# FutureOS 🚀
My take on how Operating Systems will be built in the future.
<img src="https://github.com/user-attachments/assets/4bdf7e12-0f9b-4f68-a588-4aed67ec6d45" alt="FutureOS Interface" />
## 🚀 Getting Started
```bash
# Install
pip install futureos
# Launch
python -m futureos
```
> [!IMPORTANT]
> To use FutureOS, ensure you have Ollama installed with the `gemma2:2b` model. For optimal performance, CUDA should be enabled on your system.
## 🌟 The Challenge
Currently, there are several powerful AI tools for both Windows and macOS:
- 🤖 [Claude computer use](https://www.anthropic.com/news/3-5-models-and-computer-use)
- 👨💻 [GitHub Copilot](https://copilot.github.com)
- 🗣️ New Siri on Mac
However, these tools implement their own methods of indexing and searching files, and they're not integrated with the OS.
## 💡 The Vision
What if:
- Every utility in the OS had access to `vector search` for collections of all `files`, `directories`, `pictures`, and more?
- All content was automatically indexed by the OS whenever a file is created or modified?
- Every utility had access to an `AI chain` (like LangChain) deeply integrated into the OS?
- Every command could leverage these AI capabilities?
Let's go further... 🚀
Why force users to remember specific commands with their parameters and arguments? By creating descriptive docstrings for commands, we can build a vector store of all commands and automatically choose the best one for user input. Users don't even need to know exact command names!
## ✅ Quality Assurance
"But without proper testing, users wouldn't like it."
I know! That's why there are [multiple tests](/tests/test_create_collections.py) for every command with output that's easy to parse by LLM.
## 🏗️ Architecture
### Command System
Every command inherits from a base `Command` class that provides:
```python
def get_file(self, question: str, max_distance: float) -> str:
"""Find relevant files using vector similarity search"""
def get_directory(self, question: str, max_distance: float) -> str:
"""Find relevant directories using vector similarity search"""
def run_chain(self, chain: RunnablePassthrough, input: dict) -> Any:
"""Execute LangChain operations for content processing"""
```
### 🔍 How Commands Work
Let's break down how a typical command (cd) is constructed:
```python
class cd(Command):
"""
NAME
cd - change the current working directory
SYNOPSIS
cd [directory]
DESCRIPTION
Change the current working directory to the specified directory.
NATURAL LANGUAGE COMMANDS
- Change directory to X
- Go to folder X
- Navigate to directory X
"""
```
The docstring serves multiple purposes:
1. **Vector Embedding Source**: Used to create embeddings for command matching
2. **Natural Language Patterns**: Defines common ways users might express the command
3. **Documentation**: Serves as built-in help text
### Command Implementation
```python
def execute(self, args: Any) -> None:
try:
# Handle both traditional and natural language input
directory = self.get_directory(args.query) if args.query else args.directory
# Resolve and validate path
target_path = resolve_path(directory)
# Execute directory change
if target_path.is_dir():
constants.CURRENT_DIRECTORY = target_path
self.print(f"Changed to {target_path}", style="green")
except Exception as e:
self.print(f"Error: {str(e)}", style="red")
```
## 🤝 Contributing
All contributors are welcome!
## 📄 License
MIT License - See [LICENSE](LICENSE) file for details.
Raw data
{
"_id": null,
"home_page": null,
"name": "futureos",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": null,
"keywords": "AI, Operating System, ollama, python, chromadb, langchain",
"author": null,
"author_email": "Michal Skibinski <michalskibinski109@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/68/4a/1be21cbc19e4f1e4cee169a0ea795bafd0469e3d73e4762408041aed4cf6/futureos-0.3.0.tar.gz",
"platform": null,
"description": "\r\n\r\n\r\n\r\n# FutureOS \ud83d\ude80\r\n\r\nMy take on how Operating Systems will be built in the future.\r\n\r\n<img src=\"https://github.com/user-attachments/assets/4bdf7e12-0f9b-4f68-a588-4aed67ec6d45\" alt=\"FutureOS Interface\" />\r\n\r\n## \ud83d\ude80 Getting Started\r\n\r\n```bash\r\n# Install\r\npip install futureos\r\n\r\n# Launch\r\npython -m futureos\r\n```\r\n\r\n> [!IMPORTANT] \r\n> To use FutureOS, ensure you have Ollama installed with the `gemma2:2b` model. For optimal performance, CUDA should be enabled on your system.\r\n\r\n## \ud83c\udf1f The Challenge\r\n\r\nCurrently, there are several powerful AI tools for both Windows and macOS:\r\n\r\n- \ud83e\udd16 [Claude computer use](https://www.anthropic.com/news/3-5-models-and-computer-use)\r\n- \ud83d\udc68\u200d\ud83d\udcbb [GitHub Copilot](https://copilot.github.com)\r\n- \ud83d\udde3\ufe0f New Siri on Mac\r\n\r\nHowever, these tools implement their own methods of indexing and searching files, and they're not integrated with the OS.\r\n\r\n## \ud83d\udca1 The Vision\r\n\r\nWhat if:\r\n\r\n- Every utility in the OS had access to `vector search` for collections of all `files`, `directories`, `pictures`, and more?\r\n- All content was automatically indexed by the OS whenever a file is created or modified?\r\n- Every utility had access to an `AI chain` (like LangChain) deeply integrated into the OS?\r\n- Every command could leverage these AI capabilities?\r\n\r\nLet's go further... \ud83d\ude80\r\n\r\nWhy force users to remember specific commands with their parameters and arguments? By creating descriptive docstrings for commands, we can build a vector store of all commands and automatically choose the best one for user input. Users don't even need to know exact command names!\r\n\r\n## \u2705 Quality Assurance\r\n\r\n\"But without proper testing, users wouldn't like it.\"\r\n\r\nI know! That's why there are [multiple tests](/tests/test_create_collections.py) for every command with output that's easy to parse by LLM.\r\n\r\n## \ud83c\udfd7\ufe0f Architecture\r\n\r\n### Command System\r\n\r\nEvery command inherits from a base `Command` class that provides:\r\n\r\n```python\r\ndef get_file(self, question: str, max_distance: float) -> str:\r\n \"\"\"Find relevant files using vector similarity search\"\"\"\r\n\r\ndef get_directory(self, question: str, max_distance: float) -> str:\r\n \"\"\"Find relevant directories using vector similarity search\"\"\"\r\n\r\ndef run_chain(self, chain: RunnablePassthrough, input: dict) -> Any:\r\n \"\"\"Execute LangChain operations for content processing\"\"\"\r\n```\r\n\r\n### \ud83d\udd0d How Commands Work\r\n\r\nLet's break down how a typical command (cd) is constructed:\r\n\r\n```python\r\nclass cd(Command):\r\n \"\"\"\r\n NAME\r\n cd - change the current working directory\r\n\r\n SYNOPSIS\r\n cd [directory]\r\n\r\n DESCRIPTION\r\n Change the current working directory to the specified directory.\r\n\r\n NATURAL LANGUAGE COMMANDS\r\n - Change directory to X\r\n - Go to folder X\r\n - Navigate to directory X\r\n \"\"\"\r\n```\r\n\r\nThe docstring serves multiple purposes:\r\n\r\n1. **Vector Embedding Source**: Used to create embeddings for command matching\r\n2. **Natural Language Patterns**: Defines common ways users might express the command\r\n3. **Documentation**: Serves as built-in help text\r\n\r\n### Command Implementation\r\n\r\n```python\r\ndef execute(self, args: Any) -> None:\r\n try:\r\n # Handle both traditional and natural language input\r\n directory = self.get_directory(args.query) if args.query else args.directory\r\n\r\n # Resolve and validate path\r\n target_path = resolve_path(directory)\r\n\r\n # Execute directory change\r\n if target_path.is_dir():\r\n constants.CURRENT_DIRECTORY = target_path\r\n self.print(f\"Changed to {target_path}\", style=\"green\")\r\n except Exception as e:\r\n self.print(f\"Error: {str(e)}\", style=\"red\")\r\n```\r\n\r\n## \ud83e\udd1d Contributing\r\n\r\nAll contributors are welcome!\r\n\r\n## \ud83d\udcc4 License\r\n\r\nMIT License - See [LICENSE](LICENSE) file for details.\r\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "First fully AI based Operating system ",
"version": "0.3.0",
"project_urls": {
"Homepage": "https://github.com/miskibin/futureOS",
"Repository": "https://github.com/miskibin/futureOS"
},
"split_keywords": [
"ai",
" operating system",
" ollama",
" python",
" chromadb",
" langchain"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "1c28086d3ebf330b9577011a9c5d7bd933cc385f8529f5f90c255086cd828890",
"md5": "7ca34f1ead1eeee467791ad72a7658c2",
"sha256": "5ea9d0be2a682e1f35468543e9abffe3ff986e580520582cd12809a8e6a32813"
},
"downloads": -1,
"filename": "futureos-0.3.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "7ca34f1ead1eeee467791ad72a7658c2",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.10",
"size": 23013,
"upload_time": "2024-11-14T14:17:38",
"upload_time_iso_8601": "2024-11-14T14:17:38.367694Z",
"url": "https://files.pythonhosted.org/packages/1c/28/086d3ebf330b9577011a9c5d7bd933cc385f8529f5f90c255086cd828890/futureos-0.3.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "684a1be21cbc19e4f1e4cee169a0ea795bafd0469e3d73e4762408041aed4cf6",
"md5": "978ebc87bd47992a339da9073036bdf1",
"sha256": "205b1fbf5587461e89ec0fa0695ee89e1fe2a3397f569c2ea2185ead4168a470"
},
"downloads": -1,
"filename": "futureos-0.3.0.tar.gz",
"has_sig": false,
"md5_digest": "978ebc87bd47992a339da9073036bdf1",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 22572,
"upload_time": "2024-11-14T14:17:39",
"upload_time_iso_8601": "2024-11-14T14:17:39.511354Z",
"url": "https://files.pythonhosted.org/packages/68/4a/1be21cbc19e4f1e4cee169a0ea795bafd0469e3d73e4762408041aed4cf6/futureos-0.3.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-11-14 14:17:39",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "miskibin",
"github_project": "futureOS",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "futureos"
}
JSON
