moderation: add rule results to dict and eval

invenio.cfg

@@ -76,7 +76,7 @@ from zenodo_rdm.github.schemas import CitationMetadataSchema
 from zenodo_rdm.legacy.resources import record_serializers
 from zenodo_rdm.metrics.config import METRICS_CACHE_UPDATE_INTERVAL
 from zenodo_rdm.moderation.errors import UserBlockedException
-from zenodo_rdm.moderation.handlers import CommunityScoreHandler, RecordScoreHandler
+from zenodo_rdm.moderation.handlers import CommunityModerationHandler, RecordModerationHandler
 from zenodo_rdm.openaire.records.components import OpenAIREComponent
 from zenodo_rdm.permissions import (
     ZenodoCommunityPermissionPolicy,
@@ -817,11 +817,11 @@ RDM_RECORDS_SERVICE_COMPONENTS = DefaultRecordsComponents + [
 """Addd OpenAIRE component to records service."""
 
 RDM_CONTENT_MODERATION_HANDLERS = [
-    RecordScoreHandler(),
+    RecordModerationHandler(),
 ]
 """Records content moderation handlers."""
 RDM_COMMUNITY_CONTENT_MODERATION_HANDLERS = [
-    CommunityScoreHandler(),
+    CommunityModerationHandler(),
 ]
 """Community content moderation handlers."""
 
@@ -1062,3 +1062,5 @@ COMMUNITIES_SHOW_BROWSE_MENU_ENTRY = True
 
 JOBS_ADMINISTRATION_ENABLED = True
 """Enable Jobs administration view."""
+
+SPAM_DETECTOR_MODEL="spam-scikit:1.0.0"

site/zenodo_rdm/moderation/config.py

@@ -41,21 +41,21 @@
 MODERATION_SPAM_FILE_EXTS = {"jpg", "jpeg", "pdf", "png", "jfif", "docx", "webp"}
 """Frequest spam file extensions."""
 
-MODERATION_RECORD_SCORE_RULES = [
-    verified_user_rule,
-    links_rule,
-    files_rule,
-    text_sanitization_rule,
-    match_query_rule,
-]
+MODERATION_RECORD_SCORE_RULES = {
+    "verified-user-rule": verified_user_rule,
+    "links-rule": links_rule,
+    "files-rule": files_rule,
+    "text-sanitization-rule": text_sanitization_rule,
+    "match-query-rule": match_query_rule,
+}
 """Scoring rules for record moderation."""
 
-MODERATION_COMMUNITY_SCORE_RULES = [
-    links_rule,
-    text_sanitization_rule,
-    verified_user_rule,
-    match_query_rule,
-]
+MODERATION_COMMUNITY_SCORE_RULES = {
+    "links-rule": links_rule,
+    "text-sanitization-rule": text_sanitization_rule,
+    "verified-user-rule": verified_user_rule,
+    "match-query-rule": match_query_rule,
+}
 """Scoring rules for communtiy moderation."""
 
 MODERATION_PERCOLATOR_INDEX_PREFIX = "moderation-queries"

site/zenodo_rdm/moderation/handlers.py

@@ -61,7 +61,7 @@ def on_post_rollback(self, uow):
         self.commit_op.on_post_commit(uow)
 
 
-class BaseScoreHandler:
+class BaseModerationHandler:
     """Base handler to calculate moderation scores based on rules."""
 
     def __init__(self, rules=None):
@@ -73,14 +73,17 @@ def rules(self):
         """Get scoring rules."""
         if isinstance(self._rules, str):
             return current_app.config[self._rules]
-        return self._rules or []
+        return self._rules or {}
+
+    def evaluate_result(self, params):
+        return sum(params.values())
 
     def run(self, identity, draft=None, record=None, uow=None):
         """Calculate the moderation score for a given record or draft."""
         try:
-            score = 0
-            for rule in self.rules:
-                score += rule(identity, draft=draft, record=record)
+            results = {}
+            for name, rule in self.rules.items():
+                results[name] = rule(identity, draft=draft, record=record)
 
             # TODO: Move to caller handler method (i.e. publish, create, update, etc.)
             user_id = None
@@ -100,10 +103,11 @@ def run(self, identity, draft=None, record=None, uow=None):
             action_ctx = {
                 "user_id": user_id,
                 "record_pid": record.pid.pid_value,
-                "score": score,
+                "results": results,
             }
             apply_actions = current_app.config.get("MODERATION_APPLY_ACTIONS", False)
-            if score > current_scores.spam_threshold:
+
+            if self.evaluate_result(results) > current_scores.spam_threshold:
                 action_ctx["action"] = "block"
                 if apply_actions:
                     uow.register(
@@ -132,7 +136,7 @@ def run(self, identity, draft=None, record=None, uow=None):
                         "Block moderation action triggered",
                         extra=action_ctx,
                     )
-            elif score < current_scores.ham_threshold:
+            elif self.evaluate_result(results) < current_scores.ham_threshold:
                 # If the user is already verified, we don't need to verify again
                 if user.verified:
                     return
@@ -176,7 +180,7 @@ def run(self, identity, draft=None, record=None, uow=None):
             current_app.logger.exception("Error calculating moderation score")
 
 
-class RecordScoreHandler(BaseHandler, BaseScoreHandler):
+class RecordModerationHandler(BaseHandler, BaseModerationHandler):
     """Handler for calculating scores for records."""
 
     def __init__(self):
@@ -188,7 +192,9 @@ def publish(self, identity, draft=None, record=None, uow=None, **kwargs):
         self.run(identity, record=record, uow=uow)
 
 
-class CommunityScoreHandler(community_moderation.BaseHandler, BaseScoreHandler):
+class CommunityModerationHandler(
+    community_moderation.BaseHandler, BaseModerationHandler
+):
     """Handler for calculating scores for communities."""
 
     def __init__(self):