.. _installation_offline: Offline Installation ==================== .. meta:: :description: Offline installation guide for DataLab on systems without internet access :keywords: DataLab, offline, installation, Python, pip, wheel, no internet, Windows, Linux, macOS This guide explains how to install DataLab on a system **without internet access**. .. note:: This page focuses on **offline installation using pip** (Python package manager). For other offline installation methods, see the :ref:`installation_offline_alternatives` section below. Offline installation via pip ---------------------------- :octicon:`info;1em;sd-text-info` :bdg-info-line:`GNU/Linux` :bdg-info-line:`Windows` :bdg-info-line:`macOS` This method works on **any operating system** with **Python 3.9 to 3.14**. It assumes you have an online machine available to download packages, and a separate offline machine where DataLab will be installed. Requirements ------------ Offline machine ^^^^^^^^^^^^^^^ - GNU/Linux, Windows, or macOS - Python 3.9 to 3.14 installed with the **same major version** as the online machine - No assumptions regarding installed packages .. important:: **Version compatibility is critical:** - ✅ **Allowed:** Python 3.10.2 ↔ Python 3.10.11 (same major version) - ❌ **Not allowed:** Python 3.10.x ↔ Python 3.11.x (different major versions) SciPy, NumPy, and scikit-image are highly dependent on the Python major version. Both machines must use the same major Python version for reliable offline installation. To check your Python version, run: .. code-block:: console python --version Online machine ^^^^^^^^^^^^^^ - Windows, Linux, or macOS - Python with the **same major version** as the offline machine - ``pip`` installed Preparing packages (online machine) ----------------------------------- Download all required **Wheel packages** for offline installation. Create a folder for packages ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. tab-set:: .. tab-item:: Windows .. code-block:: console mkdir C:\datalab-offline cd C:\datalab-offline .. tab-item:: GNU/Linux / macOS .. code-block:: console mkdir ~/datalab-offline cd ~/datalab-offline Download DataLab and dependencies ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. tab-set:: .. tab-item:: Windows .. code-block:: console pip download datalab-platform --dest C:\datalab-offline .. tab-item:: GNU/Linux / macOS .. code-block:: console pip download datalab-platform --dest ~/datalab-offline This command downloads: - ``datalab-platform`` - ``sigima`` - ``plotpy`` - ``guidata`` - ``pythonqwt`` - ``numpy``, ``scipy``, ``scikit-image``, and other dependencies Expected folder structure ^^^^^^^^^^^^^^^^^^^^^^^^^ After downloading, your folder should contain files similar to: .. code-block:: text datalab-offline/ ├─ datalab_platform‑1.x.x‑py3‑none‑any.whl ├─ sigima‑1.x.x‑py3‑none‑any.whl ├─ plotpy‑x.x.x‑py3‑none‑any.whl ├─ guidata‑3.x.x‑py3‑none‑any.whl ├─ pythonqwt‑x.x.x‑py3‑none‑any.whl ├─ numpy‑...‑cp310‑cp310‑[platform].whl ├─ scipy‑...‑cp310‑cp310‑[platform].whl ├─ scikit_image‑...‑cp310‑cp310‑[platform].whl └─ ... .. note:: The ``[platform]`` suffix in the wheel filenames will vary depending on your system: ``win_amd64`` (Windows), ``manylinux`` (Linux), ``macosx`` (macOS). Transfer to offline machine ^^^^^^^^^^^^^^^^^^^^^^^^^^^ Transfer the ``datalab-offline`` folder to your offline machine using: - USB drive - Network share - Any other approved method in your organization Offline installation -------------------- On the offline machine, navigate to the folder and install DataLab. Navigate to the folder ^^^^^^^^^^^^^^^^^^^^^^ .. tab-set:: .. tab-item:: Windows .. code-block:: console cd C:\datalab-offline .. tab-item:: GNU/Linux / macOS .. code-block:: console cd ~/datalab-offline Install DataLab offline ^^^^^^^^^^^^^^^^^^^^^^^ .. code-block:: console python -m pip install --no-index --find-links . datalab-platform The ``--no-index`` option tells pip not to connect to the internet, and ``--find-links .`` tells pip to look for packages in the current directory. Verification ------------ To verify that DataLab is installed correctly, run: .. code-block:: console python -m datalab.app This should launch the DataLab application. Optional: Virtual environment ----------------------------- It is recommended to install DataLab in a virtual environment to avoid conflicts with other Python packages. Create and activate a virtual environment ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. tab-set:: .. tab-item:: Windows .. code-block:: console python -m venv C:\venvs\datalab C:\venvs\datalab\Scripts\activate .. tab-item:: GNU/Linux / macOS .. code-block:: console python -m venv ~/venvs/datalab source ~/venvs/datalab/bin/activate Install DataLab in the virtual environment ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. tab-set:: .. tab-item:: Windows .. code-block:: console python -m pip install --no-index --find-links C:\datalab-offline datalab-platform .. tab-item:: GNU/Linux / macOS .. code-block:: console python -m pip install --no-index --find-links ~/datalab-offline datalab-platform Launch DataLab ^^^^^^^^^^^^^^ .. code-block:: console python -m datalab.app WinPython for offline usage --------------------------- :octicon:`info;1em;sd-text-info` :bdg-info-line:`Windows` `WinPython `_ is particularly well-suited for offline installations as it provides a portable, self-contained Python environment. Prepare on online machine ^^^^^^^^^^^^^^^^^^^^^^^^^ 1. Download WinPython with the **same major Python version** as your target system 2. Install WinPython 3. Open the *WinPython Command Prompt* 4. Download packages: .. code-block:: console pip download datalab-platform --dest C:\datalab-offline Install on offline machine ^^^^^^^^^^^^^^^^^^^^^^^^^^ 1. Transfer WinPython and the ``datalab-offline`` folder to the offline machine 2. Open the *WinPython Command Prompt* 3. Install DataLab: .. code-block:: console cd C:\datalab-offline python -m pip install --no-index --find-links . datalab-platform Benefits of WinPython ^^^^^^^^^^^^^^^^^^^^^ - **Portable:** No installation required, can run from USB drive - **Self-contained:** All dependencies included - **Ideal for restricted environments:** No admin rights needed - **Multiple versions:** Can have multiple Python versions side by side Quick summary ------------- The complete workflow in four steps: 1. **On online machine:** Download packages .. code-block:: console pip download datalab-platform --dest /path/to/datalab-offline 2. **Transfer:** Copy the ``datalab-offline`` folder to the offline machine 3. **On offline machine:** Install offline .. code-block:: console python -m pip install --no-index --find-links /path/to/datalab-offline datalab-platform 4. **Verify:** Launch DataLab .. code-block:: console python -m datalab.app .. _installation_offline_alternatives: Other offline installation methods ---------------------------------- While this guide focuses on pip-based offline installation, DataLab can also be installed offline using other methods: Stand-alone installer (Windows) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ :octicon:`info;1em;sd-text-info` :bdg-info-line:`Windows` **Recommended for offline installation** The **all-in-one installer** is the simplest way to install DataLab offline on Windows. It does not require any Python distribution or internet connection on the target machine. **Advantages:** - No Python installation required - Single executable file - Self-contained with all dependencies - No internet connection needed during installation - Works on air-gapped systems **How to use:** 1. On a machine with internet access, download the installer from the `DataLab Releases `_ page 2. Transfer the installer executable to the offline machine (USB drive, network share, etc.) 3. Run the installer on the offline machine For more information, see :ref:`install_aioinstaller`. WinPython distribution (Windows) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ :octicon:`info;1em;sd-text-info` :bdg-info-line:`Windows` `DataLab-WinPython `_ is a portable Python distribution that includes DataLab and all its dependencies. **Advantages:** - Portable (can run from USB drive) - No installation required - Includes full Python environment - Can be extended with additional packages - Ideal for restricted environments **How to use:** 1. Download DataLab-WinPython from the `releases page `_ 2. Transfer to the offline machine 3. Extract and run For more information, see :ref:`install_winpython`. Conda offline installation ^^^^^^^^^^^^^^^^^^^^^^^^^^ :octicon:`info;1em;sd-text-info` :bdg-info-line:`GNU/Linux` :bdg-info-line:`Windows` :bdg-info-line:`macOS` Conda packages can also be installed offline using a similar approach to pip: 1. **On online machine:** Download the conda package and dependencies: .. code-block:: console conda create -n datalab-offline --download-only --json conda-forge::datalab Or use ``conda list --explicit`` to create a specification file. 2. **Transfer** the downloaded packages to the offline machine 3. **On offline machine:** Install from the local packages: .. code-block:: console conda create -n datalab --offline -c file:///path/to/packages datalab For more information about offline conda installation, see the `Conda documentation `_. For standard online installation, see :ref:`install_conda`. .. seealso:: For standard online installation methods, see :ref:`installation`.