sage-mem 0.4.2 → 0.5.0

package/lib/index.js

@@ -15,12 +15,13 @@
 
 import { Remote, TypertRemoteService } from '@deepseek-ai/dsh-typert-protocol'
 import { homedir } from 'node:os'
-import { readFile, readdir, writeFile, unlink } from 'node:fs/promises'
+import { readFile, readdir, writeFile, unlink, stat } from 'node:fs/promises'
 import { join, basename } from 'node:path'
 
 const MEMORY_DIR = process.env.SAGE_MEM_DIR || join(homedir(), '.sage-mem', 'memory')
 const MAX_RESULTS = 5
 const MAX_CHARS_PER_FILE = 1500
+const MAX_BASELINE = 5
 
 /** 提取消息文本，兼容 content 为字符串或 text-block 数组两种形态。 */
 function extractText(content) {
@@ -51,7 +52,7 @@ function score(query, text) {
   return hit / q.size
 }
 
-/** 扫 memory 目录，读回全部记忆文件的正文（跳过索引与子目录）。 */
+/** 扫 memory 目录，读回全部记忆文件（跳过索引与子目录），并解析 frontmatter 摘要。 */
 async function scanMemoryFiles() {
   try {
     const entries = await readdir(MEMORY_DIR, { withFileTypes: true })
@@ -61,7 +62,13 @@ async function scanMemoryFiles() {
     const loaded = await Promise.all(names.map(async (name) => {
       try {
         const content = await readFile(join(MEMORY_DIR, name), 'utf8')
-        return { file: name, content }
+        const meta = parseFrontmatter(content)
+        return {
+          file: name,
+          content,
+          description: meta.description || '',
+          baseline: parseBaseline(content),
+        }
       } catch {
         return null
       }
@@ -72,10 +79,22 @@ async function scanMemoryFiles() {
   }
 }
 
-/** 按得分选相关文件，降序取 top N。 */
+/**
+ * 相关性打分只用 frontmatter 的 description + 文件名，不扫全文。
+ *
+ * 旧版拿 query 的双字去全文里找，得分是「覆盖率」，分母是 query，不惩罚
+ * 文件长度 —— 越长的记忆越容易蒙中。实测 8 个典型问题：project_heartscape
+ * (5.9KB) 与 project_self-cognition (6.8KB) 命中 7 个，连「今天天气不错」都
+ * 注入 2300 token；而「论文写得怎么样了」该命中的 project_position-paper
+ * 反被挤掉。description 是「一句话说清这条是什么」的精准摘要，用它当检索
+ * 信号，噪音大幅下降。
+ */
 function selectRelevant(query, files) {
   return files
-    .map(f => ({ ...f, score: score(query, f.content) }))
+    .map(f => {
+      const hay = f.description + ' ' + f.file.replace(/\.md$/, '').replace(/[_-]/g, ' ')
+      return { ...f, score: score(query, hay) }
+    })
     .filter(f => f.score > 0)
     .sort((a, b) => b.score - a.score)
     .slice(0, MAX_RESULTS)
@@ -99,6 +118,15 @@ function parseFrontmatter(content) {
   return { name: grab('name'), description: grab('description'), type }
 }
 
+/** 解析 frontmatter 的 `baseline` 标记：会话第一回合无条件注入。 */
+function parseBaseline(content) {
+  const m = content.match(/^---\r?\n([\s\S]*?)\r?\n---/)
+  if (!m) return false
+  const hit = m[1].match(/^baseline:\s*(.+)$/m)
+  if (!hit) return false
+  return /^(true|yes|1)$/i.test(hit[1].replace(/^["']|["']$/g, '').trim())
+}
+
 /** 防路径穿越：只取 basename，且必须在 memory 目录内（.md，排除索引）。 */
 function safeName(raw) {
   const name = basename(String(raw || ''))
@@ -125,9 +153,15 @@ function markRemote(cls, method, exportName) {
 export class MemoryGateway extends TypertRemoteService {
   constructor(ctx) {
     super(ctx, 'memory')
+    // 一并挂载星图 remote（同一记忆目录，纯只读）—— starmap.* 服务
+    new StarmapGateway(ctx)
+
+    // 同一轮对话里问题不变、注入内容也不变，但 DSH 每一步都会往 session 追加
+    // 一条 context 消息——不去重的话同一份记忆会被反复追加，撑爆上下文。
+    const recallCache = new Map()
 
     // 按问题检索注入：system-prompt/assemble 是异步 waterfall，监听器可 await。
-    // 从 agent.session 拿当前 user message，扫 memory 目录按双字匹配选相关
+    // 从 agent.session 拿当前 user message，扫 memory 目录按摘要匹配选相关
     // 文件，读全文作为动态 context 段注入——让 agent 第一轮就「想起」。
     ctx.on('system-prompt/assemble', async (assembly, context, next) => {
       const assembled = await next()
@@ -136,15 +170,44 @@ export class MemoryGateway extends TypertRemoteService {
         const session = agent?.session
         if (!session || typeof session.deriveMessages !== 'function') return assembled
         const messages = [...session.deriveMessages()]
-        const lastUser = messages.reverse().find(m => m?.role === 'user')
-        const text = extractText(lastUser?.content)
-        if (!text || text.length < 2) return assembled
+
+        // 关键：DSH 0.1.5 把工具结果（source.kind === 'tool'）和注入内容
+        // （source.kind === 'plugin'）也存成 role: 'user' 的消息。只按 role 取
+        // 「最后一条」会取到 tool-result，extractText 得到空串，下面的长度检查
+        // 直接返回——表现就是全程静默不注入。只有 source.kind === 'user' 是真人输入。
+        const userMessages = messages.filter(m => m?.role === 'user' && m?.source?.kind === 'user')
+        const text = extractText(userMessages[userMessages.length - 1]?.content)
 
         const files = await scanMemoryFiles()
-        const relevant = selectRelevant(text, files)
-        if (relevant.length === 0) return assembled
+        const picked = []
+        const seen = new Set()
+        const take = (f) => {
+          if (seen.has(f.file)) return
+          seen.add(f.file)
+          picked.push(f)
+        }
+
+        // 新会话首步：注入 frontmatter 标了 `baseline: true` 的记忆。当前这条
+        // 用户消息在 assemble 时还没写进 session，所以首步这里恰好是 0 条。
+        // 这一层补「开场白失忆」——「继续吧」这类短话在 description 里找不到
+        // 匹配，纯检索会返回 0 条，而那恰恰是最需要上下文的时刻。
+        if (userMessages.length === 0) {
+          files
+            .filter(f => f.baseline)
+            .sort((a, b) => a.file.localeCompare(b.file))
+            .slice(0, MAX_BASELINE)
+            .forEach(take)
+        }
+
+        if (text && text.length >= 2) selectRelevant(text, files).forEach(take)
+        if (picked.length === 0) return assembled
 
-        const parts = relevant.map(f => {
+        const signature = text + '\u0000' + picked.map(f => f.file).join(',')
+        if (recallCache.get(session.id) === signature) return assembled
+        recallCache.set(session.id, signature)
+        if (recallCache.size > 8) recallCache.delete(recallCache.keys().next().value)
+
+        const parts = picked.map(f => {
           const body = f.content.length > MAX_CHARS_PER_FILE
             ? f.content.slice(0, MAX_CHARS_PER_FILE) + '\n…（截断）'
             : f.content
@@ -209,4 +272,69 @@ markRemote(MemoryGateway, 'readFile', 'readFile')
 markRemote(MemoryGateway, 'writeFile', 'writeFile')
 markRemote(MemoryGateway, 'deleteFile', 'deleteFile')
 
+/** 提取正文第一个 `# ` 标题作为展示标题。 */
+function extractTitle(content) {
+  const m = content.match(/^---\r?\n[\s\S]*?\r?\n---/)
+  const body = m ? content.slice(m[0].length) : content
+  const h1 = body.match(/^#\s+(.+)$/m)
+  return h1 ? h1[1].trim() : ''
+}
+
+/** 扫记忆目录，读全部 .md（跳过索引与 session-log），带体积与修改时间。 */
+async function scanStars() {
+  let entries
+  try { entries = await readdir(MEMORY_DIR, { withFileTypes: true }) } catch { return [] }
+  const names = entries
+    .filter(e => e.isFile() && e.name.endsWith('.md') && e.name !== 'MEMORY.md' && e.name !== 'session-log.md')
+    .map(e => e.name)
+  const loaded = await Promise.all(names.map(async (name) => {
+    try {
+      const full = join(MEMORY_DIR, name)
+      const [content, info] = await Promise.all([readFile(full, 'utf8'), stat(full)])
+      return { name, content, bytes: info.size, mtimeMs: info.mtimeMs }
+    } catch { return null }
+  }))
+  return loaded.filter(Boolean)
+}
+
+/**
+ * 记忆星图（host 半）—— 只读扫描记忆目录，把每条记忆解析成一颗「星」。
+ * 通过 TypertRemoteService 暴露两个 remote 方法给浏览器半：
+ *   - starmap.listStars()  全部星（元数据，不含正文）
+ *   - starmap.readFile(n)  单条记忆全文（文件名白名单校验）
+ */
+export class StarmapGateway extends TypertRemoteService {
+  constructor(ctx) { super(ctx, 'starmap') }
+
+  async listStars() {
+    const files = await scanStars()
+    const stars = files.map(f => {
+      const meta = parseFrontmatter(f.content)
+      const title = extractTitle(f.content)
+      return {
+        file: f.name,
+        kind: meta.type || 'special',
+        title: title || meta.name || (meta.description ? meta.description.slice(0, 24) : f.name.replace(/\.md$/, '')),
+        desc: meta.description || '',
+        bytes: f.bytes,
+        mtimeMs: f.mtimeMs,
+      }
+    })
+    return { count: stars.length, stars }
+  }
+
+  async readFile(name) {
+    const raw = String(name || '')
+    const safe = basename(raw)
+    if (!safe || safe !== raw || !safe.endsWith('.md') || safe === 'MEMORY.md') {
+      throw new Error('sage-starmap: invalid file name')
+    }
+    const content = await readFile(join(MEMORY_DIR, safe), 'utf8')
+    return { name: safe, content }
+  }
+}
+
+markRemote(StarmapGateway, 'listStars', 'listStars')
+markRemote(StarmapGateway, 'readFile', 'readFile')
+
 export default MemoryGateway

package/lib/typert.host.js

@@ -9,6 +9,23 @@ import { z } from 'zod'
 const fileNameSchema = z.string()
 const fileContentSchema = z.string()
 
+const starmapStarSchema = z.object({
+	'file': z.string().readonly(),
+	'kind': z.string().readonly(),
+	'title': z.string().readonly(),
+	'desc': z.string().readonly(),
+	'bytes': z.number().readonly(),
+	'mtimeMs': z.number().readonly(),
+}).readonly()
+const starmapListSchema = z.object({
+	'count': z.number().readonly(),
+	'stars': z.array(starmapStarSchema).readonly(),
+}).readonly()
+const starmapReadSchema = z.object({
+	'name': z.string().readonly(),
+	'content': z.string().readonly(),
+}).readonly()
+
 const fileSchema = z.object({
 	'file': z.string().readonly(),
 	'type': z.string().readonly(),
@@ -81,6 +98,28 @@ export const TYPERT = {
 			result: { mode: 'strict', typeSymbol: 'sage-mem/types#DeleteResult', schema: deleteResultSchema },
 			sourceLocation: { "file": "workspace/sage-mem/lib/index.js", "line": 1, "column": 1 },
 		},
+		{
+			id: 'sage-mem#starmap/listStars',
+			service: 'starmap',
+			namespace: 'starmap',
+			method: 'listStars',
+			invocation: { kind: 'direct' },
+			parameters: [],
+			result: { mode: 'strict', typeSymbol: 'sage-mem/types#StarmapList', schema: starmapListSchema },
+			sourceLocation: { "file": "workspace/sage-mem/lib/index.js", "line": 1, "column": 1 },
+		},
+		{
+			id: 'sage-mem#starmap/readFile',
+			service: 'starmap',
+			namespace: 'starmap',
+			method: 'readFile',
+			invocation: { kind: 'direct' },
+			parameters: [
+				{ name: 'name', wire: 'name', source: 'json', codec: { mode: 'strict', typeSymbol: 'sage-mem/types#FileName', schema: fileNameSchema } },
+			],
+			result: { mode: 'strict', typeSymbol: 'sage-mem/types#StarmapRead', schema: starmapReadSchema },
+			sourceLocation: { "file": "workspace/sage-mem/lib/index.js", "line": 1, "column": 1 },
+		},
 	],
 	model: {
 		"services": [],

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "sage-mem",
-  "description": "File-based memory plugin for DeepSeek Harness (DSH) — cross-session memory in transparent markdown files",
-  "version": "0.4.2",
+  "description": "File-based memory plugin for DeepSeek Harness (DSH) — cross-session memory in transparent markdown files, with a memory star-map view",
+  "version": "0.5.0",
   "type": "module",
   "main": "lib/index.js",
   "exports": {
@@ -36,7 +36,8 @@
       "dshReleases": {
         "0.1.2-rc.1": "compatible",
         "0.1.3-alpha.1": "compatible",
-        "0.1.3-alpha.2": "compatible"
+        "0.1.3-alpha.2": "compatible",
+        "0.1.5-rc.1": "compatible"
       }
     }
   },
@@ -51,7 +52,7 @@
   },
   "repository": {
     "type": "git",
-    "url": "https://github.com/gezi-wen/sage-mem.git"
+    "url": "git+https://github.com/gezi-wen/sage-mem.git"
   },
   "keywords": [
     "memory",