convert: add input object document

src/runcrate/convert.py

@@ -19,6 +19,7 @@
 import hashlib
 import json
 import re
+from io import StringIO
 from pathlib import Path
 
 import networkx as nx
@@ -37,6 +38,7 @@
 
 
 WORKFLOW_BASENAME = "packed.cwl"
+INPUTS_FILE_BASENAME = "primary-job.json"
 
 CWL_TYPE_MAP = {
     "string": "Text",
@@ -186,6 +188,8 @@ def __init__(self, root, workflow_name=None, license=None, readme=None):
         # index collections by their main entity's id
         self.collections = {}
         self.hashes = {}
+        # map source files to destination files
+        self.file_map = {}
 
     @staticmethod
     def _get_step_maps(cwl_defs):
@@ -256,6 +260,7 @@ def build(self):
         self.add_engine_run(crate)
         self.add_action(crate, self.workflow_run)
         self.patch_workflow_input_collection(crate)
+        self.add_inputs_file(crate)
         return crate
 
     def add_root_metadata(self, crate):
@@ -529,6 +534,11 @@ def convert_param(self, prov_param, crate, convert_secondary=True, parent=None):
                     "sha1": hash_,
                 })
                 self._set_alternate_name(prov_param, action_p, parent=parent)
+                try:
+                    source_k = str(source.resolve(strict=False))
+                except RuntimeError:
+                    source_k = str(source)
+                self.file_map[source_k] = dest
             return action_p
         if "ro:Folder" in type_names:
             hash_ = self.hashes[prov_param.id.localpart]
@@ -644,3 +654,34 @@ def patch_workflow_input_collection(self, crate, wf=None):
         for tool in wf.get("hasPart", []):
             if "ComputationalWorkflow" in as_list(tool.type):
                 self.patch_workflow_input_collection(crate, wf=tool)
+
+    def _map_input_data(self, data):
+        if isinstance(data, list):
+            return [self._map_input_data(_) for _ in data]
+        if isinstance(data, dict):
+            rval = {}
+            for k, v in data.items():
+                if k == "location":
+                    source = self.root / "workflow" / v
+                    try:
+                        source_k = str(source.resolve(strict=False))
+                    except RuntimeError:
+                        source_k = str(source)
+                    dest = self.file_map.get(source_k)
+                    rval[k] = str(dest) if dest else v
+                else:
+                    rval[k] = self._map_input_data(v)
+            return rval
+        return data
+
+    def add_inputs_file(self, crate):
+        path = self.root / "workflow" / INPUTS_FILE_BASENAME
+        if path.is_file():
+            with open(path) as f:
+                data = json.load(f)
+            data = self._map_input_data(data)
+            source = StringIO(json.dumps(data, indent=4))
+            crate.add_file(source, path.name, properties={
+                "name": "input object document",
+                "encodingFormat": "application/json",
+            })

tests/test_cwlprov_crate_builder.py

@@ -32,6 +32,7 @@ def test_revsort(data_dir, tmpdir):
     output = tmpdir / "revsort-run-1-crate"
     license = "Apache-2.0"
     readme = data_dir / "README.md"
+    inputs_file = root / "workflow" / "primary-job.json"
     workflow_name = "RevSort"
     builder = ProvCrateBuilder(root, workflow_name=workflow_name, license=license, readme=readme)
     crate = builder.build()
@@ -140,6 +141,22 @@ def test_revsort(data_dir, tmpdir):
     assert set(_connected(workflow)) == set([
         ("packed.cwl#sorttool.cwl/output", "packed.cwl#main/output"),
     ])
+    inputs_f = crate.get(inputs_file.name)
+    assert inputs_f
+    assert inputs_f.type == "File"
+    assert inputs_f["encodingFormat"] == "application/json"
+    with open(inputs_file) as f:
+        ro_json = json.load(f)
+    with open(output / inputs_file.name) as f:
+        crate_json = json.load(f)
+    assert set(crate_json) == {"input", "reverse_sort"}
+    assert set(crate_json["input"]) == set(ro_json["input"])
+    for k, v in crate_json["input"].items():
+        if k == "location":
+            assert v == ro_json["input"][k].rsplit("/", 1)[-1]
+        else:
+            assert v == ro_json["input"][k]
+
     # file contents
     in_text = (root / "data/32/327fc7aedf4f6b69a42a7c8b808dc5a7aff61376").read_text()
     assert (output / wf_input_file.id).read_text() == in_text