enums as URIs: enhance documentation, fix behaviour

[mr-c]

CWL v1.0 - v1.2 define the type enum as a specialization of the Salad type EnumSchema which means that each of the symbols are either full URIs or the ending parts of URIs built from the base URI for the document [0].

The text of the standards does not directly say how to serialize type: enums on the command line, but the conformance tests[1] show that a short name is computed and used. Short names are mentioned in the v1.0 version of the CWL standards but not defined explicitly until v1.1: https://www.commonwl.org/v1.1/SchemaSalad.html#Short_names

enums being a URI might make them attractive for use in specifying the output format of File types, but currently cwltool only outputs the short name[2]. So the utility of the URI is lost in this circumstance.

• CWL v1.2.1: Add an explicit explanation to https://www.commonwl.org/v1.2/CommandLineTool.html#CommandLineBinding on how to render enums on the command line (compute the short name and use it)

• CWL v1.3+: Decide if/how the full URI of a type: enum is accessible in the format field and elsewhere

[0] v1.0 Schema (the same for v1.1 and v1.2):

common-workflow-language/v1.0/Process.yml

Line 571 in c8041aa

- name: InputEnumSchema

common-workflow-language/v1.0/salad/schema_salad/metaschema/metaschema_base.yml

Line 129 in c8041aa

symbols:

[1]

common-workflow-language/v1.0/conformance_test_v1.0.yaml

Line 2569 in c8041aa

- tool: v1.0/anon_enum_inside_array.cwl

https://github.com/common-workflow-language/common-workflow-language/blob/main/v1.0/v1.0/anon_enum_inside_array.cwl
https://github.com/common-workflow-language/common-workflow-language/blob/main/v1.0/v1.0/anon_enum_inside_array.yml
The correct result is

first mus_musculus second homo_sapiens

not

first […]v1.0/v1.0/anon_enum_inside_array.cwl#second/mus_musculus second […]v1.0/v1.0/anon_enum_inside_array.cwl#second/homo_sapiens

[2] https://cwl.discourse.group/t/set-tool-output-format-string-from-workflow-inputs/524/10?u=mrc

Related issue here: #568
Related discussions:

• metaschema predicates pollute salad document namespace schema_salad#506
• https://cwl.discourse.group/t/set-tool-output-format-string-from-workflow-inputs/524/9
• https://matrix.to/#/!RQMxrGNGkeDmWHOaEs:gitter.im/$pvFcTFhyTSnFES2oZ6e8iZA2uFae_9RayIvBvWf5LZc (a.k.a https://gitter.im/common-workflow-language/common-workflow-language?at=61e9531ce1a1264f0a66ce1d )