memoryweave/docs/progress/task-phase1-1-models-queue.md

183 lines
4.8 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

## OpenCode 任务 - Phase 1.1 数据模型 + 持久化队列
**日期**: 2026-05-24
**项目**: 织忆 (MemoryWeave)
**代码仓库**: ~/projects/zhiyi/
---
### 背景
织忆是独立记忆服务,把对话日志蒸馏成结构化记忆。此任务是 Phase 1.1 的核心入口。
---
### 具体任务
#### 任务 1: Episode 模型
文件: `src/models/episode.py`
```python
from dataclasses import dataclass, field
from datetime import datetime
from typing import Optional
@dataclass
class Episode:
"""原始记忆单元 — 对话日志、任务记录等"""
id: str # UUID
timestamp: datetime # 创建时间
content: str # 原始内容
entities: list[str] = field(default_factory=list) # 实体列表
facts: list[str] = field(default_factory=list) # 事实列表
metadata: dict = field(default_factory=dict) # 元数据
source: str = "hermes" # 来源: hermes/openclaw/manual
```
**要求**:
- dataclass 风格
-`to_dict()` / `from_dict()` 序列化方法
- UUID 生成用 `uuid.uuid4()`
---
#### 任务 2: Distilled 模型
文件: `src/models/distilled.py`
```python
from dataclasses import dataclass, field
from datetime import datetime
from typing import Optional
@dataclass
class Distilled:
"""蒸馏后的结构化记忆"""
id: str
episode_id: str # 来源 Episode ID
type: str # "decision" | "request" | "fact" | "pattern"
summary: str # 摘要
entities: list[str] = field(default_factory=list)
facts: list[str] = field(default_factory=list)
confidence: float = 0.5 # 置信度 0-1
status: str = "pending" # "pending" | "validated" | "deprecated"
importance: int = 0 # 重要性 0-5>=3 永不衰减
created_at: datetime = field(default_factory=datetime.now)
updated_at: datetime = field(default_factory=datetime.now)
```
**要求**:
- 同上,有序列化方法
- `status` 可选值用常量或 enum
---
#### 任务 3: 持久化队列
文件: `src/distill/queue.py`
```python
import sqlite3
import json
from pathlib import Path
from datetime import datetime
from typing import Optional
from ..models.episode import Episode
class PersistenceQueue:
"""SQLite 持久化队列 — 入队/出队/持久化"""
def __init__(self, db_path: str = "zhiyi.db"):
self.db_path = db_path
self.conn = sqlite3.connect(db_path)
self._init_db()
def _init_db(self):
"""初始化表结构"""
self.conn.execute("""
CREATE TABLE IF NOT EXISTS episode_queue (
id TEXT PRIMARY KEY,
data TEXT NOT NULL,
enqueued_at TEXT NOT NULL,
dequeued_at TEXT,
status TEXT DEFAULT 'pending'
)
""")
self.conn.commit()
def enqueue(self, episode: Episode) -> bool:
"""入队,返回是否成功"""
def dequeue(self) -> Optional[Episode]:
"""出队,返回 Episode 或 None"""
def peek(self) -> Optional[Episode]:
"""查看队首,不出队"""
def size(self) -> int:
"""队列长度"""
def is_empty(self) -> bool:
"""队列是否为空"""
def requeue(self, episode: Episode) -> bool:
"""重新入队(处理失败时)"""
```
**验收标准**:
1. 100条连续写入无丢失
2. 服务重启后队列数据恢复
3. 并发写入安全(加锁)
**测试用例**: `tests/test_queue.py`
```python
def test_queue_persistence():
q = PersistenceQueue(":memory:") # 内存测试
# 测试入队出队
ep = Episode(id="1", timestamp=datetime.now(), content="test")
q.enqueue(ep)
assert q.size() == 1
dequeued = q.dequeue()
assert dequeued.id == "1"
# 测试重连后恢复(内存队列不需要)
# 真实 db 测试需要持久化路径
```
---
### 技术要求
- **语言**: Python 3.10+
- **代码规范**: PEP8
- **无外部依赖**: 只用标准库 + sqlite3
- **测试覆盖**: 每个文件有对应测试
---
### 参考
- 设计文档: `~/mc/小唯/07-Wiki/concepts/织忆(MemoryWeave)-v2.6-完整定稿.md` 第4章
- 参考项目: `~/projects/memoryfabric-research/agent-memory-skill/memory-engine.py`(线性衰减参考队列实现)
---
### 注意事项
1. **不写设计之外的代码** — 按任务清单来
2. **有问题先问** — 不要自作主张
3. **完成后发飞书通知** — 牧尘或小唯
4. **commit 要规范**`feat: Phase 1.1 数据模型 + 持久化队列`
---
### 产出
| 文件 | 说明 |
|------|------|
| src/models/episode.py | Episode 模型 |
| src/models/distilled.py | Distilled 模型 |
| src/distill/queue.py | 持久化队列 |
| tests/test_queue.py | 队列测试 |
| tests/test_models.py | 模型测试 |