Github user cestella commented on a diff in the pull request:
https://github.com/apache/incubator-metron/pull/438#discussion_r100542070
--- Diff:
metron-platform/metron-enrichment/src/main/java/org/apache/metron/threatintel/triage/ThreatTriageProcessor.java
---
@@ -52,15 +74,36 @@ public ThreatTriageProcessor( SensorEnrichmentConfig
config
@Nullable
@Override
- public Double apply(@Nullable Map input) {
- List<Number> scores = new ArrayList<>();
+ public ThreatScore apply(@Nullable Map input) {
+
+ ThreatScore threatScore = new ThreatScore();
StellarPredicateProcessor predicateProcessor = new
StellarPredicateProcessor();
+ StellarProcessor processor = new StellarProcessor();
VariableResolver resolver = new MapVariableResolver(input,
sensorConfig.getConfiguration(), threatIntelConfig.getConfig());
+
+ // attempt to apply each rule to the threat
for(RiskLevelRule rule : threatTriageConfig.getRiskLevelRules()) {
if(predicateProcessor.parse(rule.getRule(), resolver,
functionResolver, context)) {
- scores.add(rule.getScore());
+
+ // add the rule's score to the overall threat score
+ String reason = execute(rule.getReason(), processor, resolver,
String.class);
+ RuleScore score = new RuleScore(rule, reason);
+ threatScore.addRuleScore(score);
}
}
- return threatTriageConfig.getAggregator().aggregate(scores,
threatTriageConfig.getAggregationConfig());
+
+ // calculate the aggregate threat score
+ Aggregators aggregators = threatTriageConfig.getAggregator();
+ List<Number> allScores =
threatScore.getRuleScores().stream().map(score ->
score.getRule().getScore()).collect(Collectors.toList());
+ Double aggregateScore = aggregators.aggregate(allScores,
threatTriageConfig.getAggregationConfig());
+
+ // return the overall threat score
+ threatScore.setScore(aggregateScore);
+ return threatScore;
+ }
+
+ private <T> T execute(String expression, StellarProcessor processor,
VariableResolver resolver, Class<T> clazz) {
+ Object result = processor.parse(expression, resolver,
functionResolver, context);
--- End diff --
Ok, so, yes, I have an opinion about that. Sadly, with the current
implementation of Stellar, we cannot cache the parse tree and then apply it
after the fact. It's just an artifact of how we do the parsing: we actually
execute the statement as we parse rather than constructing an AST that can then
be evaluated later given a message.
I think this would be a GREAT modification to Stellar and one that I had on
my mental backlog for some time. In fact, I took the opportunity as part of
the profiler window DSL to do it the right way (parse to an object that can
then be cached and evaluated on multiple sets of inputs without reparse).
tl;dr: We can't do it now easily, BUT we should for multiple reasons:
* code clarity - decoupling the stellar language from the generated parser
code
* performance - saving lexing and parsing
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---