wrote all default weights, actually computing the equivalence score

logging for unsupported objects, finished implementing some methods. Missing to implement patterning.

stix2/environment.py

@@ -1,10 +1,14 @@
 """Python STIX2 Environment API."""
 
 import copy
+import logging
+import math
 
 from .core import parse as _parse
 from .datastore import CompositeDataSource, DataStoreMixin
 
+logger = logging.getLogger(__name__)
+
 
 class ObjectFactory(object):
     """Easily create STIX objects with default values for certain properties.
@@ -187,13 +191,16 @@ class Environment(DataStoreMixin):
         else:
             return None
 
-    def semantically_equivalent(self, obj1, obj2):
+    @staticmethod
+    def semantically_equivalent(obj1, obj2, **weight_dict):
         """This method is meant to verify if two objects of the same type are
         semantically equivalent.
 
         Args:
             obj1: A stix2 object instance
             obj2: A stix2 object instance
+            weight_dict: A dictionary that can be used to override settings
+                in the semantic equivalence process
 
         Returns:
             float: A number between 0.0 and 1.0 as a measurement of equivalence.
@@ -206,7 +213,58 @@ class Environment(DataStoreMixin):
             see `the Committee Note <link here>`__.
 
         """
-        equivalence_score = 0.0
+        # default weights used for the semantic equivalence process
+        weigths = {
+            "attack-pattern": {
+                "name": 30,
+                "external_references": 70,
+            },
+            "campaign": {
+                "name": 60,
+                "aliases": 40,
+            },
+            "identity": {
+                "name": 60,
+                "identity_class": 20,
+                "sectors": 20,
+            },
+            "indicator": {
+                "indicator_types": 15,
+                "pattern": 80,
+                "valid_from": 5,
+            },
+            "location": {
+                "longitude_latitude": 34,
+                "region": 33,
+                "country": 33,
+            },
+            "malware": {
+                "malware_types": 20,
+                "name": 80,
+            },
+            "threat-actor": {
+                "name": 60,
+                "threat_actor_types": 20,
+                "aliases": 20,
+            },
+            "tool": {
+                "tool_types": 20,
+                "name": 80,
+            },
+            "vulnerability": {
+                "name": 30,
+                "external_references": 70,
+            },
+            "_internal": {
+                "tdelta": 1,
+            },
+        }
+
+        if weight_dict:
+            weigths.update(weight_dict)
+
+        matching_score = 0.0
+        sum_weights = 0.0
         type1, type2 = obj1["type"], obj2["type"]
 
         if type1 != type2:
@@ -217,61 +275,132 @@ class Environment(DataStoreMixin):
 
         if type1 == "attack-pattern":
             if _check_property_present("name", obj1, obj2):
-                _partial_string_based(obj1["name"], obj2["name"])
+                w = weigths["attack-pattern"]["name"]
+                sum_weights += w
+                matching_score += w * _partial_string_based(obj1["name"], obj2["name"])
             if _check_property_present("external_references", obj1, obj2):
+                w = weigths["attack-pattern"]["external_references"]
+                sum_weights += w
+                matching_score += (
+                    w *
                     _partial_external_reference_based(obj1["external_references"], obj2["external_references"])
+                )
+
         elif type1 == "campaign":
             if _check_property_present("name", obj1, obj2):
-                _partial_string_based(obj1["name"], obj2["name"])
+                w = weigths["campaign"]["name"]
+                sum_weights += w
+                matching_score += w * _partial_string_based(obj1["name"], obj2["name"])
             if _check_property_present("aliases", obj1, obj2):
-                _partial_list_based(obj1["aliases"], obj2["aliases"])
+                w = weigths["campaign"]["aliases"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["aliases"], obj2["aliases"])
+
         elif type1 == "course-of-action":
-            pass
+            logger.warning("%s type is not supported for semantic equivalence", type1)
+
         elif type1 == "identity":
             if _check_property_present("name", obj1, obj2):
-                _exact_match(obj1["name"], obj2["name"])
+                w = weigths["identity"]["name"]
+                sum_weights += w
+                matching_score += w * _exact_match(obj1["name"], obj2["name"])
             if _check_property_present("identity_class", obj1, obj2):
-                _exact_match(obj1["identity_class"], obj2["identity_class"])
+                w = weigths["identity"]["identity_class"]
+                sum_weights += w
+                matching_score += w * _exact_match(obj1["identity_class"], obj2["identity_class"])
             if _check_property_present("sectors", obj1, obj2):
-                _partial_list_based(obj1["sectors"], obj2["sectors"])
+                w = weigths["identity"]["sectors"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["sectors"], obj2["sectors"])
+
         elif type1 == "indicator":
             if _check_property_present("indicator_types", obj1, obj2):
-                _partial_list_based(obj1["indicator_types"], obj2["indicator_types"])
+                w = weigths["indicator"]["indicator_types"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["indicator_types"], obj2["indicator_types"])
             if _check_property_present("pattern", obj1, obj2):
-                pass  # TODO: needs to be done
+                w = weigths["indicator"]["pattern"]
+                sum_weights += w
+                matching_score += w * _custom_pattern_based(obj1["pattern"], obj2["pattern"])
             if _check_property_present("valid_from", obj1, obj2):
-                _partial_timestamp_based(obj1["valid_from"], obj2["valid_from"])
+                w = weigths["indicator"]["valid_from"]
+                sum_weights += w
+                matching_score += (
+                    w *
+                    _partial_timestamp_based(obj1["valid_from"], obj2["valid_from"], weigths["_internal"]["tdelta"])
+                )
+
         elif type1 == "instrusion-set":
-            pass
+            logger.warning("%s type is not supported for semantic equivalence", type1)
+
         elif type1 == "location":
-            pass
+            if _check_property_present("latitude", obj1, obj2) and _check_property_present("longitude", obj1, obj2):
+                w = weigths["location"]["longitude_latitude"]
+                sum_weights += w
+                matching_score += (
+                    w *
+                    _partial_location_distance(obj1["latitude"], obj1["longitude"], obj2["latitude"], obj2["longitude"])
+                )
+            if _check_property_present("region", obj1, obj2):
+                w = weigths["location"]["region"]
+                sum_weights += w
+                matching_score += w * _exact_match(obj1["region"], obj2["region"])
+            if _check_property_present("country", obj1, obj2):
+                w = weigths["location"]["country"]
+                sum_weights += w
+                matching_score += w * _exact_match(obj1["country"], obj2["country"])
+
         elif type1 == "malware":
             if _check_property_present("malware_types", obj1, obj2):
-                _partial_list_based(obj1["malware_types"], obj2["malware_types"])
+                w = weigths["malware"]["malware_types"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["malware_types"], obj2["malware_types"])
             if _check_property_present("name", obj1, obj2):
-                _partial_string_based(obj1["name"], obj2["name"])
+                w = weigths["malware"]["name"]
+                sum_weights += w
+                matching_score += w * _partial_string_based(obj1["name"], obj2["name"])
+
         elif type1 == "observed-data":
-            pass
+            logger.warning("%s type is not supported for semantic equivalence", type1)
+
         elif type1 == "report":
-            pass
+            logger.warning("%s type is not supported for semantic equivalence", type1)
+
         elif type1 == "threat-actor":
             if _check_property_present("name", obj1, obj2):
-                _partial_string_based(obj1["name"], obj2["name"])
+                w = weigths["threat-actor"]["name"]
+                sum_weights += w
+                matching_score += w * _partial_string_based(obj1["name"], obj2["name"])
             if _check_property_present("threat_actor_types", obj1, obj2):
-                _partial_list_based(obj1["threat_actor_types"], obj2["threat_actor_types"])
+                w = weigths["threat-actor"]["threat_actor_types"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["threat_actor_types"], obj2["threat_actor_types"])
             if _check_property_present("aliases", obj1, obj2):
-                _partial_list_based(obj1["aliases"], obj2["aliases"])
+                w = weigths["threat-actor"]["aliases"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["aliases"], obj2["aliases"])
+
         elif type1 == "tool":
             if _check_property_present("tool_types", obj1, obj2):
-                _partial_list_based(obj1["tool_types"], obj2["tool_types"])
+                w = weigths["tool"]["tool_types"]
+                sum_weights += w
+                matching_score += w * _partial_list_based(obj1["tool_types"], obj2["tool_types"])
             if _check_property_present("name", obj1, obj2):
-                _partial_string_based(obj1["name"], obj2["name"])
+                w = weigths["tool"]["name"]
+                sum_weights += w
+                matching_score += w * _partial_string_based(obj1["name"], obj2["name"])
+
         elif type1 == "vulnerability":
             if _check_property_present("name", obj1, obj2):
-                _partial_string_based(obj1["name"], obj2["name"])
+                w = weigths["vulnerability"]["name"]
+                sum_weights += w
+                matching_score += w * _partial_string_based(obj1["name"], obj2["name"])
             if _check_property_present("external_references", obj1, obj2):
-                _partial_external_reference_based(obj1["external_references"], obj2["external_references"])
+                w = weigths["vulnerability"]["external_references"]
-        # TODO: need to actually calculate the value
+                sum_weights += w
+                matching_score += w * _partial_external_reference_based(obj1["external_references"], obj2["external_references"])
+
+        equivalence_score = (matching_score / sum_weights) * 100.0
         return equivalence_score
 
 
@@ -281,16 +410,15 @@ def _check_property_present(prop, obj1, obj2):
     return False
 
 
-def _partial_timestamp_based(t1, t2):
+def _partial_timestamp_based(t1, t2, tdelta):
     from .utils import parse_into_datetime
-    tdelta = 1  # One day...
     stix_t1, stix_t2 = parse_into_datetime(t1), parse_into_datetime(t2)
     return 1 - min(abs(stix_t1.timestamp() - stix_t2.timestamp()) / (86400 * tdelta), 1)
 
 
 def _partial_list_based(l1, l2):
     l1_set, l2_set = set(l1), set(l2)
-    return len(l1_set.intersection(l2_set)) / max(len(l1_set), len(l2_set))
+    return len(l1_set.intersection(l2_set)) / max(len(l1), len(l2))
 
 
 def _exact_match(val1, val2):
@@ -304,9 +432,53 @@ def _partial_string_based(str1, str2):
     return distance.get_jaro_distance(str1, str2)
 
 
+def _custom_pattern_based(pattern1, pattern2):
+    return 0  # TODO: Needs to be implemented
+
+
 def _partial_external_reference_based(refs1, refs2):
-    pass  # TODO: needs to be done
+    allowed = set(("veris", "cve", "capec", "mitre-attack"))
+    matches = 0
+
+    if len(refs1) >= len(refs2):
+        l1 = refs1
+        l2 = refs2
+    else:
+        l1 = refs2
+        l2 = refs1
+
+    for ext_ref1 in l1:
+        for ext_ref2 in l2:
+            sn_match = False
+            ei_match = False
+            url_match = False
+            source_name = None
+
+            if _check_property_present("source_name", ext_ref1, ext_ref2):
+                if ext_ref1["source_name"] == ext_ref2["source_name"]:
+                    source_name = ext_ref1["source_name"]
+                    sn_match = True
+            if _check_property_present("external_id", ext_ref1, ext_ref2):
+                if ext_ref1["external_id"] == ext_ref2["external_id"]:
+                    ei_match = True
+            if _check_property_present("url", ext_ref1, ext_ref2):
+                if ext_ref1["url"] == ext_ref2["url"]:
+                    url_match = True
+
+            # Special case: if source_name is a STIX defined name and either
+            # external_id or url match then its a perfect match and other entries
+            # can be ignored.
+            if sn_match and (ei_match or url_match) and source_name in allowed:
+                return 1.0
+
+            # Regular check. If the source_name (not STIX-defined) or external_id or
+            # url matches then we consider the entry a match.
+            if (sn_match or ei_match or url_match) and source_name not in allowed:
+                matches += 1
+
+    return matches / max(len(refs1), len(refs2))
 
 
-def _partial_location_distance(loc1, loc2):
+def _partial_location_distance(lat1, long1, lat2, long2):
-    pass  # TODO: needs to be done
+    distance = math.sqrt(((lat2 - lat1) ** 2) + ((long2 - long1) ** 2))
+    return 1 - (distance / 1000.0)