Installation
============

zarrio can be installed in several ways depending on your needs.

Using pip
---------

The easiest way to install zarrio is using pip:

.. code-block:: bash

    pip install zarrio

This will install the latest stable version from PyPI.

Using conda
-----------

If you're using conda, you can install zarrio from conda-forge:

.. code-block:: bash

    conda install -c conda-forge zarrio

Installing from source
----------------------

To install the latest development version from source:

.. code-block:: bash

    git clone https://github.com/oceanum/zarrio.git
    cd zarrio
    pip install -e .

Dependencies
------------

zarrio requires the following dependencies:

Core dependencies:
~~~~~~~~~~~~~~~~~~

- Python >= 3.8
- xarray >= 0.18.0
- zarr >= 2.10.0
- numpy >= 1.20.0
- pandas >= 1.3.0
- click >= 8.0.0
- pyyaml >= 5.4.0
- dask >= 2021.0.0
- netCDF4 >= 1.5.0

Optional dependencies:
~~~~~~~~~~~~~~~~~~~~~~

- blosc: For Blosc compression support
- numcodecs: For additional compression codecs
- intake: For intake catalog support
- fsspec: For cloud storage support
- gcsfs: For Google Cloud Storage support

Development dependencies:
~~~~~~~~~~~~~~~~~~~~~~~~~

- pytest >= 6.2.0
- pytest-cov >= 2.12.0
- black >= 21.0.0
- flake8 >= 3.9.0
- mypy >= 0.910
- sphinx >= 4.0.0
- sphinx-rtd-theme >= 1.0.0
- pre-commit >= 2.13.0

To install development dependencies:

.. code-block:: bash

    pip install -e ".[dev]"

Verification
------------

To verify that zarrio is installed correctly, you can run:

.. code-block:: bash

    python -c "import zarrio; print(f'zarrio version: {zarrio.__version__}')"

Or using the CLI:

.. code-block:: bash

    zarrio --version

Docker Installation
-------------------

zarrio can also be used via Docker containers. See the :doc:`docker` documentation for details on building and running Docker images.

System requirements
-------------------

zarrio is designed to work on Linux, macOS, and Windows systems with Python 3.8 or higher.

For large datasets, ensure you have sufficient disk space and memory. The library can handle datasets larger than available RAM through dask's chunked array operations.