.commit()
+    return {"message": "删除成功"}

api/app/schemas/feedback.py

@@ -0,0 +1,17 @@
+from datetime import datetime
+
+from pydantic import BaseModel
+
+
+class FeedbackCreate(BaseModel):
+    name: str | None = None
+    email: str | None = None
+    content: str
+
+
+class FeedbackResponse(FeedbackCreate):
+    id: int
+    created_at: datetime
+
+    class Config:
+        from_attributes = True

api/app/schemas/link.py

@@ -0,0 +1,18 @@
+from datetime import datetime
+
+from pydantic import BaseModel
+
+
+class LinkCreate(BaseModel):
+    name: str
+    url: str
+    description: str | None = None
+    sort_order: int = 0
+
+
+class LinkResponse(LinkCreate):
+    id: int
+    created_at: datetime
+
+    class Config:
+        from_attributes = True

api/app/schemas/setting.py

@@ -0,0 +1,17 @@
+from pydantic import BaseModel, ConfigDict, Field
+
+
+class AiSettingBase(BaseModel):
+    provider: str = Field(default="nvidia")
+    api_base: str = Field(default="https://integrate.api.nvidia.com/v1")
+    api_key: str = Field(default="")
+    model_name: str = Field(default="nvidia/llama-3.1-nemotron-70b-instruct")
+    temperature: float = Field(default=0.7, ge=0, le=2)
+    max_tokens: int = Field(default=2048, ge=1)
+    crawl_enabled: bool = Field(default=False)
+    crawl_keywords: str = Field(default="")
+    max_pages_per_run: int = Field(default=3, ge=1)
+
+
+class AiSettingCreate(AiSettingBase):
+    pass

api/main.py.bak

@@ -0,0 +1,39 @@
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+
+from app.config import API_TITLE, API_VERSION
+from app.routers import jokes_router, categories_router, auth_router, admin_router, settings_router
+from app.database import Base, engine
+from app.models.setting import AiSetting  # 注册模型，确保 create_all 能看到
+
+# 创建新表（如果不存在）
+Base.metadata.create_all(bind=engine)
+
+# Create FastAPI application
+app = FastAPI(title=API_TITLE, version=API_VERSION)
+
+# Configure CORS
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Mount routers
+app.include_router(jokes_router, prefix="/api")
+app.include_router(categories_router, prefix="/api")
+app.include_router(auth_router, prefix="/api")
+app.include_router(admin_router, prefix="/api")
+app.include_router(settings_router, prefix="/api")
+
+
+@app.get("/")
+def root():
+    return {"message": "欢迎使用笑话大全 API"}
+
+
+@app.get("/health")
+def health_check():
+    return {"status": "healthy"}

api/requirements.txt

@@ -5,3 +5,6 @@ pydantic==2.5.3
 python-jose[cryptography]==3.3.0
 passlib[bcrypt]==1.7.4
 python-multipart==0.0.6
+openai==1.12.0
+crawl4ai==0.3.0
+httpx==0.27.0

api/server.log

@@ -0,0 +1,501 @@
+INFO:     Will watch for changes in these directories: ['D:\\bwstudio\\joke\\api']
+INFO:     Uvicorn running on http://127.0.0.1:8001 (Press CTRL+C to quit)
+INFO:     Started reloader process [2236] using WatchFiles
+INFO:     Started server process [1824]
+INFO:     Waiting for application startup.
+INFO:     Application startup complete.
+INFO:     127.0.0.1:61949 - "GET /health HTTP/1.1" 200 OK
+INFO:     127.0.0.1:61961 - "POST /api/auth/login HTTP/1.1" 200 OK
+INFO:     127.0.0.1:61964 - "GET /api/admin/jokes?page=1&page_size=3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62046 - "GET / HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62046 - "GET /favicon.ico HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:62086 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62088 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62090 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62151 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62154 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62158 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62161 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62164 - "GET /api/admin/jokes/1278 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62167 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62173 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62176 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62179 - "GET /api/admin/jokes/1275 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62182 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62198 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62201 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62204 - "GET /api/admin/jokes/1273 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62208 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62213 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62216 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62219 - "GET /api/admin/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62223 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62230 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62233 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62236 - "GET /api/admin/jokes/1274 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62239 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62242 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62245 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62261 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62264 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62267 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62270 - "GET /api/admin/jokes/1262 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62273 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62283 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62286 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62289 - "GET /api/admin/jokes/1265 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62292 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62295 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62298 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62302 - "GET /api/admin/settings/active HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62304 - "GET /api/admin/settings HTTP/1.1" 500 Internal Server Error
+ERROR:    Exception in ASGI application
+Traceback (most recent call last):
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\protocols\http\httptools_impl.py", line 419, in run_asgi
+    result = await app(  # type: ignore[func-returns-value]
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\middleware\proxy_headers.py", line 84, in __call__
+    return await self.app(scope, receive, send)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\applications.py", line 1054, in __call__
+    await super().__call__(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\applications.py", line 123, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 186, in __call__
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 164, in __call__
+    await self.app(scope, receive, _send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\cors.py", line 83, in __call__
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\exceptions.py", line 62, in __call__
+    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 762, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 782, in app
+    await route.handle(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 297, in handle
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 77, in app
+    await wrap_app_handling_exceptions(app, request)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 72, in app
+    response = await func(request)
+               ^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 315, in app
+    content = await serialize_response(
+              ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 160, in serialize_response
+    return field.serialize(
+           ^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\_compat.py", line 147, in serialize
+    return self._type_adapter.dump_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\pydantic\type_adapter.py", line 607, in dump_python
+    return self.serializer.to_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+pydantic_core._pydantic_core.PydanticSerializationError: Unable to serialize unknown type: <class 'app.models.setting.AiSetting'>
+INFO:     127.0.0.1:62308 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62324 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62327 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62330 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62356 - "GET /api/admin/settings/active HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62359 - "GET /api/admin/settings HTTP/1.1" 500 Internal Server Error
+ERROR:    Exception in ASGI application
+Traceback (most recent call last):
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\protocols\http\httptools_impl.py", line 419, in run_asgi
+    result = await app(  # type: ignore[func-returns-value]
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\middleware\proxy_headers.py", line 84, in __call__
+    return await self.app(scope, receive, send)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\applications.py", line 1054, in __call__
+    await super().__call__(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\applications.py", line 123, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 186, in __call__
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 164, in __call__
+    await self.app(scope, receive, _send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\cors.py", line 83, in __call__
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\exceptions.py", line 62, in __call__
+    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 762, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 782, in app
+    await route.handle(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 297, in handle
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 77, in app
+    await wrap_app_handling_exceptions(app, request)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 72, in app
+    response = await func(request)
+               ^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 315, in app
+    content = await serialize_response(
+              ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 160, in serialize_response
+    return field.serialize(
+           ^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\_compat.py", line 147, in serialize
+    return self._type_adapter.dump_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\pydantic\type_adapter.py", line 607, in dump_python
+    return self.serializer.to_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+pydantic_core._pydantic_core.PydanticSerializationError: Unable to serialize unknown type: <class 'app.models.setting.AiSetting'>
+INFO:     127.0.0.1:62362 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62365 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62368 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62371 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62383 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62386 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62389 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62392 - "GET /api/admin/settings/active HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62395 - "GET /api/admin/settings HTTP/1.1" 500 Internal Server Error
+ERROR:    Exception in ASGI application
+Traceback (most recent call last):
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\protocols\http\httptools_impl.py", line 419, in run_asgi
+    result = await app(  # type: ignore[func-returns-value]
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\middleware\proxy_headers.py", line 84, in __call__
+    return await self.app(scope, receive, send)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\applications.py", line 1054, in __call__
+    await super().__call__(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\applications.py", line 123, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 186, in __call__
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 164, in __call__
+    await self.app(scope, receive, _send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\cors.py", line 83, in __call__
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\exceptions.py", line 62, in __call__
+    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 762, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 782, in app
+    await route.handle(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 297, in handle
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 77, in app
+    await wrap_app_handling_exceptions(app, request)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 72, in app
+    response = await func(request)
+               ^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 315, in app
+    content = await serialize_response(
+              ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 160, in serialize_response
+    return field.serialize(
+           ^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\_compat.py", line 147, in serialize
+    return self._type_adapter.dump_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\pydantic\type_adapter.py", line 607, in dump_python
+    return self.serializer.to_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+pydantic_core._pydantic_core.PydanticSerializationError: Unable to serialize unknown type: <class 'app.models.setting.AiSetting'>
+INFO:     127.0.0.1:62398 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62401 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62404 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62407 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62414 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62419 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62422 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62425 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62429 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62432 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62435 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62438 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62441 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62444 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62447 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62451 - "GET /api/admin/settings/active HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62453 - "GET /api/admin/settings HTTP/1.1" 500 Internal Server Error
+ERROR:    Exception in ASGI application
+Traceback (most recent call last):
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\protocols\http\httptools_impl.py", line 419, in run_asgi
+    result = await app(  # type: ignore[func-returns-value]
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\middleware\proxy_headers.py", line 84, in __call__
+    return await self.app(scope, receive, send)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\applications.py", line 1054, in __call__
+    await super().__call__(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\applications.py", line 123, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 186, in __call__
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 164, in __call__
+    await self.app(scope, receive, _send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\cors.py", line 83, in __call__
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\exceptions.py", line 62, in __call__
+    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 762, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 782, in app
+    await route.handle(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 297, in handle
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 77, in app
+    await wrap_app_handling_exceptions(app, request)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 72, in app
+    response = await func(request)
+               ^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 315, in app
+    content = await serialize_response(
+              ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 160, in serialize_response
+    return field.serialize(
+           ^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\_compat.py", line 147, in serialize
+    return self._type_adapter.dump_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\pydantic\type_adapter.py", line 607, in dump_python
+    return self.serializer.to_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+pydantic_core._pydantic_core.PydanticSerializationError: Unable to serialize unknown type: <class 'app.models.setting.AiSetting'>
+INFO:     127.0.0.1:62456 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62459 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62462 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62465 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62468 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62471 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62474 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62477 - "GET /api/admin/settings/active HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62480 - "GET /api/admin/settings HTTP/1.1" 500 Internal Server Error
+ERROR:    Exception in ASGI application
+Traceback (most recent call last):
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\protocols\http\httptools_impl.py", line 419, in run_asgi
+    result = await app(  # type: ignore[func-returns-value]
+             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\uvicorn\middleware\proxy_headers.py", line 84, in __call__
+    return await self.app(scope, receive, send)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\applications.py", line 1054, in __call__
+    await super().__call__(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\applications.py", line 123, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 186, in __call__
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\errors.py", line 164, in __call__
+    await self.app(scope, receive, _send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\cors.py", line 83, in __call__
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\middleware\exceptions.py", line 62, in __call__
+    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 762, in __call__
+    await self.middleware_stack(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 782, in app
+    await route.handle(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 297, in handle
+    await self.app(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 77, in app
+    await wrap_app_handling_exceptions(app, request)(scope, receive, send)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 64, in wrapped_app
+    raise exc
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\_exception_handler.py", line 53, in wrapped_app
+    await app(scope, receive, sender)
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\starlette\routing.py", line 72, in app
+    response = await func(request)
+               ^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 315, in app
+    content = await serialize_response(
+              ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\routing.py", line 160, in serialize_response
+    return field.serialize(
+           ^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\fastapi\_compat.py", line 147, in serialize
+    return self._type_adapter.dump_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\PF\AppData\Roaming\Python\Python311\site-packages\pydantic\type_adapter.py", line 607, in dump_python
+    return self.serializer.to_python(
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+pydantic_core._pydantic_core.PydanticSerializationError: Unable to serialize unknown type: <class 'app.models.setting.AiSetting'>
+INFO:     127.0.0.1:62483 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62486 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62489 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62492 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62883 - "GET /api/admin/stats HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62885 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62887 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62889 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62892 - "GET /api/admin/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62895 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62898 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62901 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62904 - "GET /api/admin/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62907 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62992 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62994 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62997 - "GET /api/jokes/?page=1&page_size=30 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:62999 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63002 - "GET /api/jokes/1275 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63005 - "GET /api/jokes/?page=1&page_size=50 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63008 - "GET /api/jokes/1246 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63011 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63014 - "GET /api/jokes/?page=1&page_size=20&type_ids=1 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63017 - "GET /api/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63021 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63034 - "GET /api/jokes/?page=1&page_size=20&type_ids=12 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63037 - "GET /api/jokes/?page=1&page_size=20&type_ids=12,1 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63040 - "GET /api/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63043 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63046 - "GET /api/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63049 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63052 - "GET /api/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63055 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63059 - "GET /api/jokes/1277 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63062 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+WARNING:  WatchFiles detected changes in 'app\models\link.py'. Reloading...
+INFO:     127.0.0.1:63122 - "GET /api/jokes/?page=2&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63125 - "GET /api/jokes/?page=3&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63128 - "GET /api/jokes/?page=4&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63131 - "GET /api/jokes/?page=5&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63134 - "GET /api/jokes/?page=7&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63138 - "GET /api/jokes/1140 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63141 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63144 - "GET /api/jokes/?page=1&page_size=20&type_ids=4 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63147 - "GET /api/jokes/923 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63150 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63161 - "GET /api/jokes/?page=1&page_size=20&type_ids=3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63164 - "GET /api/jokes/?page=1&page_size=20&type_ids=3,4 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63167 - "GET /api/jokes/?page=1&page_size=20&type_ids=3,4,5 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63170 - "GET /api/jokes/?page=1&page_size=20&type_ids=3,4,5,6 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63173 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63176 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63179 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63182 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63185 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1,8 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63188 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1,8,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63191 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1,8,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63194 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1,8,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63197 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1,8,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63200 - "GET /api/jokes/?page=1&page_size=20&type_ids=4,5,6,3,2,1,8,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63204 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63208 - "GET /api/jokes/1279 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63211 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63214 - "GET /api/jokes/?page=1&page_size=20&type_ids=3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63217 - "GET /api/jokes/?page=1&page_size=20&type_ids=3,4 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63220 - "GET /api/jokes/?page=1&page_size=20&type_ids=3,4,5 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63223 - "GET /api/jokes/?page=1&page_size=20&type_ids=3,4,5,6 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63226 - "GET /api/jokes/1251 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63229 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63232 - "GET /api/jokes/?page=1&page_size=20&type_ids=2 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63235 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63238 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63241 - "GET /api/jokes/?page=1&page_size=20&crowd_ids=15 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63244 - "GET /api/jokes/?page=1&page_size=20&crowd_ids=15,16 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63264 - "GET /api/jokes/?page=1&page_size=20&type_ids=2 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63267 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63270 - "GET /api/jokes/1273 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63273 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63276 - "GET /api/jokes/1277 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63279 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63295 - "GET /api/jokes/?page=2&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63298 - "GET /api/jokes/?page=3&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63301 - "GET /api/jokes/?page=1&page_size=20&type_ids=1 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63304 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63307 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63310 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63313 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4,5 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63316 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4,5,6 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63319 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4,5,6,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63322 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4,5,6,7,8 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63325 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4,5,6,7,8,9 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63328 - "GET /api/jokes/?page=1&page_size=20&type_ids=1,2,3,4,5,6,7,8,9,10 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63331 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4,5,6,7,8,9,10 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63334 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4,5,6,7,8,9 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63337 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4,5,6,7,8 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63340 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4,5,6,7 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63343 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4,5,6 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63346 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4,5 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63349 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3,4 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63352 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63369 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63371 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63373 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63376 - "GET /api/admin/jokes?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63379 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63390 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63392 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63394 - "GET /api/jokes/?page=1&page_size=30 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63396 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63427 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63446 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63448 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63452 - "GET /api/jokes/?page=1&page_size=20&type_ids=2,3 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63450 - "GET /api/jokes/?page=1&page_size=30 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63454 - "GET /api/jokes/1263 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63456 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63466 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63468 - "GET /api/admin/feedbacks HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63471 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63473 - "GET /api/admin/feedbacks HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63476 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63479 - "GET /api/admin/feedbacks HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63482 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63486 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63495 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63498 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63503 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63501 - "GET /api/jokes/?page=1&page_size=30 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63508 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63511 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63514 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63517 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63520 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63524 - "GET /api/jokes/?page=1&page_size=30 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63565 - "GET /api/admin/feedbacks HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63568 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63571 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63574 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63577 - "GET /api/admin/feedbacks HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63580 - "GET /api/admin/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63583 - "GET /api/admin/feedbacks HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63592 - "GET /api/categories/types HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63594 - "GET /api/categories/crowds HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63598 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63596 - "GET /api/jokes/?page=1&page_size=30 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63607 - "GET /api/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63609 - "GET /api/links HTTP/1.1" 404 Not Found
+INFO:     127.0.0.1:63612 - "GET /api/jokes/?page=1&page_size=20 HTTP/1.1" 200 OK
+INFO:     127.0.0.1:63615 - "GET /api/links HTTP/1.1" 404 Not Found

crawler/__init__.py

@@ -0,0 +1 @@
+"""AI 提示词模板"""

crawler/ai_service.py

@@ -0,0 +1,71 @@
+"""LLM 处理：调用 NVIDIA NIM（OpenAI 兼容 API）进行笑话提取、改写和分类。"""
+
+import json
+import os
+from openai import OpenAI
+
+
+class AiService:
+    def __init__(self, api_base: str, api_key: str, model_name: str, temperature: float = 0.7, max_tokens: int = 2048):
+        self.client = OpenAI(base_url=api_base, api_key=api_key)
+        self.model = model_name
+        self.temperature = temperature
+        self.max_tokens = max_tokens
+
+    def extract_jokes(self, page_content: str, known_types: list[str], known_crowds: list[str]) -> list[dict]:
+        """从页面内容中提取笑话，返回结构化数据。"""
+        from crawler.prompts import EXTRACTION_SYSTEM_PROMPT, EXTRACTION_USER_PROMPT
+
+        user_prompt = EXTRACTION_USER_PROMPT.format(
+            page_content=page_content[:8000],
+            known_types=", ".join(known_types),
+            known_crowds=", ".join(known_crowds),
+        )
+
+        response = self.client.chat.completions.create(
+            model=self.model,
+            messages=[
+                {"role": "system", "content": EXTRACTION_SYSTEM_PROMPT},
+                {"role": "user", "content": user_prompt},
+            ],
+            temperature=self.temperature,
+            max_tokens=self.max_tokens,
+        )
+
+        raw = response.choices[0].message.content
+        return self._parse_json(raw)
+
+    def rewrite_joke(self, content: str) -> str:
+        """润色单条笑话内容。"""
+        from crawler.prompts import REWRITE_SYSTEM_PROMPT, REWRITE_USER_PROMPT
+
+        response = self.client.chat.completions.create(
+            model=self.model,
+            messages=[
+                {"role": "system", "content": REWRITE_SYSTEM_PROMPT},
+                {"role": "user", "content": REWRITE_USER_PROMPT.format(content=content)},
+            ],
+            temperature=0.8,
+            max_tokens=500,
+        )
+        return response.choices[0].message.content.strip()
+
+    def _parse_json(self, raw: str) -> list[dict]:
+        """安全解析 LLM 返回的 JSON。"""
+        try:
+            data = json.loads(raw)
+            if isinstance(data, dict):
+                jokes = data.get("jokes") or data.get("items") or [data]
+            else:
+                jokes = data
+            return [j for j in jokes if isinstance(j, dict) and j.get("content")]
+        except json.JSONDecodeError:
+            # 尝试提取 markdown 代码块
+            if "```json" in raw:
+                raw = raw.split("```json")[1].split("```")[0]
+            elif "```" in raw:
+                raw = raw.split("```")[1].split("```")[0]
+            try:
+                return json.loads(raw.strip())
+            except Exception:
+                return []

crawler/crawler_service.py

@@ -0,0 +1,439 @@
+"""统一网页获取：搜索笑话站点 + 深度翻页抓取。"""
+
+import asyncio
+import os
+import re
+import sys
+from urllib.parse import quote, urljoin, urlparse
+
+# Windows 下禁用 rich 控制台输出，避免 GBK 编码错误
+if sys.platform == "win32":
+    os.environ["PYTHONIOENCODING"] = "utf-8"
+    os.environ["TERM"] = "dumb"
+
+try:
+    from crawl4ai import AsyncWebCrawler
+    from crawl4ai.async_configs import CrawlerRunConfig, BrowserConfig
+    HAS_CRAWL4AI = True
+except ImportError:
+    HAS_CRAWL4AI = False
+    print("[!] crawl4ai 未安装，将使用 requests 替代（不支持 JS 渲染）")
+
+import httpx
+
+
+class CrawlerService:
+    def __init__(self, headless: bool = True):
+        self.headless = headless
+        self.session = None
+        self._crawler = None
+        # 站点级别容错跟踪
+        self.site_failures: dict[str, int] = {}
+        # 最近一次抓取的原始 HTML（供翻页链接发现使用）
+        self._last_raw_html: str | None = None
+
+    def _get_session(self) -> httpx.Client:
+        if self.session is None:
+            self.session = httpx.Client(
+                headers={
+                    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/125.0.0.0 Safari/537.36",
+                    "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
+                    "Accept-Language": "zh-CN,zh;q=0.9,en;q=0.8",
+                    "Referer": "https://www.baidu.com/",
+                },
+                timeout=30,
+                follow_redirects=True,
+            )
+        return self.session
+
+    async def _get_crawler(self):
+        """获取或创建 crawl4ai 实例"""
+        if HAS_CRAWL4AI and self._crawler is None:
+            browser_cfg = BrowserConfig(headless=self.headless, verbose=False)
+            self._crawler = AsyncWebCrawler(config=browser_cfg)
+            await self._crawler.__aenter__()
+            if not self.headless:
+                print("  [*] 浏览器窗口已打开（crawl4ai 控制）")
+        return self._crawler
+
+    async def close(self):
+        if self._crawler:
+            await self._crawler.__aexit__(None, None, None)
+            self._crawler = None
+        if self.session:
+            self.session.close()
+            self.session = None
+
+    # ============================================================
+    # 页面抓取（带重试）
+    # ============================================================
+
+    async def crawl_page_with_retry(self, url: str, max_retries: int = 2, timeout: int = 30000) -> tuple[str, bool]:
+        """抓取页面，返回 (html内容, 是否成功)，失败自动重试"""
+        for attempt in range(max_retries + 1):
+            print(f"  [~] 正在获取页面... (尝试 {attempt+1}/{max_retries+1})")
+            try:
+                html = await self._crawl_one(url, timeout)
+                if html:
+                    print(f"  [~] 页面获取成功，内容长度: {len(html)} 字符")
+                    return html, True
+                print(f"  [!] 页面内容为空")
+            except Exception as e:
+                print(f"  [!] 抓取失败 (尝试 {attempt+1}/{max_retries+1}): {url[:60]} - {e}")
+
+            if attempt < max_retries:
+                wait = 3 * (attempt + 1)
+                print(f"  [~] 等待 {wait} 秒后重试...")
+                await asyncio.sleep(wait)
+        return "", False
+
+    async def _crawl_one(self, url: str, timeout: int) -> str:
+        """单次页面抓取，返回纯文本内容（同时保存原始 HTML 供翻页发现）"""
+        if HAS_CRAWL4AI:
+            # crawl4ai 统一处理 headless / visible 两种模式
+            crawler = await self._get_crawler()
+            result = await crawler.arun(url, config=CrawlerRunConfig(verbose=False))
+            if result and result.success:
+                # fit_html 是 AI 清洗后的正文（翻页链接常被清洗掉）
+                # cleaned_html 保留完整结构（用于翻页发现）
+                self._last_raw_html = result.cleaned_html or result.html or ""
+                html = result.fit_html or result.cleaned_html or ""
+                return self._html_to_text(html)
+            return ""
+        else:
+            # 无 crawl4ai 时 fallback 到 requests
+            resp = self._get_session().get(url, timeout=timeout / 1000)
+            if resp.status_code == 200:
+                self._last_raw_html = resp.text
+                return self._html_to_text(resp.text)
+            return ""
+
+    @staticmethod
+    def _html_to_text(html: str) -> str:
+        """简易 HTML 转纯文本"""
+        # 移除 script/style 标签内容
+        text = re.sub(r'<script[^>]*>.*?</script>', '', html, flags=re.DOTALL | re.IGNORECASE)
+        text = re.sub(r'<style[^>]*>.*?</style>', '', text, flags=re.DOTALL | re.IGNORECASE)
+        # 移除 HTML 标签
+        text = re.sub(r'<[^>]+>', ' ', text)
+        # 合并空白
+        text = re.sub(r'\s+', ' ', text).strip()
+        return text
+
+    # ============================================================
+    # 搜索笑话站点
+    # ============================================================
+
+    # 已知的笑话聚合站域名黑/白名单
+    JOKE_SITE_KEYWORDS = [
+        "笑话大全", "冷笑话", "搞笑段子", "笑话集锦",
+        "幽默笑话", "爆笑笑话", "成人笑话", "小笑话",
+    ]
+
+    async def search_joke_sites(self, keywords: list[str], max_results: int = 10) -> list[dict]:
+        """搜索笑话站点，返回 [{url, title, domain}]"""
+        all_results = []
+
+        for keyword in keywords:
+            results = await self._search_and_filter(keyword, max_results=max_results)
+            all_results.extend(results)
+
+        # 去重（按域名）
+        seen_domains = set()
+        unique = []
+        for r in all_results:
+            domain = r["domain"]
+            if domain not in seen_domains:
+                seen_domains.add(domain)
+                unique.append(r)
+
+        print(f"  [*] 搜索到 {len(unique)} 个唯一站点")
+        for s in unique:
+            print(f"      - {s['domain']}: {s['title'][:40]}")
+        return unique[:max_results]
+
+    async def _search_and_filter(self, keyword: str, max_results: int = 10) -> list[dict]:
+        """搜索并过滤出疑似笑话聚合站的结果"""
+        # 用多个搜索词提高覆盖率
+        search_queries = [
+            f"{keyword} 网站",
+            f"{keyword} 大全",
+            f"{keyword} 列表",
+        ]
+
+        seen = set()
+        sites = []
+
+        for q in search_queries:
+            if len(sites) >= max_results:
+                break
+
+            search_results = await self._search_bing(q, max_pages=2)
+
+            for r in search_results:
+                if len(sites) >= max_results:
+                    break
+
+                url = r["url"]
+                domain = urlparse(url).netloc.lower()
+
+                if domain in seen:
+                    continue
+                seen.add(domain)
+
+                # 过滤：排除已知的单篇文章站点和搜索引擎
+                if self._is_joke_collection_site(url, r.get("title", "")):
+                    r["domain"] = domain
+                    sites.append(r)
+
+        return sites
+
+    def _is_joke_collection_site(self, url: str, title: str) -> bool:
+        """判断URL是否疑似笑话聚合站（不是单篇文章）"""
+        domain = urlparse(url).netloc.lower()
+        path = urlparse(url).path.lower()
+
+        # 排除项
+        exclude_domains = [
+            "bing.com", "microsoft.com", "baidu.com", "google.com",
+            "sohu.com", "sina.com", "163.com", "qq.com", "toutiao.com",
+            "weibo.com", "zhihu.com", "bilibili.com", "douban.com",
+        ]
+        if any(d in domain for d in exclude_domains):
+            return False
+
+        # 排除明显的单篇文章模式
+        single_article_patterns = [
+            r'/p/\d+', r'/article/\d+', r'/post/\d+', r'/archives/\d+',
+            r'/a/\d+', r'/\d{5,}', r'/detail/\d+', r'/read/\d+',
+            r'\.html$',  # 静态 html 文章页
+        ]
+        # 但如果域名本身含 joke 特征，不排除
+        is_joke_domain = any(kw in domain or kw in title for kw in
+                             ["joke", "xiaohua", "笑话", "段子", "幽默", "搞笑"])
+
+        for p in single_article_patterns:
+            if re.search(p, path) and not is_joke_domain:
+                return False
+
+        # 聚合站特征：域名或标题含特定词，或URL有分类/列表模式
+        collection_patterns = [
+            "joke", "xiaohua", "笑话", "段子", "幽默", "搞笑",
+            "/page/", "/list/", "/category/", "/tag/", "joke",
+        ]
+        for p in collection_patterns:
+            if p in domain or p in path or p.lower() in title:
+                return True
+
+        # 有列表/目录模式的也认为是聚合站
+        if re.search(r'(page|list|category|tag|index)', path):
+            return True
+
+        return False
+
+    # ============================================================
+    # 翻页链接发现
+    # ============================================================
+
+    def discover_page_links(self, html: str, base_url: str) -> list[str]:
+        """从页面 HTML 中发现翻页链接，返回去重排序后的 URL 列表"""
+        # 优先使用 _last_raw_html（完整 HTML 而非纯文本），兜底用传入的 html
+        raw = self._last_raw_html or html
+        base_parsed = urlparse(base_url)
+        base_domain = f"{base_parsed.scheme}://{base_parsed.netloc}"
+
+        links = set()
+
+        # 1. <link rel="next">
+        for m in re.finditer(r'<link[^>]*rel="next"[^>]*href="([^"]+)"', raw, re.IGNORECASE):
+            links.add(urljoin(base_domain, m.group(1)))
+
+        # 2. 翻页文字链接（下一页、下页、>、» 等）
+        page_text_patterns = [
+            r'<a[^>]*href="([^"]*page[^"]*)"[^>]*>\s*(?:下一页|下页|下一页»|»|›|>|Next|last)\s*</a>',
+            r'<a[^>]*>\s*(?:下一页|下页|»|›|>)\s*</a>\s*<a[^>]*href="([^"]*)"',
+        ]
+        for pattern in page_text_patterns:
+            for m in re.finditer(pattern, raw, re.IGNORECASE):
+                href = m.group(1).strip()
+                if href and href not in ("#", "javascript:void(0)"):
+                    links.add(urljoin(base_url, href))
+
+        # 3. 提取所有带数字的翻页 link（?page=N, /page/N/, index_N.html, page_N.html, &page=N）
+        page_link_patterns = [
+            r'href="([^"]*[?&]page=(\d+)[^"]*)"',      # ?page=2  &page=2
+            r'href="([^"]*/page/(\d+)[^"]*)"',           # /page/2/
+            r'href="([^"]*[?&]p=(\d+)[^"]*)"',           # ?p=2
+            r'href="([^"]*[?&]pn=(\d+)[^"]*)"',          # ?pn=2
+            r'href="([^"]*[?&]offset=(\d+)[^"]*)"',      # ?offset=10
+            r'href="([^"]*[?&]start=(\d+)[^"]*)"',       # ?start=10
+            r'href="([^"]*[?&]page_index=(\d+)[^"]*)"',  # ?page_index=2
+            # 匹配 xxx_N.html / page_N.html / list_2.html
+            r'href="([^"]*(?:page|list|index)[-_]?(\d+)\.html?)"',
+            # 匹配 /page_N/ 格式
+            r'href="([^"]*/page[-_]?(\d+)/?)"',
+        ]
+        for pattern in page_link_patterns:
+            for m in re.finditer(pattern, raw, re.IGNORECASE):
+                full_url = urljoin(base_url, m.group(1))
+                links.add(full_url)
+
+        # 4. 翻页数字链接
+        for m in re.finditer(r'<a[^>]*href="([^"]*page=(\d+)[^"]*)"[^>]*>\s*\d+\s*</a>', raw, re.IGNORECASE):
+            links.add(urljoin(base_url, m.group(1)))
+
+        # 过滤：只保留同一域名下的链接
+        result = []
+        for link in links:
+            parsed = urlparse(link)
+            if parsed.netloc and parsed.netloc != base_parsed.netloc:
+                continue  # 跨域排除
+            if parsed.path == base_parsed.path and parsed.query == base_parsed.query:
+                continue  # 排除自身
+            result.append(link)
+
+        # 去重排序
+        return sorted(set(result))
+
+    def _extract_page_number(self, url: str) -> int:
+        """从 URL 中提取页码，用于排序"""
+        nums = re.findall(r'page[=/](\d+)|[?&]p=(\d+)|index[-_]?(\d+)|/page[-_]?(\d+)', url, re.IGNORECASE)
+        for n in nums:
+            for g in n:
+                if g:
+                    return int(g)
+        return 99  # 没识别到页码的排最后
+
+    # ============================================================
+    # 分类/标签链接发现
+    # ============================================================
+
+    def discover_category_links(self, html: str, base_url: str) -> list[str]:
+        """从页面 HTML 中发现分类/标签链接，返回去重排序后的 URL 列表"""
+        raw = self._last_raw_html or html
+        base_parsed = urlparse(base_url)
+        base_domain = f"{base_parsed.scheme}://{base_parsed.netloc}"
+
+        links = set()
+
+        # 1. 匹配分类链接（category-N.html, category-N_M.html, tag-N.html 等）
+        cat_patterns = [
+            r'href="([^"]*/(?:category|cat|sort|type)[-_]?\d+(?:[-_]\d+)?\.html?)"',
+            r'href="([^"]*/(?:tag|tags)/?[-_]?\d*)["\s>]',
+            r'href="([^"]*/tags?[-_]?\d+\.html?)"',
+        ]
+        for pattern in cat_patterns:
+            for m in re.finditer(pattern, raw, re.IGNORECASE):
+                full_url = urljoin(base_domain, m.group(1))
+                parsed = urlparse(full_url)
+                # 只保留同域链接
+                if parsed.netloc and parsed.netloc != base_parsed.netloc:
+                    continue
+                links.add(full_url)
+
+        # 2. 过滤：排除单篇文章、首页、搜索页
+        single_article = re.compile(
+            r'/(?:p|post|article|archives|detail|read|xiaohua)/\d+',
+            re.IGNORECASE,
+        )
+        result = []
+        for link in links:
+            parsed = urlparse(link)
+            path = parsed.path.rstrip("/")
+            # 排除自身
+            if path == base_parsed.path.rstrip("/") and parsed.query == base_parsed.query:
+                continue
+            # 排除单篇文章
+            if single_article.search(path):
+                continue
+            # 排除明显的非分类路径（首页翻页）
+            if re.search(r'/page[-_]?\d+\.html?$', path) and 'category' not in path and 'tag' not in path:
+                continue
+            result.append(link)
+
+        return sorted(set(result))
+
+    # ============================================================
+    # Bing 搜索（复用旧逻辑）
+    # ============================================================
+
+    async def _search_bing(self, keyword: str, max_pages: int = 3) -> list[dict]:
+        """搜索 Bing，返回结果 URL 列表"""
+        results = []
+        session = self._get_session()
+
+        for page in range(max_pages):
+            first = page * 10
+            url = f"https://www.bing.com/search?q={quote(keyword)}&first={first}"
+            print(f"  [*] Bing 搜索: {keyword[:20]}")
+
+            html = None
+            try:
+                if HAS_CRAWL4AI:
+                    crawler = await self._get_crawler()
+                    result = await crawler.arun(url, config=CrawlerRunConfig(verbose=False))
+                    if result.success:
+                        html = result.html if result.html else result.cleaned_html
+                else:
+                    resp = session.get(url)
+                    html = resp.text if resp.status_code == 200 else None
+
+                if html:
+                    urls = self._extract_bing_urls(html)
+                    for item in urls:
+                        item["keyword"] = keyword
+                        results.append(item)
+                else:
+                    print(f"  [!] 获取搜索页面失败")
+
+                await asyncio.sleep(2)
+
+            except Exception as e:
+                print(f"  [!] 搜索异常: {e}")
+                continue
+
+        return results
+
+    def _extract_bing_urls(self, html: str) -> list[dict]:
+        """从 Bing 搜索结果 HTML 中提取链接和标题"""
+        results = []
+        pattern = re.compile(
+            r'<h2[^>]*>\s*<a[^>]*href="(https?[^"]+)"[^>]*>(.*?)</a>',
+            re.DOTALL,
+        )
+        for match in pattern.finditer(html):
+            href = match.group(1).strip()
+            title = re.sub(r'<[^>]+>', '', match.group(2)).strip()
+            if href and title and len(title) > 5 and "bing.com" not in href and "microsoft.com" not in href:
+                results.append({"url": href, "title": title})
+        return results
+
+    # 别名兼容
+    async def search_bing(self, keyword: str, max_pages: int = 3) -> list[dict]:
+        return await self._search_bing(keyword, max_pages)
+
+    async def search(self, keyword: str, max_pages: int = 3) -> list[dict]:
+        return await self._search_bing(keyword, max_pages)
+
+    async def search_baidu(self, keyword: str, max_pages: int = 3) -> list[dict]:
+        return await self._search_bing(keyword, max_pages)
+
+    # ============================================================
+    # 旧接口兼容（单页抓取）
+    # ============================================================
+
+    async def crawl_page(self, url: str) -> str:
+        """抓取单个页面（兼容旧接口）"""
+        content, ok = await self.crawl_page_with_retry(url)
+        if ok:
+            return content[:8000]
+        return ""
+
+    async def crawl_batch(self, urls: list[str]) -> list[tuple[str, str]]:
+        """批量抓取（兼容旧接口）"""
+        contents = []
+        for url in urls:
+            content, ok = await self.crawl_page_with_retry(url)
+            contents.append((url, content[:8000] if ok else ""))
+            await asyncio.sleep(1.5)
+        return contents

crawler/main.py

@@ -0,0 +1,85 @@
+"""
+爬虫持续采集入口（深度采集模式）。
+流程：搜索笑话站点 → 翻页深度采集 → AI 提取 → 入库。
+每批采集 20 条后休息 2-10 分钟，出错休息 3-11 分钟后继续。
+按 Ctrl+C 中断。
+
+用法:
+  python crawler/main.py                     # 默认，无头浏览器
+  python crawler/main.py --no-headless       # 显示浏览器窗口（方便测试）
+"""
+import argparse
+import asyncio
+import os
+import sys
+
+if sys.platform == "win32":
+    import shutil
+    os.environ["PYTHONIOENCODING"] = "utf-8"
+    os.environ["TERM"] = "dumb"
+    try:
+        shutil.get_terminal_size()
+    except Exception:
+        pass
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from crawler.processor import Processor
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="笑话爬虫 - 深度采集模式")
+    parser.add_argument(
+        "--no-headless",
+        action="store_true",
+        help="显示浏览器窗口（默认无头模式）",
+    )
+    return parser.parse_args()
+
+
+def main():
+    args = parse_args()
+    headless = not args.no_headless
+
+    api_base = os.getenv("API_BASE", "http://localhost:8001")
+    username = os.getenv("CRAWL_USERNAME", "admin")
+    password = os.getenv("CRAWL_PASSWORD", "admin123")
+    keyword_str = os.getenv("CRAWL_KEYWORD", "笑话大全,搞笑段子,冷笑话,幽默笑话,爆笑笑话,笑话集锦")
+    keywords = [k.strip() for k in keyword_str.split(",") if k.strip()]
+
+    if not keywords:
+        print("错误: 未设置关键词")
+        return
+
+    print(f"=" * 50)
+    print(f"笑话爬虫 - 深度采集模式")
+    print(f"搜索词: {keywords}")
+    print(f"API 地址: {api_base}")
+    print(f"浏览器: {'显示窗口' if not headless else '无头模式'}")
+    print(f"每批 20 条后休息 2-10 分钟")
+    print(f"出错后休息 3-11 分钟后重试")
+    print(f"按 Ctrl+C 终止")
+    print(f"=" * 50)
+
+    try:
+        import crawl4ai
+        print(f"crawl4ai 版本: {crawl4ai.__version__}")
+    except ImportError:
+        print("错误: crawl4ai 未安装，请先运行: pip install crawl4ai")
+        return
+
+    processor = Processor(
+        api_base=api_base,
+        username=username,
+        password=password,
+        headless=headless,
+    )
+
+    try:
+        asyncio.run(processor.run_continuous(keywords=keywords, max_pages=3, batch_size=20))
+    except KeyboardInterrupt:
+        print("\n用户中断，退出")
+
+
+if __name__ == "__main__":
+    main()

crawler/processor.py

@@ -0,0 +1,470 @@
+"""爬虫流程编排：搜索笑话站点 → 深度翻页采集 → AI 提取 → 入库"""
+
+import asyncio
+import hashlib
+import random
+import re
+import httpx
+from urllib.parse import urljoin, urlparse
+
+from crawler.ai_service import AiService
+from crawler.crawler_service import CrawlerService
+
+
+class Processor:
+    def __init__(self, api_base: str, username: str, password: str, headless: bool = True):
+        self.api_base = api_base.rstrip("/")
+        self.username = username
+        self.password = password
+        self.token = None
+        self.ai = None
+        self.crawler = CrawlerService(headless=headless)
+        self.types = []
+        self.crowds = []
+
+    # === API 认证 ===
+    def _login(self) -> str:
+        resp = httpx.post(
+            f"{self.api_base}/api/auth/login",
+            json={"username": self.username, "password": self.password},
+            timeout=30,
+        )
+        resp.raise_for_status()
+        return resp.json()["access_token"]
+
+    def _get(self, path: str) -> dict | list:
+        resp = httpx.get(
+            f"{self.api_base}{path}",
+            headers={"Authorization": f"Bearer {self.token}"},
+            timeout=30,
+        )
+        resp.raise_for_status()
+        return resp.json()
+
+    def _post(self, path: str, data: dict) -> dict:
+        resp = httpx.post(
+            f"{self.api_base}{path}",
+            json=data,
+            headers={"Authorization": f"Bearer {self.token}"},
+            timeout=30,
+        )
+        resp.raise_for_status()
+        return resp.json()
+
+    # === 初始化 ===
+    def setup(self):
+        print("[*] 正在登录...")
+        self.token = self._login()
+        print("[*] 登录成功")
+
+        ai_config = self._get("/api/admin/settings/active")
+        self.ai = AiService(
+            api_base=ai_config["api_base"],
+            api_key=ai_config["api_key"],
+            model_name=ai_config["model_name"],
+            temperature=ai_config.get("temperature", 0.7),
+            max_tokens=ai_config.get("max_tokens", 2048),
+        )
+        print(f"[*] AI 配置: {ai_config['model_name']}")
+
+        self.types = self._get("/api/categories/types")
+        self.crowds = self._get("/api/categories/crowds")
+        print(f"[*] 分类: {len(self.types)} 种类型, {len(self.crowds)} 种人群")
+
+    def _get_existing_hashes(self) -> set[str]:
+        """获取库里已有笑话的 content hash，用于去重"""
+        try:
+            data = self._get("/api/admin/jokes?page=1&page_size=1000")
+            hashes = set()
+            for j in data.get("items", []):
+                content = j.get("content", "")
+                if content:
+                    hashes.add(hashlib.md5(content.encode()).hexdigest())
+            return hashes
+        except Exception:
+            return set()
+
+    def _submit_joke(self, joke: dict) -> bool:
+        """提交单条笑话至 API（status=pending 待审核）"""
+        try:
+            # Support both old format (type/crowd) and new format (types/crowds arrays)
+            type_names = joke.get("types", [joke.get("type", "")])
+            crowd_names = joke.get("crowds", [joke.get("crowd", "")])
+            if isinstance(type_names, str):
+                type_names = [type_names] if type_names else []
+            if isinstance(crowd_names, str):
+                crowd_names = [crowd_names] if crowd_names else []
+
+            type_ids = []
+            crowd_ids = []
+            for n in type_names:
+                for t in self.types:
+                    if t.get("name") == n:
+                        type_ids.append(t.get("id"))
+                        break
+            for n in crowd_names:
+                for c in self.crowds:
+                    if c.get("name") == n:
+                        crowd_ids.append(c.get("id"))
+                        break
+
+            payload = {
+                "title": joke.get("title", "无标题"),
+                "content": joke.get("content", ""),
+                "type_ids": type_ids if type_ids else None,
+                "crowd_ids": crowd_ids if crowd_ids else None,
+                "status": "pending",
+            }
+            self._post("/api/admin/jokes", payload)
+            return True
+        except Exception as e:
+            print(f"  [!] 提交失败: {e}")
+            return False
+
+    # === 主循环 ===
+
+    async def run_continuous(self, keywords: list[str], max_pages: int = 3, batch_size: int = 20):
+        """持续深度采集循环"""
+        print(f"\n>> 深度采集模式启动")
+        print(f"    搜索词: {keywords}")
+        print(f"    每批目标: {batch_size} 条")
+        print(f"    API: {self.api_base}")
+        print(f"    按 Ctrl+C 中断\n")
+
+        self.setup()
+
+        existing_hashes = self._get_existing_hashes()
+        print(f"[*] 当前库中已有 {len(existing_hashes)} 条笑话（用于去重）")
+
+        type_names = [t.get("name", "") for t in self.types]
+        crowd_names = [c.get("name", "") for c in self.crowds]
+        total_saved = 0
+        visited_urls: set[str] = set()
+
+        try:
+            while True:
+                batch_saved = 0
+                print(f"\n{'='*50}")
+                print(f"  开始新一批深度采集 (已累计 {total_saved} 条)")
+                print(f"{'='*50}")
+
+                try:
+                    # Step 1: 搜索笑话站点
+                    print(f"\n[*] 搜索笑话站点...")
+                    sites = await self.crawler.search_joke_sites(keywords, max_results=8)
+
+                    if not sites:
+                        print(f"  [!] 未找到笑话站点，休息后重试")
+                        rest = random.randint(120, 600)
+                        print(f"  休息 {rest//60} 分 {rest%60} 秒...")
+                        await asyncio.sleep(rest)
+                        continue
+
+                    # Step 2: 逐个站点深度采集
+                    for site in sites:
+                        if batch_saved >= batch_size:
+                            break
+
+                        domain = site["domain"]
+                        site_url = site["url"]
+
+                        print(f"\n{'─'*40}")
+                        print(f"  开始采集站点: {domain}")
+                        print(f"{'─'*40}")
+
+                        # 检查该站点失败次数
+                        fail_count = self.crawler.site_failures.get(domain, 0)
+                        if fail_count >= 3:
+                            print(f"  [!] 站点 {domain} 已连续失败 {fail_count} 次，跳过")
+                            continue
+
+                        saved_from_site = await self._crawl_site_deep(
+                            site_url=site_url,
+                            domain=domain,
+                            type_names=type_names,
+                            crowd_names=crowd_names,
+                            existing_hashes=existing_hashes,
+                            visited_urls=visited_urls,
+                            target=batch_size - batch_saved,
+                        )
+                        batch_saved += saved_from_site
+                        total_saved += saved_from_site
+
+                    # Step 3: 休息
+                    rest = random.randint(120, 600)
+                    print(f"\n[OK] 本批入库 {batch_saved} 条，休息 {rest//60} 分 {rest%60} 秒...")
+                    print(f"     按 Ctrl+C 中断\n")
+
+                except Exception as e:
+                    rest = random.randint(180, 660)
+                    print(f"\n[!] 出错: {e}")
+                    print(f"    休息 {rest//60} 分 {rest%60} 秒后重试...")
+
+                await asyncio.sleep(rest)
+
+        except asyncio.CancelledError:
+            print("\n用户中断，退出")
+        finally:
+            await self.crawler.close()
+
+    async def _crawl_site_deep(
+        self,
+        site_url: str,
+        domain: str,
+        type_names: list[str],
+        crowd_names: list[str],
+        existing_hashes: set[str],
+        visited_urls: set[str],
+        target: int,
+    ) -> int:
+        """深度采集一个站点：抓取首页 → 发现翻页链接 → 逐页抓取提取笑话"""
+        saved = 0
+        pages_to_crawl = []
+
+        # 1. 抓取首页
+        print(f"  [*] 抓取首页: {site_url[:60]}")
+        html, ok = await self.crawler.crawl_page_with_retry(site_url)
+        if not ok:
+            self.crawler.site_failures[domain] = self.crawler.site_failures.get(domain, 0) + 1
+            print(f"  [!] 首页抓取失败 ({self.crawler.site_failures[domain]}/3)")
+            return 0
+
+        visited_urls.add(site_url)
+
+        # 2. 从首页提取笑话
+        try:
+            jokes = self.ai.extract_jokes(html[:8000], type_names, crowd_names)
+            saved += self._save_jokes(jokes, existing_hashes, target - saved)
+            print(f"  [+] 首页提取 {len(jokes)} 条，入库 {saved} 条")
+        except Exception as e:
+            print(f"  [!] 首页 AI 提取失败: {e}")
+
+        if saved >= target:
+            self.crawler.site_failures[domain] = 0
+            return saved
+
+        # 3. 发现翻页链接
+        page_links = self.crawler.discover_page_links(html, site_url)
+        # 过滤已访问的链接
+        page_links = [l for l in page_links if l not in visited_urls]
+        # 按页码排序
+        page_links.sort(key=lambda l: self.crawler._extract_page_number(l))
+
+        # 限制翻页深度，避免无限抓取
+        max_pages_per_site = 30
+        page_links = page_links[:max_pages_per_site]
+        print(f"  [*] 发现 {len(page_links)} 个翻页链接，开始逐页采集...")
+
+        # 4. 逐页翻页采集
+        for idx, page_url in enumerate(page_links):
+            if saved >= target:
+                break
+
+            # 检查站点是否已失效
+            if self.crawler.site_failures.get(domain, 0) >= 3:
+                print(f"  [!] 站点 {domain} 失败过多，跳过")
+                break
+
+            print(f"  [*] 翻页 {idx+1}/{len(page_links)}: {page_url[:60]}")
+
+            html, ok = await self.crawler.crawl_page_with_retry(page_url)
+            visited_urls.add(page_url)
+
+            if not ok:
+                self.crawler.site_failures[domain] = self.crawler.site_failures.get(domain, 0) + 1
+                print(f"  [!] 抓取失败 ({self.crawler.site_failures[domain]}/3)")
+                continue
+
+            # 重置失败计数
+            self.crawler.site_failures[domain] = 0
+
+            try:
+                jokes = self.ai.extract_jokes(html[:8000], type_names, crowd_names)
+                new_saved = self._save_jokes(jokes, existing_hashes, target - saved)
+                if new_saved > 0:
+                    saved += new_saved
+                    print(f"  [+] 提取 {len(jokes)} 条，入库 {new_saved} 条 (累计 {saved}/{target})")
+                else:
+                    print(f"  [*] 提取 {len(jokes)} 条（均为重复）")
+            except Exception as e:
+                print(f"  [!] AI 提取失败: {e}")
+
+            await asyncio.sleep(random.uniform(1, 3))
+
+        self.crawler.site_failures[domain] = 0
+
+        # 5. 发现分类链接并逐个深度采集
+        cat_links = self.crawler.discover_category_links(html, site_url)
+        cat_links = [l for l in cat_links if l not in visited_urls]
+        # 限制分类数量
+        max_categories = 20
+        cat_links = cat_links[:max_categories]
+        if cat_links:
+            print(f"  [*] 发现 {len(cat_links)} 个分类链接，开始逐类采集...")
+            for cat_url in cat_links:
+                if saved >= target:
+                    break
+                if self.crawler.site_failures.get(domain, 0) >= 3:
+                    print(f"  [!] 站点 {domain} 失败过多，跳过分类")
+                    break
+
+                saved += await self._crawl_category(
+                    cat_url=cat_url,
+                    domain=domain,
+                    type_names=type_names,
+                    crowd_names=crowd_names,
+                    existing_hashes=existing_hashes,
+                    visited_urls=visited_urls,
+                    target=target - saved,
+                )
+
+        print(f"  [*] 站点 {domain} 采集完成，共入库 {saved} 条")
+        return saved
+
+    async def _crawl_category(
+        self,
+        cat_url: str,
+        domain: str,
+        type_names: list[str],
+        crowd_names: list[str],
+        existing_hashes: set[str],
+        visited_urls: set[str],
+        target: int,
+    ) -> int:
+        """深度采集一个分类页及其翻页"""
+        saved = 0
+        cat_name = cat_url.split("/")[-1].split(".")[0]
+        print(f"\n  {'─'*36}")
+        print(f"  分类采集 [{cat_name}]: {cat_url[:60]}")
+        print(f"  {'─'*36}")
+
+        # 1. 抓取分类首页
+        html, ok = await self.crawler.crawl_page_with_retry(cat_url)
+        visited_urls.add(cat_url)
+        if not ok:
+            self.crawler.site_failures[domain] = self.crawler.site_failures.get(domain, 0) + 1
+            print(f"  [!] 分类首页抓取失败 ({self.crawler.site_failures[domain]}/3)")
+            return 0
+
+        # 2. AI 提取笑话
+        try:
+            jokes = self.ai.extract_jokes(html[:8000], type_names, crowd_names)
+            saved += self._save_jokes(jokes, existing_hashes, target - saved)
+            print(f"  [+] 分类首页提取 {len(jokes)} 条，入库 {saved} 条")
+        except Exception as e:
+            print(f"  [!] 分类首页 AI 提取失败: {e}")
+
+        if saved >= target:
+            return saved
+
+        # 3. 发现该分类的翻页链接
+        # 先尝试通用翻页模式，再尝试分类特定翻页（category-5_2.html）
+        page_links = self.crawler.discover_page_links(html, cat_url)
+        # 从当前分类 URL 派生出分类翻页模式（e.g. category-5 → category-5_2.html）
+        cat_base = cat_url.split("/")[-1].replace(".html", "")
+        cat_page_pattern = re.compile(
+            rf'href="([^"]*{re.escape(cat_base)}[-_]?(\d+)\.html?)"',
+            re.IGNORECASE,
+        )
+        raw = self.crawler._last_raw_html or html
+        for m in cat_page_pattern.finditer(raw):
+            full_url = urljoin(cat_url, m.group(1))
+            page_links.append(full_url)
+
+        page_links = [l for l in page_links if l not in visited_urls]
+        page_links = list(set(page_links))  # 去重
+        page_links.sort(key=lambda l: self.crawler._extract_page_number(l))
+
+        # 限制翻页深度
+        max_pages_per_cat = 20
+        page_links = page_links[:max_pages_per_cat]
+        if page_links:
+            print(f"  [*] 发现 {len(page_links)} 个翻页链接，开始逐页采集...")
+
+        # 4. 逐页抓取
+        for idx, page_url in enumerate(page_links):
+            if saved >= target:
+                break
+            if self.crawler.site_failures.get(domain, 0) >= 3:
+                print(f"  [!] 站点 {domain} 失败过多，跳过本分类")
+                break
+
+            print(f"  [*] 翻页 {idx+1}/{len(page_links)}: {page_url[:60]}")
+
+            html, ok = await self.crawler.crawl_page_with_retry(page_url)
+            visited_urls.add(page_url)
+
+            if not ok:
+                self.crawler.site_failures[domain] = self.crawler.site_failures.get(domain, 0) + 1
+                print(f"  [!] 分类翻页抓取失败 ({self.crawler.site_failures[domain]}/3)")
+                continue
+
+            self.crawler.site_failures[domain] = 0
+
+            try:
+                jokes = self.ai.extract_jokes(html[:8000], type_names, crowd_names)
+                new_saved = self._save_jokes(jokes, existing_hashes, target - saved)
+                if new_saved > 0:
+                    saved += new_saved
+                    print(f"  [+] 提取 {len(jokes)} 条，入库 {new_saved} 条 (累计 {saved}/{target})")
+                else:
+                    print(f"  [*] 提取 {len(jokes)} 条（均为重复）")
+            except Exception as e:
+                print(f"  [!] AI 提取失败: {e}")
+
+            await asyncio.sleep(random.uniform(1, 3))
+
+        print(f"  [*] 分类 [{cat_name}] 采集完成，入库 {saved} 条")
+        return saved
+
+    def _save_jokes(self, jokes: list[dict], existing_hashes: set[str], limit: int) -> int:
+        """去重并入库笑话，返回成功入库数"""
+        saved = 0
+        for joke in jokes:
+            if saved >= limit:
+                break
+            content_text = joke.get("content", "")
+            if not content_text:
+                continue
+            h = hashlib.md5(content_text.encode()).hexdigest()
+            if h in existing_hashes:
+                continue
+            existing_hashes.add(h)
+            if self._submit_joke(joke):
+                saved += 1
+                print(f"  [+] 入库: {joke.get('title', '')[:30]}")
+        return saved
+
+    # === 单站点采集（供 site_crawler.py 调用） ===
+
+    async def crawl_site(self, site_url: str, batch_size: int = 9999):
+        """初始化后深度采集单个站点"""
+        print(f"\n>> 单站点采集: {site_url}\n")
+        self.setup()
+
+        existing_hashes = self._get_existing_hashes()
+        print(f"[*] 当前库中已有 {len(existing_hashes)} 条笑话（用于去重）")
+
+        type_names = [t.get("name", "") for t in self.types]
+        crowd_names = [c.get("name", "") for c in self.crowds]
+        domain = urlparse(site_url).netloc.lower()
+        visited_urls: set[str] = set()
+
+        try:
+            saved = await self._crawl_site_deep(
+                site_url=site_url,
+                domain=domain,
+                type_names=type_names,
+                crowd_names=crowd_names,
+                existing_hashes=existing_hashes,
+                visited_urls=visited_urls,
+                target=batch_size,
+            )
+            print(f"\n[OK] 站点采集完成，共入库 {saved} 条笑话")
+        finally:
+            await self.crawler.close()
+
+    # === 旧接口兼容 ===
+    async def run(self, keywords: list[str], max_pages: int = 3):
+        """单轮爬取（旧接口，内部调用 run_continuous）"""
+        await self.run_continuous(keywords, max_pages, batch_size=9999)

crawler/prompts.py

@@ -0,0 +1,41 @@
+"""AI 提示词模板。"""
+
+# ===== 笑话提取 =====
+EXTRACTION_SYSTEM_PROMPT = """你是一个笑话提取专家。从给定的网页文本中识别并提取所有笑话、幽默段子或有趣内容。
+要求：
+1. 只返回真正的笑话内容，不要提取普通文章或新闻
+2. 每条笑话需要包含：title（简短标题）、content（完整笑话内容）、type（类型）、crowd（人群）
+3. 如果网页中没有笑话，返回空数组 []
+4. 永远返回合法的 JSON 格式，根节点为数组或包含 jokes 键的对象"""
+
+EXTRACTION_USER_PROMPT = """网页内容：
+---
+{page_content}
+---
+
+已知笑话类型：{known_types}
+已知人群分类：{known_crowds}
+
+请提取所有笑话，以 JSON 格式返回，示例：
+[
+  {{"title": "程序员的幽默", "content": "程序员去相亲...", "types": ["谐音梗", "段子"], "crowds": ["职场", "大学生"]}},
+  {{"title": "...", "content": "...", "types": ["..."], "crowds": ["..."]}}
+]
+
+注意：types 和 crowds 是数组，可以填多个。
+只返回 JSON，不要其他文字。"""
+
+
+# ===== 笑话改写 =====
+REWRITE_SYSTEM_PROMPT = """你是一个幽默作家，负责润色和改写笑话。
+要求：
+1. 保持笑话的核心笑点不变
+2. 语言更通顺、更幽默
+3. 字数控制在原内容的 80%-120% 之间
+4. 不要添加任何解释说明"""
+
+REWRITE_USER_PROMPT = """请润色以下笑话：
+
+{content}
+
+只返回润色后的笑话文字，不要其他内容。"""

crawler/site_crawler.py

@@ -0,0 +1,165 @@
+"""
+单站点深度采集工具。
+从 JSON 文件按编号加载站点，或直接指定 URL，深度采集该站点所有笑话。
+
+用法:
+  python crawler/site_crawler.py --id 1                    # 采集 site_finder 发现的 #1 站点
+  python crawler/site_crawler.py --id 1,2,3               # 批量采集多个站点
+  python crawler/site_crawler.py --url https://...         # 直接采集指定 URL
+  python crawler/site_crawler.py --id 1 --no-headless      # 显示浏览器窗口
+"""
+import argparse
+import asyncio
+import json
+import os
+import sys
+
+if sys.platform == "win32":
+    import shutil
+    os.environ["PYTHONIOENCODING"] = "utf-8"
+    os.environ["TERM"] = "dumb"
+    try:
+        shutil.get_terminal_size()
+    except Exception:
+        pass
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from crawler.processor import Processor
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="单站点深度采集工具")
+    parser.add_argument(
+        "--id",
+        type=str,
+        default=None,
+        help="站点编号（从 site_finder 生成的 JSON 读取），多个用逗号分隔",
+    )
+    parser.add_argument(
+        "--url",
+        type=str,
+        default=None,
+        help="直接指定站点 URL（与 --id 二选一）",
+    )
+    parser.add_argument(
+        "--sites-file",
+        type=str,
+        default="joke_sites.json",
+        help="站点列表 JSON 文件路径（默认: joke_sites.json）",
+    )
+    parser.add_argument(
+        "--api-base",
+        type=str,
+        default=os.getenv("API_BASE", "http://localhost:8001"),
+        help="API 服务地址",
+    )
+    parser.add_argument(
+        "--username",
+        type=str,
+        default=os.getenv("CRAWL_USERNAME", "admin"),
+        help="管理员用户名",
+    )
+    parser.add_argument(
+        "--password",
+        type=str,
+        default=os.getenv("CRAWL_PASSWORD", "admin123"),
+        help="管理员密码",
+    )
+    parser.add_argument(
+        "--no-headless",
+        action="store_true",
+        help="显示浏览器窗口",
+    )
+    return parser.parse_args()
+
+
+def load_sites(path: str) -> list[dict]:
+    """从 JSON 文件加载站点列表"""
+    if not os.path.exists(path):
+        print(f"错误: 站点文件 {path} 不存在，请先运行 site_finder.py")
+        sys.exit(1)
+    try:
+        with open(path, "r", encoding="utf-8") as f:
+            data = json.load(f)
+        if not data:
+            print(f"错误: 站点文件 {path} 为空")
+            sys.exit(1)
+        return data
+    except Exception as e:
+        print(f"错误: 读取站点文件失败: {e}")
+        sys.exit(1)
+
+
+def resolve_sites(args) -> list[str]:
+    """解析 --id 或 --url 参数，返回待采集的 URL 列表"""
+    if args.url:
+        return [args.url]
+
+    if not args.id:
+        print("错误: 请指定 --id 或 --url")
+        print("  例如: python crawler/site_crawler.py --id 1")
+        print("  例如: python crawler/site_crawler.py --url https://xiaohua.com")
+        sys.exit(1)
+
+    # 解析编号列表 "1,2,3" → [1, 2, 3]
+    try:
+        ids = [int(x.strip()) for x in args.id.split(",") if x.strip()]
+    except ValueError:
+        print("错误: --id 参数必须是数字，多个用逗号分隔")
+        sys.exit(1)
+
+    sites = load_sites(args.sites_file)
+    found = []
+    for sid in ids:
+        match = [s for s in sites if s["id"] == sid]
+        if match:
+            found.append(match[0])
+            print(f"  [*] 站点 #{sid}: {match[0]['domain']} — {match[0]['title'][:40]}")
+        else:
+            print(f"  [!] 站点 #{sid} 未找到（可用编号: {[s['id'] for s in sites[:10]]}...)")
+
+    if not found:
+        print("错误: 没有找到有效的站点编号")
+        sys.exit(1)
+
+    return [s["url"] for s in found]
+
+
+def main():
+    args = parse_args()
+    headless = not args.no_headless
+    urls = resolve_sites(args)
+
+    print(f"=" * 50)
+    print(f"单站点深度采集工具")
+    print(f"目标站点: {len(urls)} 个")
+    for u in urls:
+        print(f"  - {u}")
+    print(f"API 地址: {args.api_base}")
+    print(f"浏览器: {'显示窗口' if not headless else '无头模式'}")
+    print(f"=" * 50)
+
+    try:
+        import crawl4ai
+        print(f"crawl4ai 版本: {crawl4ai.__version__}")
+    except ImportError:
+        print("错误: crawl4ai 未安装，请先运行: pip install crawl4ai")
+        return
+
+    processor = Processor(
+        api_base=args.api_base,
+        username=args.username,
+        password=args.password,
+        headless=headless,
+    )
+
+    try:
+        for url in urls:
+            asyncio.run(processor.crawl_site(url))
+    except KeyboardInterrupt:
+        print("\n用户中断，退出")
+
+
+if __name__ == "__main__":
+    main()

crawler/site_finder.py

@@ -0,0 +1,128 @@
+"""
+笑话站点发现工具。
+搜索 Bing 获取笑话聚合网站列表，保存到 JSON 文件。
+
+用法:
+  python crawler/site_finder.py
+  python crawler/site_finder.py --keywords "笑话大全,冷笑话" --output my_sites.json
+  python crawler/site_finder.py --no-headless
+"""
+import argparse
+import asyncio
+import json
+import os
+import sys
+from datetime import datetime
+
+if sys.platform == "win32":
+    import shutil
+    os.environ["PYTHONIOENCODING"] = "utf-8"
+    os.environ["TERM"] = "dumb"
+    try:
+        shutil.get_terminal_size()
+    except Exception:
+        pass
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from crawler.crawler_service import CrawlerService
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="笑话站点发现工具")
+    parser.add_argument(
+        "--keywords",
+        type=str,
+        default=os.getenv("CRAWL_KEYWORD", "笑话大全,搞笑段子,冷笑话,幽默笑话,爆笑笑话,笑话集锦"),
+        help="搜索关键词，逗号分隔",
+    )
+    parser.add_argument(
+        "--output",
+        type=str,
+        default="joke_sites.json",
+        help="输出 JSON 文件路径（默认: joke_sites.json）",
+    )
+    parser.add_argument(
+        "--max-sites",
+        type=int,
+        default=15,
+        help="最多保留几个站点（默认: 15）",
+    )
+    parser.add_argument(
+        "--no-headless",
+        action="store_true",
+        help="显示浏览器窗口",
+    )
+    return parser.parse_args()
+
+
+def load_existing_sites(path: str) -> list[dict]:
+    """加载已有站点列表"""
+    if os.path.exists(path):
+        try:
+            with open(path, "r", encoding="utf-8") as f:
+                return json.load(f)
+        except Exception:
+            pass
+    return []
+
+
+def save_sites(path: str, sites: list[dict]):
+    """保存站点列表到 JSON 文件"""
+    with open(path, "w", encoding="utf-8") as f:
+        json.dump(sites, f, ensure_ascii=False, indent=2)
+    print(f"\n[OK] 已保存 {len(sites)} 个站点到 {path}")
+
+
+def main():
+    args = parse_args()
+    headless = not args.no_headless
+    keywords = [k.strip() for k in args.keywords.split(",") if k.strip()]
+
+    print(f"=" * 50)
+    print(f"笑话站点发现工具")
+    print(f"搜索词: {keywords}")
+    print(f"最大站点数: {args.max_sites}")
+    print(f"输出文件: {args.output}")
+    print(f"浏览器: {'显示窗口' if not headless else '无头模式'}")
+    print(f"=" * 50)
+
+    # 加载已存在的站点（保留已有编号）
+    existing = load_existing_sites(args.output)
+    existing_domains = {s["domain"] for s in existing}
+    next_id = max([s["id"] for s in existing], default=0) + 1
+    print(f"[*] 已有 {len(existing)} 个站点记录，新编号从 {next_id} 开始")
+
+    async def run():
+        nonlocal next_id
+        crawler = CrawlerService(headless=headless)
+        try:
+            new_sites = await crawler.search_joke_sites(keywords, max_results=args.max_sites)
+        finally:
+            await crawler.close()
+
+        # 合并新旧站点（去重）
+        added = 0
+        for site in new_sites:
+            domain = site["domain"]
+            if domain not in existing_domains:
+                site["id"] = next_id
+                site["found_at"] = datetime.now().isoformat(timespec="seconds")
+                existing.append(site)
+                existing_domains.add(domain)
+                next_id += 1
+                added += 1
+                print(f"  [+] 新增 #{site['id']}: {site['domain']} — {site['title'][:40]}")
+
+        save_sites(args.output, existing)
+
+        if existing:
+            print(f"\n站点列表：")
+            for s in existing:
+                print(f"  #{s['id']:2d}  {s['domain']:30s} {s['title'][:35]}")
+
+    asyncio.run(run())
+
+
+if __name__ == "__main__":
+    main()

docs/superpowers/plans/2026-06-02-joke-generator-plan.md

@@ -0,0 +1,754 @@
+# 智能笑话生成器实现计划
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** 实现用户通过选择场景和输入关键词，调用 AI 生成笑话的功能，支持收藏到本地。
+
+**Architecture:** 后端新增 `/api/generate` 接口调用 AI 服务；前端新增 `/generate` 页面，复用现有笑话卡片样式，支持 localStorage 收藏。
+
+**Tech Stack:** FastAPI + OpenAI SDK (NVIDIA NIM) / Vue3 + Element Plus + localStorage
+
+---
+
+## 文件变更概览
+
+### 新增文件
+- `api/app/routers/generate.py` - 生成器 API 路由
+- `api/app/schemas/joke.py` - 添加 GenerateRequest/GenerateResponse schema
+- `web/src/views/generate/index.vue` - 生成器页面
+- `web/src/api/generate.js` - 前端 API 模块
+
+### 修改文件
+- `api/main.py` - 注册新路由
+- `web/src/router/index.js` - 添加 /generate 路由
+- `web/src/components/AppHeader.vue` - 导航栏添加 AI 生成入口按钮
+
+---
+
+## Task 1: 后端 Schema 定义
+
+**Files:**
+- Modify: `api/app/schemas/joke.py:1-45`
+
+- [ ] **Step 1: 添加 GenerateRequest 和 GenerateResponse schema**
+
+在 `api/app/schemas/joke.py` 文件末尾添加：
+
+```python
+class GenerateRequest(BaseModel):
+    """笑话生成请求"""
+    keywords: list[str] = []
+    scenarios: list[str] = []
+
+    class Config:
+        json_schema_extra = {
+            "example": {
+                "keywords": ["加班"],
+                "scenarios": ["职场"]
+            }
+        }
+
+
+class GenerateResponse(BaseModel):
+    """笑话生成响应"""
+    title: str
+    content: str
+    created_at: datetime | None = None
+
+    class Config:
+        from_attributes = True
+```
+
+- [ ] **Step 2: 验证文件语法**
+
+Run: `cd api && python -c "from app.schemas.joke import GenerateRequest, GenerateResponse; print('OK')"`
+Expected: `OK`
+
+- [ ] **Step 3: 提交**
+
+```bash
+git add api/app/schemas/joke.py
+git commit -m "feat(api): add GenerateRequest/GenerateResponse schemas"
+```
+
+---
+
+## Task 2: 后端 AI 生成路由
+
+**Files:**
+- Create: `api/app/routers/generate.py`
+- Modify: `api/main.py:26-33`
+
+- [ ] **Step 1: 创建生成器路由文件**
+
+创建 `api/app/routers/generate.py`:
+
+```python
+"""智能笑话生成器 API"""
+
+from datetime import datetime
+from fastapi import APIRouter, HTTPException
+from openai import OpenAI
+from sqlalchemy.orm import Session
+
+from app.database import get_db
+from app.models.setting import AiSetting
+from app.schemas.joke import GenerateRequest, GenerateResponse
+
+
+router = APIRouter(prefix="/generate", tags=["生成器"])
+
+
+# AI Prompt
+GENERATION_PROMPT = """你是一位幽默大师，专门创作轻松搞笑的短笑话。
+
+{context}
+
+要求：
+1. 根据场景和关键词创作一条原创笑话
+2. 笑话要有反转或意外结局
+3. 语言简洁，30-150字
+4. 直接输出笑话内容，不需要解释
+
+格式：
+标题：xxx
+内容：xxx
+"""
+
+
+def _build_prompt(scenarios: list[str], keywords: list[str]) -> str:
+    """构建 AI prompt"""
+    parts = []
+    if scenarios:
+        parts.append(f"场景：{', '.join(scenarios)}")
+    if keywords:
+        parts.append(f"关键词：{', '.join(keywords)}")
+    if not parts:
+        parts.append("场景：日常生活的各种趣事（不指定具体场景）")
+    return GENERATION_PROMPT.format(context="\n".join(parts))
+
+
+@router.post("", response_model=GenerateResponse)
+def generate_joke(
+    req: GenerateRequest,
+    db: Session = Depends(get_db),
+):
+    """调用 AI 生成笑话"""
+    # 获取激活的 AI 配置
+    setting = db.query(AiSetting).filter(AiSetting.is_active == True).first()
+    if not setting:
+        # 如果没有配置，尝试返回默认配置
+        setting = db.query(AiSetting).first()
+    
+    if not setting or not setting.api_key:
+        raise HTTPException(status_code=503, detail="AI 服务未配置，请联系管理员")
+
+    # 调用 AI
+    try:
+        client = OpenAI(base_url=setting.api_base, api_key=setting.api_key)
+        prompt = _build_prompt(req.scenarios, req.keywords)
+        
+        response = client.chat.completions.create(
+            model=setting.model_name,
+            messages=[{"role": "user", "content": prompt}],
+            temperature=setting.temperature,
+            max_tokens=setting.max_tokens,
+        )
+        
+        raw = response.choices[0].message.content
+        return _parse_response(raw)
+        
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"生成失败: {str(e)}")
+
+
+def _parse_response(raw: str) -> GenerateResponse:
+    """解析 AI 返回内容，提取标题和内容"""
+    title = ""
+    content = raw
+    
+    # 尝试提取 "标题：xxx" 或 "标题:xxx"
+    for line in raw.split("\n"):
+        line = line.strip()
+        if line.startswith("标题：") or line.startswith("标题:"):
+            title = line.split("：", 1)[-1].split(":", 1)[-1].strip()
+            content = content.replace(line, "").strip()
+            break
+    
+    # 如果没有提取到标题，取第一行或前20字
+    if not title:
+        first_line = raw.split("\n")[0].strip()
+        if first_line.startswith("标题"):
+            first_line = first_line.split("：", 1)[-1].split(":", 1)[-1].strip()
+        title = first_line[:30] if len(first_line) > 30 else first_line
+    
+    return GenerateResponse(
+        title=title or "生成的笑话",
+        content=content.strip(),
+        created_at=datetime.now(),
+    )
+```
+
+- [ ] **Step 2: 修改 main.py 注册路由**
+
+在 `api/main.py` 第 5 行添加导入：
+
+```python
+from app.routers import jokes_router, categories_router, auth_router, admin_router, settings_router, links_router, feedback_router, generate_router
+```
+
+在 `api/main.py` 第 32 行后添加：
+
+```python
+app.include_router(generate_router, prefix="/api")
+```
+
+- [ ] **Step 3: 手动验证 API**
+
+Run: `cd api && python -m uvicorn main:app --reload --port 8001`
+Expected: 服务启动无错误
+
+访问 `http://localhost:8001/docs` 验证 `/api/generate` 接口存在
+
+- [ ] **Step 4: 测试无 AI 配置情况**
+
+Run: `curl -X POST http://localhost:8001/api/generate -H "Content-Type: application/json" -d "{}"`
+Expected: `{"detail":"AI 服务未配置，请联系管理员"}`
+
+- [ ] **Step 5: 提交**
+
+```bash
+git add api/app/routers/generate.py api/main.py
+git commit -m "feat(api): add joke generate endpoint with AI"
+```
+
+---
+
+## Task 3: 前端 API 模块
+
+**Files:**
+- Create: `web/src/api/generate.js`
+
+- [ ] **Step 1: 创建前端 API 模块**
+
+创建 `web/src/api/generate.js`:
+
+```javascript
+import request from './request'
+
+/**
+ * 调用 AI 生成笑话
+ * @param {string[]} keywords - 关键词列表
+ * @param {string[]} scenarios - 场景列表
+ * @returns {Promise<{title: string, content: string, created_at: string}>}
+ */
+export const generateJoke = (keywords = [], scenarios = []) => {
+  return request.post('/generate', {
+    keywords,
+    scenarios
+  })
+}
+```
+
+- [ ] **Step 2: 验证语法**
+
+Run: `cd web && node -c src/api/generate.js`
+Expected: 无语法错误（注意：ES module 语法可能有警告，可跳过）
+
+- [ ] **Step 3: 提交**
+
+```bash
+git add web/src/api/generate.js
+git commit -m "feat(web): add generateJoke API function"
+```
+
+---
+
+## Task 4: 前端路由配置
+
+**Files:**
+- Modify: `web/src/router/index.js:1-17`
+
+- [ ] **Step 1: 添加 /generate 路由**
+
+在 `web/src/router/index.js` 第 7 行后添加：
+
+```javascript
+{
+  path: '/generate',
+  name: 'Generate',
+  component: () => import('@/views/generate/index.vue')
+},
+```
+
+- [ ] **Step 2: 提交**
+
+```bash
+git add web/src/router/index.js
+git commit -m "feat(web): add /generate route"
+```
+
+---
+
+## Task 5: 头部导航 AI 入口按钮
+
+**Files:**
+- Modify: `web/src/components/AppHeader.vue`
+
+- [ ] **Step 1: 在导航栏添加 AI 生成按钮**
+
+在 `web/src/components/AppHeader.vue` 的 `header-nav` 区域，找到 `random-btn` 后添加：
+
+```vue
+<router-link
+  to="/generate"
+  class="nav-item generate-btn"
+  :class="{ active: $route.path === '/generate' }"
+  title="AI 生成笑话"
+>
+  <span>&#x2728;</span>
+</router-link>
+```
+
+并在样式区域添加（在 `.random-btn:hover` 后）：
+
+```css
+.generate-btn {
+  font-size: 16px;
+  padding: 8px 12px;
+}
+.generate-btn:hover {
+  background: rgba(147, 51, 234, 0.1);
+}
+.generate-btn.active {
+  background: rgba(147, 51, 234, 0.15);
+  color: #9333ea;
+}
+```
+
+- [ ] **Step 2: 提交**
+
+```bash
+git add web/src/components/AppHeader.vue
+git commit -m "feat(web): add AI generate button to header nav"
+```
+
+---
+
+## Task 6: 生成器页面
+
+**Files:**
+- Create: `web/src/views/generate/index.vue`
+
+- [ ] **Step 1: 创建生成器页面**
+
+创建 `web/src/views/generate/index.vue`:
+
+```vue
+<template>
+  <div class="generate-page">
+    <h1 class="page-title">
+      <span class="emoji">&#x2728;</span>
+      智能笑话生成器
+    </h1>
+
+    <!-- 场景选择 -->
+    <div class="section">
+      <div class="section-label">选择场景（可多选）</div>
+      <div class="scenario-chips">
+        <el-check-tag
+          v-for="s in predefinedScenarios"
+          :key="s"
+          :checked="selectedScenarios.includes(s)"
+          @change="toggleScenario(s)"
+          class="scenario-chip"
+        >
+          {{ s }}
+        </el-check-tag>
+      </div>
+    </div>
+
+    <!-- 关键词输入 -->
+    <div class="section">
+      <div class="section-label">添加关键词（可选）</div>
+      <el-input
+        v-model="keywordInput"
+        placeholder="输入关键词，如：加班、相亲、熊孩子..."
+        @keyup.enter="addKeyword"
+        clearable
+      >
+        <template #append>
+          <el-button @click="addKeyword">添加</el-button>
+        </template>
+      </el-input>
+      <div v-if="keywords.length" class="keyword-tags">
+        <el-tag
+          v-for="(kw, i) in keywords"
+          :key="i"
+          closable
+          @close="removeKeyword(i)"
+          class="keyword-tag"
+        >
+          {{ kw }}
+        </el-tag>
+      </div>
+    </div>
+
+    <!-- 生成按钮 -->
+    <div class="generate-actions">
+      <el-button
+        type="primary"
+        size="large"
+        :loading="loading"
+        :disabled="loading"
+        @click="handleGenerate"
+        class="generate-btn"
+      >
+        <span v-if="!loading">&#x1F3B2; 开始生成</span>
+        <span v-else>生成中...</span>
+      </el-button>
+      <el-button v-if="generatedJoke" size="large" @click="resetForm">
+        清空
+      </el-button>
+    </div>
+
+    <!-- 生成结果 -->
+    <div v-if="generatedJoke" class="result-card">
+      <h2 class="result-title">{{ generatedJoke.title }}</h2>
+      <pre class="result-content">{{ generatedJoke.content }}</pre>
+      <div class="result-actions">
+        <el-button type="primary" @click="handleGenerate">
+          &#x1F504; 重新生成
+        </el-button>
+        <el-button @click="handleFavorite" :type="isFavorited ? 'danger' : 'default'">
+          <span v-if="!isFavorited">&#x2764; 收藏</span>
+          <span v-else>&#x2714; 已收藏</span>
+        </el-button>
+      </div>
+    </div>
+
+    <!-- 收藏列表 -->
+    <div v-if="favorites.length" class="favorites-section">
+      <h3>&#x1F4EC; 我的收藏 ({{ favorites.length }})</h3>
+      <div class="favorite-list">
+        <div
+          v-for="(fav, i) in favorites"
+          :key="fav.id"
+          class="favorite-item"
+        >
+          <div class="fav-title">{{ fav.title }}</div>
+          <div class="fav-content">{{ fav.content.substring(0, 100) }}...</div>
+          <div class="fav-meta">
+            <span class="fav-date">{{ formatDate(fav.created_at) }}</span>
+            <el-button type="danger" size="small" text @click="removeFavorite(i)">
+              删除
+            </el-button>
+          </div>
+        </div>
+      </div>
+    </div>
+  </div>
+</template>
+
+<script setup>
+import { ref, computed } from 'vue'
+import { ElMessage } from 'element-plus'
+import { generateJoke } from '@/api/generate'
+
+const predefinedScenarios = ['职场', '校园', '社交', '家庭', '情感', '搞笑日常']
+
+const selectedScenarios = ref([])
+const keywordInput = ref('')
+const keywords = ref([])
+const loading = ref(false)
+const generatedJoke = ref(null)
+const isFavorited = ref(false)
+
+// 收藏
+const favorites = ref(JSON.parse(localStorage.getItem('joke_favorites') || '[]'))
+
+const STORAGE_KEY = 'joke_favorites'
+
+function toggleScenario(s) {
+  const idx = selectedScenarios.value.indexOf(s)
+  if (idx >= 0) {
+    selectedScenarios.value.splice(idx, 1)
+  } else {
+    selectedScenarios.value.push(s)
+  }
+}
+
+function addKeyword() {
+  const kw = keywordInput.value.trim()
+  if (kw && !keywords.value.includes(kw)) {
+    keywords.value.push(kw)
+    keywordInput.value = ''
+  }
+}
+
+function removeKeyword(i) {
+  keywords.value.splice(i, 1)
+}
+
+async function handleGenerate() {
+  loading.value = true
+  isFavorited.value = false
+  try {
+    const res = await generateJoke(keywords.value, selectedScenarios.value)
+    generatedJoke.value = res
+    // 检查是否已收藏
+    const exists = favorites.value.find(f => f.content === res.content)
+    isFavorited.value = !!exists
+  } catch (e) {
+    ElMessage.error(e.detail || '生成失败，请重试')
+    generatedJoke.value = null
+  } finally {
+    loading.value = false
+  }
+}
+
+function handleFavorite() {
+  if (!generatedJoke.value) return
+  // 检查是否已存在
+  const exists = favorites.value.findIndex(f => f.content === generatedJoke.value.content)
+  if (exists >= 0) {
+    ElMessage.info('已经收藏过了')
+    return
+  }
+  favorites.value.unshift({
+    id: Date.now().toString(),
+    title: generatedJoke.value.title,
+    content: generatedJoke.value.content,
+    created_at: new Date().toISOString()
+  })
+  localStorage.setItem(STORAGE_KEY, JSON.stringify(favorites.value))
+  isFavorited.value = true
+  ElMessage.success('收藏成功')
+}
+
+function removeFavorite(i) {
+  favorites.value.splice(i, 1)
+  localStorage.setItem(STORAGE_KEY, JSON.stringify(favorites.value))
+  // 检查当前生成的是否也从收藏中被删除了
+  if (generatedJoke.value) {
+    const stillExists = favorites.value.find(f => f.content === generatedJoke.value.content)
+    isFavorited.value = !!stillExists
+  }
+}
+
+function resetForm() {
+  generatedJoke.value = null
+  isFavorited.value = false
+}
+
+function formatDate(dateStr) {
+  if (!dateStr) return ''
+  return new Date(dateStr).toLocaleDateString('zh-CN')
+}
+</script>
+
+<style scoped>
+.generate-page {
+  max-width: 720px;
+  margin: 0 auto;
+  padding: 40px 24px;
+}
+
+.page-title {
+  font-size: 28px;
+  font-weight: 800;
+  color: var(--text-primary);
+  text-align: center;
+  margin-bottom: 40px;
+}
+.emoji {
+  display: inline-block;
+  margin-right: 10px;
+}
+
+.section {
+  margin-bottom: 28px;
+}
+.section-label {
+  font-size: 15px;
+  font-weight: 600;
+  color: var(--text-secondary);
+  margin-bottom: 12px;
+}
+
+.scenario-chips {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 10px;
+}
+:deep(.scenario-chip) {
+  padding: 8px 18px;
+  border-radius: 20px;
+  background: var(--bg-card);
+  border: 1.5px solid var(--border);
+  cursor: pointer;
+  font-size: 14px;
+  transition: all 0.2s;
+}
+:deep(.scenario-chip:hover) {
+  border-color: #9333ea;
+}
+:deep(.scenario-chip.is-checked) {
+  background: #9333ea;
+  border-color: #9333ea;
+  color: white;
+}
+
+.keyword-tags {
+  display: flex;
+  flex-wrap: wrap;
+  gap: 8px;
+  margin-top: 12px;
+}
+.keyword-tag {
+  background: #f3e8ff;
+  border-color: #9333ea;
+  color: #9333ea;
+}
+
+.generate-actions {
+  display: flex;
+  justify-content: center;
+  gap: 16px;
+  margin-bottom: 32px;
+}
+:deep(.generate-btn) {
+  padding: 12px 40px;
+  font-size: 16px;
+  background: linear-gradient(135deg, #9333ea, #7c3aed);
+  border: none;
+  border-radius: 24px;
+}
+:deep(.generate-btn:hover) {
+  box-shadow: 0 4px 20px rgba(147, 51, 234, 0.4);
+}
+
+.result-card {
+  background: var(--bg-card);
+  border-radius: var(--radius-lg);
+  padding: 32px;
+  margin-bottom: 32px;
+  box-shadow: var(--shadow);
+  border: 1px solid var(--border);
+}
+.result-title {
+  font-size: 22px;
+  font-weight: 700;
+  color: var(--text-primary);
+  margin-bottom: 16px;
+  text-align: center;
+}
+.result-content {
+  font-size: 16px;
+  line-height: 1.9;
+  color: var(--text-secondary);
+  white-space: pre-wrap;
+  margin-bottom: 24px;
+}
+.result-actions {
+  display: flex;
+  justify-content: center;
+  gap: 12px;
+}
+
+.favorites-section {
+  margin-top: 40px;
+}
+.favorites-section h3 {
+  font-size: 18px;
+  font-weight: 600;
+  color: var(--text-primary);
+  margin-bottom: 16px;
+  padding-bottom: 12px;
+  border-bottom: 1px solid var(--border);
+}
+
+.favorite-list {
+  display: flex;
+  flex-direction: column;
+  gap: 12px;
+}
+.favorite-item {
+  background: var(--bg-card);
+  border-radius: var(--radius);
+  padding: 16px 20px;
+  border: 1px solid var(--border);
+}
+.fav-title {
+  font-weight: 600;
+  color: var(--text-primary);
+  margin-bottom: 6px;
+}
+.fav-content {
+  font-size: 14px;
+  color: var(--text-muted);
+  margin-bottom: 8px;
+  line-height: 1.5;
+}
+.fav-meta {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+}
+.fav-date {
+  font-size: 12px;
+  color: var(--text-muted);
+}
+</style>
+```
+
+- [ ] **Step 2: 验证页面加载**
+
+启动服务后访问 `http://localhost:3000/generate`，验证：
+- 页面正常显示
+- 场景选择显示正常
+- 关键词输入正常
+
+- [ ] **Step 3: 提交**
+
+```bash
+git add web/src/views/generate/index.vue
+git commit -m "feat(web): add AI joke generator page with favorites"
+```
+
+---
+
+## Task 7: 更新 CLAUDE.md API 文档
+
+**Files:**
+- Modify: `CLAUDE.md`
+
+- [ ] **Step 1: 添加新 API 文档**
+
+在 API 接口一览部分添加：
+
+```
+- `POST /api/generate` — 公开：AI 生成笑话（keywords + scenarios）
+```
+
+- [ ] **Step 2: 提交**
+
+```bash
+git add CLAUDE.md
+git commit -m "docs: add /api/generate to API docs"
+```
+
+---
+
+## 验证清单
+
+完成所有任务后，请验证：
+
+- [ ] API 文档 `http://localhost:8001/docs` 显示 `/api/generate` 接口
+- [ ] 头部导航显示 AI 生成按钮
+- [ ] 访问 `/generate` 页面正常
+- [ ] 选择场景 + 输入关键词可正常调用 AI 生成
+- [ ] 生成结果正确显示
+- [ ] 收藏按钮可用，收藏数据正确存储到 localStorage
+- [ ] 收藏列表正确显示，可删除收藏
+- [ ] 页面样式正常（浅色/深色主题）

docs/superpowers/specs/2026-06-02-joke-generator-design.md

@@ -0,0 +1,151 @@
+# 智能笑话生成器设计文档
+
+**日期：** 2026-06-02  
+**状态：** 已批准
+
+## 功能概述
+
+用户点击导航栏的「AI生成」按钮，进入生成器页面，输入关键词和/或选择场景，自动调用 AI 生成一条相关笑话。用户可查看、收藏生成的笑话。
+
+## 用户交互流程
+
+1. 用户点击导航栏「AI生成」图标 → 跳转 `/generate` 页面
+2. 用户选择场景（可选，支持多选）→ 用户输入关键词（可选）
+3. 点击「开始生成」→ 调用 AI → 显示生成的笑话
+4. 用户可执行以下操作：
+   - 重新生成（使用相同参数）
+   - 收藏到本地
+   - 查看我的收藏
+   - 删除收藏
+
+## 页面结构
+
+```
+/generate
+├── 场景选择区（chips，可多选 / 手动输入）
+├── 关键词输入框
+├── 生成按钮 [开始生成]
+├── 结果展示区（生成后显示笑话卡片）
+│   ├── 笑话内容
+│   └── 操作按钮（重新生成 / 收藏 / 再来一条）
+└── 我的收藏（localStorage 存储）
+    └── 收藏列表（可删除）
+```
+
+## 预设场景选项
+
+- 职场
+- 校园
+- 社交
+- 家庭
+- 情感
+- 搞笑日常
+
+## API 设计
+
+### POST /api/generate
+
+生成 AI 笑话
+
+**请求体：**
+```json
+{
+  "keywords": ["加班", "老板"],
+  "scenario": "职场"
+}
+```
+
+**响应：**
+```json
+{
+  "title": "办公室的一天",
+  "content": "老板问：为什么每天早上都迟到？\n我说：因为堵车。\n老板说：那你为什么不早点出门？\n我说：因为早出门也会堵。\n老板沉默了。",
+  "created_at": "2026-06-02T12:00:00Z"
+}
+```
+
+**错误响应：**
+- 401：`{ "detail": "AI 服务未配置" }`
+- 500：`{ "detail": "生成失败，请重试" }`
+
+## AI Prompt
+
+```
+你是一位幽默大师，专门创作轻松搞笑的短笑话。
+
+场景：{场景}
+关键词：{关键词}
+
+要求：
+1. 根据场景和关键词创作一条原创笑话
+2. 笑话要有反转或意外结局
+3. 语言简洁，30-150字
+4. 直接输出笑话内容，不需要解释
+
+格式：
+标题：xxx
+内容：xxx
+```
+
+## 数据存储
+
+**收藏数据（localStorage）**
+```json
+{
+  "joke_favorites": [
+    {
+      "id": "uuid",
+      "title": "笑话标题",
+      "content": "笑话内容",
+      "created_at": "2026-06-02T12:00:00Z"
+    }
+  ]
+}
+```
+
+## 组件清单
+
+| 组件 | 说明 |
+|------|------|
+| HeaderNav 生成按钮 | 导航栏添加 AI 生成图标入口 |
+| GeneratePage | 生成器主页面 |
+| ScenarioSelector | 场景选择 chips 组件 |
+| KeywordInput | 关键词输入框 |
+| JokeCard | 生成的笑话展示卡片（复用现有 JokeCard 样式） |
+| FavoriteList | 我的收藏列表 |
+| FavoriteItem | 收藏项（带删除按钮） |
+
+## 错误处理
+
+| 场景 | 处理方式 |
+|------|----------|
+| AI 未配置 | Toast 提示「系统暂未配置 AI 服务」 |
+| 生成失败 | Toast 提示「生成失败，请重试」 |
+| 网络错误 | Toast 提示「网络错误，请检查网络连接」 |
+| 加载中 | 按钮显示 loading spinner，禁用点击 |
+
+## 技术实现
+
+### 后端
+- 新增 `POST /api/generate` 接口
+- 使用现有 AiSetting 配置调用 AI
+- 返回结构化笑话数据
+
+### 前端
+- 新增 `web/src/views/generate/index.vue` 页面
+- 新增 `web/src/api/generate.js` API 模块
+- 复用现有组件样式
+- localStorage 存储收藏数据
+
+## 文件变更清单
+
+### 新增文件
+- `api/app/routers/generate.py` - 生成器后端路由
+- `api/app/schemas/joke.py` - 添加 GenerateRequest/GenerateResponse schema
+- `web/src/views/generate/index.vue` - 生成器页面
+- `web/src/api/generate.js` - 前端 API
+
+### 修改文件
+- `api/main.py` - 注册新路由
+- `web/src/router/index.js` - 添加 /generate 路由
+- `web/src/components/Header.vue` - 添加生成器入口按钮（如有）

docs/采集工具操作文档.md

@@ -0,0 +1,149 @@
+# 采集工具操作文档
+
+## 概述
+
+采集工具分为两个独立程序：
+
+| 程序 | 功能 | 输出 |
+|------|------|------|
+| `site_finder.py` | 搜索 Bing，发现笑话聚合网站 | `joke_sites.json`（站点编号列表） |
+| `site_crawler.py` | 按编号或 URL，深度翻页采集笑话 | 直接提交到 API（数据库） |
+
+---
+
+## 一、程序 1：发现笑话站点
+
+搜索 Bing 获取笑话聚合网站，保存到 JSON 文件。
+
+```bash
+# 搜索并保存站点列表（默认输出到 joke_sites.json）
+python crawler/site_finder.py
+
+# 指定关键词
+python crawler/site_finder.py --keywords "笑话大全,冷笑话,段子"
+
+# 指定输出文件
+python crawler/site_finder.py --output my_sites.json
+
+# 限制最多 10 个站点
+python crawler/site_finder.py --max-sites 10
+
+# 显示浏览器窗口
+python crawler/site_finder.py --no-headless
+```
+
+**参数说明：**
+
+| 参数 | 默认值 | 说明 |
+|------|--------|------|
+| `--keywords` | 笑话大全,搞笑段子,冷笑话,... | 搜索关键词，逗号分隔 |
+| `--output` | `joke_sites.json` | 输出 JSON 文件路径 |
+| `--max-sites` | 15 | 最多保留几个站点 |
+| `--no-headless` | — | 显示浏览器窗口（不加则无头模式） |
+
+**输出格式（joke_sites.json）：**
+
+```json
+[
+  {
+    "id": 1,
+    "url": "https://xiaohua.example.com",
+    "domain": "xiaohua.example.com",
+    "title": "笑话大全 - 爆笑来袭",
+    "found_at": "2026-05-31T10:30:00"
+  },
+  {
+    "id": 2,
+    "url": "https://joke.example.net",
+    "domain": "joke.example.net",
+    "title": "每日一笑",
+    "found_at": "2026-05-31T10:30:05"
+  }
+]
+```
+
+已有站点会自动保留编号，新站点追加，按域名去重。
+
+---
+
+## 二、程序 2：深度采集指定站点
+
+从 JSON 文件按编号加载站点，或直接指定 URL，深度采集该站点所有笑话并提交到 API。
+
+### 按编号采集（推荐）
+
+```bash
+# 采集 #1 站点
+python crawler/site_crawler.py --id 1
+
+# 批量采集多个站点
+python crawler/site_crawler.py --id 1,2,3
+
+# 指定站点文件
+python crawler/site_crawler.py --id 1 --sites-file my_sites.json
+```
+
+### 直接指定 URL
+
+```bash
+# 直接采集任意 URL
+python crawler/site_crawler.py --url https://xiaohua.example.com
+```
+
+python crawler/site_crawler.py --url https://www.52xiaohua.com  --no-headless
+
+### 其他选项
+
+```bash
+# 显示浏览器窗口
+python crawler/site_crawler.py --id 1 --no-headless
+
+# 指定 API 地址
+python crawler/site_crawler.py --id 1 --api-base http://192.168.1.10:8001
+
+# 指定管理员账号
+python crawler/site_crawler.py --id 1 --username admin --password admin123
+```
+
+**参数说明：**
+
+| 参数 | 默认值 | 说明 |
+|------|--------|------|
+| `--id` | — | 站点编号，多个用逗号分隔 |
+| `--url` | — | 直接指定 URL（与 --id 二选一） |
+| `--sites-file` | `joke_sites.json` | 站点列表文件路径 |
+| `--api-base` | `http://localhost:8001` | API 服务地址 |
+| `--username` | `admin` | 管理员用户名 |
+| `--password` | `admin123` | 管理员密码 |
+| `--no-headless` | — | 显示浏览器窗口 |
+
+---
+
+## 三、配合使用流程
+
+```bash
+# 1. 先启动 API
+cd api && uvicorn main:app --reload --port 8001
+
+# 2. 发现笑话站点（建议第一次用 --no-headless 观察效果）
+cd crawler && python site_finder.py --no-headless
+
+# 3. 查看生成的站点列表
+cat joke_sites.json
+
+# 4. 选择站点编号，开始深度采集
+python site_crawler.py --id 1 --no-headless
+
+# 5. 继续采集其他站点
+python site_crawler.py --id 2,3 --no-headless
+```
+
+---
+
+## 四、采集说明
+
+- **翻页采集**：自动发现首页的翻页链接（`?page=N`、`/page/N/`、`index_N.html`、`下一页` 等），逐页抓取所有笑话
+- **去重**：基于内容 MD5 哈希去重，同一笑话不会重复入库
+- **容错**：单个页面失败自动重试 2 次；同一站点连续失败 3 次自动跳过
+- **翻页限制**：每个站点最多采 30 页，避免无限抓取
+- **状态**：采集的笑话默认 `pending`（待审核），需在后台管理页面审核通过后才会在前台展示

joke_sites.json

@@ -0,0 +1,98 @@
+[
+  {
+    "url": "https://www.52xiaohua.com/xiaohua/1105.html",
+    "title": "100个经典幽默笑话大全，笑到肚子疼超级爆笑 - 爆笑笑话 ...",
+    "keyword": "笑话大全 网站",
+    "domain": "www.52xiaohua.com",
+    "id": 1,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://xiaohua.zol.com.cn/",
+    "title": "笑话大全-爆笑经典笑话 冷笑话【ZOL笑话库】",
+    "keyword": "笑话大全 网站",
+    "domain": "xiaohua.zol.com.cn",
+    "id": 2,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://www.yjbys.com/xiaohua/baoxiao/439774.html",
+    "title": "经典笑话大全爆笑100条",
+    "keyword": "笑话大全 网站",
+    "domain": "www.yjbys.com",
+    "id": 3,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://zhilezhi.com/xiaohua",
+    "title": "笑话大全 - 幽默笑话和搞笑段子网站 | 知乐知",
+    "keyword": "笑话大全 网站",
+    "domain": "zhilezhi.com",
+    "id": 4,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://xiaohua.hao86.com/",
+    "title": "笑话大全爆笑简短_史上最强笑话十则2026",
+    "keyword": "笑话大全 网站",
+    "domain": "xiaohua.hao86.com",
+    "id": 5,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://www.douyin.com/video/7457017269722402067",
+    "title": "搞笑视频合集。 每日一笑，烦恼忘掉。#搞笑 #搞笑视频 #看 ...",
+    "keyword": "搞笑段子 网站",
+    "domain": "www.douyin.com",
+    "id": 6,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://www.iqiyi.com/v_fhial680gw.html",
+    "title": "搞笑沙雕短片合集恶搞整蛊让你乐不停-爱奇艺",
+    "keyword": "搞笑段子 网站",
+    "domain": "www.iqiyi.com",
+    "id": 7,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://www.qiushidabaike.com/",
+    "title": "笑话大全_恶搞囧事 - 糗事大百科",
+    "keyword": "搞笑段子 网站",
+    "domain": "www.qiushidabaike.com",
+    "id": 8,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://doc.quark.cn/preview/tongyongshenghuo-shejiaoyule-lengxiaohua%2Fxiaohua/C247E8B69B4BFB30FA3A859CC1A0C94B",
+    "title": "100个经典冷笑话大全爆笑",
+    "keyword": "冷笑话 大全",
+    "domain": "doc.quark.cn",
+    "id": 9,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://www.meiwengushi.com/100gejingdianlengxiaohuadaquan",
+    "title": "100个经典冷笑话大全",
+    "keyword": "冷笑话 大全",
+    "domain": "www.meiwengushi.com",
+    "id": 10,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://www.thepaper.cn/newsDetail_forward_17848009",
+    "title": "20个幽默冷笑话，个个爆笑不断，记得分享给身边朋友们",
+    "keyword": "幽默笑话 网站",
+    "domain": "www.thepaper.cn",
+    "id": 11,
+    "found_at": "2026-05-31T10:17:46"
+  },
+  {
+    "url": "https://baike.sogou.com/v65352.htm",
+    "title": "幽默 _ 百科",
+    "keyword": "幽默笑话 网站",
+    "domain": "baike.sogou.com",
+    "id": 12,
+    "found_at": "2026-05-31T10:17:46"
+  }
+]

optimizer/main.py

@@ -0,0 +1,100 @@
+"""
+笑话优化 CLI 工具。
+循环检测所有笑话 → AI 质量检测 → AI 润色 → AI 评价分类。
+
+用法:
+  python optimizer/main.py                           # 处理所有 pending/approved 笑话
+  python optimizer/main.py --status rejected          # 只处理已拒绝的
+  python optimizer/main.py --limit 10                 # 只处理前 10 条
+  python optimizer/main.py --id 1,2,3                 # 指定 ID 处理
+  python optimizer/main.py --api-url http://localhost:8001
+"""
+
+import argparse
+import os
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from optimizer.optimizer import Optimizer
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="笑话优化工具 — 质量检测 / AI 润色 / 评价分类")
+
+    parser.add_argument(
+        "--api-url",
+        type=str,
+        default=os.getenv("API_URL", "http://localhost:8001"),
+        help="API 地址（默认: http://localhost:8001）",
+    )
+    parser.add_argument(
+        "--username",
+        type=str,
+        default=os.getenv("API_USERNAME", "admin"),
+        help="管理员用户名（默认: admin）",
+    )
+    parser.add_argument(
+        "--password",
+        type=str,
+        default=os.getenv("API_PASSWORD", "admin123"),
+        help="管理员密码（默认: admin123）",
+    )
+    parser.add_argument(
+        "--status",
+        type=str,
+        default=None,
+        choices=["pending", "approved", "rejected"],
+        help="按状态筛选（默认: 全部）",
+    )
+    parser.add_argument(
+        "--limit",
+        type=int,
+        default=None,
+        help="最多处理 N 条（默认: 全部）",
+    )
+    parser.add_argument(
+        "--id",
+        type=str,
+        default=None,
+        dest="ids",
+        help="指定笑话 ID，逗号分隔（如: 1,2,3）",
+    )
+
+    return parser.parse_args()
+
+
+def main():
+    args = parse_args()
+
+    ids = None
+    if args.ids:
+        ids = [int(x.strip()) for x in args.ids.split(",") if x.strip()]
+
+    print(f"=" * 50)
+    print(f"笑话优化工具")
+    print(f"API 地址: {args.api_url}")
+    print(f"筛选状态: {args.status or '全部'}")
+    if args.limit:
+        print(f"处理上限: {args.limit} 条")
+    if ids:
+        print(f"指定 ID: {ids}")
+    print(f"=" * 50)
+
+    optimizer = Optimizer(
+        api_base=args.api_url,
+        username=args.username,
+        password=args.password,
+    )
+
+    try:
+        optimizer.run(status=args.status, limit=args.limit, ids=ids)
+    except KeyboardInterrupt:
+        print("\n用户中断")
+    except Exception as e:
+        print(f"\n[!] 运行错误: {e}")
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()

optimizer/optimizer.py

@@ -0,0 +1,341 @@
+"""笑话优化核心逻辑：质量检测 → AI 润色 → 评价分类。"""
+
+import hashlib
+import json
+import sys
+import time
+
+import httpx
+from openai import OpenAI
+
+from optimizer.prompts import (
+    QUALITY_CHECK_SYSTEM_PROMPT,
+    QUALITY_CHECK_USER_PROMPT,
+    POLISH_SYSTEM_PROMPT,
+    POLISH_USER_PROMPT,
+    EVALUATE_SYSTEM_PROMPT,
+    EVALUATE_USER_PROMPT,
+)
+
+
+class Optimizer:
+    def __init__(self, api_base: str, username: str, password: str):
+        self.api_base = api_base.rstrip("/")
+        self.username = username
+        self.password = password
+        self.token = None
+        self.ai_client = None
+        self.model_name = ""
+        self.types = []
+        self.crowds = []
+        # 统计
+        self.stats = {"checked": 0, "rejected": 0, "polished": 0, "evaluated": 0, "skipped": 0}
+
+    # === API 认证 ===
+    def _login(self) -> str:
+        resp = httpx.post(
+            f"{self.api_base}/api/auth/login",
+            json={"username": self.username, "password": self.password},
+            timeout=30,
+        )
+        resp.raise_for_status()
+        return resp.json()["access_token"]
+
+    def _get(self, path: str) -> dict | list:
+        resp = httpx.get(
+            f"{self.api_base}{path}",
+            headers={"Authorization": f"Bearer {self.token}"},
+            timeout=30,
+        )
+        resp.raise_for_status()
+        return resp.json()
+
+    def _put(self, path: str, data: dict) -> dict:
+        resp = httpx.put(
+            f"{self.api_base}{path}",
+            json=data,
+            headers={"Authorization": f"Bearer {self.token}"},
+            timeout=30,
+        )
+        resp.raise_for_status()
+        return resp.json()
+
+    # === 初始化 ===
+    def setup(self):
+        print("[*] 正在登录...")
+        self.token = self._login()
+        print("[*] 登录成功")
+
+        ai_config = self._get("/api/admin/settings/active")
+        self.ai_client = OpenAI(
+            base_url=ai_config["api_base"],
+            api_key=ai_config["api_key"],
+        )
+        self.model_name = ai_config["model_name"]
+        print(f"[*] AI 模型: {self.model_name}")
+
+        self.types = self._get("/api/categories/types")
+        self.crowds = self._get("/api/categories/crowds")
+        print(f"[*] 分类: {len(self.types)} 种类型, {len(self.crowds)} 种人群")
+
+    # === 读取笑话 ===
+    def get_jokes(self, status: str | None = None, limit: int | None = None,
+                  ids: list[int] | None = None) -> list[dict]:
+        """从 API 分页读取笑话"""
+        if ids:
+            jokes = []
+            for jid in ids:
+                try:
+                    j = self._get(f"/api/admin/jokes/{jid}")
+                    jokes.append(j)
+                except Exception as e:
+                    print(f"  [!] 获取笑话 #{jid} 失败: {e}")
+            return jokes
+
+        page = 1
+        page_size = 100
+        all_jokes = []
+
+        while True:
+            try:
+                path = f"/api/admin/jokes?page={page}&page_size={page_size}"
+                if status:
+                    path += f"&status={status}"
+                data = self._get(path)
+                items = data.get("items", [])
+                if not items:
+                    break
+                all_jokes.extend(items)
+                print(f"  [*] 已读取 {len(all_jokes)} 条...")
+                if limit and len(all_jokes) >= limit:
+                    all_jokes = all_jokes[:limit]
+                    break
+                page += 1
+            except Exception as e:
+                print(f"  [!] 分页读取失败 (page={page}): {e}")
+                break
+
+        return all_jokes
+
+    # === Stage 1: 质量检测 ===
+    def quality_check(self, content: str) -> dict:
+        """判断笑话是否有笑点，返回 {"has_punchline": bool, "reason": str}"""
+        resp = self.ai_client.chat.completions.create(
+            model=self.model_name,
+            messages=[
+                {"role": "system", "content": QUALITY_CHECK_SYSTEM_PROMPT},
+                {"role": "user", "content": QUALITY_CHECK_USER_PROMPT.format(content=content[:2000])},
+            ],
+            temperature=0.3,
+            max_tokens=200,
+        )
+        raw = resp.choices[0].message.content.strip()
+        return self._parse_json(raw, {"has_punchline": True, "reason": ""})
+
+    # === Stage 2: AI 润色 ===
+    def polish(self, content: str) -> str:
+        """润色笑话内容"""
+        resp = self.ai_client.chat.completions.create(
+            model=self.model_name,
+            messages=[
+                {"role": "system", "content": POLISH_SYSTEM_PROMPT},
+                {"role": "user", "content": POLISH_USER_PROMPT.format(content=content)},
+            ],
+            temperature=0.8,
+            max_tokens=1024,
+        )
+        return resp.choices[0].message.content.strip()
+
+    # === Stage 3: 评价分类 ===
+    def evaluate(self, content: str) -> dict:
+        """评价并分类，返回 {"types": [...], "crowds": [...], "score": int, "comment": str}"""
+        type_names = [t.get("name", "") for t in self.types]
+        crowd_names = [c.get("name", "") for c in self.crowds]
+
+        resp = self.ai_client.chat.completions.create(
+            model=self.model_name,
+            messages=[
+                {"role": "system", "content": EVALUATE_SYSTEM_PROMPT},
+                {"role": "user", "content": EVALUATE_USER_PROMPT.format(
+                    content=content[:2000],
+                    known_types=", ".join(type_names),
+                    known_crowds=", ".join(crowd_names),
+                )},
+            ],
+            temperature=0.3,
+            max_tokens=300,
+        )
+        raw = resp.choices[0].message.content.strip()
+        result = self._parse_json(raw, {"types": [], "crowds": [], "score": 5, "comment": ""})
+        # Backward compatibility: if LLM returns old single format, convert to array
+        if isinstance(result.get("types"), str):
+            result["types"] = [result["types"]] if result["types"] else []
+        if isinstance(result.get("crowds"), str):
+            result["crowds"] = [result["crowds"]] if result["crowds"] else []
+        return result
+
+    # === 辅助方法 ===
+    def _parse_json(self, raw: str, default: dict) -> dict:
+        """安全解析 LLM 返回的 JSON"""
+        try:
+            data = json.loads(raw)
+            if isinstance(data, dict):
+                return data
+            return default
+        except json.JSONDecodeError:
+            if "```json" in raw:
+                raw = raw.split("```json")[1].split("```")[0]
+            elif "```" in raw:
+                raw = raw.split("```")[1].split("```")[0]
+            try:
+                return json.loads(raw.strip())
+            except Exception:
+                return default
+
+    def _get_type_id(self, type_name: str) -> int | None:
+        for t in self.types:
+            if t.get("name") == type_name:
+                return t.get("id")
+        return None
+
+    def _get_crowd_id(self, crowd_name: str) -> int | None:
+        for c in self.crowds:
+            if c.get("name") == crowd_name:
+                return c.get("id")
+        return None
+
+    # === 单条笑话处理（3 阶段） ===
+    def process_joke(self, joke: dict) -> bool:
+        """处理单条笑话：质量检测 → 润色 → 评价分类，返回是否成功"""
+        joke_id = joke.get("id")
+        title = joke.get("title", "")
+        content = joke.get("content", "")
+
+        if not content:
+            print(f"  [!] #{joke_id} 内容为空，跳过")
+            self.stats["skipped"] += 1
+            return False
+
+        print(f"\n  {'='*40}")
+        print(f"  处理 #{joke_id}: {title[:30]}")
+        print(f"  {'='*40}")
+
+        # Stage 1: 质量检测
+        print(f"  [1/3] 质量检测...")
+        try:
+            check = self.quality_check(content)
+            if not check.get("has_punchline", True):
+                reason = check.get("reason", "无笑点")
+                print(f"  [!] 无笑点: {reason}")
+                # 标记为 rejected
+                self._put(f"/api/admin/jokes/{joke_id}", {"status": "rejected"})
+                self.stats["rejected"] += 1
+                self.stats["checked"] += 1
+                return True  # 处理完成（已拒绝）
+            print(f"  [OK] 有笑点: {check.get('reason', '')}")
+        except Exception as e:
+            print(f"  [!] 质量检测失败: {e}，跳过本条")
+            self.stats["skipped"] += 1
+            return False
+
+        self.stats["checked"] += 1
+
+        # Stage 2: AI 润色
+        print(f"  [2/3] AI 润色...")
+        try:
+            polished = self.polish(content)
+            if polished and polished != content:
+                print(f"  [OK] 润色完成 ({len(content)} -> {len(polished)} 字)")
+            else:
+                print(f"  [*] 润色后无变化")
+        except Exception as e:
+            print(f"  [!] 润色失败: {e}")
+            polished = content  # 润色失败时使用原文
+
+        # Stage 3: 评价分类
+        print(f"  [3/3] 评价分类...")
+        try:
+            eval_result = self.evaluate(polished)
+            type_names = eval_result.get("types", [])
+            crowd_names = eval_result.get("crowds", [])
+            score = eval_result.get("score", 5)
+            print(f"  [OK] 类型={type_names}, 人群={crowd_names}, 评分={score}/10")
+        except Exception as e:
+            print(f"  [!] 评价分类失败: {e}")
+            type_names = []
+            crowd_names = []
+            score = None
+
+        # 保存更新
+        try:
+            update = {
+                "polished_content": polished,
+                "status": "approved" if (score or 5) >= 4 else "pending",
+            }
+            if type_names:
+                update["type_ids"] = []
+                for n in type_names:
+                    tid = self._get_type_id(n)
+                    if tid:
+                        update["type_ids"].append(tid)
+                if not update["type_ids"]:
+                    del update["type_ids"]
+            if crowd_names:
+                update["crowd_ids"] = []
+                for n in crowd_names:
+                    cid = self._get_crowd_id(n)
+                    if cid:
+                        update["crowd_ids"].append(cid)
+                if not update["crowd_ids"]:
+                    del update["crowd_ids"]
+
+            self._put(f"/api/admin/jokes/{joke_id}", update)
+            self.stats["polished"] += 1
+            self.stats["evaluated"] += 1
+            print(f"  [OK] 更新成功")
+            return True
+        except Exception as e:
+            print(f"  [!] 更新失败: {e}")
+            return False
+
+    # === 主循环 ===
+    def run(self, status: str | None = None, limit: int | None = None,
+            ids: list[int] | None = None):
+        """主入口：读取笑话并逐个处理"""
+        print(f"\n>> 笑话优化模式启动")
+        print(f"    筛选状态: {status or '全部'}")
+        if limit:
+            print(f"    处理数量: {limit}")
+        if ids:
+            print(f"    指定 ID: {ids}")
+
+        self.setup()
+
+        jokes = self.get_jokes(status, limit, ids)
+        print(f"\n[*] 共读取 {len(jokes)} 条笑话，开始处理")
+
+        for idx, joke in enumerate(jokes):
+            print(f"\n  --- 进度 {idx+1}/{len(jokes)} ---")
+            try:
+                self.process_joke(joke)
+            except KeyboardInterrupt:
+                print("\n用户中断")
+                break
+            except Exception as e:
+                print(f"  [!] 处理异常: {e}")
+                self.stats["skipped"] += 1
+
+            # 每条间稍等，避免 API 限流
+            if idx < len(jokes) - 1:
+                time.sleep(1)
+
+        # 输出统计
+        print(f"\n{'='*40}")
+        print(f"  处理完成")
+        print(f"  {'='*40}")
+        print(f"  检查: {self.stats['checked']} 条")
+        print(f"  拒绝(无笑点): {self.stats['rejected']} 条")
+        print(f"  润色: {self.stats['polished']} 条")
+        print(f"  评价分类: {self.stats['evaluated']} 条")
+        print(f"  跳过(失败): {self.stats['skipped']} 条")
+        print(f"{'='*40}")

optimizer/prompts.py

@@ -0,0 +1,58 @@
+"""AI 提示词模板 — 笑话质量检测、润色、评价分类。"""
+
+# ===== Stage 1: 质量检测 =====
+QUALITY_CHECK_SYSTEM_PROMPT = """你是一个幽默内容审核专家。判断以下内容是否是一个合格的笑话/段子。
+
+合格标准（满足任一即可）：
+1. 有明确的笑点或反转（punchline）
+2. 有幽默的语言表达或双关
+3. 有意外结局或情理之中意料之外
+
+不合格标准（符合任一即判定不合格）：
+1. 纯粹的事实陈述，没有任何幽默元素
+2. 只是对话片段，没有笑点
+3. 普通故事或叙事，没有幽默设计
+4. 说教或道理阐述
+5. 内容不完整或难以理解
+
+始终返回 JSON 格式：{"has_punchline": true/false, "reason": "简要说明判断理由"}"""
+
+QUALITY_CHECK_USER_PROMPT = """请判断以下内容是否为合格笑话：
+
+{content}
+
+返回 JSON 格式。"""
+
+# ===== Stage 2: AI 润色 =====
+POLISH_SYSTEM_PROMPT = """你是一个专业的幽默文案编辑。请润色以下笑话，要求：
+1. 保持核心笑点不变
+2. 优化语言表达，使其更通顺、更精炼
+3. 增强节奏感和幽默效果，但不改变原意
+4. 字数控制在原内容的 80%-120%
+5. 不要添加额外解释或评论
+6. 直接输出润色后的内容，不要加任何前缀"""
+
+POLISH_USER_PROMPT = """请润色以下笑话：
+
+{content}
+
+只输出润色后的笑话内容。"""
+
+# ===== Stage 3: 评价分类 =====
+EVALUATE_SYSTEM_PROMPT = """你是一个笑话分类和评价专家。对给定的笑话进行分析，返回 JSON 格式的分类和评分结果。
+
+要求：
+1. types: 从提供的类型列表中选择所有匹配的类型名称（数组，可以选多个）
+2. crowds: 从提供的人群列表中选择所有匹配的人群名称（数组，可以选多个）
+3. score: 1-10 分，基于幽默程度、创意和表达效果
+4. comment: 简短评语（10字以内）
+
+始终返回 JSON 格式。"""
+
+EVALUATE_USER_PROMPT = """笑话内容：
+{content}
+
+可选类型：{known_types}
+可选人群：{known_crowds}
+
+返回 JSON 格式：{{"types": ["类型1", "类型2"], "crowds": ["人群1", "人群2"], "score": 8, "comment": "简短评语"}}"""