JWST schemas

This package contains schemas of the following types:

Data model schemas
Reference file schemas
Transform schemas (not covered here)

Data model schemas

JWST datamodels are in part defined by an ASDF schema. For example RampModel uses the schema found in ramp.schema.yaml. These data model schemas typically contain many references to other schemas to allow common structures to be shared across data models. Here is a (partial) example:

%YAML 1.1
---
$schema: "http://stsci.edu/schemas/fits-schema/fits-schema"
id: "http://stsci.edu/schemas/jwst_datamodel/ramp.schema"
allOf:
- $ref: core.schema
- $ref: bunit.schema
- $ref: photometry.schema
- $ref: wcsinfo.schema
- type: object

Each $ref above will pull in the common structure defined in the referenced schema. All data model schemas reference core.schema.

Reference file schemas

JWST reference file schemas are similar to the data model schemas but use a different set of shared schemas.

%YAML 1.1
---
$schema: "http://stsci.edu/schemas/fits-schema/fits-schema"
id: "http://stsci.edu/schemas/jwst_datamodel/dark.schema"
title: Dark current data model
allOf:
- $ref: referencefile.schema
- $ref: keyword_exptype.schema
- $ref: keyword_readpatt.schema

Note that reference file schemas $ref referencefile.schema.

Reference file keywords and use by CRDS

Reference file schemas often contain references to keyword_* schemas (for example keyword_exptype.schema above). These define standard keywords that are used for reference file selection by CRDS. For the above example, the exptype from a science file is matched with the CRDS parkey of the same name to determine the appropriate reference file. When crafting (or updating) a reference file schema it’s important to make sure that the referenced keyword schemas match those expected by CRDS.

This can involve adding “pattern” keywords (for example keyword_pexptype.schema) when a reference file might be used for several keyword values. For example, if a single reference file matches all filters, it can reference keyword_pfilter.schema and then CRDS can use a “pattern” to avoid hosting copies of the same file for every filter. See the CRDS docs for more details about patterns.