Handle 'None' string in fence config

fence/config-default.yaml

@@ -558,7 +558,7 @@ dbGaP:
     #
     # If `parse_consent_code` is true, then a user will be given access to the exact
     # same consent codes in the child studies
-    parent_to_child_studies_mapping:
+    parent_to_child_studies_mapping: {}
       # 'phs001194': ['phs000571', 'phs001843']
     # A consent of "c999" can indicate access to that study's "exchange area data"
     # and when a user has access to one study's exchange area data, they

fence/config.py

@@ -6,6 +6,8 @@
 from gen3config import Config
 
 from cdislogging import get_logger
+from collections import Counter
+from typing import List, Dict, Any
 
 logger = get_logger(__name__)
 
@@ -150,28 +152,76 @@ def post_process(self):
                 logger.warning(
                     f"IdP '{idp_id}' is using multifactor_auth_claim_info '{mfa_info['claim']}', which is neither AMR or ACR. Unable to determine if a user used MFA. Fence will continue and assume they have not used MFA."
                 )
+        self._validate_dbgap_config(self._configs["dbGaP"])
 
-        self._validate_parent_child_studies(self._configs["dbGaP"])
+    @staticmethod
+    def find_duplicates(collection):
+        """
+
+        Args:
+            collection: any data type accepted by the Counter object
+
+        Returns:
+            a set of items that were found more than once in the collection
+
+        note: if collection is a set, this function will never find duplicates
+        (per the definition of a set)
+        """
+        item_with_occurrences = Counter(collection)
+        duplicates = {item
+                      for item, occurrences in item_with_occurrences.items()
+                      if occurrences > 1}
+        return duplicates
+
+    @staticmethod
+    def get_parent_studies_safely(dbgap_config):
+        """
+        Args:
+            dbgap_config: { parent_to_child_studies_mapping: { k1: v1, ... } }
+        Returns:
+            [k1, k2, ...]
+        """
+        study_mapping = dbgap_config.get('parent_to_child_studies_mapping', {})
+        # study mapping could be 'None'
+        safe_studies = list(study_mapping.keys()) if isinstance(study_mapping, dict) else []
+        return safe_studies
 
     @staticmethod
-    def _validate_parent_child_studies(dbgap_configs):
+    def validate_parent_child_studies(dbgap_configs: List[Dict[str, Any]]) -> None:
+        """
+        Given a list of dictionaries that contain a parent -> child study mapping,
+        this function will check that all parents are unique and not duplicated
+        Args:
+            dbgap_configs: [ { parent_to_child_studies_mapping: { k1: v1, ... } }, ... ]
+
+        Returns:
+            Nothing. This function will throw if duplicates are found.
+        """
+        safe_list_of_parent_studies = []
+        for dbgap_config in dbgap_configs:
+            safe_parent_studies = FenceConfig.get_parent_studies_safely(dbgap_config)
+            safe_list_of_parent_studies.extend(safe_parent_studies)
+
+        duplicates = FenceConfig.find_duplicates(safe_list_of_parent_studies)
+        if len(duplicates) > 0:
+            raise Exception(
+                f"{duplicates} are duplicate parent study ids found in parent_to_child_studies_mapping for "
+                f"multiple dbGaP configurations.")
+
+    @staticmethod
+    def enforce_is_array(dbgap_configs):
         if isinstance(dbgap_configs, list):
-            configs = dbgap_configs
+            corrected_dbgap_configs = dbgap_configs
+        elif isinstance(dbgap_configs, dict):
+            corrected_dbgap_configs = [dbgap_configs]
         else:
-            configs = [dbgap_configs]
-
-        all_parent_studies = set()
-        for dbgap_config in configs:
-            parent_studies = dbgap_config.get(
-                "parent_to_child_studies_mapping", {}
-            ).keys()
-            conflicts = parent_studies & all_parent_studies
-            if len(conflicts) > 0:
-                raise Exception(
-                    f"{conflicts} are duplicate parent study ids found in parent_to_child_studies_mapping for "
-                    f"multiple dbGaP configurations."
-                )
-            all_parent_studies.update(parent_studies)
+            raise Exception("The dbgap configuration is not a recognized data structure, aborting!")
+        return corrected_dbgap_configs
+
+    @staticmethod
+    def _validate_dbgap_config(dbgap_configs):
+        corrected_dbgap_configs = FenceConfig.enforce_is_array(dbgap_configs)
+        FenceConfig.validate_parent_child_studies(corrected_dbgap_configs)
 
 
 config = FenceConfig(DEFAULT_CFG_PATH)

tests/test_app_config.py

@@ -120,6 +120,20 @@ def test_app_config():
         patcher.stop()
 
 
+def test_validate_parent_child_studies_against_none_config():
+    """
+    our dbgap config can have nothing for the parent_to_child_studies_mapping property
+    this test ensures the validator handles that case
+    """
+    none_string_config = [{"parent_to_child_studies_mapping": 'None'},
+                          {"parent_to_child_studies_mapping": 'None'},]
+
+    try:
+        FenceConfig.validate_parent_child_studies(none_string_config)
+    except Exception:
+        pytest.fail("Study validation failed when given 'None' mapping!")
+
+
 def test_app_config_parent_child_study_mapping(monkeypatch):
     invalid_dbgap_configs = [
         {
@@ -136,4 +150,23 @@ def test_app_config_parent_child_study_mapping(monkeypatch):
         },
     ]
     with pytest.raises(Exception):
-        FenceConfig._validate_parent_child_studies(invalid_dbgap_configs)
+        FenceConfig.validate_parent_child_studies(invalid_dbgap_configs)
+
+    valid_dbgap_configs = [
+        {
+            "parent_to_child_studies_mapping": {
+                "phs001194": ["phs000571", "phs001843"],
+                "phs001193": ["phs000572", "phs001844"],
+            }
+        },
+        {
+            "parent_to_child_studies_mapping": {
+                "phs001195": ["phs0015623"],
+                "phs001192": ["phs0001", "phs002"],
+            }
+        },
+    ]
+    try:
+        FenceConfig.validate_parent_child_studies(valid_dbgap_configs)
+    except Exception:
+        pytest.fail("Study validation failed when it should have passed!")