feat: add smart extraction trigger mvp

README.md

@@ -45,7 +45,6 @@ ST-BME（ST-Bionic-Memory-Ecology）是一个运行在 SillyTavern 第三方扩
 
 以下方向在现有代码中仅有预留开关、设计痕迹或路线图描述，尚不应视为完整能力：
 
-- **惊奇度分割 / 智能触发提取**
 - **完整端到端测试与稳定性验证**
 - **图谱可视化面板**
 - **自定义 Schema 编辑体验增强**
@@ -293,7 +292,7 @@ SillyTavern/public/scripts/extensions/third-party/ST-BME/
 ## 路线图
 
 - [ ] 补充端到端测试与典型场景回归样例
-- [ ] 完善惊奇度分割与智能触发提取
+- [x] 完善惊奇度分割与智能触发提取（轻量 MVP）
 - [ ] 完成反思条目生成与注入策略
 - [ ] 提供图谱可视化界面
 - [ ] 增强 Schema 配置与编辑体验

index.js

@@ -94,6 +94,7 @@ const defaultSettings = {
   // ① 惊奇度分割（P2）
   enableSmartTrigger: false, // 启用惊奇度分割
   triggerPatterns: "", // 自定义触发正则
+  smartTriggerThreshold: 2, // 轻量触发阈值
 
   // ⑤ 主动遗忘（P2）
   enableSleepCycle: false, // 启用主动遗忘
@@ -168,6 +169,103 @@ function saveGraphToChat() {
 
 // ==================== 核心流程 ====================
 
+const DEFAULT_TRIGGER_KEYWORDS = [
+  "突然",
+  "没想到",
+  "原来",
+  "其实",
+  "发现",
+  "背叛",
+  "死亡",
+  "复活",
+  "恢复记忆",
+  "失忆",
+  "告白",
+  "暴露",
+  "秘密",
+  "计划",
+  "规则",
+  "契约",
+  "位置",
+  "地点",
+  "离开",
+  "来到",
+];
+
+function getSmartTriggerDecision(chat, lastProcessed, settings) {
+  const pendingMessages = chat
+    .slice(lastProcessed + 1)
+    .filter((msg) => !msg.is_system)
+    .map((msg) => ({
+      role: msg.is_user ? "user" : "assistant",
+      content: msg.mes || "",
+    }));
+
+  if (pendingMessages.length === 0) {
+    return { triggered: false, score: 0, reasons: [] };
+  }
+
+  const reasons = [];
+  let score = 0;
+  const combinedText = pendingMessages.map((m) => m.content).join("\n");
+
+  const keywordHits = DEFAULT_TRIGGER_KEYWORDS.filter((keyword) =>
+    combinedText.includes(keyword),
+  );
+  if (keywordHits.length > 0) {
+    score += Math.min(2, keywordHits.length);
+    reasons.push(`关键词: ${keywordHits.slice(0, 3).join(", ")}`);
+  }
+
+  const customPatterns = String(settings.triggerPatterns || "")
+    .split(/\r?\n|,/)
+    .map((s) => s.trim())
+    .filter(Boolean);
+  for (const pattern of customPatterns) {
+    try {
+      const regex = new RegExp(pattern, "i");
+      if (regex.test(combinedText)) {
+        score += 2;
+        reasons.push(`自定义触发: ${pattern}`);
+        break;
+      }
+    } catch {
+      // 忽略无效正则，避免影响主流程
+    }
+  }
+
+  const roleSwitchCount = pendingMessages.reduce((count, message, index) => {
+    if (index === 0) return count;
+    return count + (message.role !== pendingMessages[index - 1].role ? 1 : 0);
+  }, 0);
+  if (roleSwitchCount >= 2) {
+    score += 1;
+    reasons.push("多轮往返互动");
+  }
+
+  const punctuationHits = (combinedText.match(/[!?！？]/g) || []).length;
+  if (punctuationHits >= 2) {
+    score += 1;
+    reasons.push("情绪/冲突波动");
+  }
+
+  const entityLikeHits =
+    combinedText.match(
+      /[A-Z][a-z]{2,}|[\u4e00-\u9fff]{2,6}(先生|小姐|王国|城|镇|村|学院|组织|公司|小队|军团)/g,
+    ) || [];
+  if (entityLikeHits.length > 0) {
+    score += 1;
+    reasons.push("疑似新实体/新地点");
+  }
+
+  const threshold = Math.max(1, settings.smartTriggerThreshold || 2);
+  return {
+    triggered: score >= threshold,
+    score,
+    reasons,
+  };
+}
+
 /**
  * 提取管线：处理未提取的对话楼层
  */
@@ -194,8 +292,17 @@ async function runExtraction() {
 
   if (unprocessedStarts.length === 0) return;
 
-  // 按 extractEvery 批次处理
+  const smartTriggerDecision = settings.enableSmartTrigger
-  if (unprocessedStarts.length < settings.extractEvery) return;
+    ? getSmartTriggerDecision(chat, lastProcessed, settings)
+    : { triggered: false, score: 0, reasons: [] };
+
+  // 按 extractEvery 批次处理；若启用智能触发，则允许提前提取
+  if (
+    unprocessedStarts.length < settings.extractEvery &&
+    !smartTriggerDecision.triggered
+  ) {
+    return;
+  }
 
   isExtracting = true;
 
@@ -219,7 +326,12 @@ async function runExtraction() {
       });
     }
 
-    console.log(`[ST-BME] 开始提取: 楼层 ${startIdx}-${endIdx}`);
+    console.log(
+      `[ST-BME] 开始提取: 楼层 ${startIdx}-${endIdx}` +
+        (smartTriggerDecision.triggered
+          ? ` [智能触发 score=${smartTriggerDecision.score}; ${smartTriggerDecision.reasons.join(" / ")}]`
+          : ""),
+    );
 
     const result = await extractMemories({
       graph: currentGraph,
@@ -697,6 +809,12 @@ function bindSettingsUI() {
       settings.enableSmartTrigger = $(this).prop("checked");
       saveSettingsDebounced();
     });
+  $("#st_bme_trigger_patterns")
+    .val(settings.triggerPatterns || "")
+    .on("input", function () {
+      settings.triggerPatterns = $(this).val();
+      saveSettingsDebounced();
+    });
 
   // P2: 主动遗忘
   $("#st_bme_sleep_cycle")

settings.html

@@ -2,10 +2,11 @@
   <div class="inline-drawer">
     <div class="inline-drawer-toggle inline-drawer-header">
       <b>ST-BME 图谱记忆</b>
-            <div class="inline-drawer-icon fa-solid fa-circle-chevron-down down"></div>
+      <div
+        class="inline-drawer-icon fa-solid fa-circle-chevron-down down"
+      ></div>
     </div>
     <div class="inline-drawer-content">
-
       <!-- 基础设置 -->
       <div class="st-bme-section">
         <div class="st-bme-row">
@@ -17,7 +18,14 @@
 
         <div class="st-bme-row">
           <label for="st_bme_extract_every">每 N 条回复提取</label>
-                    <input type="number" id="st_bme_extract_every" class="text_pole" min="1" max="10" value="1" />
+          <input
+            type="number"
+            id="st_bme_extract_every"
+            class="text_pole"
+            min="1"
+            max="10"
+            value="1"
+          />
         </div>
       </div>
 
@@ -43,7 +51,14 @@
 
         <div class="st-bme-row">
           <label for="st_bme_inject_depth">注入深度</label>
-                    <input type="number" id="st_bme_inject_depth" class="text_pole" min="0" max="9999" value="4" />
+          <input
+            type="number"
+            id="st_bme_inject_depth"
+            class="text_pole"
+            min="0"
+            max="9999"
+            value="4"
+          />
         </div>
       </div>
 
@@ -55,17 +70,41 @@
 
         <div class="st-bme-row">
           <label for="st_bme_graph_weight">图扩散权重</label>
-                    <input type="number" id="st_bme_graph_weight" class="text_pole" min="0" max="1" step="0.1" value="0.6" />
+          <input
+            type="number"
+            id="st_bme_graph_weight"
+            class="text_pole"
+            min="0"
+            max="1"
+            step="0.1"
+            value="0.6"
+          />
         </div>
 
         <div class="st-bme-row">
           <label for="st_bme_vector_weight">向量权重</label>
-                    <input type="number" id="st_bme_vector_weight" class="text_pole" min="0" max="1" step="0.1" value="0.3" />
+          <input
+            type="number"
+            id="st_bme_vector_weight"
+            class="text_pole"
+            min="0"
+            max="1"
+            step="0.1"
+            value="0.3"
+          />
         </div>
 
         <div class="st-bme-row">
           <label for="st_bme_importance_weight">重要性权重</label>
-                    <input type="number" id="st_bme_importance_weight" class="text_pole" min="0" max="1" step="0.1" value="0.1" />
+          <input
+            type="number"
+            id="st_bme_importance_weight"
+            class="text_pole"
+            min="0"
+            max="1"
+            step="0.1"
+            value="0.1"
+          />
         </div>
       </div>
 
@@ -89,7 +128,14 @@
           </div>
           <div class="st-bme-row st-bme-indent">
             <label for="st_bme_evo_neighbors">近邻数量</label>
-                        <input type="number" id="st_bme_evo_neighbors" class="text_pole" min="1" max="10" value="5" />
+            <input
+              type="number"
+              id="st_bme_evo_neighbors"
+              class="text_pole"
+              min="1"
+              max="10"
+              value="5"
+            />
           </div>
 
           <div class="st-bme-row">
@@ -100,7 +146,15 @@
           </div>
           <div class="st-bme-row st-bme-indent">
             <label for="st_bme_conflict_threshold">对照阈值</label>
-                        <input type="number" id="st_bme_conflict_threshold" class="text_pole" min="0.5" max="0.99" step="0.05" value="0.85" />
+            <input
+              type="number"
+              id="st_bme_conflict_threshold"
+              class="text_pole"
+              min="0.5"
+              max="0.99"
+              step="0.05"
+              value="0.85"
+            />
           </div>
 
           <div class="st-bme-row">
@@ -111,7 +165,14 @@
           </div>
           <div class="st-bme-row st-bme-indent">
             <label for="st_bme_synopsis_every">每 N 次提取更新</label>
-                        <input type="number" id="st_bme_synopsis_every" class="text_pole" min="1" max="20" value="5" />
+            <input
+              type="number"
+              id="st_bme_synopsis_every"
+              class="text_pole"
+              min="1"
+              max="20"
+              value="5"
+            />
           </div>
         </div>
 
@@ -144,6 +205,15 @@
               <span>惊奇度分割 <small class="st-bme-hint">EM-LLM</small></span>
             </label>
           </div>
+          <div class="st-bme-row st-bme-indent">
+            <label for="st_bme_trigger_patterns">自定义触发模式</label>
+            <input
+              type="text"
+              id="st_bme_trigger_patterns"
+              class="text_pole"
+              placeholder="突然,真相,秘密,背叛 或 正则"
+            />
+          </div>
 
           <div class="st-bme-row">
             <label class="checkbox_label" for="st_bme_sleep_cycle">
@@ -153,7 +223,15 @@
           </div>
           <div class="st-bme-row st-bme-indent">
             <label for="st_bme_forget_threshold">遗忘阈值</label>
-                        <input type="number" id="st_bme_forget_threshold" class="text_pole" min="0.1" max="1.0" step="0.1" value="0.5" />
+            <input
+              type="number"
+              id="st_bme_forget_threshold"
+              class="text_pole"
+              min="0.1"
+              max="1.0"
+              step="0.1"
+              value="0.5"
+            />
           </div>
 
           <div class="st-bme-row">
@@ -164,7 +242,15 @@
           </div>
           <div class="st-bme-row st-bme-indent">
             <label for="st_bme_prob_chance">触发概率</label>
-                        <input type="number" id="st_bme_prob_chance" class="text_pole" min="0.01" max="0.5" step="0.05" value="0.15" />
+            <input
+              type="number"
+              id="st_bme_prob_chance"
+              class="text_pole"
+              min="0.01"
+              max="0.5"
+              step="0.05"
+              value="0.15"
+            />
           </div>
 
           <div class="st-bme-row">
@@ -175,7 +261,14 @@
           </div>
           <div class="st-bme-row st-bme-indent">
             <label for="st_bme_reflect_every">每 N 次提取反思</label>
-                        <input type="number" id="st_bme_reflect_every" class="text_pole" min="3" max="30" value="10" />
+            <input
+              type="number"
+              id="st_bme_reflect_every"
+              class="text_pole"
+              min="3"
+              max="30"
+              value="10"
+            />
           </div>
         </div>
       </div>
@@ -188,17 +281,32 @@
 
         <div class="st-bme-row">
           <label for="st_bme_embed_url">API 地址</label>
-                    <input type="text" id="st_bme_embed_url" class="text_pole" placeholder="https://api.openai.com/v1" />
+          <input
+            type="text"
+            id="st_bme_embed_url"
+            class="text_pole"
+            placeholder="https://api.openai.com/v1"
+          />
         </div>
 
         <div class="st-bme-row">
           <label for="st_bme_embed_key">API Key</label>
-                    <input type="password" id="st_bme_embed_key" class="text_pole" placeholder="sk-..." />
+          <input
+            type="password"
+            id="st_bme_embed_key"
+            class="text_pole"
+            placeholder="sk-..."
+          />
         </div>
 
         <div class="st-bme-row">
           <label for="st_bme_embed_model">模型</label>
-                    <input type="text" id="st_bme_embed_model" class="text_pole" placeholder="text-embedding-3-small" />
+          <input
+            type="text"
+            id="st_bme_embed_model"
+            class="text_pole"
+            placeholder="text-embedding-3-small"
+          />
         </div>
 
         <div class="st-bme-row">
@@ -241,7 +349,6 @@
           </button>
         </div>
       </div>
-
     </div>
   </div>
 </div>

tests/smart-trigger.mjs

@@ -0,0 +1,132 @@
+import assert from "node:assert/strict";
+
+function getSmartTriggerDecision(chat, lastProcessed, settings) {
+  const DEFAULT_TRIGGER_KEYWORDS = [
+    "突然",
+    "没想到",
+    "原来",
+    "其实",
+    "发现",
+    "背叛",
+    "死亡",
+    "复活",
+    "恢复记忆",
+    "失忆",
+    "告白",
+    "暴露",
+    "秘密",
+    "计划",
+    "规则",
+    "契约",
+    "位置",
+    "地点",
+    "离开",
+    "来到",
+  ];
+
+  const pendingMessages = chat
+    .slice(lastProcessed + 1)
+    .filter((msg) => !msg.is_system)
+    .map((msg) => ({
+      role: msg.is_user ? "user" : "assistant",
+      content: msg.mes || "",
+    }));
+
+  if (pendingMessages.length === 0) {
+    return { triggered: false, score: 0, reasons: [] };
+  }
+
+  const reasons = [];
+  let score = 0;
+  const combinedText = pendingMessages.map((m) => m.content).join("\n");
+
+  const keywordHits = DEFAULT_TRIGGER_KEYWORDS.filter((keyword) =>
+    combinedText.includes(keyword),
+  );
+  if (keywordHits.length > 0) {
+    score += Math.min(2, keywordHits.length);
+    reasons.push(`关键词: ${keywordHits.slice(0, 3).join(", ")}`);
+  }
+
+  const customPatterns = String(settings.triggerPatterns || "")
+    .split(/\r?\n|,/)
+    .map((s) => s.trim())
+    .filter(Boolean);
+  for (const pattern of customPatterns) {
+    try {
+      const regex = new RegExp(pattern, "i");
+      if (regex.test(combinedText)) {
+        score += 2;
+        reasons.push(`自定义触发: ${pattern}`);
+        break;
+      }
+    } catch {
+      // ignore invalid regex
+    }
+  }
+
+  const roleSwitchCount = pendingMessages.reduce((count, message, index) => {
+    if (index === 0) return count;
+    return count + (message.role !== pendingMessages[index - 1].role ? 1 : 0);
+  }, 0);
+  if (roleSwitchCount >= 2) {
+    score += 1;
+    reasons.push("多轮往返互动");
+  }
+
+  const punctuationHits = (combinedText.match(/[!?！？]/g) || []).length;
+  if (punctuationHits >= 2) {
+    score += 1;
+    reasons.push("情绪/冲突波动");
+  }
+
+  const entityLikeHits =
+    combinedText.match(
+      /[A-Z][a-z]{2,}|[\u4e00-\u9fff]{2,6}(先生|小姐|王国|城|镇|村|学院|组织|公司|小队|军团)/g,
+    ) || [];
+  if (entityLikeHits.length > 0) {
+    score += 1;
+    reasons.push("疑似新实体/新地点");
+  }
+
+  const threshold = Math.max(1, settings.smartTriggerThreshold || 2);
+  return {
+    triggered: score >= threshold,
+    score,
+    reasons,
+  };
+}
+
+const noTrigger = getSmartTriggerDecision(
+  [
+    { is_user: true, mes: "今天天气不错。" },
+    { is_user: false, mes: "是的，我们继续赶路。" },
+  ],
+  -1,
+  { triggerPatterns: "", smartTriggerThreshold: 3 },
+);
+assert.equal(noTrigger.triggered, false);
+
+const keywordTrigger = getSmartTriggerDecision(
+  [
+    { is_user: true, mes: "我们突然发现城堡地下有秘密。" },
+    { is_user: false, mes: "原来失踪的人都被关在这里！" },
+  ],
+  -1,
+  { triggerPatterns: "", smartTriggerThreshold: 2 },
+);
+assert.equal(keywordTrigger.triggered, true);
+assert.ok(keywordTrigger.score >= 2);
+
+const customTrigger = getSmartTriggerDecision(
+  [
+    { is_user: true, mes: "她轻声说出真相。" },
+    { is_user: false, mes: "所有人都沉默了。" },
+  ],
+  -1,
+  { triggerPatterns: "真相|背叛", smartTriggerThreshold: 2 },
+);
+assert.equal(customTrigger.triggered, true);
+assert.ok(customTrigger.reasons.some((r) => r.includes("自定义触发")));
+
+console.log("smart-trigger tests passed");