Ling0925
/
suna
mirror of https://github.com/kortix-ai/suna.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

README 0.1.8

This commit is contained in:
marko-kraemer 2024-11-18 07:14:40 +01:00
parent b20c074ede
commit 9935daabe3
5 changed files with 155 additions and 513 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
CHANGELOG.md
View File

@ -1,9 +1,14 @@
0.1.8 0.1.8
- Tool Parser Base Class - Added base processor classes for extensible tool handling:
- Tool Executor Base Class - ToolParserBase: Abstract base class for parsing LLM responses
- Execute_tools_on_stream – execute tools while the response is streaming - ToolExecutorBase: Abstract base class for tool execution strategies
- Docstring docs - ResultsAdderBase: Abstract base class for managing results
- Added dual support for OpenAPI and XML tool calling patterns:
- XML schema decorator for XML-based tool definitions
- XML-specific processors for parsing and execution
- Standard processors for OpenAPI function calling
- Enhanced streaming capabilities:
- execute_tools_on_stream: Execute tools in real-time during streaming
0.1.7 0.1.7
- Streaming Responses with Tool Calls - v1 streaming responses

196
README.md
View File

@ -2,10 +2,47 @@
AgentPress is a collection of _simple, but powerful_ utilities that serve as building blocks for creating AI agents. *Plug, play, and customize.* AgentPress is a collection of _simple, but powerful_ utilities that serve as building blocks for creating AI agents. *Plug, play, and customize.*
- **Threads**: Simple message thread handling utilities ## How It Works
- **Tools**: Flexible tool definition and automatic execution
- **State Management**: Simple JSON key-value state management Each AI agent iteration follows a clear, modular flow:
1. **Message & LLM Handling**
- Messages are managed in threads via `ThreadManager`
- LLM API calls are made through a unified interface (`llm.py`)
- Supports streaming responses for real-time interaction
2. **Response Processing**
- LLM returns both content and tool calls
- Content is streamed in real-time
- Tool calls are parsed using either:
- Standard OpenAPI function calling
- XML-based tool definitions
- Custom parsers (extend `ToolParserBase`)
3. **Tool Execution**
- Tools are executed either:
- In real-time during streaming (`execute_tools_on_stream`)
- After complete response
- In parallel or sequential order
- Supports both standard and XML tool formats
- Extensible through `ToolExecutorBase`
4. **Results Management**
- Results from both content and tool executions are handled
- Supports different result formats (standard/XML)
- Customizable through `ResultsAdderBase`
This modular architecture allows you to:
- Use standard OpenAPI function calling
- Switch to XML-based tool definitions
- Create custom processors by extending base classes
- Mix and match different approaches
- **Threads**: Simple message thread handling utilities with streaming support
- **Tools**: Flexible tool definition with both OpenAPI and XML formats
- **State Management**: Thread-safe JSON key-value state management
- **LLM Integration**: Provider-agnostic LLM calls via LiteLLM - **LLM Integration**: Provider-agnostic LLM calls via LiteLLM
- **Response Processing**: Support for both standard and XML-based tool calling
## Installation & Setup ## Installation & Setup
@ -19,7 +56,7 @@ pip install agentpress
agentpress init agentpress init
``` ```
Creates a `agentpress` directory with all the core utilities. Creates a `agentpress` directory with all the core utilities.
Check out [File Overview](#file-overview) for explanations of the generated util files. Check out [File Overview](#file-overview) for explanations of the generated files.
3. If you selected the example agent during initialization: 3. If you selected the example agent during initialization:
- Creates an `agent.py` file with a web development agent example - Creates an `agent.py` file with a web development agent example
@ -31,24 +68,31 @@ Check out [File Overview](#file-overview) for explanations of the generated util
## Quick Start ## Quick Start
1. Set up your environment variables (API keys, etc.) in a `.env` file. 1. Set up your environment variables in a `.env` file:
- OPENAI_API_KEY, ANTHROPIC_API_KEY, GROQ_API_KEY, etc... Whatever LLM you want to use, we use LiteLLM (https://litellm.ai) (Call 100+ LLMs using the OpenAI Input/Output Format) – set it up in your `.env` file.. Also check out the agentpress/llm.py and modify as needed to support your wanted LLM. ```bash
OPENAI_API_KEY=your_key_here
ANTHROPIC_API_KEY=your_key_here
GROQ_API_KEY=your_key_here
```
2. Create a calculator_tool.py 2. Create a calculator tool with OpenAPI schema:
```python ```python
from agentpress.tool import Tool, ToolResult, tool_schema from agentpress.tool import Tool, ToolResult, openapi_schema
class CalculatorTool(Tool): class CalculatorTool(Tool):
@tool_schema({ @openapi_schema({
"name": "add", "type": "function",
"description": "Add two numbers", "function": {
"parameters": { "name": "add",
"type": "object", "description": "Add two numbers",
"properties": { "parameters": {
"a": {"type": "number"}, "type": "object",
"b": {"type": "number"} "properties": {
}, "a": {"type": "number"},
"required": ["a", "b"] "b": {"type": "number"}
},
"required": ["a", "b"]
}
} }
}) })
async def add(self, a: float, b: float) -> ToolResult: async def add(self, a: float, b: float) -> ToolResult:
@ -59,7 +103,29 @@ class CalculatorTool(Tool):
return self.fail_response(f"Failed to add numbers: {str(e)}") return self.fail_response(f"Failed to add numbers: {str(e)}")
``` ```
3. Use the Thread Manager, create a new thread – or access an existing one. Then Add the Calculator Tool, and run the thread. It will automatically use & execute the python function associated with the tool: 3. Or create a tool with XML schema:
```python
from agentpress.tool import Tool, ToolResult, xml_schema
class FilesTool(Tool):
@xml_schema(
tag_name="create-file",
mappings=[
{"param_name": "file_path", "node_type": "attribute", "path": "."},
{"param_name": "file_contents", "node_type": "content", "path": "."}
],
example='''
<create-file file_path="path/to/file">
File contents go here
</create-file>
'''
)
async def create_file(self, file_path: str, file_contents: str) -> ToolResult:
# Implementation here
pass
```
4. Use the Thread Manager with streaming and tool execution:
```python ```python
import asyncio import asyncio
from agentpress.thread_manager import ThreadManager from agentpress.thread_manager import ThreadManager
@ -71,67 +137,93 @@ async def main():
manager.add_tool(CalculatorTool) manager.add_tool(CalculatorTool)
# Create a new thread # Create a new thread
# Alternatively, you could use an existing thread_id like:
# thread_id = "existing-thread-uuid"
thread_id = await manager.create_thread() thread_id = await manager.create_thread()
# Add your custom logic here # Add your message
await manager.add_message(thread_id, { await manager.add_message(thread_id, {
"role": "user", "role": "user",
"content": "What's 2 + 2?" "content": "What's 2 + 2?"
}) })
# Run with streaming and tool execution
response = await manager.run_thread( response = await manager.run_thread(
thread_id=thread_id, thread_id=thread_id,
system_message={ system_message={
"role": "system", "role": "system",
"content": "You are a helpful assistant with calculation abilities." "content": "You are a helpful assistant with calculation abilities."
}, },
model_name="gpt-4", model_name="anthropic/claude-3-5-sonnet-latest",
use_tools=True, stream=True,
execute_tool_calls=True native_tool_calling=True,
execute_tools=True,
execute_tools_on_stream=True
) )
print("Response:", response)
# Handle streaming response
if isinstance(response, AsyncGenerator):
async for chunk in response:
if hasattr(chunk.choices[0], 'delta'):
delta = chunk.choices[0].delta
if hasattr(delta, 'content') and delta.content:
print(delta.content, end='', flush=True)
asyncio.run(main()) asyncio.run(main())
``` ```
4. Autonomous Web Developer Agent (the standard example) 5. View conversation threads in a web UI:
When you run `agentpress init` and select the example agent – you will get code for a simple implementation of an AI Web Developer Agent that leverages architecture similar to platforms like our own [Softgen](https://softgen.ai/) Platform.
- **Files Tool**: Allows the agent to create, read, update, and delete files within the workspace.
- **Terminal Tool**: Enables the agent to execute terminal commands.
- **State Workspace Management**: The agent has access to a workspace whose state is stored and sent on every request. This state includes all file contents, ensuring the agent knows what it is editing.
- **User Interaction via CLI**: After each action, the agent pauses and allows the user to provide further instructions through the CLI.
You can find the complete implementation in our [example-agent](agentpress/examples/example-agent/agent.py) directory.
5. Thread Viewer
Run the thread viewer to view messages of threads in a stylised web UI:
```bash ```bash
streamlit run agentpress/thread_viewer_ui.py streamlit run agentpress/thread_viewer_ui.py
``` ```
## File Overview ## File Overview
### agentpress/llm.py ### Core Components
Core LLM API interface using LiteLLM. Supports 100+ LLMs using the OpenAI Input/Output Format. Easy to extend for custom model configurations and API endpoints. `make_llm_api_call()` can be imported to make LLM calls.
### agentpress/thread_manager.py #### agentpress/llm.py
Orchestrates conversations between users, LLMs, and tools. Manages message history and automatically handles tool execution when LLMs request them. Tools registered here become available for LLM function calls. LLM API interface using LiteLLM. Supports 100+ LLMs with OpenAI-compatible format. Includes streaming, retry logic, and error handling.
### agentpress/tool.py #### agentpress/thread_manager.py
Base infrastructure for LLM-compatible tools. Inherit from `Tool` class and use `@tool_schema` decorator to create tools that are automatically registered for LLM function calling. Returns standardized `ToolResult` responses. Manages conversation threads with support for:
- Message history management
- Tool registration and execution
- Streaming responses
- Both OpenAPI and XML tool calling patterns
### agentpress/tool_registry.py #### agentpress/tool.py
Central registry for tool management. Keeps track of available tools and their schemas, allowing selective function registration. Works with `thread_manager.py` to expose tools to LLMs. Base infrastructure for tools with:
- OpenAPI schema decorator for standard function calling
- XML schema decorator for XML-based tool calls
- Standardized ToolResult responses
### agentpress/state_manager.py #### agentpress/tool_registry.py
Simple key-value based state persistence using JSON files. For maintaining environment state, settings, or other persistent data. Central registry for tool management:
- Registers both OpenAPI and XML tools
- Maintains tool schemas and implementations
- Provides tool lookup and validation
#### agentpress/state_manager.py
Thread-safe state persistence:
- JSON-based key-value storage
- Atomic operations with locking
- Automatic file handling
### Response Processing
#### agentpress/llm_response_processor.py
Handles LLM response processing with support for:
- Streaming and complete responses
- Tool call extraction and execution
- Result formatting and message management
#### Standard Processing
- `standard_tool_parser.py`: Parses OpenAPI function calls
- `standard_tool_executor.py`: Executes standard tool calls
- `standard_results_adder.py`: Manages standard results
#### XML Processing
- `xml_tool_parser.py`: Parses XML-formatted tool calls
- `xml_tool_executor.py`: Executes XML tool calls
- `xml_results_adder.py`: Manages XML results
## Philosophy ## Philosophy
- **Plug & Play**: Start with our defaults, then customize to your needs. - **Plug & Play**: Start with our defaults, then customize to your needs.
@ -160,7 +252,7 @@ pip install poetry
poetry install poetry install
``` ```
3. For quick testing, you can install directly from the current directory: 3. For quick testing:
```bash ```bash
pip install -e . pip install -e .
``` ```

94
agentpress/agents/simple_web_dev/workspace/index.html
View File

@ -1,94 +0,0 @@
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Modern Landing Page</title>
<link rel="stylesheet" href="styles.css">
</head>
<body>
<header>
<nav class="navbar">
<div class="logo">Brand</div>
<ul class="nav-links">
<li><a href="#home">Home</a></li>
<li><a href="#features">Features</a></li>
<li><a href="#contact">Contact</a></li>
</ul>
<div class="hamburger">
<span></span>
<span></span>
<span></span>
</div>
</nav>
</header>
<main>
<section id="home" class="hero">
<div class="hero-content">
<h1>Welcome to the Future</h1>
<p>Experience innovation at its finest</p>
<button class="cta-button">Get Started</button>
</div>
</section>
<section id="testimonials" class="testimonials">
<h2>What Our Clients Say</h2>
<div class="testimonial-grid">
<div class="testimonial-card">
<div class="testimonial-avatar">👤</div>
<p class="testimonial-text">"Amazing service! The team went above and beyond."</p>
<p class="testimonial-author">- John Doe, CEO</p>
</div>
<div class="testimonial-card">
<div class="testimonial-avatar">👤</div>
<p class="testimonial-text">"Incredible results. Would highly recommend!"</p>
<p class="testimonial-author">- Jane Smith, Designer</p>
</div>
<div class="testimonial-card">
<div class="testimonial-avatar">👤</div>
<p class="testimonial-text">"Professional and efficient service."</p>
<p class="testimonial-author">- Mike Johnson, Developer</p>
</div>
</div>
</section>
<section id="features" class="features">
<h2>Our Features</h2>
<div class="feature-grid">
<div class="feature-card">
<div class="feature-icon">🚀</div>
<h3>Fast Performance</h3>
<p>Lightning-quick loading times</p>
</div>
<div class="feature-card">
<div class="feature-icon">🎨</div>
<h3>Beautiful Design</h3>
<p>Stunning visuals and animations</p>
</div>
<div class="feature-card">
<div class="feature-icon">📱</div>
<h3>Responsive</h3>
<p>Works on all devices</p>
</div>
</div>
</section>
<section id="contact" class="contact">
<h2>Contact Us</h2>
<form id="contact-form">
<input type="text" placeholder="Name" required>
<input type="email" placeholder="Email" required>
<textarea placeholder="Message" required></textarea>
<button type="submit">Send Message</button>
</form>
</section>
</main>
<footer>
<p>&copy; 2024 Brand. All rights reserved.</p>
</footer>
<script src="script.js"></script>
</body>
</html>

69
agentpress/agents/simple_web_dev/workspace/script.js
View File

@ -1,69 +0,0 @@
document.addEventListener('DOMContentLoaded', () => {
const handleIntersection = (entries, observer) => {
entries.forEach(entry => {
if (entry.isIntersecting) {
entry.target.classList.add('animate');
observer.unobserve(entry.target);
}
});
};
const observerOptions = {
threshold: 0.2,
rootMargin: '0px'
};
const animationObserver = new IntersectionObserver(handleIntersection, observerOptions);
document.querySelectorAll('.feature-card, .testimonial-card').forEach(element => {
animationObserver.observe(element);
});
const hamburger = document.querySelector('.hamburger');
const navLinks = document.querySelector('.nav-links');
const links = document.querySelectorAll('.nav-links a');
hamburger.addEventListener('click', () => {
navLinks.classList.toggle('active');
});
links.forEach(link => {
link.addEventListener('click', (e) => {
e.preventDefault();
const targetId = link.getAttribute('href');
const targetSection = document.querySelector(targetId);
targetSection.scrollIntoView({
behavior: 'smooth'
});
if (window.innerWidth <= 768) {
navLinks.style.display = 'none';
}
});
});
const contactForm = document.getElementById('contact-form');
contactForm.addEventListener('submit', (e) => {
e.preventDefault();
const formData = new FormData(contactForm);
const formObject = Object.fromEntries(formData);
alert('Message sent successfully!');
contactForm.reset();
});
const observer = new IntersectionObserver((entries) => {
entries.forEach(entry => {
if (entry.isIntersecting) {
entry.target.style.opacity = '1';
entry.target.style.transform = 'translateY(0)';
}
});
}, { threshold: 0.1 });
document.querySelectorAll('.feature-card').forEach(card => {
card.style.opacity = '0';
card.style.transform = 'translateY(20px)';
observer.observe(card);
});
});

292
agentpress/agents/simple_web_dev/workspace/styles.css
View File

@ -1,292 +0,0 @@
:root {
--primary-color: #2563eb;
--secondary-color: #1e40af;
--text-color: #1f2937;
--background-color: #ffffff;
--accent-color: #dbeafe;
}
* {
margin: 0;
padding: 0;
box-sizing: border-box;
}
body {
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif;
line-height: 1.6;
color: var(--text-color);
}
.navbar {
display: flex;
justify-content: space-between;
align-items: center;
padding: 1rem 5%;
position: fixed;
width: 100%;
background: rgba(255, 255, 255, 0.95);
backdrop-filter: blur(10px);
z-index: 1000;
}
.logo {
font-size: 1.5rem;
font-weight: bold;
color: var(--primary-color);
}
.nav-links {
display: flex;
gap: 2rem;
list-style: none;
}
@media (max-width: 768px) {
.nav-links {
position: fixed;
top: 70px;
left: 0;
right: 0;
flex-direction: column;
background: rgba(255, 255, 255, 0.98);
padding: 2rem;
gap: 1.5rem;
text-align: center;
transform: translateY(-100%);
transition: transform 0.3s ease;
}
.nav-links.active {
transform: translateY(0);
}
}
.nav-links a {
text-decoration: none;
color: var(--text-color);
transition: color 0.3s ease;
}
.nav-links a:hover {
color: var(--primary-color);
}
.hamburger {
display: none;
flex-direction: column;
gap: 4px;
cursor: pointer;
}
.hamburger span {
width: 25px;
height: 3px;
background: var(--text-color);
transition: 0.3s ease;
}
.hero {
min-height: 100vh;
display: flex;
align-items: center;
justify-content: center;
background: linear-gradient(135deg, var(--accent-color), var(--background-color));
padding: 2rem;
}
.hero-content {
text-align: center;
max-width: 800px;
}
.hero h1 {
font-size: 3.5rem;
margin-bottom: 1rem;
animation: fadeInUp 1s ease;
}
.hero p {
font-size: 1.25rem;
margin-bottom: 2rem;
animation: fadeInUp 1s ease 0.2s;
opacity: 0;
animation-fill-mode: forwards;
}
.cta-button {
padding: 1rem 2rem;
font-size: 1.1rem;
background: var(--primary-color);
color: white;
border: none;
border-radius: 5px;
cursor: pointer;
transition: background 0.3s ease;
animation: fadeInUp 1s ease 0.4s;
opacity: 0;
animation-fill-mode: forwards;
}
.cta-button:hover {
background: var(--secondary-color);
}
.features {
padding: 5rem 2rem;
background: var(--background-color);
}
.features h2 {
text-align: center;
margin-bottom: 3rem;
}
.feature-grid {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(250px, 1fr));
gap: 2rem;
max-width: 1200px;
margin: 0 auto;
}
.feature-card {
padding: 2rem;
text-align: center;
background: white;
border-radius: 10px;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
transition: transform 0.3s ease;
}
.feature-card:hover {
transform: translateY(-5px);
}
.testimonials {
padding: 5rem 2rem;
background: linear-gradient(135deg, var(--accent-color) 0%, var(--background-color) 100%);
}
.testimonial-grid {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(300px, 1fr));
gap: 2rem;
max-width: 1200px;
margin: 0 auto;
}
.testimonial-card {
background: white;
padding: 2rem;
border-radius: 10px;
box-shadow: 0 4px 15px rgba(0, 0, 0, 0.1);
text-align: center;
transition: transform 0.3s ease;
}
.testimonial-card:hover {
transform: translateY(-5px);
}
.testimonial-avatar {
font-size: 3rem;
margin-bottom: 1rem;
}
.testimonial-text {
font-style: italic;
margin-bottom: 1rem;
color: var(--text-color);
}
.testimonial-author {
font-weight: bold;
color: var(--primary-color);
}
.feature-icon {
font-size: 2.5rem;
margin-bottom: 1rem;
}
.contact {
padding: 5rem 2rem;
background: var(--accent-color);
}
.contact h2 {
text-align: center;
margin-bottom: 3rem;
}
#contact-form {
display: flex;
flex-direction: column;
gap: 1rem;
max-width: 600px;
margin: 0 auto;
}
#contact-form input,
#contact-form textarea {
padding: 1rem;
border: 1px solid #ddd;
border-radius: 5px;
font-size: 1rem;
}
#contact-form textarea {
height: 150px;
resize: vertical;
}
#contact-form button {
padding: 1rem;
background: var(--primary-color);
color: white;
border: none;
border-radius: 5px;
cursor: pointer;
transition: background 0.3s ease;
}
#contact-form button:hover {
background: var(--secondary-color);
}
footer {
text-align: center;
padding: 2rem;
background: var(--text-color);
color: white;
}
@keyframes fadeInUp {
from {
opacity: 0;
transform: translateY(20px);
}
to {
opacity: 1;
transform: translateY(0);
}
}
@media (max-width: 768px) {
.nav-links {
display: none;
}
.hamburger {
display: flex;
}
.hero h1 {
font-size: 2.5rem;
}
.feature-grid {
grid-template-columns: 1fr;
}
}