Definitions

On the use of non-relational database

The data accessible from the REST API is stored in tables within the Apache HBase database. HBase is a non-relational database (i.e., it is not SQL-like). This has many advantages, notably that it does not rely on a fixed schema, allowing for easy schema migration. However, this also means that we must handle the description of the data ourselves when you run a query

To give you a more practical view, a long-lived variable object that emits alerts for years will consist of alert data with different fields: some fields may have been created, while others may have been deleted over the course of the event. This is handled at the query runtime.

Fields definition

What are the available fields in Fink?

You can also browse fields per endpoint online at https://lsst.fink-portal.org/schemas, or you can programmatically access them. For example the fields returned by the endpoint /api/v1/sources:

curl -H "Content-Type: application/json" -X POST -d '{"endpoint": "/api/v1/sources"}'\
    https://api.lsst.fink-portal.org/api/v1/schema -o sources_schema.json

Prefix meaning (r:, f:, ...)

The fields are all prefixed with a letter determining the provenance of the data:

r: original data from LSST
f: added values from Fink Science modules
v: added values from Fink generated at the query runtime
b: cutout data

These are called column families.

Tag definition

What are the available user-defined in Fink?

The list of Fink class can be found at https://api.lsst.fink-portal.org/api/v1/tags. We recommend also to read how the classification scheme is built.

You can programmatically access the list of all the Fink tags using e.g.:

curl -H "Content-Type: application/json" -X GET \
    https://api.lsst.fink-portal.org/api/v1/tags -o fink_tags.json