Skip to main content

OS Helper

🇫🇷 · 🇬🇧

CI License: BSD-3-Clause Python Local-first

OS Helper belongs to a collection of libraries called AI Helpers developed for building Artificial Intelligence.

🌍 AI Helpers

logo

OS Helper is a Python library that provides utility functions for working with different operating systems. It offers a set of tools to simplify common system operations, file handling, and OS-specific tasks.

The Promise

os-helper is part of a local-first, sovereignty-minded suite. Rather than market that, here is the honest, case-by-case reality:

  1. Guaranteed local. os-helper is a pure local filesystem / utility toolbox. Nothing is uploaded, there is no telemetry, and there is no account. The optional Tree Radar GUI reads your disk and renders the treemap locally in your browser (the server binds to 127.0.0.1 only); your paths, sizes, and content hashes never leave the machine.

  2. Not possible to be local: the caveats. Two helpers make outbound HTTP by design, because fetching something is their whole purpose: download_file() (it downloads a URL you hand it) and the URL-liveness checks (is_working_url() / check_url). get_user_ip() also calls a public echo service on purpose. These are the only network calls in the library, and you only trigger them by explicitly calling them.

  3. Your decision. Nothing here forces the cloud. temporary_remote_file() can stage to S3/GCS/SFTP, but only when you wire it to a remote. If you build network behavior on top of os-helper, that is your choice, never a default.

Documentation

💻 Documentation

🗺️ Landscape

📋 Examples

🔐 Credentials Management

Features

Everything is a thin, well-typed, well-documented wrapper: no heavy system dependency, pure-Python across macOS / Linux / Windows.

  • OS detection: windows(), linux(), macos(), unix().
  • Hardware inspection: hardware_info() (one-call snapshot), plus its building blocks: cpu_count_logical(), cpu_count_physical(), cpu_model(), ram_gb(), gpu_vendor() (apple/nvidia/amd/intel/cpu), gpus() (name + VRAM per discrete GPU), apple_chip_name() / apple_unified_memory_gb() for Apple Silicon's shared memory pool. Live figures too: cpu_percent(), available_ram_gb(), disk_usage_gb(), gpu_utilization_percent() (Apple via IOKit, no sudo/powermetrics needed; NVIDIA via nvidia-smi; AMD via rocm-smi).
  • Process and command execution: system() (shell-free subprocess, captured stdout/stderr, optional exit-code and expected-output checks), openfile() (open with the OS default app), getpid(), get_nb_workers() (scikit-learn n_jobs convention, NB_WORKERS-overridable).
  • Paths: join(), folder_name_ext() (splits on the last dot only, so archive.tar.gz decomposes as archive.tar + gz, never collapsed to one multi-part extension), absolute2relative_path(), relative2absolute_path(), path_without_home(), recursive_glob().
  • Files and directories: file_exists(), dir_exists() (with emptiness checks), size_file(), checkfile(), copyfile(), make_directory(), remove_directory(), remove_files() (best-effort batch).
  • Temporary resources: temporary_filename() (context-managed, optional target directory), temporary_folder(), make_temporary_directory() (persistent, caller-owned cleanup), temporary_remote_file() (stage to S3/GCS/SFTP/anywhere with guaranteed remote cleanup).
  • Hashing: hash_string(), hashfile(), hashfolder() (RIPEMD-160 when available, BLAKE2b fallback; stable 40-char hex digests cross-platform).
  • Configuration loading: get_config() with a deterministic fallback order, JSON/YAML file (or folder), then .env files, then process environment.
  • Strings: emptystring() (None / empty / whitespace), asciistring() (accent-folding, filesystem-safe slugs).
  • Downloads and networking: download_file() (streaming, flat memory, adaptive block size, progress bar, returns {path, content_type, bytes}), progress_bar() (shared byte-scaled tqdm factory, auto-quiet off-TTY), is_working_url(), get_user_ip().
  • Folder reporting and archiving: folder_description() (size map, Bootstrap index.html, and description.json), zip_folder().
  • Durations and timestamps: now_string(), format_size(), time2str(), str2time().
  • Timing and profiling: wall_timer(), cpu_timer(), gpu_timer() (CUDA events / Apple-Silicon MPS, lazy torch), and MATLAB-style tic() / toc().
  • Logging surface: init_logging() (colored console + file, named-logger and live-stream modes), verbosity() (integer level get/set), and debug() / info() / warning() / error() / critical() / check().
  • Multiple surfaces, one codebase: importable library, an os-helper argparse CLI (always installed), an os-helper-click twin (via the [cli] extra), an HTTP API (os-helper-api, via [api]), and MCP tools (os-helper-mcp, via [mcp]). The API/MCP surfaces expose only the safe, side-effect-free subset (hardware info, hashing, ASCII, formatting, URL check, config loading; no filesystem mutation).

Installation

Prerequisites: Python 3.10–3.13 and git, cross-platform (os-helper needs no heavy system dependency):

  • 🍎 macOS (Homebrew): brew install python git
  • 🐧 Ubuntu/Debian: sudo apt update && sudo apt install -y python3 python3-pip git
  • 🪟 Windows (PowerShell): winget install Python.Python.3.12 Git.Git

We recommend using Python environments. Check this link if you're unfamiliar with setting one up: 🥸 Tech tips.

# Core utilities (library + argparse CLI)
pip install os-helper

# Optional click-based CLI twin
pip install "os-helper[cli]"

From source (no PyPI)

git clone https://github.com/warith-harchaoui/os-helper.git
cd os-helper

# Core utilities (library + argparse CLI)
pip install -e .

# Optional click-based CLI twin
pip install -e ".[cli]"

Usage

Below are examples demonstrating how to use various features of the os_helper library. Make sure to import the library as osh before starting.

import os_helper as osh
  1. Set Verbosity and Check Operating System
# Set verbosity level to display debugging messages
osh.verbosity(3)

# Check if the system is Unix-based (Linux or macOS)
if osh.unix():
    osh.info("You are running on a Unix-based system.")
else:
    osh.info("You are not running on a Unix-based system.")
  1. Timestamp and File Existence Check
# Generate a formatted timestamp for logging
timestamp = osh.now_string("log")
osh.info(f"Current timestamp (log format): {timestamp}")

# Check if a file exists and is not empty
test_file = "example.txt"
if osh.file_exists(test_file, check_empty=True):
    osh.info(f"File {test_file} exists and is not empty.")
else:
    osh.error(f"File {test_file} does not exist or is empty.")
  1. Directory Creation and File Search
# Create a directory
test_dir = "test_directory"
osh.make_directory(test_dir)
osh.info(f"Directory {test_dir} created.")

# Perform recursive search for '.txt' files in the directory
matching_files = osh.recursive_glob(test_dir, "*.txt")
osh.info(f"Matching files: {matching_files}")
  1. Copy and Remove Files
# Copy a file from source to destination
source_file = "source.txt"
destination_file = "backup_source.txt"
osh.copyfile(source_file, destination_file)
osh.info(f"File {source_file} copied to {destination_file}")

# Remove the copied file (each removal is logged at INFO level)
osh.remove_files([destination_file])
  1. Decompose a Path and Temporary File Creation
# Decompose a file path into folder, basename, and extension
folder, basename, ext = osh.folder_name_ext("/path/to/myfile.tar.gz")
osh.info(f"Folder: {folder}, Basename: {basename}, Extension: {ext}")

# Create and write to a temporary file
with osh.temporary_filename(suffix=".log") as temp_log:
    osh.info(f"Temporary file created at: {temp_log}")
    with open(temp_log, "w") as log_file:
        log_file.write("This is a temporary log entry.")
  1. Running System Commands
# Execute a system command and capture its output
cmd_output = osh.system("echo 'Hello, World!'")
osh.info(f"Command output: {cmd_output['out']}")
  1. Hashing Files and Strings
# Hash the contents of a file
file_to_hash = "testfile.txt"
if osh.file_exists(file_to_hash):
    file_hash = osh.hashfile(file_to_hash)
    osh.info(f"Hash of {file_to_hash}: {file_hash}")

# Hash a string with a specific length
hashed_string = osh.hash_string("MyTestString", size=8)
osh.info(f"Hashed string: {hashed_string}")
  1. ASCII String Conversion and Process ID
# Convert a string into a safe ASCII format
safe_string = osh.asciistring("Café-Con-Leche!", replacement_char="_")
osh.info(f"Safe ASCII string: {safe_string}")

# Get the current process ID
pid = osh.getpid()
osh.info(f"Current Process ID: {pid}")
  1. Check URL Validity and Zip Folder
# Check if a URL is valid and reachable
url = "https://www.example.com"
if osh.is_working_url(url):
    osh.info(f"The URL {url} is valid and reachable.")
else:
    osh.error(f"The URL {url} is not reachable.")

# Zip a folder
folder_to_zip = "my_folder"
zip_output = "my_folder_backup.zip"
osh.zip_folder(folder_to_zip, zip_output)
osh.info(f"Folder {folder_to_zip} zipped into {zip_output}")

Multi-surface exposure

os-helper is not just a library: the same functions are exposed as a Python import, an argparse CLI, a click CLI twin, an HTTP API, and MCP tools:

# Python library (default)
import os_helper as osh

# argparse-based CLI (installed automatically)
os-helper os system
os-helper hardware info
os-helper path exists ~/.zshrc
os-helper hash string hello --size 8
os-helper misc format-size 12345678
os-helper misc now --fmt filename

# click-based CLI twin (needs the [cli] extra)
pip install "os-helper[cli]"
os-helper-click hash file ./pyproject.toml
os-helper-click hardware info

HTTP API + MCP

The library's safe, side-effect-free operations (hardware info, hashing, ASCII normalization, size/time formatting, URL reachability, config loading; deliberately no filesystem-mutation endpoint) are also reachable over HTTP, and as MCP tools for any MCP-aware agent host:

pip install "os-helper[api]"
os-helper-api                    # -> http://127.0.0.1:8010 (docs at /docs)
curl http://127.0.0.1:8010/hardware

pip install "os-helper[mcp]"
os-helper-mcp                    # same app + an /mcp endpoint (fastapi-mcp)

Optional Tree Radar GUI

A first, real slice of the GUI.md plan ships as an optional surface: Tree Radar, a local disk-usage treemap dashboard. Each rectangle is a file or folder (area = size), colored by age, hash-dedupe status, or type family. It reads your disk and renders it in your browser; nothing is uploaded.

The GUI's web stack lives behind the [gui] extra so the core import os_helper stays lean (no FastAPI in the default install):

pip install "os-helper[gui]"

# Launch the local dashboard (loopback only), then open http://127.0.0.1:8017/gui
os-helper gui --root ~/Downloads
# or the dedicated entry point:
os-helper-gui --root ~/Downloads

The remaining GUI milestones (Dedupe Lens actions, Config Explorer) stay described in GUI.md.

Author

Acknowledgements

Special thanks to Mohamed Chelali and Bachir Zerroug for fruitful discussions.

License

This project is licensed under the BSD-3-Clause License; see the LICENSE file for details.

Metadata

Release files for os-helper 2.3.4

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for os-helper 2.3.4
File Size Uploaded
os_helper-2.3.4.tar.gz 122.9 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for os-helper 2.3.4
File Interpreter ABI Platform
os_helper-2.3.4-py3-none-any.whl Python 3 none any Details

Total release size: 222.0 kB

Release files / os_helper-2.3.4.tar.gz

Download URL os_helper-2.3.4.tar.gz
Size 122.9 kB
Tags Source
SHA-256 checksum
How to use checksums
adee0ed5e0b2a591d0b6e6895c7a014345c706f31a9377c7895fbf2e211c6ef2
BLAKE2b-256 checksum
How to use checksums
288e7a69881b6f230059867eda305b6aa2a909a32f7c063fd0f8fd8dfb191fae
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.13

Release files / os_helper-2.3.4-py3-none-any.whl

Download URL os_helper-2.3.4-py3-none-any.whl
Size 99.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
0748b581ea2d7a574554c553b7bbeec9f6985f4b020388dfb549c42d0d513267
BLAKE2b-256 checksum
How to use checksums
36dd85ff966ceeba878b5d4a15c9f58172af1d4c8aa8a547df9e5c707dc13ae1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.13.13

Release history Release notifications | RSS feed

This release

2.3.4 This release

2 release files

2.3.3

2 release files

2.3.2

2 release files

2.3.1

2 release files

2.3.0

2 release files

2.2.0

2 release files

2.0.1

2 release files

2.0.0

2 release files

1.8.0

2 release files

1.7.2

2 release files

1.5.2

2 release files

1.5.1

2 release files

1.5.0

2 release files

1.4.2

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page