easy-agent-sdk 0.1.2__tar.gz

easy_agent_sdk-0.1.2/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Yiran Peng
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

easy_agent_sdk-0.1.2/MANIFEST.in

@@ -0,0 +1,6 @@
+include LICENSE
+include README.md
+include README_CN.md
+include requirements.txt
+recursive-include easyagent *.yaml
+recursive-include easyagent py.typed

easy_agent_sdk-0.1.2/PKG-INFO

@@ -0,0 +1,220 @@
+Metadata-Version: 2.4
+Name: easy-agent-sdk
+Version: 0.1.2
+Summary: A lightweight AI Agent framework built on LiteLLM with ReAct reasoning, tool calling, and smart memory.
+Author-email: Yiran Peng <amagipeng@gmail.com>
+Maintainer-email: Yiran Peng <amagipeng@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/SNHuan/EasyAgent
+Project-URL: Documentation, https://github.com/SNHuan/EasyAgent#readme
+Project-URL: Repository, https://github.com/SNHuan/EasyAgent
+Project-URL: Bug Tracker, https://github.com/SNHuan/EasyAgent/issues
+Project-URL: Changelog, https://github.com/SNHuan/EasyAgent/releases
+Keywords: agent,llm,react,ai,litellm,openai,claude,gpt,tool-calling
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Typing :: Typed
+Requires-Python: >=3.12
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: litellm>=1.80.0
+Requires-Dist: pydantic>=2.12.5
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.23; extra == "dev"
+Requires-Dist: build; extra == "dev"
+Requires-Dist: twine; extra == "dev"
+Dynamic: license-file
+
+# EasyAgent
+
+[![PyPI version](https://badge.fury.io/py/easy-agent-sdk.svg)](https://badge.fury.io/py/easy-agent-sdk)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+[![Python 3.12+](https://img.shields.io/badge/python-3.12+-blue.svg)](https://www.python.org/downloads/)
+
+English | [简体中文](README_CN.md)
+
+A lightweight AI Agent framework built on LiteLLM, featuring multi-model support, tool calling, and intelligent memory management.
+
+> **~809 lines of code, production-ready Agent capabilities** — Multi-model adapters, tool calling, smart memory, ReAct reasoning, DAG pipelines, debug tracing.
+
+## Features
+
+- **Multi-Model Support** - Unified interface via LiteLLM for OpenAI, Anthropic, Gemini, and more
+- **Tool Calling** - Protocol-based tool definition with `@register_tool` decorator
+- **Memory** - Sliding window + auto-summarization strategies for context management
+- **ReAct Loop** - Standard think → act → observe reasoning cycle
+- **DAG Pipeline** - Directed Acyclic Graph workflow orchestration with parallel execution
+- **Debug Friendly** - Colored logging, token usage and cost tracking
+
+## Installation
+
+```bash
+pip install easy-agent-sdk
+```
+
+**From source:**
+
+```bash
+git clone https://github.com/SNHuan/EasyAgent.git
+cd EasyAgent
+pip install -e ".[dev]"
+```
+
+## Quick Start
+
+### 1. Configuration
+
+Create a config file `config.yaml`:
+
+```yaml
+debug: true
+summary_model: gpt-4o-mini
+
+models:
+  gpt-4o-mini:
+    api_type: openai
+    base_url: https://api.openai.com/v1
+    api_key: sk-xxx
+```
+
+Set environment variable:
+
+```bash
+export EA_DEFAULT_CONFIG=/path/to/config.yaml
+```
+
+### 2. Define Tools
+
+```python
+from easyagent.tool import register_tool
+
+@register_tool
+class GetWeather:
+    name = "get_weather"
+    type = "function"
+    description = "Get the weather for a city."
+    parameters = {
+        "type": "object",
+        "properties": {"city": {"type": "string", "description": "City name"}},
+        "required": ["city"],
+    }
+
+    def init(self) -> None:
+        pass
+
+    def execute(self, city: str) -> str:
+        return f"The weather in {city} is sunny, 25°C."
+```
+
+### 3. Create Agent
+
+```python
+import asyncio
+from easyagent.agent import ReactAgent
+from easyagent.config.base import ModelConfig
+from easyagent.model.litellm_model import LiteLLMModel
+
+config = ModelConfig.load()
+model = LiteLLMModel(**config.get_model("gpt-4o-mini"))
+
+agent = ReactAgent(
+    model=model,
+    tools=["get_weather"],
+    system_prompt="You are a helpful assistant.",
+    max_iterations=10,
+)
+
+result = asyncio.run(agent.run("What's the weather in Beijing?"))
+print(result)
+```
+
+## Core Components
+
+### Agent
+
+| Class | Description |
+|-------|-------------|
+| `ReactAgent` | ReAct loop: think → act → observe |
+| `ToolAgent` | Tool registration and execution |
+
+### Memory
+
+```python
+from easyagent.memory import SlidingWindowMemory, SummaryMemory
+
+# Sliding window
+memory = SlidingWindowMemory(max_messages=20, max_tokens=4000)
+
+# Auto-summary for long tasks
+memory = SummaryMemory(task_id="task_001", reserve_ratio=0.3)
+```
+
+### Pipeline
+
+DAG-based workflow with parallel execution:
+
+```python
+import asyncio
+from easyagent.pipeline.base import BaseNode, BasePipeline, NodeContext
+
+class FetchData(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.data = "raw_data"
+
+class ProcessA(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.result_a = f"{ctx.data}_A"
+
+class ProcessB(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.result_b = f"{ctx.data}_B"
+
+fetch = FetchData()
+process_a = ProcessA()
+process_b = ProcessB()
+
+fetch >> [process_a, process_b]  # Parallel branches
+
+pipeline = BasePipeline(root=fetch)
+ctx = asyncio.run(pipeline.run())
+```
+
+### Debug
+
+```python
+from easyagent.debug.log import LogCollector, Logger
+
+log = Logger("MyApp")
+
+with LogCollector() as collector:
+    log.info("Step 1")
+    log.info("Step 2")
+
+print(collector.to_text())
+```
+
+## Project Structure
+
+```
+easyagent/
+├── agent/          # ReactAgent, ToolAgent
+├── model/          # LiteLLMModel, Message, ToolCall
+├── memory/         # SlidingWindowMemory, SummaryMemory
+├── tool/           # ToolManager, @register_tool
+├── pipeline/       # BaseNode, BasePipeline
+├── config/         # ModelConfig
+├── prompt/         # Prompt templates
+└── debug/          # Logger, LogCollector
+```
+
+## License
+
+[MIT License](LICENSE) © 2025 Yiran Peng

easy_agent_sdk-0.1.2/README.md

@@ -0,0 +1,185 @@
+# EasyAgent
+
+[![PyPI version](https://badge.fury.io/py/easy-agent-sdk.svg)](https://badge.fury.io/py/easy-agent-sdk)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+[![Python 3.12+](https://img.shields.io/badge/python-3.12+-blue.svg)](https://www.python.org/downloads/)
+
+English | [简体中文](README_CN.md)
+
+A lightweight AI Agent framework built on LiteLLM, featuring multi-model support, tool calling, and intelligent memory management.
+
+> **~809 lines of code, production-ready Agent capabilities** — Multi-model adapters, tool calling, smart memory, ReAct reasoning, DAG pipelines, debug tracing.
+
+## Features
+
+- **Multi-Model Support** - Unified interface via LiteLLM for OpenAI, Anthropic, Gemini, and more
+- **Tool Calling** - Protocol-based tool definition with `@register_tool` decorator
+- **Memory** - Sliding window + auto-summarization strategies for context management
+- **ReAct Loop** - Standard think → act → observe reasoning cycle
+- **DAG Pipeline** - Directed Acyclic Graph workflow orchestration with parallel execution
+- **Debug Friendly** - Colored logging, token usage and cost tracking
+
+## Installation
+
+```bash
+pip install easy-agent-sdk
+```
+
+**From source:**
+
+```bash
+git clone https://github.com/SNHuan/EasyAgent.git
+cd EasyAgent
+pip install -e ".[dev]"
+```
+
+## Quick Start
+
+### 1. Configuration
+
+Create a config file `config.yaml`:
+
+```yaml
+debug: true
+summary_model: gpt-4o-mini
+
+models:
+  gpt-4o-mini:
+    api_type: openai
+    base_url: https://api.openai.com/v1
+    api_key: sk-xxx
+```
+
+Set environment variable:
+
+```bash
+export EA_DEFAULT_CONFIG=/path/to/config.yaml
+```
+
+### 2. Define Tools
+
+```python
+from easyagent.tool import register_tool
+
+@register_tool
+class GetWeather:
+    name = "get_weather"
+    type = "function"
+    description = "Get the weather for a city."
+    parameters = {
+        "type": "object",
+        "properties": {"city": {"type": "string", "description": "City name"}},
+        "required": ["city"],
+    }
+
+    def init(self) -> None:
+        pass
+
+    def execute(self, city: str) -> str:
+        return f"The weather in {city} is sunny, 25°C."
+```
+
+### 3. Create Agent
+
+```python
+import asyncio
+from easyagent.agent import ReactAgent
+from easyagent.config.base import ModelConfig
+from easyagent.model.litellm_model import LiteLLMModel
+
+config = ModelConfig.load()
+model = LiteLLMModel(**config.get_model("gpt-4o-mini"))
+
+agent = ReactAgent(
+    model=model,
+    tools=["get_weather"],
+    system_prompt="You are a helpful assistant.",
+    max_iterations=10,
+)
+
+result = asyncio.run(agent.run("What's the weather in Beijing?"))
+print(result)
+```
+
+## Core Components
+
+### Agent
+
+| Class | Description |
+|-------|-------------|
+| `ReactAgent` | ReAct loop: think → act → observe |
+| `ToolAgent` | Tool registration and execution |
+
+### Memory
+
+```python
+from easyagent.memory import SlidingWindowMemory, SummaryMemory
+
+# Sliding window
+memory = SlidingWindowMemory(max_messages=20, max_tokens=4000)
+
+# Auto-summary for long tasks
+memory = SummaryMemory(task_id="task_001", reserve_ratio=0.3)
+```
+
+### Pipeline
+
+DAG-based workflow with parallel execution:
+
+```python
+import asyncio
+from easyagent.pipeline.base import BaseNode, BasePipeline, NodeContext
+
+class FetchData(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.data = "raw_data"
+
+class ProcessA(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.result_a = f"{ctx.data}_A"
+
+class ProcessB(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.result_b = f"{ctx.data}_B"
+
+fetch = FetchData()
+process_a = ProcessA()
+process_b = ProcessB()
+
+fetch >> [process_a, process_b]  # Parallel branches
+
+pipeline = BasePipeline(root=fetch)
+ctx = asyncio.run(pipeline.run())
+```
+
+### Debug
+
+```python
+from easyagent.debug.log import LogCollector, Logger
+
+log = Logger("MyApp")
+
+with LogCollector() as collector:
+    log.info("Step 1")
+    log.info("Step 2")
+
+print(collector.to_text())
+```
+
+## Project Structure
+
+```
+easyagent/
+├── agent/          # ReactAgent, ToolAgent
+├── model/          # LiteLLMModel, Message, ToolCall
+├── memory/         # SlidingWindowMemory, SummaryMemory
+├── tool/           # ToolManager, @register_tool
+├── pipeline/       # BaseNode, BasePipeline
+├── config/         # ModelConfig
+├── prompt/         # Prompt templates
+└── debug/          # Logger, LogCollector
+```
+
+## License
+
+[MIT License](LICENSE) © 2025 Yiran Peng

easy_agent_sdk-0.1.2/README_CN.md

@@ -0,0 +1,185 @@
+# EasyAgent
+
+[![PyPI version](https://badge.fury.io/py/easy-agent-sdk.svg)](https://badge.fury.io/py/easy-agent-sdk)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+[![Python 3.12+](https://img.shields.io/badge/python-3.12+-blue.svg)](https://www.python.org/downloads/)
+
+[English](README.md) | 中文
+
+轻量级 AI Agent 框架，基于 LiteLLM 构建，支持多模型、工具调用和智能记忆管理。
+
+> **~809 行代码，完整实现生产级 Agent 能力** — 多模型适配、工具调用、智能记忆、ReAct 推理、DAG 流水线、调试追踪。
+
+## 特性
+
+- **多模型支持** - 通过 LiteLLM 统一接口，支持 OpenAI、Anthropic、Gemini 等
+- **工具调用** - 基于 Protocol 的工具定义，`@register_tool` 装饰器自动注册
+- **记忆模块** - 滑动窗口 + 自动摘要两种策略，自动管理上下文长度
+- **ReAct 循环** - think → act → observe 标准推理循环
+- **DAG Pipeline** - 基于有向无环图的流水线编排，支持节点并行执行
+- **调试友好** - 彩色日志输出，token 消耗和成本追踪
+
+## 安装
+
+```bash
+pip install easy-agent-sdk
+```
+
+**从源码安装：**
+
+```bash
+git clone https://github.com/SNHuan/EasyAgent.git
+cd EasyAgent
+pip install -e ".[dev]"
+```
+
+## 快速开始
+
+### 1. 配置
+
+创建配置文件 `config.yaml`：
+
+```yaml
+debug: true
+summary_model: gpt-4o-mini
+
+models:
+  gpt-4o-mini:
+    api_type: openai
+    base_url: https://api.openai.com/v1
+    api_key: sk-xxx
+```
+
+设置环境变量：
+
+```bash
+export EA_DEFAULT_CONFIG=/path/to/config.yaml
+```
+
+### 2. 定义工具
+
+```python
+from easyagent.tool import register_tool
+
+@register_tool
+class GetWeather:
+    name = "get_weather"
+    type = "function"
+    description = "获取城市天气"
+    parameters = {
+        "type": "object",
+        "properties": {"city": {"type": "string", "description": "城市名"}},
+        "required": ["city"],
+    }
+
+    def init(self) -> None:
+        pass
+
+    def execute(self, city: str) -> str:
+        return f"{city}天气晴朗，25°C。"
+```
+
+### 3. 创建 Agent
+
+```python
+import asyncio
+from easyagent.agent import ReactAgent
+from easyagent.config.base import ModelConfig
+from easyagent.model.litellm_model import LiteLLMModel
+
+config = ModelConfig.load()
+model = LiteLLMModel(**config.get_model("gpt-4o-mini"))
+
+agent = ReactAgent(
+    model=model,
+    tools=["get_weather"],
+    system_prompt="你是一个有帮助的助手。",
+    max_iterations=10,
+)
+
+result = asyncio.run(agent.run("北京天气怎么样？"))
+print(result)
+```
+
+## 核心组件
+
+### Agent
+
+| 类 | 说明 |
+|---|------|
+| `ReactAgent` | ReAct 循环：think → act → observe |
+| `ToolAgent` | 工具注册和执行 |
+
+### Memory
+
+```python
+from easyagent.memory import SlidingWindowMemory, SummaryMemory
+
+# 滑动窗口
+memory = SlidingWindowMemory(max_messages=20, max_tokens=4000)
+
+# 自动摘要（长任务）
+memory = SummaryMemory(task_id="task_001", reserve_ratio=0.3)
+```
+
+### Pipeline
+
+基于 DAG 的流水线，支持并行执行：
+
+```python
+import asyncio
+from easyagent.pipeline.base import BaseNode, BasePipeline, NodeContext
+
+class FetchData(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.data = "raw_data"
+
+class ProcessA(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.result_a = f"{ctx.data}_A"
+
+class ProcessB(BaseNode):
+    async def execute(self, ctx: NodeContext) -> None:
+        ctx.result_b = f"{ctx.data}_B"
+
+fetch = FetchData()
+process_a = ProcessA()
+process_b = ProcessB()
+
+fetch >> [process_a, process_b]  # 并行分支
+
+pipeline = BasePipeline(root=fetch)
+ctx = asyncio.run(pipeline.run())
+```
+
+### Debug
+
+```python
+from easyagent.debug.log import LogCollector, Logger
+
+log = Logger("MyApp")
+
+with LogCollector() as collector:
+    log.info("Step 1")
+    log.info("Step 2")
+
+print(collector.to_text())
+```
+
+## 项目结构
+
+```
+easyagent/
+├── agent/          # ReactAgent, ToolAgent
+├── model/          # LiteLLMModel, Message, ToolCall
+├── memory/         # SlidingWindowMemory, SummaryMemory
+├── tool/           # ToolManager, @register_tool
+├── pipeline/       # BaseNode, BasePipeline
+├── config/         # ModelConfig
+├── prompt/         # Prompt 模板
+└── debug/          # Logger, LogCollector
+```
+
+## 许可证
+
+[MIT License](LICENSE) © 2025 Yiran Peng