Add prefix-based inheritance to codegen (#1028)

Adds the possibility to specify a parent parser for
each prefix in a Schema SALAD specification. The idea is that each
extensions could come with a different namespace, and would define
its own parser for its classes.

With this commit, a user can leverage the `--codegen-parent` option
to explicitly map a prefix to a parent parser (e.g., a Python module)
with a "prefix=package" syntax. For example:

```bash
--codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema"
```

Inheritance is not able to automatically reconstruct the whole class
hierarchy, meaning that if B inherits from A and C inherits from B,
the user must manually specify the mapping also for namespace A.
However, we do not expect huge parser hierarchies for schema SALAD
specifications.

Co-authored-by: Michael R. Crusoe <1330696+mr-c@users.noreply.github.com>

Author: GlassOfWhiskey

.github/workflows/ci-tests.yml

@@ -61,7 +61,7 @@ jobs:
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@e79a6962e0d4c0c17b229090214935d2e33f8354 # v6
         with:
-          fail_ci_if_error: true
+          fail_ci_if_error: false
         env:
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
 

Makefile

@@ -122,7 +122,7 @@ diff_pylint_report: pylint_report.txt
 	diff-quality --compare-branch=origin/main --violations=pylint pylint_report.txt
 
 .coverage:
-	pytest --cov --cov-config=.coveragerc --cov-report= ${PYTEST_EXTRA}
+	pytest --cov --cov-config=.coveragerc --cov-report= --junitxml=junit.xml -o junit_family=legacy ${PYTEST_EXTRA}
 	$(COVBASE) -m schema_salad.main \
 		--print-jsonld-context src/schema_salad/metaschema/metaschema.yml \
 		> /dev/null
@@ -226,7 +226,7 @@ release:
 flake8: FORCE
 	flake8 $(PYSOURCES)
 
-src/schema_salad/metaschema.py: src/schema_salad/codegen_base.py src/schema_salad/python_codegen_support.py src/schema_salad/python_codegen.py src/schema_salad/metaschema/*.yml
+src/schema_salad/metaschema.py: src/schema_salad/codegen_base.py src/schema_salad/runtime.py src/schema_salad/python_codegen.py src/schema_salad/metaschema/*.yml
 	schema-salad-tool --codegen python src/schema_salad/metaschema/metaschema.yml > $@
 
 vpath %.yml src/schema_salad/tests/cpp_tests

README.rst

@@ -51,17 +51,20 @@ Schema salad can be used as a command line tool or imported as a Python module::
 
    $ schema-salad-tool
    usage: schema-salad-tool [-h] [--rdf-serializer RDF_SERIALIZER] [--skip-schemas]
-                         [--strict-foreign-properties] [--print-jsonld-context]
-                         [--print-rdfs] [--print-avro] [--print-rdf] [--print-pre]
-                         [--print-index] [--print-metadata] [--print-inheritance-dot]
-                         [--print-fieldrefs-dot] [--codegen language] [--codegen-target CODEGEN_TARGET]
-                         [--codegen-examples directory] [--codegen-package dotted.package]
-                         [--codegen-copyright copyright_string] [--print-oneline]
-                         [--print-doc] [--strict | --non-strict]
+                         [--strict-foreign-properties] [--version]
+                         [--print-jsonld-context | --print-rdfs | --print-avro |
+                              --print-rdf | --print-pre | --print-index | --print-metadata |
+                              --print-inheritance-dot | --print-fieldrefs-dot | --codegen LANGUAGE |
+                              --print-oneline | --print-doc]
+                         [--codegen-target CODEGEN_TARGET] [--codegen-examples DIRECTORY]
+                         [--codegen-package DOTTED.PACKAGE] [--codegen-parent PARENTS_MAP]
+                         [--codegen-copyright COPYRIGHT_STRING]
+                         [--codegen-spdx-copyright-text SPDX_COPYRIGHT_TEXT [SPDX_COPYRIGHT_TEXT ...]]
+                         [--codegen-spdx-license-identifier SPDX_LICENSE_IDENTIFIER]
+                         [--codegen-parser-info PARSER_INFO] [--strict | --non-strict]
                          [--verbose | --quiet | --debug] [--only ONLY] [--redirect REDIRECT]
                          [--brand BRAND] [--brandlink BRANDLINK] [--brandstyle BRANDSTYLE]
-                         [--brandinverse] [--primtype PRIMTYPE] [--version]
-                         [schema] [document]
+                         [--brandinverse] [--primtype PRIMTYPE] [schema] [document ...]
 
    $ python
    >>> import schema_salad

src/schema_salad/codegen.py

@@ -31,6 +31,7 @@ def codegen(
     examples: str | None = None,
     package: str | None = None,
     copyright: str | None = None,
+    parents_map: dict[str, str] | None = None,
     spdx_copyright_text: list[str] | None = None,
     spdx_license_identifier: str | None = None,
     parser_info: str | None = None,
@@ -87,7 +88,11 @@ def codegen(
                     return
                 case "python":
                     gen = PythonCodeGen(
-                        dest, copyright=copyright, parser_info=info, salad_version=salad_version
+                        dest,
+                        copyright=copyright,
+                        parents_map=parents_map,
+                        parser_info=info,
+                        salad_version=salad_version,
                     )
         case "java":
             gen = JavaCodeGen(

src/schema_salad/main.py

@@ -29,6 +29,15 @@
 _logger: Final = logging.getLogger("salad")
 
 
+def parse_kv(pair: str) -> tuple[str, str]:
+    """Parse `key=value` arguments into a tuple of `key` and `value`."""
+    try:
+        k, v = pair.split("=", 1)
+        return k, v
+    except ValueError as e:
+        raise argparse.ArgumentTypeError(f"Invalid format: '{pair}', expected key=value") from e
+
+
 def printrdf(
     workflow: str,
     wf: CommentedMap | CommentedSeq,
@@ -147,6 +156,15 @@ def arg_parser() -> argparse.ArgumentParser:
         "from the base URL (Java/TypeScript/.Net/Dlang only).",
     )
 
+    codegen_opts.add_argument(
+        "--codegen-parent",
+        type=parse_kv,
+        metavar="PARENTS_MAP",
+        action="append",
+        default=None,
+        help="Optional mapping of prefix into parent package for generated classes (Python only).",
+    )
+
     codegen_opts.add_argument(
         "--codegen-copyright",
         type=str,
@@ -351,6 +369,7 @@ def main(argsl: list[str] | None = None) -> int:
             examples=args.codegen_examples,
             package=args.codegen_package,
             copyright=args.codegen_copyright,
+            parents_map=dict(args.codegen_parent or []),
             spdx_license_identifier=args.codegen_spdx_license_identifier,
             spdx_copyright_text=args.codegen_spdx_copyright_text,
             parser_info=args.codegen_parser_info,