123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207 |
- .. _installation:
- Installation
- ------------
- JupyterLab can be installed using ``conda``, ``pip``, ``pipenv`` or ``docker``.
- conda
- ~~~~~
- If you use ``conda``, you can install it with:
- .. code:: bash
- conda install -c conda-forge jupyterlab
- pip
- ~~~
- If you use ``pip``, you can install it with:
- .. code:: bash
- pip install jupyterlab
- If installing using ``pip install --user``, you must add the user-level
- ``bin`` directory to your ``PATH`` environment variable in order to launch
- ``jupyter lab``.
- pipenv
- ~~~~~~
- If you use ``pipenv``, you can install it as:
- .. code:: bash
- pipenv install jupyterlab
- pipenv shell
- or from a git checkout:
- .. code:: bash
- pipenv install git+git://github.com/jupyterlab/jupyterlab.git#egg=jupyterlab
- pipenv shell
- When using ``pipenv``, in order to launch ``jupyter lab``, you must activate the project's virtualenv.
- For example, in the directory where ``pipenv``'s ``Pipfile`` and ``Pipfile.lock`` live (i.e., where you ran the above commands):
- .. code:: bash
- pipenv shell
- jupyter lab
- Docker
- ~~~~~~
- If you have `Docker installed <https://docs.docker.com/install/>`__, you can install and use JupyterLab by selecting one
- of the many `ready-to-run Docker images <https://jupyter-docker-stacks.readthedocs.io/en/latest/using/selecting.html>`__
- maintained by the Jupyter Team. Follow the instructions in the `Quick Start Guide <https://jupyter-docker-stacks.readthedocs.io/en/latest/>`__
- to deploy the chosen Docker image. NOTE: Ensure your docker command includes the `-e JUPYTER_ENABLE_LAB=yes` flag to ensure
- JupyterLab is enabled in your container.
- Installing with Previous Versions of Notebook
- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- If you are using a version of Jupyter Notebook earlier than 5.3, then you must also run the following command to enable the JupyterLab
- server extension:
- .. code:: bash
- jupyter serverextension enable --py jupyterlab --sys-prefix
- Prerequisites
- ~~~~~~~~~~~~~
- JupyterLab requires the Jupyter Notebook version 4.3 or later. To check
- the version of the ``notebook`` package that you have installed:
- .. code:: bash
- jupyter notebook --version
- Usage with JupyterHub
- ~~~~~~~~~~~~~~~~~~~~~
- Install JupyterLab and JupyterHub.
- In ``jupyterhub_config.py``, configure the ``Spawner`` to tell the single-user notebook servers to default to JupyterLab:
- ``c.Spawner.default_url = '/lab'``
- Supported browsers
- ~~~~~~~~~~~~~~~~~~
- The latest versions of the following browsers are currently known to work:
- - Firefox
- - Chrome
- - Safari
- Earlier browser versions may also work, but come with no guarantees.
- JupyterLab uses CSS Variables for styling, which is one reason for the
- minimum versions listed above. IE 11+ or Edge 14 do not support
- CSS Variables, and are not directly supported at this time.
- A tool like `postcss <https://postcss.org/>`__ can be used to convert the CSS files in the
- ``jupyterlab/build`` directory manually if desired.
- Installation problems
- ~~~~~~~~~~~~~~~~~~~~~
- If your computer is behind corporate proxy or firewall,
- you may encounter HTTP and SSL errors due to custom security profiles managed by corporate IT departments.
- Example of typical error, when conda cannot connect to own repositories:
- + `CondaHTTPError: HTTP 000 CONNECTION FAILED for url <https://repo.anaconda.com/pkgs/main/win-64/current_repodata.json>`
- This may happen because your company can block connections to widely-used repositories in Python and JavaScript communities.
- Here are some widely-used sites that host packages in the Python and JavaScript open-source ecosystem. Your network adminstrator may be able to allow http and https connections to these:
- - \*.pypi.org
- - \*.pythonhosted.org
- - \*.continuum.io
- - \*.anaconda.com
- - \*.conda.io
- - \*.github.com
- - \*.githubusercontent.com
- - \*.npmjs.com
- - \*.yarnpkg.com
- Alternatively you can specify proxy user (mostly domain user with password),
- that is allowed to communicate via network. This can be easily achieved
- by setting two common environment variables: `HTTP_PROXY` and `HTTPS_PROXY`.
- These variables are automatically used by many open-source tools (like ``conda``) if set correctly.
- .. code:: bash
- # For Windows
- set HTTP_PROXY=http://USER:PWD@proxy.company.com:PORT
- set HTTPS_PROXY=https://USER:PWD@proxy.comp any.com:PORT
- # For Linux / MacOS
- export HTTP_PROXY=http://USER:PWD@proxy.company.com:PORT
- export HTTPS_PROXY=https://USER:PWD@proxy.company.com:PORT
- In case you can communicate via HTTP, but installation with ``conda`` fails
- on connectivity problems to HTTPS servers, you can disable using SSL for ``conda``.
- .. warning:: Disabling SSL in communication is generally not recommended and involves potential security risk.
- .. code:: bash
- # Configure npm to not use SSL
- conda config --set ssl_verify False
- You can do a similar thing for ``pip``.
- The approach here is to mark repository servers as trusted hosts,
- which means, SSL communication will not be required for downloading Python libraries.
- .. code:: bash
- # Install pandas (without SSL)
- pip install --trusted-host pypi.org --trusted-host files.pythonhosted.org pandas
- Using the tips from above, you can handle many network problems
- related to installing Python libraries.
- Many Jupyter extensions require having a working ``npm`` command,
- which is required for downloading useful Jupyter extensions or other JavaScript dependencies.
- Example of typical error message, when ``npm`` cannot connect to own repositories:
- + `ValueError: "@jupyterlab/toc" is not a valid npm package`
- .. code:: bash
- # Set proxy for NPM
- npm config set proxy http://USER:PWD@proxy.company.com:PORT
- npm config set proxy https://USER:PWD@proxy.company.com:PORT
- # Set default registry for NPM (optional, useful in case if common JavaScript libs cannot be found)
- npm config set registry http://registry.npmjs.org/
- In case you can communicate via HTTP, but installation with ``npm`` fails
- on connectivity problems to HTTPS servers, you can disable using SSL for ``npm``.
- .. warning:: Disabling SSL in communication is generally not recommended and involves potential security risk.
- .. code:: bash
- # Configure npm to not use SSL
- npm set strict-ssl False
|