Allow selecting subset of collection by group tag

lib/galaxy/tools/parameters/basic.py

@@ -834,7 +834,7 @@ def get_legal_values(self, trans, other_values):
         else:
             return self.legal_values
 
-    def from_json(self, value, trans, other_values={}):
+    def from_json(self, value, trans, other_values={}, require_legal_value=True):
         legal_values = self.get_legal_values(trans, other_values)
         workflow_building_mode = trans.workflow_building_mode
         for context_value in other_values.values():
@@ -858,7 +858,7 @@ def from_json(self, value, trans, other_values={}):
             return value
         if (not legal_values or value is None) and self.optional:
             return None
-        if not legal_values:
+        if not legal_values and require_legal_value:
             raise ValueError("Parameter %s requires a value, but has no legal values defined." % self.name)
         if isinstance(value, list):
             if not self.multiple:
@@ -877,7 +877,7 @@ def from_json(self, value, trans, other_values={}):
                         return []
                     else:
                         raise ValueError("No option was selected for %s but input is not optional." % self.name)
-            if value not in legal_values:
+            if value not in legal_values and require_legal_value:
                 raise ValueError("An invalid option was selected for %s, %r, please verify." % (self.name, value))
             return value
 
@@ -1016,6 +1016,104 @@ def _get_dbkey_names(self, trans=None):
         return self.tool.app.genome_builds.get_genome_build_names(trans=trans)
 
 
+class SelectTagParameter(SelectToolParameter):
+    """
+    Select set that is composed of a set of tags available for an input.
+    """
+    def __init__(self, tool, input_source):
+        input_source = ensure_input_source(input_source)
+        SelectToolParameter.__init__(self, tool, input_source)
+        self.tool = tool
+        self.tag_key = input_source.get("group", False)
+        self.optional = input_source.get("optional", False)
+        self.multiple = input_source.get("multiple", False)
+        self.accept_default = input_source.get_bool("accept_default", False)
+        if self.accept_default:
+            self.optional = True
+        self.data_ref = input_source.get("data_ref", None)
+        self.ref_input = None
+        # Legacy style default value specification...
+        self.default_value = input_source.get("default_value", None)
+        if self.default_value is None:
+            # Newer style... more in line with other parameters.
+            self.default_value = input_source.get("value", None)
+        self.is_dynamic = True
+
+    def from_json(self, value, trans, other_values={}):
+        """
+        Label convention prepends column number with a 'c', but tool uses the integer. This
+        removes the 'c' when entered into a workflow.
+        """
+        if self.multiple:
+            tag_list = set()
+            # split on newline and ,
+            if isinstance(value, list) or isinstance(value, string_types):
+                if not isinstance(value, list):
+                    value = value.split('\n')
+                for tag_str in value:
+                    for tag in str(tag_str).split(','):
+                        tag = tag.strip()
+                        if tag:
+                            tag_list.add(tag)
+            value = list(tag_list)
+        else:
+            if not value:
+                value = None
+        # We skip requiring legal values -- this is similar to optional, but allows only subset of datasets to be positive
+        # TODO: May not actually be required for (nested) collection input ?
+        return super(SelectTagParameter, self).from_json(value, trans, other_values, require_legal_value=False)
+
+    def get_tag_list(self, other_values):
+        """
+        Generate a select list containing the columns of the associated
+        dataset (if found).
+        """
+        # Get the value of the associated data reference (a dataset)
+        history_items = other_values.get(self.data_ref, None)
+        # Check if a dataset is selected
+        if not history_items:
+            return []
+        tags = set()
+        for history_item in util.listify(history_items):
+            if hasattr(history_item, 'dataset_instances'):
+                for dataset in history_item.dataset_instances:
+                    for tag in dataset.tags:
+                        if tag.user_tname == 'group':
+                            tags.add(tag.user_value)
+            else:
+                for tag in history_item.tags:
+                    if tag.user_tname == 'group':
+                        tags.add(tag.user_value)
+        return list(tags)
+
+    def get_options(self, trans, other_values):
+        """
+        Show tags
+        """
+        options = []
+        for tag in self.get_tag_list(other_values):
+            options.append(('Tags: ' + tag, tag, False))
+        return options
+
+    def get_initial_value(self, trans, other_values):
+        if self.default_value is not None:
+            return self.default_value
+        return SelectToolParameter.get_initial_value(self, trans, other_values)
+
+    def get_legal_values(self, trans, other_values):
+        if self.data_ref not in other_values:
+            raise ValueError("Value for associated data reference not found (data_ref).")
+        return set(self.get_tag_list(other_values))
+
+    def get_dependencies(self):
+        return [self.data_ref]
+
+    def to_dict(self, trans, other_values={}):
+        d = super(SelectTagParameter, self).to_dict(trans, other_values=other_values)
+        d['data_ref'] = self.data_ref
+        return d
+
+
 class ColumnListParameter(SelectToolParameter):
     """
     Select list that consists of either the total number of columns or only
@@ -2103,6 +2201,7 @@ def to_dict(self, trans, other_values=None):
     genomebuild=GenomeBuildParameter,
     select=SelectToolParameter,
     color=ColorToolParameter,
+    select_tag=SelectTagParameter,
     data_column=ColumnListParameter,
     hidden=HiddenToolParameter,
     hidden_data=HiddenDataToolParameter,

lib/galaxy/tools/wrappers.py

@@ -233,6 +233,7 @@ def __init__(self, dataset, datatypes_registry=None, tool=None, name=None, datas
             self.unsanitized = dataset
             self.dataset = wrap_with_safe_string(dataset, no_wrap_classes=ToolParameterValueWrapper)
             self.metadata = self.MetadataWrapper(dataset.metadata)
+            self.groups = {tag.user_value.lower() for tag in dataset.tags if tag.user_tname == 'group'}
         self.datatypes_registry = datatypes_registry
         self.false_path = getattr(dataset_path, "false_path", None)
         self.false_extra_files_path = getattr(dataset_path, "false_extra_files_path", None)
@@ -356,6 +357,9 @@ class DatasetCollectionWrapper(ToolParameterValueWrapper, HasDatasets):
     def __init__(self, job_working_directory, has_collection, dataset_paths=[], **kwargs):
         super(DatasetCollectionWrapper, self).__init__()
         self.job_working_directory = job_working_directory
+        self._dataset_elements_cache = {}
+        self.dataset_paths = dataset_paths
+        self.kwargs = kwargs
 
         if has_collection is None:
             self.__input_supplied = False
@@ -374,6 +378,7 @@ def __init__(self, job_working_directory, has_collection, dataset_paths=[], **kw
         else:
             collection = has_collection
             self.name = None
+        self.collection = collection
 
         elements = collection.elements
         element_instances = odict.odict()
@@ -394,6 +399,12 @@ def __init__(self, job_working_directory, has_collection, dataset_paths=[], **kw
         self.__element_instances = element_instances
         self.__element_instance_list = element_instance_list
 
+    def get_datasets_for_group(self, group):
+        group = group.lower()
+        if not self._dataset_elements_cache.get(group):
+            self._dataset_elements_cache[group] = [self._dataset_wrapper(e.element_object, self.dataset_paths, identifier=e.element_identifier, **self.kwargs) for e in self.collection.dataset_elements if any((t for t in e.dataset_instance.tags if t.user_tname.lower() == 'group' and t.value.lower() == group))]
+        return self._dataset_elements_cache[group]
+
     def keys(self):
         if not self.__input_supplied:
             return []

test/unit/test_objectstore.py

@@ -224,6 +224,7 @@ class MockDataset(object):
     def __init__(self, id):
         self.id = id
         self.object_store_id = None
+        self.tags = []
 
 
 # Poor man's mocking. Need to get a real mocking library as real Galaxy development

test/unit/tools/test_wrappers.py

@@ -168,6 +168,7 @@ def __init__(self):
         self.file_name = MOCK_DATASET_PATH
         self.extra_files_path = MOCK_DATASET_EXTRA_FILES_PATH
         self.ext = MOCK_DATASET_EXT
+        self.tags = []
 
 
 class MockTool(object):