memorix 0.8.0 → 0.9.0

package/CHANGELOG.md

@@ -2,6 +2,23 @@
 
 All notable changes to this project will be documented in this file.
 
+## [0.9.0] — 2026-02-24
+
+### Added
+- **Memory Consolidation** (`memorix_consolidate`) — Find and merge similar observations to reduce memory bloat. Uses Jaccard text similarity to cluster observations by entity+type, then merges them preserving all facts, files, and concepts. Supports `preview` (dry run) and `execute` modes with configurable similarity threshold.
+- **Temporal Queries** — `memorix_search` now supports `since` and `until` parameters for date range filtering. Example: "What auth decisions did we make last week?"
+- **Explainable Recall** — Search results now include a `Matched` column showing which fields matched the query (title, entity, concept, narrative, fact, file, or fuzzy). Helps understand why each result was found.
+- **Export/Import** — Two new tools for team collaboration:
+  - `memorix_export` — Export project observations and sessions as JSON (importable) or Markdown (human-readable for PRs/docs)
+  - `memorix_import` — Import from JSON export, re-assigns IDs, skips duplicate topicKeys
+- **Dashboard Sessions Panel** — New "Sessions" tab in the web dashboard with timeline view, active/completed counts, agent info, and session summaries. Bilingual (EN/中文).
+- **Auto sessionId** — `memorix_store` now automatically associates the current active session's ID with stored observations.
+- **16 new tests** — 8 consolidation + 8 export/import (484 total).
+
+### Stats
+- **MCP Tools:** 20 → 24 (memorix_consolidate, memorix_export, memorix_import + dashboard sessions API)
+- **Tests:** 484/484 passing
+
 ## [0.8.0] — 2026-02-24
 
 ### Added

package/README.md

@@ -164,19 +164,59 @@ After 2 weeks of development, you have 50+ observations:
   → New team members' AI instantly knows your project's patterns!
 ```
 
+### Scenario 6: Session Lifecycle (v0.8.0)
+
+```
+Morning — Start a new session in Windsurf:
+  → memorix_session_start auto-injects:
+    📋 Previous Session: "Implemented JWT auth middleware"
+    🔴 JWT tokens expire silently (gotcha)
+    🟤 Use Docker for deployment (decision)
+  → AI instantly knows what you did yesterday!
+
+Evening — End the session:
+  → memorix_session_end saves structured summary
+  → Next session (any agent!) gets this context automatically
+```
+
+### Scenario 7: Topic Key Upsert — No More Duplicates (v0.8.0)
+
+```
+You update your architecture decision 3 times over a week:
+
+  Day 1: memorix_store(topicKey="architecture/auth-model", ...)
+  → Creates observation #42 (rev 1)
+
+  Day 3: memorix_store(topicKey="architecture/auth-model", ...)
+  → Updates #42 in-place (rev 2) — NOT a new #43!
+
+  Day 5: memorix_store(topicKey="architecture/auth-model", ...)
+  → Updates #42 again (rev 3)
+
+  Result: 1 observation with latest content, not 3 duplicates!
+```
+
 ---
 
 ## 🧠 What Memorix Can Do
 
-### Smart Memory (17 MCP Tools)
+### Smart Memory (24 MCP Tools)
 
 | What You Say | What Memorix Does |
 |-------------|-------------------|
-| "Remember this architecture decision" | `memorix_store` — Classifies as 🟤 decision, extracts entities, creates graph relations |
+| "Remember this architecture decision" | `memorix_store` — Classifies as 🟤 decision, extracts entities, creates graph relations, auto-associates session |
 | "What did we decide about auth?" | `memorix_search` → `memorix_detail` — 3-layer progressive disclosure, ~10x token savings |
+| "What auth decisions last week?" | `memorix_search` with `since`/`until` — Temporal queries with date range filtering |
 | "What happened around that bug fix?" | `memorix_timeline` — Shows chronological context before/after |
-| "Show me the knowledge graph" | `memorix_dashboard` — Opens interactive web UI with D3.js graph |
+| "Show me the knowledge graph" | `memorix_dashboard` — Opens interactive web UI with D3.js graph + sessions panel |
 | "Which memories are getting stale?" | `memorix_retention` — Exponential decay scores, identifies archive candidates |
+| "Start a new session" | `memorix_session_start` — Tracks session lifecycle, auto-injects previous session summaries + key memories |
+| "End this session" | `memorix_session_end` — Saves structured summary (Goal/Discoveries/Accomplished/Files) for next session |
+| "What did we do last session?" | `memorix_session_context` — Retrieves session history and key observations |
+| "Suggest a topic key for this" | `memorix_suggest_topic_key` — Generates stable keys for deduplication (e.g. `architecture/auth-model`) |
+| "Clean up duplicate memories" | `memorix_consolidate` — Find & merge similar observations by text similarity, preserving all facts |
+| "Export this project's memories" | `memorix_export` — JSON (importable) or Markdown (human-readable for PRs/docs) |
+| "Import memories from teammate" | `memorix_import` — Restore from JSON export, re-assigns IDs, deduplicates by topicKey |
 
 ### Cross-Agent Workspace Sync
 
@@ -298,7 +338,7 @@ With vector search, queries like "authentication" also match memories about "log
 Install Memorix in both IDEs. They share the same local memory directory — architecture decisions made in Cursor are instantly searchable in Claude Code. No cloud sync needed.
 
 **How do I prevent my AI from forgetting previous sessions?**
-Memorix stores observations persistently on disk. Next session, the AI calls `memorix_search` and retrieves prior decisions, gotchas, and knowledge. With auto-memory hooks, it even captures context automatically.
+Use `memorix_session_start` at the beginning of each session — it automatically injects previous session summaries and key observations (gotchas, decisions, discoveries). Use `memorix_session_end` to save a structured summary before leaving. All observations persist on disk and are searchable via `memorix_search` anytime.
 
 **How do I sync MCP server configs between IDEs?**
 Run `memorix_workspace_sync` with action `"migrate"` and your target IDE. It scans source configs and generates compatible configs for the target — merges, never overwrites.

package/README.zh-CN.md

@@ -164,19 +164,59 @@ Antigravity 会将工作目录设为自身安装路径（如 `G:\Antigravity`）
   → 新同事的 AI 立即掌握你项目的最佳实践！
 ```
 
+### 场景 6：会话生命周期（v0.8.0）
+
+```
+早上 — 在 Windsurf 中开始新会话：
+  → memorix_session_start 自动注入：
+    📋 上次会话: "实现了 JWT 认证中间件"
+    🔴 JWT token 静默过期（踩坑）
+    🟤 使用 Docker 部署（决策）
+  → AI 立即知道你昨天做了什么！
+
+晚上 — 结束会话：
+  → memorix_session_end 保存结构化摘要
+  → 下次会话（任何 Agent！）自动获取这些上下文
+```
+
+### 场景 7：Topic Key 去重 — 不再有重复记忆（v0.8.0）
+
+```
+你在一周内更新了 3 次架构决策：
+
+  第 1 天: memorix_store(topicKey="architecture/auth-model", ...)
+  → 创建观察记录 #42（rev 1）
+
+  第 3 天: memorix_store(topicKey="architecture/auth-model", ...)
+  → 原地更新 #42（rev 2）— 不是新建 #43！
+
+  第 5 天: memorix_store(topicKey="architecture/auth-model", ...)
+  → 再次更新 #42（rev 3）
+
+  结果：1 条记录包含最新内容，而不是 3 条重复！
+```
+
 ---
 
 ## 🧠 Memorix 能做什么
 
-### 智能记忆（17 个 MCP 工具）
+### 智能记忆（24 个 MCP 工具）
 
 | 你说的 | Memorix 做的 |
 |--------|-------------|
-| "记住这个架构决策" | `memorix_store` — 分类为 🟤 决策，提取实体，创建图关系 |
+| "记住这个架构决策" | `memorix_store` — 分类为 🟤 决策，提取实体，创建图关系，自动关联会话 |
 | "我们之前关于认证决定了什么？" | `memorix_search` → `memorix_detail` — 3 层渐进式展示，节省约 10 倍 token |
+| "上周关于认证的决策？" | `memorix_search` + `since`/`until` — 时序查询，按日期范围过滤 |
 | "那个 bug 修复前后发生了什么？" | `memorix_timeline` — 按时间显示前后上下文 |
-| "显示知识图谱" | `memorix_dashboard` — 打开交互式 Web UI，含 D3.js 图谱 |
+| "显示知识图谱" | `memorix_dashboard` — 打开交互式 Web UI，含 D3.js 图谱 + 会话面板 |
 | "哪些记忆快过期了？" | `memorix_retention` — 指数衰减评分，识别归档候选 |
+| "开始新会话" | `memorix_session_start` — 跟踪会话生命周期，自动注入上次会话摘要 + 关键记忆 |
+| "结束这次会话" | `memorix_session_end` — 保存结构化摘要（目标/发现/完成/文件），供下次会话使用 |
+| "上次我们做了什么？" | `memorix_session_context` — 检索会话历史和关键观察记录 |
+| "给这个建议一个 topic key" | `memorix_suggest_topic_key` — 生成稳定的去重键（如 `architecture/auth-model`） |
+| "清理重复记忆" | `memorix_consolidate` — 按文本相似度查找并合并相似观察记录，保留所有事实 |
+| "导出这个项目的记忆" | `memorix_export` — JSON（可导入）或 Markdown（人类可读，适合 PR/文档） |
+| "导入队友的记忆" | `memorix_import` — 从 JSON 导出恢复，重新分配 ID，按 topicKey 去重 |
 
 ### 跨 Agent 工作区同步
 
@@ -298,7 +338,7 @@ npm install -g @huggingface/transformers
 在两个 IDE 都安装 Memorix。它们共享相同的本地记忆目录 — 在 Cursor 中做的架构决策在 Claude Code 中立即可搜索，无需云同步。
 
 **如何防止 AI 忘记之前的会话？**
-Memorix 将观察记录持久存储在磁盘上。下次会话，AI 调用 `memorix_search` 即可检索之前的决策、踩坑和知识。配合自动记忆 Hook，甚至可以自动捕获上下文。
+在每次会话开始时调用 `memorix_session_start` — 它会自动注入上次会话的摘要和关键观察记录（踩坑、决策、发现）。会话结束时调用 `memorix_session_end` 保存结构化摘要。所有观察记录持久存储在磁盘上，随时可通过 `memorix_search` 搜索。
 
 **如何在 IDE 之间同步 MCP 服务器配置？**
 运行 `memorix_workspace_sync`，设置 action 为 `"migrate"`，指定目标 IDE。它会扫描源配置并生成兼容的目标配置 — 合并，永不覆盖。