mode-based failure policy API

Author: networmix

docs/reference/api-full.md

@@ -2068,13 +2068,7 @@ The main entry point is `apply_failures`, which:
   4) Collect the union of all failed entities across all rules.
   5) Optionally expand failures by shared-risk groups or sub-risks.
 
-Large-scale performance:
 
-- If you set `use_cache=True`, matched sets for each rule are cached,
-
-    so repeated calls to `apply_failures` can skip re-matching if the
-    network hasn't changed. If your network changes between calls,
-    you should clear the cache or re-initialize the policy.
 
 Example YAML configuration:
     ```yaml
@@ -2129,10 +2123,6 @@ Attributes:
     fail_risk_group_children (bool):
         If True, and if a risk_group is marked as failed, expand to
         children risk_groups recursively.
-    use_cache (bool):
-        If True, match results for each rule are cached to speed up
-        repeated calls. If the network changes, the cached results
-        may be stale.
     seed (Optional[int]):
         Seed for reproducible random operations. If None, operations
         will be non-deterministic.
@@ -2142,10 +2132,8 @@ Attributes:
 - `attrs` (Dict[str, Any]) = {}
 - `fail_risk_groups` (bool) = False
 - `fail_risk_group_children` (bool) = False
-- `use_cache` (bool) = False
 - `seed` (Optional[int])
 - `modes` (List[FailureMode]) = []
-- `_match_cache` (Dict[int, Set[str]]) = {}
 
 **Methods:**
 

docs/reference/dsl.md

@@ -384,7 +384,6 @@ failure_policy_set:
 
   random_failures:
     fail_risk_groups: true
-    use_cache: true
     rules:
       - entity_scope: "node"
         rule_type: "random"

ngraph/failure/policy.py

@@ -120,12 +120,6 @@ class FailurePolicy:
       4) Collect the union of all failed entities across all rules.
       5) Optionally expand failures by shared-risk groups or sub-risks.
 
-    Large-scale performance:
-      - If you set `use_cache=True`, matched sets for each rule are cached,
-        so repeated calls to `apply_failures` can skip re-matching if the
-        network hasn't changed. If your network changes between calls,
-        you should clear the cache or re-initialize the policy.
-
     Example YAML configuration:
         ```yaml
         failure_policy:
@@ -173,10 +167,6 @@ class FailurePolicy:
         fail_risk_group_children (bool):
             If True, and if a risk_group is marked as failed, expand to
             children risk_groups recursively.
-        use_cache (bool):
-            If True, match results for each rule are cached to speed up
-            repeated calls. If the network changes, the cached results
-            may be stale.
         seed (Optional[int]):
             Seed for reproducible random operations. If None, operations
             will be non-deterministic.
@@ -186,13 +176,9 @@ class FailurePolicy:
     attrs: Dict[str, Any] = field(default_factory=dict)
     fail_risk_groups: bool = False
     fail_risk_group_children: bool = False
-    use_cache: bool = False
     seed: Optional[int] = None
     modes: List[FailureMode] = field(default_factory=list)
 
-    # Internal cache for matched sets:  (rule_index -> set_of_entities)
-    _match_cache: Dict[int, Set[str]] = field(default_factory=dict, init=False)
-
     def apply_failures(
         self,
         network_nodes: Dict[str, Any],
@@ -286,9 +272,6 @@ def _match_scope(
         """Get the set of IDs matched by the given rule, either from cache
         or by performing a fresh match over the relevant entity type.
         """
-        if self.use_cache and rule_idx in self._match_cache:
-            return self._match_cache[rule_idx]
-
         # Decide which mapping to iterate
         if rule.entity_scope == "node":
             matched = self._match_entities(network_nodes, rule.conditions, rule.logic)
@@ -298,9 +281,6 @@ def _match_scope(
             matched = self._match_entities(
                 network_risk_groups, rule.conditions, rule.logic
             )
-
-        if self.use_cache:
-            self._match_cache[rule_idx] = matched
         return matched
 
     def _match_entities(
@@ -610,7 +590,6 @@ def to_dict(self) -> Dict[str, Any]:
             "attrs": self.attrs,
             "fail_risk_groups": self.fail_risk_groups,
             "fail_risk_group_children": self.fail_risk_group_children,
-            "use_cache": self.use_cache,
             "seed": self.seed,
         }
         if self.modes:

ngraph/scenario.py

@@ -240,14 +240,13 @@ def _build_failure_policy(
     ) -> FailurePolicy:
         """Constructs a FailurePolicy from data that may specify multiple rules plus
         optional top-level fields like fail_risk_groups, fail_risk_group_children,
-        use_cache, and attrs.
+        and attrs.
 
         Example:
             failure_policy_set:
               default:
                 fail_risk_groups: true
                 fail_risk_group_children: false
-                use_cache: true
                 attrs:
                   custom_key: custom_val
                 rules:
@@ -273,7 +272,6 @@ def _build_failure_policy(
         """
         fail_srg = fp_data.get("fail_risk_groups", False)
         fail_rg_children = fp_data.get("fail_risk_group_children", False)
-        use_cache = fp_data.get("use_cache", False)
         attrs = normalize_yaml_dict_keys(fp_data.get("attrs", {}))
 
         def build_rules(rule_dicts: List[Dict[str, Any]]) -> List[FailureRule]:
@@ -333,7 +331,6 @@ def build_rules(rule_dicts: List[Dict[str, Any]]) -> List[FailureRule]:
             attrs=attrs,
             fail_risk_groups=fail_srg,
             fail_risk_group_children=fail_rg_children,
-            use_cache=use_cache,
             seed=policy_seed,
             modes=modes,
         )

scenarios/nsfnet.yaml

@@ -162,7 +162,6 @@ failure_policy_set:
         ENSS router has ~99.95 % availability (p=0.0005 failure).
     fail_risk_groups: false
     fail_risk_group_children: false
-    use_cache: false
     modes:
         - weight: 1.0
           rules:

schemas/scenario.json

@@ -346,10 +346,7 @@
               "type": "boolean",
               "description": "Whether to recursively fail risk group children"
             },
-            "use_cache": {
-              "type": "boolean",
-              "description": "Whether to use caching for failure calculations"
-            },
+
                 "modes": {
                   "type": "array",
                   "description": "Weighted mode list; exactly one mode is chosen per iteration.",

tests/dsl/test_examples.py

@@ -255,7 +255,6 @@ def test_failure_policy_example():
   default:
     fail_risk_groups: true
     fail_risk_group_children: false
-    use_cache: true
     attrs:
       custom_key: "value"
     modes:

tests/failure/test_policy_set.py

@@ -60,7 +60,6 @@ def test_to_dict_serialization(self):
             modes=[FailureMode(weight=1.0, rules=[rule])],
             attrs={"name": "test_policy", "description": "Test policy"},
             fail_risk_groups=True,
-            use_cache=False,
         )
 
         fps.add("test", policy)
@@ -71,7 +70,6 @@ def test_to_dict_serialization(self):
         assert "modes" in result["test"]
         assert "attrs" in result["test"]
         assert result["test"]["fail_risk_groups"] is True
-        assert result["test"]["use_cache"] is False
         # Modes present
         assert "modes" in result["test"] and len(result["test"]["modes"]) == 1
 

tests/scenario/test_scenario.py

@@ -98,7 +98,6 @@ def valid_scenario_yaml() -> str:
       description: "Testing modal policy."
     fail_risk_groups: false
     fail_risk_group_children: false
-    use_cache: false
     modes:
       - weight: 1.0
         rules:
@@ -291,7 +290,7 @@ def test_scenario_from_yaml_valid(valid_scenario_yaml: str) -> None:
     assert isinstance(simple_policy, FailurePolicy)
     assert not simple_policy.fail_risk_groups
     assert not simple_policy.fail_risk_group_children
-    assert not simple_policy.use_cache
+    # use_cache was removed
 
     assert len(simple_policy.modes) == 1
     assert simple_policy.attrs.get("name") == "multi_rule_example"