Make parsers inherit from schema_salad.metaschema

Makefile

@@ -194,19 +194,22 @@ flake8: FORCE
 src/cwl_utils/parser/cwl_v1_0.py: FORCE
 	schema-salad-tool --codegen python \
 		--codegen-parser-info "org.w3id.cwl.v1_0" \
-		https://github.com/common-workflow-language/common-workflow-language/raw/codegen/v1.0/extensions.yml \
+		--codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema" \
+		https://github.com/common-workflow-language/common-workflow-language/raw/codegen/v1.0/CommonWorkflowLanguage.yml \
 		> $@
 
 src/cwl_utils/parser/cwl_v1_1.py: FORCE
 	schema-salad-tool --codegen python \
 		--codegen-parser-info "org.w3id.cwl.v1_1" \
-		https://github.com/common-workflow-language/cwl-v1.1/raw/codegen/extensions.yml \
+		--codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema" \
+		https://github.com/common-workflow-language/cwl-v1.1/raw/codegen/CommonWorkflowLanguage.yml \
 		> $@
 
 src/cwl_utils/parser/cwl_v1_2.py: FORCE
 	schema-salad-tool --codegen python \
 		--codegen-parser-info "org.w3id.cwl.v1_2" \
-		https://github.com/common-workflow-language/cwl-v1.2/raw/codegen/extensions.yml \
+		--codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema" \
+		https://github.com/common-workflow-language/cwl-v1.2/raw/codegen/CommonWorkflowLanguage.yml \
 		> $@
 
 regen_parsers: src/cwl_utils/parser/cwl_v1_*.py

README.rst

@@ -160,13 +160,13 @@ Regenerate parsers
 To regenerate install the ``schema_salad`` package and run:
 
 ``cwl_utils/parser/cwl_v1_0.py`` was created via
-``schema-salad-tool --codegen python https://github.com/common-workflow-language/common-workflow-language/raw/codegen/v1.0/extensions.yml --codegen-parser-info "org.w3id.cwl.v1_0" > cwl_utils/parser/cwl_v1_0.py``
+``schema-salad-tool --codegen python https://github.com/common-workflow-language/common-workflow-language/raw/codegen/v1.0/CommonWorkflowLanguage.yml --codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema" --codegen-parser-info "org.w3id.cwl.v1_0" > cwl_utils/parser/cwl_v1_0.py``
 
 ``cwl_utils/parser/cwl_v1_1.py`` was created via
-``schema-salad-tool --codegen python https://github.com/common-workflow-language/cwl-v1.1/raw/codegen/extensions.yml --codegen-parser-info "org.w3id.cwl.v1_1" > cwl_utils/parser/cwl_v1_1.py``
+``schema-salad-tool --codegen python https://github.com/common-workflow-language/cwl-v1.1/raw/codegen/CommonWorkflowLanguage.yml --codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema" --codegen-parser-info "org.w3id.cwl.v1_1" > cwl_utils/parser/cwl_v1_1.py``
 
 ``cwl_utils/parser/cwl_v1_2.py`` was created via
-``schema-salad-tool --codegen python https://github.com/common-workflow-language/cwl-v1.2/raw/codegen/extensions.yml --codegen-parser-info "org.w3id.cwl.v1_2" > cwl_utils/parser/cwl_v1_2.py``
+``schema-salad-tool --codegen python https://github.com/common-workflow-language/cwl-v1.2/raw/codegen/CommonWorkflowLanguage.yml --codegen-parent "https://w3id.org/cwl/salad=schema_salad.metaschema" --codegen-parser-info "org.w3id.cwl.v1_2" > cwl_utils/parser/cwl_v1_2.py``
 
 Release
 ~~~~~~~

pyproject.toml

@@ -36,7 +36,7 @@ dependencies = [
     "packaging",
     "rdflib",
     "requests",
-    "schema-salad >= 8.8.20250205075315,<9",
+    "schema-salad @ git+https://github.com/common-workflow-language/schema_salad@refs/pull/1028/head",
     "ruamel.yaml >= 0.17.6, < 0.20",
     "typing_extensions >= 4.10.0",
 ]
@@ -73,6 +73,9 @@ cwl-inputs-schema-gen = "cwl_utils.inputs_schema_gen:main"
 [tool.pytest.ini_options]
 addopts = "-rsx -n auto --pyargs cwl_utils.tests"
 
+[tool.hatch.metadata]
+allow-direct-references = true
+
 [tool.hatch.version]
 path = "src/cwl_utils/__meta__.py"
 

src/cwl_utils/cwl_v1_0_expression_refactor.py

@@ -11,6 +11,8 @@
 from typing import Any, cast
 
 from ruamel import yaml
+from schema_salad.metaschema import ArraySchema
+from schema_salad.runtime import LoadingOptions, save
 from schema_salad.sourceline import SourceLine
 from schema_salad.utils import json_dumps
 
@@ -41,7 +43,7 @@ def expand_stream_shortcuts(process: cwl.CommandLineTool) -> cwl.CommandLineTool
             stdout_path = process.stdout
             if not stdout_path:
                 stdout_path = hashlib.sha1(  # nosec
-                    json_dumps(cwl.save(process)).encode("utf-8")
+                    json_dumps(save(process)).encode("utf-8")
                 ).hexdigest()
                 result.stdout = stdout_path
             result.outputs[index].type_ = "File"
@@ -57,11 +59,11 @@ def escape_expression_field(contents: str) -> str:
 
 
 def clean_type_ids(
-    cwltype: cwl.ArraySchema | cwl.InputRecordSchema,
-) -> cwl.ArraySchema | cwl.InputRecordSchema:
+    cwltype: ArraySchema | cwl.InputRecordSchema,
+) -> ArraySchema | cwl.InputRecordSchema:
     """Simplify type identifiers."""
     result = copy.deepcopy(cwltype)
-    if isinstance(result, cwl.ArraySchema):
+    if isinstance(result, ArraySchema):
         if isinstance(result.items, MutableSequence):
             for item in result.items:
                 if hasattr(item, "id"):
@@ -340,8 +342,8 @@ def generate_etool_from_expr(
             self_type = target
         if isinstance(self_type, list):
             new_type: (
-                list[cwl.ArraySchema | cwl.InputRecordSchema]
-                | cwl.ArraySchema
+                list[ArraySchema | cwl.InputRecordSchema]
+                | ArraySchema
                 | cwl.InputRecordSchema
             ) = [clean_type_ids(t.type_) for t in self_type if t.type_]
         elif self_type.type_:
@@ -1218,7 +1220,7 @@ def process_level_reqs(
 
 
 def add_input_to_process(
-    process: cwl.Process, name: str, inptype: Any, loadingOptions: cwl.LoadingOptions
+    process: cwl.Process, name: str, inptype: Any, loadingOptions: LoadingOptions
 ) -> None:
     """Add a new InputParameter to the given CommandLineTool."""
     if isinstance(process, cwl.CommandLineTool):
@@ -1347,7 +1349,7 @@ def traverse_CommandLineTool(
                     modified = True
                     inp_id = "_{}_glob".format(outp.id.split("#")[-1])
                     etool_id = f"_expression_{step_id}{inp_id}"
-                    glob_target_type = ["string", cwl.ArraySchema("string", "array")]
+                    glob_target_type = ["string", ArraySchema("string", "array")]
                     target = cwl.InputParameter(id=None, type_=glob_target_type)
                     replace_step_clt_expr_with_etool(
                         expression, etool_id, parent, target, step, replace_etool
@@ -1843,7 +1845,7 @@ def traverse_step(
                                     source_types.append(temp_type)
                         source_type = cwl.InputParameter(
                             id=None,
-                            type_=cwl.ArraySchema(source_types, "array"),
+                            type_=ArraySchema(source_types, "array"),
                         )
                     else:
                         input_source_id = inp.source.split("#")[-1]

src/cwl_utils/cwl_v1_1_expression_refactor.py

@@ -11,6 +11,8 @@
 from typing import Any, cast
 
 from ruamel import yaml
+from schema_salad.metaschema import ArraySchema
+from schema_salad.runtime import LoadingOptions, save
 from schema_salad.sourceline import SourceLine
 from schema_salad.utils import json_dumps
 
@@ -41,7 +43,7 @@ def expand_stream_shortcuts(process: cwl.CommandLineTool) -> cwl.CommandLineTool
             stdout_path = process.stdout
             if not stdout_path:
                 stdout_path = hashlib.sha1(  # nosec
-                    json_dumps(cwl.save(process)).encode("utf-8")
+                    json_dumps(save(process)).encode("utf-8")
                 ).hexdigest()
                 result.stdout = stdout_path
             result.outputs[index].type_ = "File"
@@ -57,11 +59,11 @@ def escape_expression_field(contents: str) -> str:
 
 
 def clean_type_ids(
-    cwltype: cwl.ArraySchema | cwl.InputRecordSchema,
-) -> cwl.ArraySchema | cwl.InputRecordSchema:
+    cwltype: ArraySchema | cwl.InputRecordSchema,
+) -> ArraySchema | cwl.InputRecordSchema:
     """Simplify type identifiers."""
     result = copy.deepcopy(cwltype)
-    if isinstance(result, cwl.ArraySchema):
+    if isinstance(result, ArraySchema):
         if isinstance(result.items, MutableSequence):
             for item in result.items:
                 if hasattr(item, "id"):
@@ -340,8 +342,8 @@ def generate_etool_from_expr(
             self_type = target
         if isinstance(self_type, list):
             new_type: (
-                list[cwl.ArraySchema | cwl.InputRecordSchema]
-                | cwl.ArraySchema
+                list[ArraySchema | cwl.InputRecordSchema]
+                | ArraySchema
                 | cwl.InputRecordSchema
             ) = [clean_type_ids(t.type_) for t in self_type]
         else:
@@ -1220,7 +1222,7 @@ def process_level_reqs(
 
 
 def add_input_to_process(
-    process: cwl.Process, name: str, inptype: Any, loadingOptions: cwl.LoadingOptions
+    process: cwl.Process, name: str, inptype: Any, loadingOptions: LoadingOptions
 ) -> None:
     """Add a new InputParameter to the given CommandLineTool."""
     if isinstance(process, cwl.CommandLineTool):
@@ -1349,7 +1351,7 @@ def traverse_CommandLineTool(
                     modified = True
                     inp_id = "_{}_glob".format(outp.id.split("#")[-1])
                     etool_id = f"_expression_{step_id}{inp_id}"
-                    glob_target_type = ["string", cwl.ArraySchema("string", "array")]
+                    glob_target_type = ["string", ArraySchema("string", "array")]
                     target = cwl.WorkflowInputParameter(id=None, type_=glob_target_type)
                     replace_step_clt_expr_with_etool(
                         expression, etool_id, parent, target, step, replace_etool
@@ -1851,7 +1853,7 @@ def traverse_step(
                                     source_types.append(temp_type)
                         source_type = cwl.WorkflowInputParameter(
                             id=None,
-                            type_=cwl.ArraySchema(source_types, "array"),
+                            type_=ArraySchema(source_types, "array"),
                         )
                     else:
                         input_source_id = inp.source.split("#")[-1]

src/cwl_utils/cwl_v1_2_expression_refactor.py

@@ -11,6 +11,8 @@
 from typing import Any, cast
 
 from ruamel import yaml
+from schema_salad.metaschema import ArraySchema
+from schema_salad.runtime import LoadingOptions, save
 from schema_salad.sourceline import SourceLine
 from schema_salad.utils import json_dumps
 
@@ -40,7 +42,7 @@ def expand_stream_shortcuts(process: cwl.CommandLineTool) -> cwl.CommandLineTool
             stdout_path = process.stdout
             if not stdout_path:
                 stdout_path = hashlib.sha1(  # nosec
-                    json_dumps(cwl.save(process)).encode("utf-8")
+                    json_dumps(save(process)).encode("utf-8")
                 ).hexdigest()
                 result.stdout = stdout_path
             result.outputs[index].type_ = "File"
@@ -56,11 +58,11 @@ def escape_expression_field(contents: str) -> str:
 
 
 def clean_type_ids(
-    cwltype: cwl.ArraySchema | cwl.InputRecordSchema,
-) -> cwl.ArraySchema | cwl.InputRecordSchema:
+    cwltype: ArraySchema | cwl.InputRecordSchema,
+) -> ArraySchema | cwl.InputRecordSchema:
     """Simplify type identifiers."""
     result = copy.deepcopy(cwltype)
-    if isinstance(result, cwl.ArraySchema):
+    if isinstance(result, ArraySchema):
         if isinstance(result.items, MutableSequence):
             for item in result.items:
                 if hasattr(item, "id"):
@@ -339,8 +341,8 @@ def generate_etool_from_expr(
             self_type = target
         if isinstance(self_type, list):
             new_type: (
-                list[cwl.ArraySchema | cwl.InputRecordSchema]
-                | cwl.ArraySchema
+                list[ArraySchema | cwl.InputRecordSchema]
+                | ArraySchema
                 | cwl.InputRecordSchema
             ) = [clean_type_ids(t.type_) for t in self_type]
         else:
@@ -715,7 +717,7 @@ def process_workflow_inputs_and_outputs(
                 else:
                     sources = [s.split("#")[-1] for s in param2.outputSource]
                 source_type_items = utils.type_for_source(workflow, sources)
-                if isinstance(source_type_items, cwl.ArraySchema):
+                if isinstance(source_type_items, ArraySchema):
                     if isinstance(source_type_items.items, list):
                         if "null" not in source_type_items.items:
                             source_type_items.items.append("null")
@@ -1323,7 +1325,7 @@ def process_level_reqs(
 
 
 def add_input_to_process(
-    process: cwl.Process, name: str, inptype: Any, loadingOptions: cwl.LoadingOptions
+    process: cwl.Process, name: str, inptype: Any, loadingOptions: LoadingOptions
 ) -> None:
     """Add a new InputParameter to the given CommandLineTool."""
     if isinstance(process, cwl.CommandLineTool):
@@ -1452,7 +1454,7 @@ def traverse_CommandLineTool(
                     modified = True
                     inp_id = "_{}_glob".format(outp.id.split("#")[-1])
                     etool_id = f"_expression_{step_id}{inp_id}"
-                    glob_target_type = ["string", cwl.ArraySchema("string", "array")]
+                    glob_target_type = ["string", ArraySchema("string", "array")]
                     target = cwl.WorkflowInputParameter(id=None, type_=glob_target_type)
                     replace_step_clt_expr_with_etool(
                         expression, etool_id, parent, target, step, replace_etool
@@ -1954,7 +1956,7 @@ def traverse_step(
                                     source_types.append(temp_type)
                         source_type = cwl.WorkflowInputParameter(
                             id=None,
-                            type_=cwl.ArraySchema(source_types, "array"),
+                            type_=ArraySchema(source_types, "array"),
                         )
                     else:
                         input_source_id = inp.source.split("#")[-1]

src/cwl_utils/expression_refactor.py

@@ -13,6 +13,7 @@
 
 from ruamel.yaml.main import YAML
 from ruamel.yaml.scalarstring import walk_tree
+from schema_salad.runtime import save
 
 from cwl_utils import (
     cwl_v1_0_expression_refactor,
@@ -111,15 +112,12 @@ def refactor(args: argparse.Namespace) -> int:
                 traverse: Callable[[Any, bool, bool, bool, bool], tuple[Any, bool]] = (
                     cwl_v1_0_expression_refactor.traverse
                 )
-                save: saveCWL = cwl_v1_0.save
             case "v1.1":
                 top = cwl_v1_1.load_document_by_yaml(result, uri)
                 traverse = cwl_v1_1_expression_refactor.traverse
-                save = cwl_v1_1.save
             case "v1.2":
                 top = cwl_v1_2.load_document_by_yaml(result, uri)
                 traverse = cwl_v1_2_expression_refactor.traverse
-                save = cwl_v1_2.save
             case _:
                 _logger.error(
                     "Sorry, %s is not a supported CWL version by this tool.",

src/cwl_utils/normalizer.py

@@ -12,12 +12,13 @@
 
 from cwlupgrader import main as cwlupgrader
 from ruamel import yaml
+from schema_salad.runtime import save
 from schema_salad.sourceline import add_lc_filename
 
 from cwl_utils import cwl_v1_2_expression_refactor
 from cwl_utils.loghandler import _logger as _cwlutilslogger
 from cwl_utils.pack import pack
-from cwl_utils.parser.cwl_v1_2 import load_document_by_yaml, save
+from cwl_utils.parser.cwl_v1_2 import load_document_by_yaml
 
 _logger = logging.getLogger("cwl-normalizer")  # pylint: disable=invalid-name
 defaultStreamHandler = logging.StreamHandler()  # pylint: disable=invalid-name