Autogenerate JSON Schema #28

alexiswl · 2024-02-19T05:37:32Z

From #27

Summary

Run typescript-json-schema over the Workflow schema.
Refine the schema (see .github/scripts/curate_autogenerated_schema.py)

Schema refining process steps

Remove loadingOptions and extensionFields properties
Rename class_, type_ etc to class, type
Add $import and $include options and update all requirements such that an item in a requirement can use $include or $import if that property accepts a list.
Fix inputs / outputs / steps, so not just an array but can be an object where the key is the id attribute.
Fix Requirements / Hints, so not just an array but can be an object where the key is the class attribute.
Fix descriptions by removing Auto-generated... from the start
Generate the CWLFile definition, which can be one of Workflow, CommandLineTool, or ExpressionTool and also fits the CWLMetadata
Fix issues with additionalProperties and nested definitions as shown here https://stoic-agnesi-d0ac4a.netlify.app/1 Needed for the CWLFIle and CWLGraphFile attributes
Generate the CWLGraph definition, which is an array of CWLFile objects with CWLMetadata
Write out the schema as JSON
To convert the schema to yaml, use the yq binary to generate the schema as yaml. ruamel.yaml doing some weird things on values under patternProperties regexes

Also added the GitHub actions template to autogenerate this file

…ditionalProperties is set to false

codecov · 2024-02-19T05:38:29Z

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 55.62%. Comparing base (0c14cfc) to head (538c987).

❗ Current head 538c987 differs from pull request most recent head 19cd344. Consider uploading reports for the commit 19cd344 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #28   +/-   ##
=======================================
  Coverage   55.62%   55.62%           
=======================================
  Files         127      127           
  Lines        7837     7837           
  Branches     1572     1572           
=======================================
  Hits         4359     4359           
  Misses       3111     3111           
  Partials      367      367

Flag	Coverage Δ
unittests	`55.62% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

alexiswl · 2024-02-19T23:31:07Z

Moved back to graph and standard all-in-one schema, now auto-generates a yaml file as well, both files under json_schema

mr-c · 2024-02-19T08:00:26Z

.github/scripts/validate_schema.py

+CWL12_REPO_PARENT_TMPDIR_OBJ = TemporaryDirectory()
+CWL12_REPO_DIR = Path(CWL12_REPO_PARENT_TMPDIR_OBJ.name) / "cwl-v1.2"
+
+IGNORED_FILES = [


Thanks for checking against the conformance tests! I feel like there are more files to ignore.

Look for the json_schema_invalid tag https://github.com/search?q=repo%3Acommon-workflow-language%2Fcwl-v1.2%20json_schema_invalid&type=code

Are these 'json_schema_invalid' related to the inputs of the workflow or the workflow itself?

I think this one can be resolved

mr-c · 2024-02-20T11:30:48Z

.github/scripts/curate_autogenerated_schema.py

+def fix_secondary_file_schema(schema_dict: Dict) -> Dict:
+    """
+    SecondaryFileSchema should allow for a single string entry to represent the secondary file pattern
+    Although not recommended, it is allowed in the CWL spec


Why is it not recommended?

It makes the requirement attribute ambiguous (and dependent on the implementation), success of the tool then depends on the workflow engine running the code - See common-workflow-language/cwltool#1322

Is this okay to keep in or should I remove the commentary?

mr-c · 2024-02-20T11:32:31Z

.github/scripts/curate_autogenerated_schema.py

+
+def fix_inline_javascript_requirement(schema_dict: Dict) -> Dict:
+    """
+    Fix the InlineJavascriptRequirement.expressionLib array to allow for $include


Shouldn't this be more universal, given all the places a $include can occur?

What are all the places an include can occur?

Can an $include occur in any requirement?

Can an $include occur in any requirement?

Yes, and every other place: $include can be a key name at any level in a CWL document

https://www.commonwl.org/v1.2/Workflow.html#Document_preprocessing
https://www.commonwl.org/v1.2/SchemaSalad.html#Include

Does this also apply for '$import'?

If so, I'll just add them to the patternProperties of each of the definitions

I've added $include and $import availabilities to all requirements (and hints), including top level of requirements

So https://github.com/common-workflow-language/cwl-v1.2/blob/b60a42e3cc417c5b75b88fd7c6681abcc7ff5b89/tests/schemadef-wf.cwl#L7 would pass

as would https://github.com/common-workflow-language/cwl-v1.2/blob/b60a42e3cc417c5b75b88fd7c6681abcc7ff5b89/tests/schemadef_types_with_import-tool.cwl#L9

Please see 19705a2 / b31d8ba

mr-c · 2024-02-20T11:34:25Z

.github/scripts/curate_autogenerated_schema.py

+                "patternProperties": {
+                    "^s:.*$": {
+                        "type": "object"
+                    },
+                    # Or the full version
+                    "^https://schema.org/.*$": {
+                        "type": "object"
+                    },
+                    # dct terms should also be permitted
+                    # see https://github.com/common-workflow-language/cwl-v1.2/blob/main/tests/metadata.cwl
+                    "^dct:.*$": {
+                        "type": "object"
+                    },
+                    "^http://purl.org/dc/terms/.*$": {
+                        "type": "object"
+                    }


it is valid CWL to have additional prefixed/namespaced elements at any level, of any namespace; not just schema.org, DCT, etc..

Okay, so could we use the regex "^\w+:.*$" for shorthand notation + "^\w+:\/\/.* for urls?

I'm reluctant to just put in a "^.*$" for pattern properties here, it means spelling mistakes or syntax errors aren't necessarily picked up by the schema. CWLVersion, cwlVersion, cwl_version would all be valid entries

I agree that accepting any string anywhere is bad for validation. This is where having a full CWL language server like Benten can go beyond what JSON Schema can do. If you can allow keys that have a colon (:) in them, then that should suffice.

Resolved in 554c0f2 / 9867428

alexiswl · 2024-03-21T22:02:33Z

@mr-c please re-review when you have the chance, would be great to get this into the JSON schema store. Even if we just have this file generated into the cwl-ts-auto repository, we can generate a PR into the [JSON Schema store repository] (https://github.com/SchemaStore/schemastore) (like this one - https://github.com/SchemaStore/schemastore/pull/3645/files).

If we want to then move this file into the cwl-v1.2 repository, we can then just generate a new PR that points to a new url

mr-c · 2024-03-21T22:14:44Z

Did you want to rewrite the commits, or shall I squash into one?

alexiswl · 2024-03-22T02:31:25Z

I can squash into one and force push if you'd like?

Uses TypeScript to generate generate JSON Schema, We then manually fix up with Python

alexiswl · 2024-03-25T07:06:21Z

Done

alexiswl and others added 2 commits February 12, 2024 18:22

Added script to autogenerate the JSON schema

5729375

Also added the GitHub actions template to autogenerate this file

Conform to other schema validators that don't recurse through when ad…

7891f77

…ditionalProperties is set to false

alexiswl requested a review from mr-c February 19, 2024 05:43

alexiswl self-assigned this Feb 19, 2024

alexiswl added the enhancement New feature or request label Feb 19, 2024

mr-c reviewed Feb 20, 2024

View reviewed changes

Create CWL JSON Schema

538c987

Uses TypeScript to generate generate JSON Schema, We then manually fix up with Python

alexiswl force-pushed the enhancement/auto-generate-json-schema-from-ts branch from 9867428 to 538c987 Compare March 25, 2024 07:05

Updating CWL and CWLGraph Schema

19cd344

mr-c merged commit f2fc4e5 into main Mar 25, 2024

mr-c deleted the enhancement/auto-generate-json-schema-from-ts branch March 25, 2024 07:16

alexiswl mentioned this pull request Mar 25, 2024

Added CWL json schema to catalog SchemaStore/schemastore#3670

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Autogenerate JSON Schema #28

Autogenerate JSON Schema #28

alexiswl commented Feb 19, 2024 •

edited

Loading

codecov bot commented Feb 19, 2024 •

edited

Loading

alexiswl commented Feb 19, 2024

mr-c Feb 19, 2024

alexiswl Feb 21, 2024

alexiswl Mar 19, 2024

mr-c Feb 20, 2024

alexiswl Feb 21, 2024 •

edited

Loading

alexiswl Mar 19, 2024

mr-c Feb 20, 2024

alexiswl Feb 21, 2024

mr-c Feb 21, 2024 •

edited

Loading

alexiswl Feb 21, 2024

alexiswl Feb 21, 2024

alexiswl Mar 19, 2024

alexiswl Mar 19, 2024

mr-c Feb 20, 2024

alexiswl Feb 21, 2024

alexiswl Feb 21, 2024

mr-c Feb 21, 2024

alexiswl Mar 19, 2024

alexiswl commented Mar 21, 2024

mr-c commented Mar 21, 2024

alexiswl commented Mar 22, 2024

alexiswl commented Mar 25, 2024

Autogenerate JSON Schema #28

Autogenerate JSON Schema #28

Conversation

alexiswl commented Feb 19, 2024 • edited Loading

Summary

Schema refining process steps

codecov bot commented Feb 19, 2024 • edited Loading

Codecov Report

alexiswl commented Feb 19, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexiswl Feb 21, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mr-c Feb 21, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alexiswl commented Mar 21, 2024

mr-c commented Mar 21, 2024

alexiswl commented Mar 22, 2024

alexiswl commented Mar 25, 2024

alexiswl commented Feb 19, 2024 •

edited

Loading

codecov bot commented Feb 19, 2024 •

edited

Loading

alexiswl Feb 21, 2024 •

edited

Loading

mr-c Feb 21, 2024 •

edited

Loading