HAPI Client for Python

Basic usage examples for various HAPI servers are given in hapi_demo.py and the Examples section of a Jupyter Notebook hosted on Google Colab: .

Installation

```bash pip install hapiclient --upgrade

or

pip install 'git+https://github.com/hapi-server/client-python' --upgrade ```

The optional hapiplot package provides basic preview plotting capabilities of data from a HAPI server. The Plotting section of the hapiclient Jupyter Notebook shows how to plot the output of hapiclient using many other plotting libraries.

To install hapiplot, use

```bash python -m pip install hapiplot --upgrade

or

python -m pip install 'git+https://github.com/hapi-server/plot-python' --upgrade ```

See the Appendix for a fail-safe installation method.

Basic Example

```python

Get Dst index from CDAWeb HAPI server

from hapiclient import hapi

See http://hapi-server.org/servers/ for a list of

other HAPI servers and datasets.

server = 'https://cdaweb.gsfc.nasa.gov/hapi' dataset = 'OMNI2H0MRG1HR' start = '2003-09-01T00:00:00' stop = '2003-12-01T00:00:00' parameters = 'DST1800' opts = {'logging': True}

Get data

data, meta = hapi(server, dataset, parameters, start, stop, **opts) print(meta) print(data)

Plot all parameters

from hapiplot import hapiplot hapiplot(data, meta) ```

Documentation

Basic usage examples for various HAPI servers are given in hapi_demo.py and the Examples section of a Jupyter Notebook hosted on Google Colab.

See http://hapi-server.org/servers/ for a list of HAPI servers and datasets.

All of the features are extensively demonstrated in hapi_demo.ipynb, a Jupyter Notebook that can be viewed an executed on Google Colab.

Metadata Model

See also the examples in the Metadata Model section of the hapiclient Jupyter Notebook.

The HAPI client metadata model is intentionally minimal and closely follows that of the HAPI metadata model. We expect that another library will be developed that allows high-level search and grouping of information from HAPI servers. See also issue #106.

Data Model and Time Format

See also the examples in the Data Model section of the hapiclient Jupyter Notebook. The examples include

1. Fast and well-tested conversion from ISO 8601 timestamp strings to Python datetime objects
2. Putting the content of data in a Pandas DataFrame object
3. Creating an Astropy NDArray

A request for data of the form data, meta = hapi(server, dataset, parameters, start, stop)

returns the Numpy N-D array data and a Python dictionary meta from a HAPI-compliant data server server. The structure of data and meta mirrors the structure of a response from a HAPI server.

The HAPI client data model is intentionally basic. There is an ongoing discussion of a data model for Heliophysics data among the PyHC community. When this data model is complete, a function that converts data and meta to that data model will be included in the hapiclient package.

Development

bash git clone https://github.com/hapi-server/client-python cd client-python; python -m pip install -e .

or, create an isolated Anaconda installation (downloads and installs latest Miniconda3) using

``` bash make install PYTHON=python3.6

Execute command displayed to activate isolated environment

```

The command pip install -e . creates symlinks so that the local package is used instead of an installed package. You may need to execute python -m pip uninstall hapiclient to ensure the local package is used. To check the version installed, use python -m pip list | grep hapiclient.

To run tests before a commit, execute

bash make repository-test

To run an individual unit test in a Python session, use, e.g.,

python from hapiclient.test.test_hapi import test_reader_short test_reader_short()

Contact

Submit bug reports and feature requests on the repository issue tracker.