Revert "feat(extension): 多模态视频理解开关 + 抽帧/拼图参数（对齐 web NoteForm）"

feat(extension): 多模态视频理解开关 + 抽帧/拼图参数（对齐 web NoteForm）

BillNote_extension/src/background/main.ts

@@ -87,9 +87,6 @@ async function startTask(url: string): Promise<{ ok: boolean, taskId?: string, e
         link: formats.includes('link'),
         style: settings.style || undefined,
         extras: settings.extras || undefined,
-        video_understanding: settings.video_understanding || undefined,
-        video_interval: settings.video_understanding ? settings.video_interval : undefined,
-        grid_size: settings.video_understanding ? settings.grid_size : undefined,
         prefetched_transcript: prefetched ?? undefined,
       }),
     })

BillNote_extension/src/logic/constants.ts

@@ -12,9 +12,6 @@ export const DEFAULT_SETTINGS: Settings = {
   link: false,
   style: 'minimal',
   extras: '',
-  video_understanding: false,
-  video_interval: 6,
-  grid_size: [2, 2],
 }
 
 export const MAX_TASKS = 30

BillNote_extension/src/logic/types.ts

@@ -40,9 +40,6 @@ export interface GenerateRequest {
   format?: string[]
   style?: string
   extras?: string
-  video_understanding?: boolean
-  video_interval?: number
-  grid_size?: [number, number]
   // 客户端在浏览器里直接抓到的字幕，跳过后端的 download_subtitles + 音频转写
   prefetched_transcript?: {
     language: string
@@ -120,13 +117,6 @@ export interface Settings {
   link: boolean
   style: NoteStyle
   extras: string
-  // 多模态视频理解：抽帧拼图喂给视觉模型，提升画面相关问题的回答质量
-  // 要求所选 model 是视觉模型（如 gpt-4o / gemini / claude-opus 系列），文字模型会忽略图片
-  video_understanding: boolean
-  // 抽帧间隔（秒），范围 1-30，默认 6
-  video_interval: number
-  // 拼图网格 [rows, cols]，每张拼图最多 rows*cols 帧。默认 [2,2]
-  grid_size: [number, number]
 }
 
 export interface ProviderUpdatePayload {

BillNote_extension/src/options/pages/General.vue

@@ -165,39 +165,5 @@ onMounted(async () => {
         />
       </label>
     </section>
-
-    <section class="section-card">
-      <h2 class="font-semibold">视频理解（多模态）</h2>
-      <p class="text-xs text-gray-500">
-        启用后会按抽帧间隔截取视频帧拼成网格图，连同字幕一起喂给视觉模型，提升画面相关问题的回答质量。
-        <strong class="text-amber-700">需要选择视觉模型</strong>（GPT-4o / Gemini / Claude 等），文字模型会忽略图片。
-      </p>
-      <label class="flex items-center gap-2 text-sm">
-        <input v-model="settings.video_understanding" type="checkbox">
-        启用视频理解
-      </label>
-      <div v-if="settings.video_understanding" class="grid grid-cols-3 gap-3 text-sm">
-        <label class="flex flex-col gap-1">
-          <span class="text-gray-600">抽帧间隔(秒, 1-30)</span>
-          <input v-model.number="settings.video_interval" type="number" min="1" max="30" class="input">
-        </label>
-        <label class="flex flex-col gap-1">
-          <span class="text-gray-600">拼图行 (1-10)</span>
-          <input
-            :value="settings.grid_size?.[0] ?? 2"
-            type="number" min="1" max="10" class="input"
-            @input="settings.grid_size = [Number(($event.target as HTMLInputElement).value) || 2, settings.grid_size?.[1] ?? 2]"
-          >
-        </label>
-        <label class="flex flex-col gap-1">
-          <span class="text-gray-600">拼图列 (1-10)</span>
-          <input
-            :value="settings.grid_size?.[1] ?? 2"
-            type="number" min="1" max="10" class="input"
-            @input="settings.grid_size = [settings.grid_size?.[0] ?? 2, Number(($event.target as HTMLInputElement).value) || 2]"
-          >
-        </label>
-      </div>
-    </section>
   </div>
 </template>

BillNote_extension/src/popup/Popup.vue

@@ -80,9 +80,6 @@ async function start() {
       link: formats.includes('link'),
       style: settings.value.style || undefined,
       extras: settings.value.extras || undefined,
-      video_understanding: settings.value.video_understanding || undefined,
-      video_interval: settings.value.video_understanding ? settings.value.video_interval : undefined,
-      grid_size: settings.value.video_understanding ? settings.value.grid_size : undefined,
       prefetched_transcript: prefetched ?? undefined,
     })
     activeTaskId.value = task_id
@@ -228,41 +225,6 @@ onUnmounted(() => {
             placeholder="例如：重点关注游戏开发部分；保留所有专业术语原文"
           />
         </label>
-        <label class="flex items-center gap-2 mt-2">
-          <input v-model="settings.video_understanding" type="checkbox">
-          <span class="text-gray-600">启用视频理解（抽帧拼图喂视觉模型）</span>
-        </label>
-        <div v-if="settings.video_understanding" class="grid grid-cols-3 gap-2 mt-2">
-          <label class="flex flex-col gap-1">
-            <span class="text-gray-600">抽帧间隔(秒)</span>
-            <input
-              v-model.number="settings.video_interval"
-              type="number" min="1" max="30"
-              class="border rounded px-1 py-0.5"
-            >
-          </label>
-          <label class="flex flex-col gap-1">
-            <span class="text-gray-600">拼图行</span>
-            <input
-              :value="settings.grid_size?.[0] ?? 2"
-              type="number" min="1" max="10"
-              class="border rounded px-1 py-0.5"
-              @input="settings.grid_size = [Number(($event.target as HTMLInputElement).value) || 2, settings.grid_size?.[1] ?? 2]"
-            >
-          </label>
-          <label class="flex flex-col gap-1">
-            <span class="text-gray-600">拼图列</span>
-            <input
-              :value="settings.grid_size?.[1] ?? 2"
-              type="number" min="1" max="10"
-              class="border rounded px-1 py-0.5"
-              @input="settings.grid_size = [settings.grid_size?.[0] ?? 2, Number(($event.target as HTMLInputElement).value) || 2]"
-            >
-          </label>
-        </div>
-        <p v-if="settings.video_understanding" class="text-amber-700 mt-1">
-          ⚠ 需要选择视觉模型（GPT-4o / Gemini / Claude 等），文字模型会忽略图片
-        </p>
       </details>
 
       <div class="text-xs text-gray-600">