lybic-guiagents 0.1.0__tar.gz → 0.2.0__tar.gz

{lybic_guiagents-0.1.0/lybic_guiagents.egg-info → lybic_guiagents-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lybic-guiagents
-Version: 0.1.0
+Version: 0.2.0
 Summary: An open-source agentic framework that enables AI to use computers like humans and can provide a multi-agent runtime environment as an infrastructure capability
 Author: Lybic Development Team
 Author-email: Lybic Development Team <lybic@tingyutech.com>
@@ -33,7 +33,7 @@ Requires-Dist: certifi
 Requires-Dist: pytesseract
 Requires-Dist: google-genai
 Requires-Dist: python-dotenv
-Requires-Dist: Pillow~=11.0.0
+Requires-Dist: Pillow>=11.3.0
 Requires-Dist: fabric
 Requires-Dist: gymnasium~=0.28.1
 Requires-Dist: requests~=2.31.0
@@ -88,9 +88,12 @@ Requires-Dist: azure-mgmt-network
 Requires-Dist: docker
 Requires-Dist: loguru
 Requires-Dist: dotenv
+Requires-Dist: lybic>=0.7.0
 Requires-Dist: pyobjc; platform_system == "Darwin"
 Requires-Dist: pywinauto; platform_system == "Windows"
 Requires-Dist: pywin32; platform_system == "Windows"
+Provides-Extra: dev
+Requires-Dist: pytest; extra == "dev"
 Dynamic: author
 Dynamic: license-file
 Dynamic: requires-python
@@ -109,6 +112,19 @@ Dynamic: requires-python
   Lybic GUI Agent: <small>An open-source agentic framework for Computer Use Agents</small> 
 </h1>
 
+<p align="center">
+    <small>Supported OS:</small>
+    <img src="https://img.shields.io/badge/OS-Windows-blue?logo=windows&logoColor=white" alt="Windows">
+    <img src="https://img.shields.io/badge/OS-macOS-black?logo=apple&logoColor=white" alt="macOS">
+    <img src="https://img.shields.io/badge/OS-Linux-yellow?logo=linux&logoColor=black" alt="Linux">
+    <br/>
+    <small>Latest Version:</small><a href="https://pypi.org/project/lybic-guiagents/"><img alt="PyPI" src="https://img.shields.io/pypi/v/lybic-guiagents"></a>
+    &nbsp;
+    <a href="https://github.com/lybic/agent/blob/main/LICENSE"><img alt="License" src="https://img.shields.io/pypi/l/lybic-guiagents"></a>
+    &nbsp;
+    <a href="https://github.com/lybic/agent"><img alt="Stars" src="https://img.shields.io/github/stars/lybic/agent?style=social"></a>
+</p>
+
 ## What is Lybic GUI Agent?
 
 Lybic platform placeholder - comprehensive AI platform for building and deploying intelligent agents
@@ -124,6 +140,7 @@ Lybic GUI Agent is an open-source framework that enables developers and business
 </div>
 
 ## 🥳 Updates
+- [x] **2025/09/14**: The paper has been accepted by [arxiv](https://arxiv.org/abs/2509.11067)
 - [x] **2025/09/09**: We achieved the world's first place in the 50-step length of [OS-world](https://os-world.github.io/)!
 - [x] **2025/08/08**: Released v0.1.0 of [Lybic GUI Agent](https://github.com/lybic/agent) library, with support for Windows, Mac, Ubuntu and Lybic API!
 
@@ -132,6 +149,8 @@ Lybic GUI Agent is an open-source framework that enables developers and business
 1. [💡 Introduction](#-introduction)
 2. [🛠️ Installation & Setup](#%EF%B8%8F-installation--setup) 
 3. [🚀 Usage](#-usage)
+4. [🔧 Troubleshooting](#-troubleshooting)
+5. [💬 Citations](#-citations)
 
 ## 💡 Introduction
 
@@ -161,7 +180,7 @@ Lybic GUI Agent is an open-source framework that enables developers and business
     <tr>
       <td>OSWorld Verified (50 step)</td>
       <td><b>57.1%</b></td>
-      <td>53.1%</td>
+      <td>54.2%</td>
     </tr>
   </table>
 </div>
@@ -176,7 +195,15 @@ Lybic GUI Agent is an open-source framework that enables developers and business
 > [!WARNING]
 > To leverage the full potential of Lybic GUI Agent, we support multiple model providers including OpenAI, Anthropic, Gemini, and Doubao. For the best visual grounding performance, we recommend using UI-TARS models.
 
-### Installation
+### Installation(from pip)
+
+You can install Lybic GUI Agent by using pip:
+
+```bash
+pip install lybic-guiagents
+```
+
+### Installation(from source code)
 
 You can use [UV](https://docs.astral.sh/uv/getting-started/installation/) (a modern Python package manager) version 0.8.5 for installation:
 
@@ -277,6 +304,7 @@ Run in interactive mode with the `lybic` backend:
 python gui_agents/cli_app.py --backend lybic
 ```
 
+
 Run a single query with the `pyautogui` backend and a maximum of 20 steps:
 ```sh
 python gui_agents/cli_app.py --backend pyautogui --query "Find the result of 8 × 7 on a calculator" --max-steps 20
@@ -308,6 +336,25 @@ LYBIC_MAX_LIFE_SECONDS=3600
 > LYBIC_PRECREATE_SID=SBX-XXXXXXXXXXXXXXX
 > ```
 
+### Use as a service
+
+After installing lybic-guiagents, you can run it as a service.
+
+Main Components:
+- AgentService: High-level service interface (recommended for most users)
+- AgentS2, AgentSFast: Core agent implementations  
+- HardwareInterface: Hardware abstraction layer
+- ServiceConfig: Configuration management
+
+Quick Start:
+
+```python
+from gui_agents import AgentService
+service = AgentService()
+result = service.execute_task("Take a screenshot")
+print(f"Task completed: {result.status}")
+```
+
 ### VMware Configuration
 
 To use PyAutoGUI with VMware, you need to install [VMware Workstation Pro](https://www.vmware.com/products/desktop-hypervisor/workstation-and-fusion) (on Windows) and create a virtual machine. 
@@ -358,6 +405,7 @@ USE_PRECREATE_VM=Ubuntu
   ```bash
   uv sync
   uv pip install -e .
+  # uv pip install -e . -i  https://pypi.tuna.tsinghua.edu.cn/simple
   ```
 
 #### 3. Lybic Sandbox Connection Issues
@@ -406,6 +454,23 @@ If you encounter issues not covered here:
    - Complete error messages
    - Steps to reproduce the issue
 
+## 💬 Citations
+
+If you find this codebase useful, please cite:
+
+```bibtex
+@misc{guo2025agenticlybicmultiagentexecution,
+      title={Agentic Lybic: Multi-Agent Execution System with Tiered Reasoning and Orchestration}, 
+      author={Liangxuan Guo and Bin Zhu and Qingqian Tao and Kangning Liu and Xun Zhao and Xianzhe Qin and Jin Gao and Guangfu Hao},
+      year={2025},
+      eprint={2509.11067},
+      archivePrefix={arXiv},
+      primaryClass={cs.AI},
+      url={https://arxiv.org/abs/2509.11067}, 
+}
+```
+
+
 ## Stargazers over time
 
 [![Stargazers over time](https://starchart.cc/lybic/agent.svg)](https://starchart.cc/lybic/agent)

{lybic_guiagents-0.1.0 → lybic_guiagents-0.2.0}/README.md

@@ -12,6 +12,19 @@
   Lybic GUI Agent: <small>An open-source agentic framework for Computer Use Agents</small> 
 </h1>
 
+<p align="center">
+    <small>Supported OS:</small>
+    <img src="https://img.shields.io/badge/OS-Windows-blue?logo=windows&logoColor=white" alt="Windows">
+    <img src="https://img.shields.io/badge/OS-macOS-black?logo=apple&logoColor=white" alt="macOS">
+    <img src="https://img.shields.io/badge/OS-Linux-yellow?logo=linux&logoColor=black" alt="Linux">
+    <br/>
+    <small>Latest Version:</small><a href="https://pypi.org/project/lybic-guiagents/"><img alt="PyPI" src="https://img.shields.io/pypi/v/lybic-guiagents"></a>
+    &nbsp;
+    <a href="https://github.com/lybic/agent/blob/main/LICENSE"><img alt="License" src="https://img.shields.io/pypi/l/lybic-guiagents"></a>
+    &nbsp;
+    <a href="https://github.com/lybic/agent"><img alt="Stars" src="https://img.shields.io/github/stars/lybic/agent?style=social"></a>
+</p>
+
 ## What is Lybic GUI Agent?
 
 Lybic platform placeholder - comprehensive AI platform for building and deploying intelligent agents
@@ -27,6 +40,7 @@ Lybic GUI Agent is an open-source framework that enables developers and business
 </div>
 
 ## 🥳 Updates
+- [x] **2025/09/14**: The paper has been accepted by [arxiv](https://arxiv.org/abs/2509.11067)
 - [x] **2025/09/09**: We achieved the world's first place in the 50-step length of [OS-world](https://os-world.github.io/)!
 - [x] **2025/08/08**: Released v0.1.0 of [Lybic GUI Agent](https://github.com/lybic/agent) library, with support for Windows, Mac, Ubuntu and Lybic API!
 
@@ -35,6 +49,8 @@ Lybic GUI Agent is an open-source framework that enables developers and business
 1. [💡 Introduction](#-introduction)
 2. [🛠️ Installation & Setup](#%EF%B8%8F-installation--setup) 
 3. [🚀 Usage](#-usage)
+4. [🔧 Troubleshooting](#-troubleshooting)
+5. [💬 Citations](#-citations)
 
 ## 💡 Introduction
 
@@ -64,7 +80,7 @@ Lybic GUI Agent is an open-source framework that enables developers and business
     <tr>
       <td>OSWorld Verified (50 step)</td>
       <td><b>57.1%</b></td>
-      <td>53.1%</td>
+      <td>54.2%</td>
     </tr>
   </table>
 </div>
@@ -79,7 +95,15 @@ Lybic GUI Agent is an open-source framework that enables developers and business
 > [!WARNING]
 > To leverage the full potential of Lybic GUI Agent, we support multiple model providers including OpenAI, Anthropic, Gemini, and Doubao. For the best visual grounding performance, we recommend using UI-TARS models.
 
-### Installation
+### Installation(from pip)
+
+You can install Lybic GUI Agent by using pip:
+
+```bash
+pip install lybic-guiagents
+```
+
+### Installation(from source code)
 
 You can use [UV](https://docs.astral.sh/uv/getting-started/installation/) (a modern Python package manager) version 0.8.5 for installation:
 
@@ -180,6 +204,7 @@ Run in interactive mode with the `lybic` backend:
 python gui_agents/cli_app.py --backend lybic
 ```
 
+
 Run a single query with the `pyautogui` backend and a maximum of 20 steps:
 ```sh
 python gui_agents/cli_app.py --backend pyautogui --query "Find the result of 8 × 7 on a calculator" --max-steps 20
@@ -211,6 +236,25 @@ LYBIC_MAX_LIFE_SECONDS=3600
 > LYBIC_PRECREATE_SID=SBX-XXXXXXXXXXXXXXX
 > ```
 
+### Use as a service
+
+After installing lybic-guiagents, you can run it as a service.
+
+Main Components:
+- AgentService: High-level service interface (recommended for most users)
+- AgentS2, AgentSFast: Core agent implementations  
+- HardwareInterface: Hardware abstraction layer
+- ServiceConfig: Configuration management
+
+Quick Start:
+
+```python
+from gui_agents import AgentService
+service = AgentService()
+result = service.execute_task("Take a screenshot")
+print(f"Task completed: {result.status}")
+```
+
 ### VMware Configuration
 
 To use PyAutoGUI with VMware, you need to install [VMware Workstation Pro](https://www.vmware.com/products/desktop-hypervisor/workstation-and-fusion) (on Windows) and create a virtual machine. 
@@ -261,6 +305,7 @@ USE_PRECREATE_VM=Ubuntu
   ```bash
   uv sync
   uv pip install -e .
+  # uv pip install -e . -i  https://pypi.tuna.tsinghua.edu.cn/simple
   ```
 
 #### 3. Lybic Sandbox Connection Issues
@@ -309,6 +354,23 @@ If you encounter issues not covered here:
    - Complete error messages
    - Steps to reproduce the issue
 
+## 💬 Citations
+
+If you find this codebase useful, please cite:
+
+```bibtex
+@misc{guo2025agenticlybicmultiagentexecution,
+      title={Agentic Lybic: Multi-Agent Execution System with Tiered Reasoning and Orchestration}, 
+      author={Liangxuan Guo and Bin Zhu and Qingqian Tao and Kangning Liu and Xun Zhao and Xianzhe Qin and Jin Gao and Guangfu Hao},
+      year={2025},
+      eprint={2509.11067},
+      archivePrefix={arXiv},
+      primaryClass={cs.AI},
+      url={https://arxiv.org/abs/2509.11067}, 
+}
+```
+
+
 ## Stargazers over time
 
 [![Stargazers over time](https://starchart.cc/lybic/agent.svg)](https://starchart.cc/lybic/agent)

lybic_guiagents-0.2.0/gui_agents/__init__.py

@@ -0,0 +1,67 @@
+"""
+GUI Agents - A comprehensive GUI automation framework
+
+This package provides both low-level agent components and a high-level service interface
+for GUI automation tasks across different platforms and backends.
+
+Main Components:
+- AgentService: High-level service interface (recommended for most users)
+- AgentS2, AgentSFast: Core agent implementations  
+- HardwareInterface: Hardware abstraction layer
+- ServiceConfig: Configuration management
+
+Quick Start:
+    from gui_agents import AgentService
+    
+    service = AgentService()
+    result = service.execute_task("Take a screenshot")
+    print(f"Task completed: {result.status}")
+"""
+
+# High-level service interface (recommended)
+from .service import (
+    AgentService,
+    ServiceConfig, 
+    TaskRequest,
+    TaskResult,
+    TaskStatus,
+    ExecutionStats,
+    AgentServiceError,
+    ConfigurationError,
+    TaskExecutionError
+)
+
+# Core agent classes (for advanced users)
+from .agents.agent_s import AgentS2, AgentSFast
+from .agents.hardware_interface import HardwareInterface
+from .store.registry import Registry
+from .agents.global_state import GlobalState
+
+try:
+    from importlib.metadata import version
+    __version__ = version("lybic-guiagents")
+except Exception:  # during editable installs or missing meta
+    __version__ = "0.0.0+dev"
+
+# Primary exports (what users should typically use)
+__all__ = [
+    # High-level service interface
+    "AgentService",
+    "ServiceConfig",
+    "TaskRequest", 
+    "TaskResult",
+    "TaskStatus", 
+    "ExecutionStats",
+    
+    # Exceptions
+    "AgentServiceError",
+    "ConfigurationError", 
+    "TaskExecutionError",
+    
+    # Core classes (for advanced usage)
+    "AgentS2",
+    "AgentSFast", 
+    "HardwareInterface",
+    "Registry",
+    "GlobalState",
+]

lybic_guiagents-0.2.0/gui_agents/agents/Backend/ADBBackend.py

@@ -0,0 +1,62 @@
+# ---------------------------------------------------------------------------
+# 2) Android device backend (ADB)
+# ---------------------------------------------------------------------------
+from gui_agents.agents.Action import (
+    Action,
+    Click,
+    Drag,
+    TypeText,
+    Scroll,
+    Hotkey,
+    Wait,
+)
+
+from gui_agents.agents.Backend.Backend import Backend
+import time
+import subprocess
+
+class ADBBackend(Backend):
+    """Very light‑weight ADB backend (tap / swipe / text / keyevent)."""
+
+    _supported = {Click, Drag, TypeText, Hotkey, Wait}
+
+    def __init__(self, serial: str | None = None):
+        self.serial = serial  # specify target device; None = default
+
+    # ------------------------------------------------------------------
+    def execute(self, action: Action) -> None:
+        if not self.supports(type(action)):
+            raise NotImplementedError
+
+        prefix = ["adb"]
+        if self.serial:
+            prefix += ["-s", self.serial]
+        prefix.append("shell")
+
+        if isinstance(action, Click):
+            cmd = prefix
+            # cmd = prefix + ["input", "tap", str(action.xy[0]), str(action.xy[1])]
+        elif isinstance(action, Drag):
+            cmd = prefix + [
+                "input", "swipe",
+                # str(action.start[0]), str(action.start[1]),
+                # str(action.end[0]), str(action.end[1]),
+                # str(int(action.duration * 1000)), # type: ignore
+            ]
+        elif isinstance(action, TypeText):
+            text = action.text.replace(" ", "%s")  # escape spaces
+            cmd = prefix + ["input", "text", text]
+            # if action.press_enter:
+            #     subprocess.run(prefix + ["input", "keyevent", "ENTER"], check=True)
+            #     return
+        elif isinstance(action, Hotkey):
+            # Map first key for demo purposes
+            key = action.keys[0].upper()
+            cmd = prefix + ["input", "keyevent", key]
+        elif isinstance(action, Wait):
+            time.sleep(action.seconds) # type: ignore
+            return
+        else:
+            raise NotImplementedError
+
+        subprocess.run(cmd, check=True)

lybic_guiagents-0.2.0/gui_agents/agents/Backend/Backend.py

@@ -0,0 +1,28 @@
+# Abstract backend base‑class
+# ---------------------------------------------------------------------------
+from abc import ABC, abstractmethod
+from typing import Any, List, Type, Dict, Set
+from gui_agents.agents.Action import (
+    Action
+)
+
+
+class Backend(ABC):
+    """Abstract base for platform back‑ends."""
+
+    #: Each backend advertises which Action subclasses it supports.
+    _supported: Set[Type[Action]] = set()
+
+    # ---------------------------------------------------------------------
+    def supports(self, action_type: Type[Action]) -> bool:
+        return action_type in self._supported
+
+    # ---------------------------------------------------------------------
+    @abstractmethod
+    def execute(self, action: Action) ->  Any:
+        """Translate an *Action* into concrete commands.
+
+        Should raise **NotImplementedError** if the *action* type is not in
+        `self._supported`, so upper layers can decide how to degrade / retry.
+        """
+