dsh-bili-watch 0.2.26 → 0.2.28

package/index.js

@@ -433,7 +433,11 @@ function pickFolderDialog() {
  * 缓存自动清理：先按天数删过期文件，再按总量从最旧删到阈值内。
  * 返回被删统计，供面板显示。
  */
-function autoCleanCache() {
+let lastCleanAt = 0
+
+function autoCleanCache(opts = {}) {
+  // T31-P2-6：面板每 8 秒轮询 /status，不许每次都全量 walk 缓存目录
+  if (!opts.force && Date.now() - lastCleanAt < 10 * 60 * 1000) return { removed: 0, bytes: 0, skipped: true }
   if (!existsSync(CACHE)) return { removed: 0, bytes: 0 }
   const { days, maxMb } = cacheCfg()
   const cutoff = Date.now() - days * 86400000
@@ -466,6 +470,7 @@ function autoCleanCache() {
       if (total <= limit) break
     }
   }
+  lastCleanAt = Date.now()
   return { removed, bytes }
 }
 
@@ -756,7 +761,7 @@ async function lookup({ command, bv, extra = [], badOut = false }) {
 function apply(ctx) {
   // 启动时按「天数 + 总量阈值」清一次缓存，不点按钮也不会无限涨
   try {
-    const r = autoCleanCache()
+    const r = autoCleanCache({ force: true })
     if (r.removed) ctx.logger?.info?.(`[dsh-bili-watch] 自动清理缓存 ${r.removed} 个文件 / ${(r.bytes / 1048576).toFixed(1)} MB`)
   } catch {}
 
@@ -780,6 +785,7 @@ function apply(ctx) {
   try {
     if (createUserMessage && uiCfg().autoHint !== false) {
       const hinted = new Map() // message.id → true，同一条消息只提示一次
+      const HINT_CAP = 500 // T31-P2-5：常驻 Map 不能无限涨，满了整体清空（只影响重复提示）
       const hint = (msg) => {
         const { ok, text } = textOfUserMessage(msg)
         if (!ok) return null
@@ -806,6 +812,7 @@ function apply(ctx) {
             const extra = hint(msg)
             if (!extra) continue
             hinted.set(msg.id, true)
+            if (hinted.size > HINT_CAP) hinted.clear()
             fresh.push(extra)
           }
           if (!fresh.length) return decision

package/package.json

@@ -1,6 +1,6 @@
 {
     "name":  "dsh-bili-watch",
-    "version":  "0.2.26",
+    "version":  "0.2.28",
     "private":  false,
     "description":  "把 B 站视频读成文本给 DSH agent 用：元数据、评论、弹幕、字幕、本地转写、抽帧 OCR；支持 b23.tv 短链与 av 号，抽帧缓存目录可换。只读命令零依赖。",
     "type":  "module",

package/tools/look.mjs

@@ -93,7 +93,17 @@ const sleep = (ms) => new Promise((r) => setTimeout(r, ms))
 const jitter = (base, spread) => base + Math.floor(Math.random() * spread)
 
 async function jget(url, headers = H, timeoutMs = 30000) {
-  const r = await fetch(url, { headers, signal: AbortSignal.timeout(timeoutMs) })
+  // T36-P1-2：只对网络层抖动重试一次（B 站偶发 ECONNRESET）；业务 code（-352/-404）不重试
+  let r
+  for (let attempt = 0; ; attempt++) {
+    try {
+      r = await fetch(url, { headers, signal: AbortSignal.timeout(timeoutMs) })
+      if (!(r.status >= 500) || attempt >= 1) break
+    } catch (e) {
+      if (attempt >= 1) throw e
+    }
+    await sleep(300 + Math.floor(Math.random() * 400))
+  }
   const t = await r.text()
   if (!r.ok) return { code: -r.status, message: `HTTP ${r.status}` }
   try {
@@ -1366,6 +1376,18 @@ async function cmdSheet(bv, opts = {}) {
   mkdirSync(dirname(out), { recursive: true })
   execFileSync(bin, ['-y', '-threads', '2', '-i', join(tmp, '%02d.jpg'), '-vf', `scale=640:-1,tile=${cols}x${rows}`, '-frames:v', '1', out], { stdio: 'ignore', timeout: 300000 })
   console.log(`拼图: ${out}（${(statSync(out).size / 1024).toFixed(0)} KiB，共 ${picked.length} 帧）`)
+  // T42：九宫格没有时间轴，模型看图不知道「第几格是第几秒」；写一份 manifest 并打一行
+  const stepUsed = Number((/-([0-9]+)s$/.exec(r.dir) || [])[1] || opts.step || 0)
+  const cellSec = (f, i) => {
+    const m = /-([0-9]+)\.jpg$/.exec(f)
+    const n = m ? Number(m[1]) : i + 1
+    return stepUsed > 0 ? Math.max(0, (n - 1) * stepUsed) : null
+  }
+  const cells = picked.map((f, i) => { const sec = cellSec(f, i); return { cell: i + 1, file: basename(f), sec, mmss: sec == null ? '?' : fmtSec(sec) } })
+  try {
+    writeFileSync(out.replace(/\.jpg$/, '.manifest.json'), JSON.stringify({ bv, grid, step: stepUsed || null, cells }, null, 2) + '\n', 'utf8')
+  } catch {}
+  console.log('每格时间：' + cells.map((x) => `#${x.cell}=${x.mmss}`).join('  '))
   console.log('读这张图就能过一遍画面；要精确读字用 ocr <BV>。')
 }
 
@@ -1781,7 +1803,7 @@ async function cmdText(bv, credPath, opts = {}) {
       // 口播类正常 4–6 字/秒；低于 2 字/秒多半是背景音乐、外语歌或纯画面，
       // 这时候转写出来的「字」只是把杂音听成了字，继续往下走 OCR 更靠谱。
       const rate = v.dur > 0 ? a.chars / v.dur : 0
-      if (rate >= 2) {
+      if (rate >= 2 && !(a.verify && a.verify.suspect)) {
         // T40：ASR 也会把杂音/背景音听成字，主入口同样要过诱饵信号（原先只有字幕分支有）
         const asrBait = baitSignals(a.text, v, credibility(v, a.text))
         if (asrBait.suspect) {
@@ -1800,6 +1822,8 @@ async function cmdText(bv, credPath, opts = {}) {
         printText(a.text, opts, bv)
         done()
         return
+      } else if (a.verify && a.verify.suspect) {
+        say(`[track] 转写: ⚠ 疑似幻觉（${(a.verify.reasons || []).join('；')}），降级到抽帧/OCR`)
       }
       say(`[track] 转写: 质量低（${a.chars} 字，${rate.toFixed(1)} 字/秒），继续往下试`)
     } else {