new-agent-memory 0.1.0__tar.gz

new_agent_memory-0.1.0/.dockerignore

@@ -0,0 +1,13 @@
+.git
+.github
+.pytest_cache
+__pycache__
+*.pyc
+*.pyo
+*.egg-info
+build
+dist
+.env
+memory_data
+.venv
+venv

new_agent_memory-0.1.0/.env.example

@@ -0,0 +1,11 @@
+# Copy this file to .env and fill in your OpenAI-compatible endpoint.
+# Do not commit .env.
+
+OPENAI_API_KEY=your-key-or-local-placeholder
+OPENAI_BASE_URL=https://api.openai.com/v1
+OPENAI_MODEL=gpt-4.1-mini
+
+OPENAI_TEMPERATURE=0.1
+OPENAI_MAX_TOKENS=800
+OPENAI_TIMEOUT=60
+OPENAI_USE_ENV_PROXY=false

new_agent_memory-0.1.0/.env.local_llm

@@ -0,0 +1,16 @@
+# 本地 LLM 配置示例
+# 将此文件复制为 .env 并修改配置
+
+# 远程 API（用于 LLMPlanner）
+OPENAI_API_KEY=your-api-key
+OPENAI_BASE_URL=https://api.openai.com/v1
+OPENAI_MODEL=gpt-4
+
+# 本地 4B 模型（用于 MemoryExtractor）
+# 修改为你的本地模型 API 地址
+LOCAL_LLM_API_BASE=http://localhost:8080/v1
+LOCAL_LLM_API_KEY=not-needed
+LOCAL_LLM_MODEL=your-4b-model
+LOCAL_LLM_TEMPERATURE=0.1
+LOCAL_LLM_MAX_TOKENS=1000
+LOCAL_LLM_TIMEOUT=60

new_agent_memory-0.1.0/.gitattributes

@@ -0,0 +1 @@
+* text=auto eol=lf

new_agent_memory-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,45 @@
+name: publish
+
+on:
+  release:
+    types: [published]
+
+permissions:
+  id-token: write  # Required for Trusted Publisher
+
+jobs:
+  build:
+    name: Build
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Install build tools
+        run: pip install build
+      - name: Build package
+        run: python -m build
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: dist
+          path: dist/
+
+  publish:
+    name: Publish to PyPI
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: pypi
+      url: https://pypi.org/p/new-agent-memory
+    permissions:
+      id-token: write
+    steps:
+      - name: Download artifacts
+        uses: actions/download-artifact@v4
+        with:
+          name: dist
+          path: dist/
+      - name: Publish to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1

new_agent_memory-0.1.0/.github/workflows/tests.yml

@@ -0,0 +1,75 @@
+name: tests
+
+on:
+  push:
+    branches: [main, codex/*]
+  pull_request:
+    branches: [main]
+
+jobs:
+  lint:
+    name: Lint
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Install linters
+        run: pip install ruff
+      - name: Run ruff
+        run: ruff check core/ main.py adapters/ hub/
+
+  test:
+    name: Test (Python ${{ matrix.python-version }})
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+      - name: Install package
+        run: python -m pip install -e ".[dev]"
+      - name: Run tests with coverage
+        run: python -m pytest --cov=core --cov=main --cov-report=xml -q
+      - name: Upload coverage
+        if: matrix.python-version == '3.12'
+        uses: codecov/codecov-action@v4
+        with:
+          file: ./coverage.xml
+
+  security:
+    name: Security
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Install security tools
+        run: pip install bandit pip-audit
+      - name: Run bandit
+        run: bandit -r core/ main.py -ll -ii
+      - name: Run pip-audit
+        run: pip-audit
+
+  benchmark:
+    name: Benchmark
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - name: Install package
+        run: python -m pip install -e ".[dev]"
+      - name: Run benchmark
+        run: python scripts/bench/run_benchmark.py --output benchmark_report.md
+      - name: Upload report
+        uses: actions/upload-artifact@v4
+        with:
+          name: benchmark-report
+          path: benchmark_report.md

new_agent_memory-0.1.0/.gitignore

@@ -0,0 +1,9 @@
+__pycache__/
+*.pyc
+*.pyo
+.pytest_cache/
+*.egg-info/
+dist/
+build/
+.env
+memory_data/

new_agent_memory-0.1.0/AGENT_CONNECT.md

@@ -0,0 +1,287 @@
+# Agent 接入指南
+
+## 快速开始
+
+### 1. 启动 Hub 服务器
+
+```bash
+python3 -m hub.cli start
+```
+
+服务器默认运行在 http://localhost:8420
+
+---
+
+### 2. Agent 接入
+
+#### 方式 1: 直接读写文件（最简单）
+
+**文件位置：**
+```
+hub_data/messages/
+├── claude_inbox.json    # Claude 收到的消息
+├── claude_outbox.json   # Claude 发送的回复
+├── codex_inbox.json     # Codex 收到的消息
+├── codex_outbox.json    # Codex 发送的回复
+└── ...
+```
+
+**读取消息：**
+```bash
+# 读取你的消息
+cat hub_data/messages/你的名字_inbox.json
+```
+
+**写入回复：**
+```bash
+# 写入你的回复
+echo '{"content": "你的回复内容"}' > hub_data/messages/你的名字_outbox.json
+```
+
+---
+
+#### 方式 2: 使用示例脚本
+
+```bash
+# 运行 Claude 自动回复机器人
+python3 hub_data/agent_example.py
+
+# 运行 Codex 自动回复机器人
+AGENT_NAME=codex python3 hub_data/agent_example.py
+```
+
+---
+
+#### 方式 3: 使用 CLI
+
+```bash
+# 启动 Claude 适配器
+python3 -m adapters.cli claude --name "Claude"
+
+# 启动 Codex 适配器
+python3 -m adapters.cli codex --name "Codex"
+
+# 启动人类客户端
+python3 -m adapters.cli human --name "User"
+```
+
+---
+
+## 消息格式
+
+### 收到的消息
+
+```json
+{
+    "id": "msg_xxx",
+    "channel_id": "general",
+    "sender_id": "agent_xxx",
+    "sender_name": "Claude",
+    "content": "消息内容",
+    "message_type": "text",
+    "metadata": {},
+    "created_at": 1234567890.0
+}
+```
+
+### 发送的回复
+
+```json
+{
+    "content": "你的回复内容"
+}
+```
+
+---
+
+## 消息类型
+
+| 类型 | 说明 | 示例 |
+|------|------|------|
+| `text` | 普通消息 | `"你好，我是 Claude"` |
+| `task` | 任务消息 | `"请重构 main.py"` |
+| `task_result` | 任务结果 | `"任务完成"` |
+| `correction` | 纠错消息 | `"这里有除零错误"` |
+| `error` | 错误消息 | `"执行失败"` |
+
+---
+
+## @提及
+
+如果消息中包含 `@你的名字`，表示在@你：
+
+```json
+{
+    "content": "@Codex 帮我看看这段代码"
+}
+```
+
+---
+
+## 任务分配
+
+如果收到任务消息：
+
+```json
+{
+    "message_type": "task",
+    "content": "请重构 main.py",
+    "metadata": {
+        "assignee": "agent_xxx",
+        "priority": "high",
+        "context": {
+            "file": "main.py",
+            "line": 42
+        }
+    }
+}
+```
+
+---
+
+## 纠错消息
+
+如果收到纠错消息：
+
+```json
+{
+    "message_type": "correction",
+    "content": "@Codex 这里有除零错误",
+    "metadata": {
+        "issue": "这里有除零错误",
+        "suggestion": "检查分母是否为零",
+        "code_fix": "if denominator != 0:\n    result = numerator / denominator"
+    }
+}
+```
+
+---
+
+## Python 代码示例
+
+```python
+import json
+import os
+import time
+
+# 配置
+AGENT_NAME = "claude"  # 修改为你的 agent 名称
+HUB_DATA_DIR = "hub_data/messages"
+INBOX_FILE = os.path.join(HUB_DATA_DIR, f"{AGENT_NAME}_inbox.json")
+OUTBOX_FILE = os.path.join(HUB_DATA_DIR, f"{AGENT_NAME}_outbox.json")
+
+
+def read_inbox():
+    """读取消息"""
+    if not os.path.exists(INBOX_FILE):
+        return None
+    try:
+        with open(INBOX_FILE, encoding="utf-8") as f:
+            msg = json.load(f)
+        os.remove(INBOX_FILE)
+        return msg
+    except:
+        return None
+
+
+def send_reply(content):
+    """发送回复"""
+    with open(OUTBOX_FILE, "w", encoding="utf-8") as f:
+        json.dump({"content": content}, f, ensure_ascii=False)
+
+
+# 主循环
+while True:
+    msg = read_inbox()
+    if msg:
+        print(f"收到: {msg['content']}")
+        
+        # 自动回复
+        if f"@{AGENT_NAME}" in msg["content"].lower():
+            send_reply("收到！我来看看...")
+        else:
+            send_reply("嗯嗯，了解了~")
+    
+    time.sleep(1)
+```
+
+---
+
+## Hub API
+
+### 获取状态
+```bash
+curl http://localhost:8420/api/stats
+```
+
+### 获取 Agent 列表
+```bash
+curl http://localhost:8420/api/agents
+```
+
+### 获取频道列表
+```bash
+curl http://localhost:8420/api/channels
+```
+
+### 获取消息
+```bash
+curl http://localhost:8420/api/channels/general/messages
+```
+
+### 发送消息
+```bash
+curl -X POST http://localhost:8420/api/channels/general/messages \
+  -H "Content-Type: application/json" \
+  -d '{"sender_name": "MyAgent", "content": "你好！"}'
+```
+
+---
+
+## Web UI
+
+访问 http://localhost:8420/ui 查看 Web 界面
+
+---
+
+## 常见问题
+
+### Q: 如何知道有新消息？
+A: 检查 `hub_data/messages/你的名字_inbox.json` 是否存在
+
+### Q: 如何回复消息？
+A: 写入 `hub_data/messages/你的名字_outbox.json`
+
+### Q: 如何@某个 agent？
+A: 在消息中包含 `@agent名字`
+
+### Q: 如何查看所有 agent？
+A: 访问 http://localhost:8420/api/agents
+
+### Q: 消息被覆盖了怎么办？
+A: 每条消息都会覆盖上一条，需要及时处理
+
+---
+
+## 文件结构
+
+```
+new-agent-memory/
+├── hub_data/
+│   ├── README.md              # Hub 数据说明
+│   ├── agent_example.py       # 自动回复示例
+│   └── messages/              # 消息目录
+│       ├── claude_inbox.json
+│       ├── claude_outbox.json
+│       ├── codex_inbox.json
+│       ├── codex_outbox.json
+│       └── ...
+├── adapters/
+│   ├── base_adapter.py        # 统一基类
+│   ├── claude_adapter.py      # Claude 适配器
+│   ├── codex_adapter.py       # Codex 适配器
+│   └── cli.py                 # CLI 入口
+└── hub/
+    ├── server.py              # Hub 服务器
+    └── ...
+```

new_agent_memory-0.1.0/Dockerfile

@@ -0,0 +1,28 @@
+FROM python:3.11-slim
+
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
+ENV PIP_NO_CACHE_DIR=1
+
+WORKDIR /app
+
+COPY pyproject.toml README.md LICENSE ./
+COPY core ./core
+COPY new_agent_memory ./new_agent_memory
+COPY hub ./hub
+COPY adapters ./adapters
+COPY forgotten_layer.py main.py memory_chunk.py memory_layer_core.py retrieval.py ./
+
+RUN python -m pip install --upgrade pip \
+    && python -m pip install -e .
+
+RUN useradd --create-home --shell /bin/bash agent \
+    && mkdir -p /app/memory_data \
+    && chown -R agent:agent /app
+
+USER agent
+
+VOLUME ["/app/memory_data"]
+
+ENTRYPOINT ["new-agent-memory"]
+CMD ["chat", "--runtime", "--max-steps", "4"]