From 83b99b468af6792d1313f18ff741f3d32a7c3585 Mon Sep 17 00:00:00 2001
From: "Michael R. Crusoe" <michael.crusoe@gmail.com>
Date: Fri, 29 Sep 2023 16:49:23 +0200
Subject: [PATCH] extend validation to the inputs of the top level proces

---
 cwltool/context.py                   |  3 +++
 cwltool/executors.py                 | 12 +++++++++
 cwltool/main.py                      |  6 ++++-
 tests/test_validate.py               | 40 +++++++++++++++++++++++++++-
 tests/wf/1st-workflow_bad_inputs.yml |  4 +++
 5 files changed, 63 insertions(+), 2 deletions(-)
 create mode 100644 tests/wf/1st-workflow_bad_inputs.yml

diff --git a/cwltool/context.py b/cwltool/context.py
index 7f30e4c30..cbfa24102 100644
--- a/cwltool/context.py
+++ b/cwltool/context.py
@@ -2,6 +2,7 @@
 import copy
 import os
 import shutil
+import sys
 import tempfile
 import threading
 from typing import (
@@ -197,6 +198,8 @@ def __init__(self, kwargs: Optional[Dict[str, Any]] = None) -> None:
         self.mpi_config: MpiConfig = MpiConfig()
         self.default_stdout: Optional[Union[IO[bytes], TextIO]] = None
         self.default_stderr: Optional[Union[IO[bytes], TextIO]] = None
+        self.validate_only: bool = False
+        self.validate_stdout: Union[IO[bytes], TextIO, IO[str]] = sys.stdout
         super().__init__(kwargs)
         if self.tmp_outdir_prefix == "":
             self.tmp_outdir_prefix = self.tmpdir_prefix
diff --git a/cwltool/executors.py b/cwltool/executors.py
index 3cbac72c7..02b4a40d8 100644
--- a/cwltool/executors.py
+++ b/cwltool/executors.py
@@ -143,6 +143,8 @@ def check_for_abstract_op(tool: CWLObjectType) -> None:
                 process.requirements.append(req)
 
         self.run_jobs(process, job_order_object, logger, runtime_context)
+        if runtime_context.validate_only is True:
+            return (None, "ValidationSuccess")
 
         if self.final_output and self.final_output[0] is not None and finaloutdir is not None:
             self.final_output[0] = relocateOutputs(
@@ -238,6 +240,16 @@ def run_jobs(
                             process_run_id = prov_obj.record_process_start(process, job)
                             runtime_context = runtime_context.copy()
                         runtime_context.process_run_id = process_run_id
+                    if runtime_context.validate_only is True:
+                        if isinstance(job, WorkflowJob):
+                            name = job.tool.lc.filename
+                        else:
+                            name = getattr(job, "name", str(job))
+                        print(
+                            f"{name} is valid CWL. No errors detected in the inputs.",
+                            file=runtime_context.validate_stdout,
+                        )
+                        return
                     job.run(runtime_context)
                 else:
                     logger.error("Workflow cannot make any more progress.")
diff --git a/cwltool/main.py b/cwltool/main.py
index 965f863a0..527275f4c 100755
--- a/cwltool/main.py
+++ b/cwltool/main.py
@@ -1134,7 +1134,7 @@ def main(
                 make_template(tool, stdout)
                 return 0
 
-            if args.validate:
+            if len(args.job_order) == 0 and args.validate:
                 print(f"{args.workflow} is valid CWL.", file=stdout)
                 return 0
 
@@ -1294,10 +1294,14 @@ def main(
                     use_biocontainers=args.beta_use_biocontainers,
                     container_image_cache_path=args.beta_dependencies_directory,
                 )
+            runtimeContext.validate_only = args.validate
+            runtimeContext.validate_stdout = stdout
 
             (out, status) = real_executor(
                 tool, initialized_job_order_object, runtimeContext, logger=_logger
             )
+            if runtimeContext.validate_only is True:
+                return 0
 
             if out is not None:
                 if runtimeContext.research_obj is not None:
diff --git a/tests/test_validate.py b/tests/test_validate.py
index e809df386..be50d357d 100644
--- a/tests/test_validate.py
+++ b/tests/test_validate.py
@@ -1,5 +1,5 @@
 """Tests --validation."""
-
+import re
 
 from .util import get_data, get_main_output
 
@@ -14,3 +14,41 @@ def test_validate_graph_with_no_default() -> None:
     assert "packed_no_main.cwl#cat is valid CWL" in stdout
     assert "packed_no_main.cwl#collision is valid CWL" in stdout
     assert "tests/wf/packed_no_main.cwl is valid CWL" in stdout
+
+
+def test_validate_with_valid_input_object() -> None:
+    """Ensure that --validate with a valid input object."""
+    exit_code, stdout, stderr = get_main_output(
+        [
+            "--validate",
+            get_data("tests/wf/1st-workflow.cwl"),
+            "--inp",
+            get_data("tests/wf/1st-workflow.cwl"),
+            "--ex",
+            "FOO",
+        ]
+    )
+    assert exit_code == 0
+    assert "tests/wf/1st-workflow.cwl is valid CWL. No errors detected in the inputs." in stdout
+
+
+def test_validate_with_invalid_input_object() -> None:
+    """Ensure that --validate with an invalid input object."""
+    exit_code, stdout, stderr = get_main_output(
+        [
+            "--validate",
+            get_data("tests/wf/1st-workflow.cwl"),
+            get_data("tests/wf/1st-workflow_bad_inputs.yml"),
+        ]
+    )
+    assert exit_code == 1
+    stderr = re.sub(r"\s\s+", " ", stderr)
+    assert "Invalid job input record" in stderr
+    assert (
+        "tests/wf/1st-workflow_bad_inputs.yml:2:1: * the 'ex' field is not "
+        "valid because the value is not string" in stderr
+    )
+    assert (
+        "tests/wf/1st-workflow_bad_inputs.yml:1:1: * the 'inp' field is not "
+        "valid because is not a dict. Expected a File object." in stderr
+    )
diff --git a/tests/wf/1st-workflow_bad_inputs.yml b/tests/wf/1st-workflow_bad_inputs.yml
new file mode 100644
index 000000000..d95783be2
--- /dev/null
+++ b/tests/wf/1st-workflow_bad_inputs.yml
@@ -0,0 +1,4 @@
+inp: 42
+ex:
+  class: File
+  path: 1st-workflow.cwl