Implement reporting backend MongoDBReports.summaries

Author: josh-feather

lib/cuckoo/core/reporting/api.py

@@ -53,8 +53,7 @@ def iocs(self, task_id: int) -> dict:
     def summary(self, task_id: int) -> schema.Summary | None:
         raise NotImplementedError()
 
-    # TODO(jf) return list/generator of all schema.Summary's known to reporting backends
-    def summaries(self) -> Generator[schema.Summary]:
+    def summaries(self) -> Generator[schema.Summary, None, None]:
         raise NotImplementedError()
 
     def recent_suricata_alerts(self, minutes=60) -> list:

lib/cuckoo/core/reporting/backends/elasticsearch.py

@@ -1,4 +1,4 @@
-from typing import Optional
+from typing import Generator, Optional
 
 from lib.cuckoo.common import config
 from lib.cuckoo.core.reporting import api, schema
@@ -92,6 +92,9 @@ def summary(self, task_id: int) -> Optional[schema.Summary]:
         else:
             return None
 
+    def summaries(self) -> Generator[schema.Summary, None, None]:
+        pass
+
     def recent_suricata_alerts(self, minutes=60) -> list:
         pass
 

lib/cuckoo/core/reporting/backends/mongodb.py

@@ -1,7 +1,7 @@
 import itertools
 import logging
 from datetime import datetime, timedelta
-from typing import Any, Mapping, Optional, cast, TypeAlias
+from typing import Any, Generator, Mapping, Optional, cast, TypeAlias
 from collections.abc import Iterable
 
 import pymongo
@@ -150,6 +150,17 @@ def summary(self, task_id: int) -> Optional[schema.Summary]:
             return schema.Summary(**report)
         return None
 
+    def summaries(self) -> Generator[schema.Summary, None, None]:
+        tasks = self._analysis_collection.find({}, {_id: 0, _info_id: 1})
+        for task in tasks:
+            task_id = task.get("info", {}).get("id")
+            if task_id is None:
+                continue
+            if summary := self.summary(task_id):
+                yield summary
+            else:
+                continue
+
     def recent_suricata_alerts(self, minutes=60) -> list:
         gen_time = datetime.now() - timedelta(minutes=minutes)
         dummy_id = ObjectId.from_datetime(gen_time)

tests/lib/cuckoo/core/reporting/backends/test_mongodb.py

@@ -111,6 +111,17 @@ def test_summary(self):
         assert result.vt_url_summary == "20/30"
         assert result.clamav == []
 
+    @pytest.mark.usefixtures("mongodb_populate_test_data")
+    def test_summaries(self):
+        """Retrieve Summaries from MongoDB."""
+        mongo = mongodb.MongoDBReports(self.cfg)
+        results = mongo.summaries()
+        result_count = 0
+        for result in results:
+            result_count += 1
+            assert isinstance(result, schema.Summary)
+        assert result_count == len(TEST_TASK_IDS)
+
     @pytest.mark.usefixtures("mongodb_populate_test_data")
     def test_cape_configs(self):
         """Retrieve analysis configs from MongoDB."""