Keeping Everyone HAPI: Achieving Interoperability for Heliophysics and
Planetary Time Series Data
Abstract
The ability to access time series data with one API would significantly
enhance science data interoperability. The Heliophysics Application
Programmers Interface (HAPI) is a simple, standardized mechanism for
exposing time series data through a service. HAPI is being adopted by
data centers within the Planetary and Heliophysics communities,
especially for plasma, particle and field datasets. At the recent COSPAR
meeting, the Panel on Space Weather passed a resolution encouraging data
providers to have at minimum a HAPI server to deliver time series data.
The COSPAR is now considering the resolution for full organizational
endorsement. HAPI standardizes the two key parts of a data service: the
request interface and the result format. The request interface is very
simple and captures the common features of many existing data access
services. For result formats, the HAPI specification allows several
options, all of them streaming. Servers must provide a Comma Separated
Value (CSV) result format, but may optionally provide a JSON or binary
stream as well. The details of the request and result formats are
described in the current version of the specification document, which is
available at GitHub: https://github.com/hapi-server/data-specification.
Several institutions have recently added HAPI-compliant access. These
include the large Heliophysics archive at Goddard’s Coordinated Data
Analysis Web (CDAWeb), as well as the Planetary Plasma Interactions node
of the Planetary Data System, the Laboratory for Atmospheric and Space
Physics at CU Boulder, the University of Iowa, George Mason University,
and the Johns Hopkins University Applied Physics Lab. Multiple client
options are available for accessing HAPI data from the growing number of
servers. Autoplot (Faden, et al, 2010) and SPEDAS
(http://spedas.org/wiki) both read HAPI data, and other clients (Java,
Python, Matlab, IDL) can be downloaded from the HAPI Github project. The
ease with which various providers have adapted existing servers to
create a HAPI-compliant capability shows that it does capture a useful
way to represent time series data. Because clients for reading HAPI data
are also easy to create, we anticipate significant growth and interest
in this emerging standard. Faden, et al, Earth Sci Inform (2010)
3:41–49, DOI 10.1007/s12145-010-0049-0