Github user nickwallen commented on a diff in the pull request:

    https://github.com/apache/incubator-metron/pull/438#discussion_r100442785
  
    --- 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 --
    
    It is a cool idea @justinleet.  
    
    If we could build the abstract syntax tree and then cache that, we could 
certainly improve performance.  Of course, we'd have to confirm how much time 
we spend building the AST versus execution, but I'd be willing to bet you are 
right.
    
    I don't know that we have the ability to do this today though.  At least, I 
have not seen us do something like this elsewhere.  But I could be wrong.  
    
    Might be a good general strategy to explore for improving performance in a 
lot of places considering how much we rely on Stellar executions nowadays.


---
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.
---

Reply via email to