iSukces.Cli.Python 1.26.713.1

iSukces.Cli.Python

Polska wersja

.NET library for managing Python environments from C# code. It supports virtual environment creation, pip package installation, PyTorch and CUDA Toolkit detection, HuggingFace token handling, and process environment variable management.

Requirements

• .NET 10
• Windows with the Python py launcher

External tools and AI models

• Python, the py launcher, pip, optional uv, and created venv environments are external dependencies supplied outside this package.
• PyTorch (torch, torchvision, torchaudio) and CUDA Toolkit are external libraries/tools installed in the Python environment or in the operating system.
• HuggingFace tokens and resources, private package indexes, and AI models downloaded or used by user code are not part of this package.
• This package does not include AI models and does not guarantee the behavior of models, Python libraries, GPU drivers, or tools installed outside this package.

Installation

NuGet package:

dotnet add package iSukces.Cli.Python

Usage

Basic Python process configuration:

var result = await new Cli
{
    FileName = "py",
    Arguments = ["--version"]
}.WithPythonCli().RunAsync();

Console.WriteLine(result.Output);

Main features

Virtual environment management

• PythonVenv represents a Python virtual environment. It can create environments with py -m venv or uv venv and configure process environment variables such as VIRTUAL_ENV, PYTHONHOME, and PATH.
• PythonVenv.WithVenv() applies virtual environment settings to a Cli process call.
• PythonCliExtension.WithPythonCli() sets PYTHONIOENCODING=utf-8 for started processes.

Package installation

• Installer installs and uninstalls pip packages in a virtual environment. It supports private indexes through --index-url and HuggingFace tokens through HF_TOKEN.
• The touched-files mechanism stores operation flags in .prv-flags and prevents repeated installation work when an operation has already been completed.

PyTorch and CUDA detection

• PyTorchUtils runs Python code with torch and reads the PyTorch version, CUDA availability, and CUDA version reported by PyTorch.
• CudaToolkit detects installed CUDA Toolkit versions in C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA.
• PythonEnvironmentInfo aggregates PyTorch and CUDA information and can create an EnvironmentUpdater that sets CUDA_PATH and updates PATH.

Environment diagnostics

• PythonEnvironmentProblemFinder detects environment problems, including missing or broken PyTorch installation, missing GPU acceleration, and CUDA Toolkit version mismatch.
• PythonEnvironmentProblem describes a single problem, suggested remedy, and affected areas.

Installed Python versions

• PyUtils.GetInstalledVersions() runs py -0 and parses installed Python versions.
• PyLauncherParser parses Windows Python launcher output and detects version, architecture, and default interpreter.

Environment variables

• EnvironmentUpdater modifies environment variables, including PATH and CUDA_PATH, with temporary activation and restoration through IDisposable.

Dependencies

• iSukces.CliTools provides base CLI process execution and argument collection utilities.

Build

dotnet build ..\iSukces.Cli.sln

Project configuration

• Target framework: net10.0
• Package license: MIT

Disclaimer

This package only provides a .NET layer that starts external Python processes and tools in the operating system. It acts as a bridge/facade for tools supplied and installed outside this package.

This package does not include the executed tools, AI models, Python libraries, GPU drivers, or CUDA Toolkit, does not modify how they work, and is not responsible for their results, failures, environment requirements, resource usage, or any damages resulting from the use of those external tools.