Implement default locations for data and collection parameters.

Works for both files and collections. Workflow defaults override tool defaults.

TODO:

- Unit test case to ensure this only works for non-default, non-multi data parameters.
- Implement XSD once syntax is finalized.

lib/galaxy/managers/workflows.py

@@ -1686,6 +1686,11 @@ def __module_from_dict(
 
         if "in" in step_dict:
             for input_name, input_dict in step_dict["in"].items():
+                # This is just a bug in gxformat? I think the input
+                # defaults should be called input to match the input modules's
+                # input parameter name.
+                if input_name == "default":
+                    input_name = "input"
                 step_input = step.get_or_add_input(input_name)
                 NO_DEFAULT_DEFINED = object()
                 default = input_dict.get("default", NO_DEFAULT_DEFINED)

lib/galaxy/model/__init__.py

@@ -7062,10 +7062,20 @@ def input_type(self):
 
     @property
     def input_default_value(self):
-        tool_state = self.tool_inputs
-        default_value = tool_state.get("default")
-        if default_value:
-            default_value = json.loads(default_value)["value"]
+        self.get_input_default_value(None)
+
+    def get_input_default_value(self, default_default):
+        # parameter_input and the data parameters handle this slightly differently
+        # unfortunately.
+        if self.type == "parameter_input":
+            tool_state = self.tool_inputs
+            default_value = tool_state.get("default", default_default)
+        else:
+            default_value = default_default
+            for step_input in self.inputs:
+                if step_input.name == "input" and step_input.default_value_set:
+                    default_value = step_input.default_value
+                    break
         return default_value
 
     @property

lib/galaxy/tool_util/parser/interface.py

@@ -7,6 +7,7 @@
 )
 from os.path import join
 from typing import (
+    Any,
     Dict,
     List,
     Optional,
@@ -382,6 +383,9 @@ def parse_test_input_source(self) -> "InputSource":
     def parse_when_input_sources(self):
         raise NotImplementedError(NOT_IMPLEMENTED_MESSAGE)
 
+    def parse_default(self) -> Optional[Dict[str, Any]]:
+        return None
+
 
 class PageSource(metaclass=ABCMeta):
     def parse_display(self):

lib/galaxy/tool_util/parser/xml.py

@@ -3,7 +3,11 @@
 import math
 import re
 import uuid
-from typing import Optional
+from typing import (
+    Any,
+    Dict,
+    Optional,
+)
 
 import packaging.version
 
@@ -1195,6 +1199,56 @@ def parse_when_input_sources(self):
             sources.append((value, case_page_source))
         return sources
 
+    def parse_default(self) -> Optional[Dict[str, Any]]:
+        def file_default_from_elem(elem):
+            # TODO: hashes, created_from_basename, etc...
+            return {"class": "File", "location": elem.get("location")}
+
+        def read_elements(collection_elem):
+            element_dicts = []
+            elements = collection_elem.findall("element")
+            for element in elements:
+                identifier = element.get("name")
+                subcollection_elem = element.find("collection")
+                if subcollection_elem:
+                    collection_type = subcollection_elem.get("collection_type")
+                    element_dicts.append(
+                        {
+                            "class": "Collection",
+                            "identifier": identifier,
+                            "collection_type": collection_type,
+                            "elements": read_elements(subcollection_elem),
+                        }
+                    )
+                else:
+                    element_dict = file_default_from_elem(element)
+                    element_dict["identifier"] = identifier
+                    element_dicts.append(element_dict)
+            return element_dicts
+
+        elem = self.input_elem
+        element_type = self.input_elem.get("type")
+        if element_type == "data":
+            default_elem = elem.find("default")
+            if default_elem is not None:
+                return file_default_from_elem(default_elem)
+            else:
+                return None
+        else:
+            default_elem = elem.find("default")
+            if default_elem is not None:
+                default_elem = elem.find("default")
+                collection_type = default_elem.get("collection_type")
+                name = default_elem.get("name", elem.get("name"))
+                return {
+                    "class": "Collection",
+                    "name": name,
+                    "collection_type": collection_type,
+                    "elements": read_elements(default_elem),
+                }
+            else:
+                return None
+
 
 class ParallelismInfo:
     """

lib/galaxy/tool_util/parser/yaml.py

@@ -1,7 +1,9 @@
 import json
 from typing import (
+    Any,
     Dict,
     List,
+    Optional,
 )
 
 import packaging.version
@@ -357,6 +359,11 @@ def parse_static_options(self):
             static_options.append((label, value, selected))
         return static_options
 
+    def parse_default(self) -> Optional[Dict[str, Any]]:
+        input_dict = self.input_dict
+        default_def = input_dict.get("default", None)
+        return default_def
+
 
 def _ensure_has(dict, defaults):
     for key, value in defaults.items():

lib/galaxy/tools/parameters/basic.py

@@ -24,12 +24,16 @@
 from galaxy.model import (
     cached_id,
     Dataset,
+    DatasetCollection,
     DatasetCollectionElement,
+    DatasetHash,
     DatasetInstance,
+    DatasetSource,
     HistoryDatasetAssociation,
     HistoryDatasetCollectionAssociation,
     LibraryDatasetDatasetAssociation,
 )
+from galaxy.model.dataset_collections import builder
 from galaxy.schema.fetch_data import FilesPayload
 from galaxy.tool_util.parser import get_input_source as ensure_input_source
 from galaxy.util import (
@@ -42,6 +46,7 @@
 )
 from galaxy.util.dictifiable import Dictifiable
 from galaxy.util.expressions import ExpressionContext
+from galaxy.util.hash_util import HASH_NAMES
 from galaxy.util.rules_dsl import RuleSet
 from . import (
     dynamic_options,
@@ -2020,6 +2025,11 @@ def __init__(self, tool, input_source, trans=None):
         self._parse_options(input_source)
         # Load conversions required for the dataset input
         self.conversions = []
+        self.default_object = input_source.parse_default()
+        if self.optional and self.default_object is not None:
+            raise ParameterValueError(
+                "Cannot specify a Galaxy tool data parameter to be both optional and have a default value.", self.name
+            )
         for name, conv_extension in input_source.parse_conversion_tuples():
             assert None not in [
                 name,
@@ -2038,9 +2048,11 @@ def from_json(self, value, trans, other_values=None):
         other_values = other_values or {}
         if trans.workflow_building_mode is workflow_building_modes.ENABLED or is_runtime_value(value):
             return None
-        if not value and not self.optional:
+        if not value and not self.optional and not self.default_object:
             raise ParameterValueError("specify a dataset of the required format / build for parameter", self.name)
         if value in [None, "None", ""]:
+            if self.default_object:
+                return raw_to_galaxy(trans, self.default_object)
             return None
         if isinstance(value, dict) and "values" in value:
             value = self.to_python(value, trans.app)
@@ -2316,6 +2328,11 @@ def __init__(self, tool, input_source, trans=None):
         self.multiple = False  # Accessed on DataToolParameter a lot, may want in future
         self.is_dynamic = True
         self._parse_options(input_source)  # TODO: Review and test.
+        self.default_object = input_source.parse_default()
+        if self.optional and self.default_object is not None:
+            raise ParameterValueError(
+                "Cannot specify a Galaxy tool data parameter to be both optional and have a default value.", self.name
+            )
 
     @property
     def collection_types(self):
@@ -2350,9 +2367,11 @@ def from_json(self, value, trans, other_values=None):
         rval: Optional[Union[DatasetCollectionElement, HistoryDatasetCollectionAssociation]] = None
         if trans.workflow_building_mode is workflow_building_modes.ENABLED:
             return None
-        if not value and not self.optional:
+        if not value and not self.optional and not self.default_object:
             raise ParameterValueError("specify a dataset collection of the correct type", self.name)
         if value in [None, "None"]:
+            if self.default_object:
+                return raw_to_galaxy(trans, self.default_object)
             return None
         if isinstance(value, dict) and "values" in value:
             value = self.to_python(value, trans.app)
@@ -2664,6 +2683,90 @@ def to_text(self, value):
             return ""
 
 
+# Code from CWL branch to massage in order to be shared across tools and workflows,
+# and for CWL artifacts as well as Galaxy ones.
+def raw_to_galaxy(trans, as_dict_value):
+    app = trans.app
+    history = trans.history
+
+    object_class = as_dict_value["class"]
+    if object_class == "File":
+        relative_to = "/"  # TODO
+        from galaxy.tool_util.cwl.util import abs_path
+
+        path = abs_path(as_dict_value.get("location"), relative_to)
+
+        name = os.path.basename(path)
+        extension = as_dict_value.get("format") or "data"
+        dataset = Dataset()
+        source = DatasetSource()
+        source.source_uri = path
+        # TODO: validate this...
+        source.transform = as_dict_value.get("transform")
+        dataset.sources.append(source)
+
+        for hash_name in HASH_NAMES:
+            # TODO: Convert md5 -> MD5 during tool parsing.
+            if hash_name in as_dict_value:
+                hash_object = DatasetHash()
+                hash_object.hash_function = hash_name
+                hash_object.hash_value = as_dict_value[hash_name]
+                dataset.hashes.append(hash_object)
+
+        if "created_from_basename" in as_dict_value:
+            dataset.created_from_basename = as_dict_value["created_from_basename"]
+
+        dataset.state = Dataset.states.DEFERRED
+        primary_data = HistoryDatasetAssociation(
+            name=name,
+            extension=extension,
+            metadata_deferred=True,
+            designation=None,
+            visible=True,
+            dbkey="?",
+            dataset=dataset,
+            flush=False,
+            sa_session=trans.sa_session,
+        )
+        primary_data.state = Dataset.states.DEFERRED
+        permissions = app.security_agent.history_get_default_permissions(history)
+        app.security_agent.set_all_dataset_permissions(primary_data.dataset, permissions, new=True, flush=False)
+        trans.sa_session.add(primary_data)
+        history.stage_addition(primary_data)
+        history.add_pending_items()
+        trans.sa_session.flush()
+        return primary_data
+    else:
+        name = as_dict_value.get("name")
+        collection_type = as_dict_value.get("collection_type")
+        collection = DatasetCollection(
+            collection_type=collection_type,
+        )
+        hdca = HistoryDatasetCollectionAssociation(
+            name=name,
+            collection=collection,
+        )
+
+        def write_elements_to_collection(has_elements, collection_builder):
+            element_dicts = has_elements.get("elements")
+            for element_dict in element_dicts:
+                element_class = element_dict["class"]
+                identifier = element_dict["identifier"]
+                if element_class == "File":
+                    hda = raw_to_galaxy(trans, element_dict)
+                    collection_builder.add_dataset(identifier, hda)
+                else:
+                    subcollection_builder = collection_builder.get_level(identifier)
+                    write_elements_to_collection(element_dict, subcollection_builder)
+
+        collection_builder = builder.BoundCollectionBuilder(collection)
+        write_elements_to_collection(as_dict_value, collection_builder)
+        collection_builder.populate()
+        trans.sa_session.add(hdca)
+        trans.sa_session.flush()
+        return hdca
+
+
 parameter_types = dict(
     text=TextToolParameter,
     integer=IntegerToolParameter,