Merge pull request #28054 from vespa-engine/geirst/target-hits-max-adjustment-factor

Control the auto-adjustment of targetHits in ANN using post-filtering.

6 files changed, 84 insertions, 14 deletions

diff --git a/config-model/src/main/java/com/yahoo/schema/RankProfile.java b/config-model/src/main/java/com/yahoo/schema/RankProfile.java
index 69f32daef4a..35ef12f077a 100644
--- a/config-model/src/main/java/com/yahoo/schema/RankProfile.java
+++ b/config-model/src/main/java/com/yahoo/schema/RankProfile.java
@@ -100,6 +100,7 @@ public class RankProfile implements Cloneable {
     private Double termwiseLimit = null;
     private Double postFilterThreshold = null;
     private Double approximateThreshold = null;
+    private Double targetHitsMaxAdjustmentFactor = null;
 
     /** The drop limit used to drop hits with rank score less than or equal to this value */
     private double rankScoreDropLimit = -Double.MAX_VALUE;
@@ -768,6 +769,7 @@ public class RankProfile implements Cloneable {
     public void setTermwiseLimit(double termwiseLimit) { this.termwiseLimit = termwiseLimit; }
     public void setPostFilterThreshold(double threshold) { this.postFilterThreshold = threshold; }
     public void setApproximateThreshold(double threshold) { this.approximateThreshold = threshold; }
+    public void setTargetHitsMaxAdjustmentFactor(double factor) { this.targetHitsMaxAdjustmentFactor = factor; }
 
     public OptionalDouble getTermwiseLimit() {
         if (termwiseLimit != null) return OptionalDouble.of(termwiseLimit);
@@ -789,6 +791,13 @@ public class RankProfile implements Cloneable {
         return uniquelyInherited(p -> p.getApproximateThreshold(), l -> l.isPresent(), "approximate-threshold").orElse(OptionalDouble.empty());
     }
 
+    public OptionalDouble getTargetHitsMaxAdjustmentFactor() {
+        if (targetHitsMaxAdjustmentFactor != null) {
+            return OptionalDouble.of(targetHitsMaxAdjustmentFactor);
+        }
+        return uniquelyInherited(p -> p.getTargetHitsMaxAdjustmentFactor(), l -> l.isPresent(), "target-hits-max-adjustment-factor").orElse(OptionalDouble.empty());
+    }
+
     /** Whether we should ignore the default rank features. Set to null to use inherited */
     public void setIgnoreDefaultRankFeatures(Boolean ignoreDefaultRankFeatures) {
         this.ignoreDefaultRankFeatures = ignoreDefaultRankFeatures;
diff --git a/config-model/src/main/java/com/yahoo/schema/derived/RawRankProfile.java b/config-model/src/main/java/com/yahoo/schema/derived/RawRankProfile.java
index 82c0c9d516a..29bd454cc62 100644
--- a/config-model/src/main/java/com/yahoo/schema/derived/RawRankProfile.java
+++ b/config-model/src/main/java/com/yahoo/schema/derived/RawRankProfile.java
@@ -153,6 +153,7 @@ public class RawRankProfile implements RankProfilesConfig.Producer {
         private final double termwiseLimit;
         private final OptionalDouble postFilterThreshold;
         private final OptionalDouble approximateThreshold;
+        private final OptionalDouble targetHitsMaxAdjustmentFactor;
         private final double rankScoreDropLimit;
         private final boolean enableNestedMultivalueGrouping;
 
@@ -197,6 +198,7 @@ public class RawRankProfile implements RankProfilesConfig.Producer {
             enableNestedMultivalueGrouping = deployProperties.featureFlags().enableNestedMultivalueGrouping();
             postFilterThreshold = compiled.getPostFilterThreshold();
             approximateThreshold = compiled.getApproximateThreshold();
+            targetHitsMaxAdjustmentFactor = compiled.getTargetHitsMaxAdjustmentFactor();
             keepRankCount = compiled.getKeepRankCount();
             rankScoreDropLimit = compiled.getRankScoreDropLimit();
             ignoreDefaultRankFeatures = compiled.getIgnoreDefaultRankFeatures();
@@ -429,6 +431,9 @@ public class RawRankProfile implements RankProfilesConfig.Producer {
             if (approximateThreshold.isPresent()) {
                 properties.add(new Pair<>("vespa.matching.global_filter.lower_limit", String.valueOf(approximateThreshold.getAsDouble())));
             }
+            if (targetHitsMaxAdjustmentFactor.isPresent()) {
+                properties.add(new Pair<>("vespa.matching.nns.target_hits_max_adjustment_factor", String.valueOf(targetHitsMaxAdjustmentFactor.getAsDouble())));
+            }
             if (matchPhaseSettings != null) {
                 properties.add(new Pair<>("vespa.matchphase.degradation.attribute", matchPhaseSettings.getAttribute()));
                 properties.add(new Pair<>("vespa.matchphase.degradation.ascendingorder", matchPhaseSettings.getAscending() + ""));
diff --git a/config-model/src/main/java/com/yahoo/schema/parser/ConvertParsedRanking.java b/config-model/src/main/java/com/yahoo/schema/parser/ConvertParsedRanking.java
index bdecf6332a0..c25d393c8bf 100644
--- a/config-model/src/main/java/com/yahoo/schema/parser/ConvertParsedRanking.java
+++ b/config-model/src/main/java/com/yahoo/schema/parser/ConvertParsedRanking.java
@@ -65,6 +65,8 @@ public class ConvertParsedRanking {
                 (value -> profile.setPostFilterThreshold(value));
         parsed.getApproximateThreshold().ifPresent
                 (value -> profile.setApproximateThreshold(value));
+        parsed.getTargetHitsMaxAdjustmentFactor().ifPresent
+                (value -> profile.setTargetHitsMaxAdjustmentFactor(value));
         parsed.getKeepRankCount().ifPresent
             (value -> profile.setKeepRankCount(value));
         parsed.getMinHitsPerThread().ifPresent
diff --git a/config-model/src/main/java/com/yahoo/schema/parser/ParsedRankProfile.java b/config-model/src/main/java/com/yahoo/schema/parser/ParsedRankProfile.java
index 2809ee0c633..1d06b993cdc 100644
--- a/config-model/src/main/java/com/yahoo/schema/parser/ParsedRankProfile.java
+++ b/config-model/src/main/java/com/yahoo/schema/parser/ParsedRankProfile.java
@@ -29,6 +29,7 @@ class ParsedRankProfile extends ParsedBlock {
     private Double termwiseLimit = null;
     private Double postFilterThreshold = null;
     private Double approximateThreshold = null;
+    private Double targetHitsMaxAdjustmentFactor = null;
     private final List<FeatureList> matchFeatures = new ArrayList<>();
     private final List<FeatureList> rankFeatures = new ArrayList<>();
     private final List<FeatureList> summaryFeatures = new ArrayList<>();
@@ -65,6 +66,7 @@ class ParsedRankProfile extends ParsedBlock {
     Optional<Double> getTermwiseLimit() { return Optional.ofNullable(this.termwiseLimit); }
     Optional<Double> getPostFilterThreshold() { return Optional.ofNullable(this.postFilterThreshold); }
     Optional<Double> getApproximateThreshold() { return Optional.ofNullable(this.approximateThreshold); }
+    Optional<Double> getTargetHitsMaxAdjustmentFactor() { return Optional.ofNullable(this.targetHitsMaxAdjustmentFactor); }
     List<FeatureList> getMatchFeatures() { return List.copyOf(this.matchFeatures); }
     List<FeatureList> getRankFeatures() { return List.copyOf(this.rankFeatures); }
     List<FeatureList> getSummaryFeatures() { return List.copyOf(this.summaryFeatures); }
@@ -231,4 +233,9 @@ class ParsedRankProfile extends ParsedBlock {
         this.approximateThreshold = threshold;
     }
 
+    void setTargetHitsMaxAdjustmentFactor(double factor) {
+        verifyThat(targetHitsMaxAdjustmentFactor == null, "already has target-hits-max-adjustment-factor");
+        this.targetHitsMaxAdjustmentFactor = factor;
+    }
+
 }
diff --git a/config-model/src/main/javacc/SchemaParser.jj b/config-model/src/main/javacc/SchemaParser.jj
index b2cb258c0ab..42eeabb5ac7 100644
--- a/config-model/src/main/javacc/SchemaParser.jj
+++ b/config-model/src/main/javacc/SchemaParser.jj
@@ -326,6 +326,7 @@ TOKEN :
 | < TERMWISE_LIMIT: "termwise-limit" >
 | < POST_FILTER_THRESHOLD: "post-filter-threshold" >
 | < APPROXIMATE_THRESHOLD: "approximate-threshold" >
+| < TARGET_HITS_MAX_ADJUSTMENT_FACTOR: "target-hits-max-adjustment-factor" >
 | < KEEP_RANK_COUNT: "keep-rank-count" >
 | < RANK_SCORE_DROP_LIMIT: "rank-score-drop-limit" >
 | < CONSTANTS: "constants" >
@@ -1727,6 +1728,7 @@ void rankProfileItem(ParsedSchema schema, ParsedRankProfile profile) : { }
       | termwiseLimit(profile)
       | postFilterThreshold(profile)
       | approximateThreshold(profile)
+      | targetHitsMaxAdjustmentFactor(profile)
       | rankFeatures(profile)
       | rankProperties(profile)
       | secondPhase(profile)
@@ -2190,6 +2192,19 @@ void approximateThreshold(ParsedRankProfile profile) :
 }
 
 /**
+ * This rule consumes a target-hits-max-adjustment-factor statement for a rank profile.
+ *
+ * @param profile the rank profile to modify
+ */
+void targetHitsMaxAdjustmentFactor(ParsedRankProfile profile) :
+{
+    double factor;
+}
+{
+    (<TARGET_HITS_MAX_ADJUSTMENT_FACTOR> <COLON> factor = floatValue()) { profile.setTargetHitsMaxAdjustmentFactor(factor); }
+}
+
+/**
  * Consumes a rank-properties block of a rank profile. There
  * is a little trick within this rule to allow the final rank property
  * to skip the terminating newline token.
@@ -2641,6 +2656,7 @@ String identifierWithDash() :
     | <SECOND_PHASE>
     | <STRUCT_FIELD>
     | <SUMMARY_TO>
+    | <TARGET_HITS_MAX_ADJUSTMENT_FACTOR>
     | <TERMWISE_LIMIT>
     | <UPPER_BOUND>
     ) { return token.image; }
diff --git a/config-model/src/test/java/com/yahoo/schema/RankProfileTestCase.java b/config-model/src/test/java/com/yahoo/schema/RankProfileTestCase.java
index 85225f0d255..380b458ea8c 100644
--- a/config-model/src/test/java/com/yahoo/schema/RankProfileTestCase.java
+++ b/config-model/src/test/java/com/yahoo/schema/RankProfileTestCase.java
@@ -459,17 +459,9 @@ public class RankProfileTestCase extends AbstractSchemaTestCase {
     }
 
     private void verifyApproximateNearestNeighborThresholdSettings(Double postFilterThreshold, Double approximateThreshold) throws ParseException {
-        var rankProfileRegistry = new RankProfileRegistry();
-        var props = new TestProperties();
-        var queryProfileRegistry = new QueryProfileRegistry();
-        var builder = new ApplicationBuilder(rankProfileRegistry, queryProfileRegistry, props);
-        builder.addSchema(createSDWithRankProfileThresholds(postFilterThreshold, approximateThreshold));
-        builder.build(true);
-
-        var schema = builder.getSchema();
-        var rankProfile = rankProfileRegistry.get(schema, "my_profile");
-        var rawRankProfile = new RawRankProfile(rankProfile, new LargeRankingExpressions(new MockFileRegistry()), queryProfileRegistry,
-                                                new ImportedMlModels(), new AttributeFields(schema), props);
+        var rp = createRankProfile(postFilterThreshold, approximateThreshold, null);
+        var rankProfile = rp.getFirst();
+        var rawRankProfile = rp.getSecond();
 
         if (postFilterThreshold != null) {
             assertEquals((double)postFilterThreshold, rankProfile.getPostFilterThreshold().getAsDouble(), 0.000001);
@@ -488,13 +480,52 @@ public class RankProfileTestCase extends AbstractSchemaTestCase {
         }
     }
 
-    private String createSDWithRankProfileThresholds(Double postFilterThreshold, Double approximateThreshold) {
+    @Test
+    void target_hits_max_adjustment_factor_is_configurable() throws ParseException {
+        verifyTargetHitsMaxAdjustmentFactor(null);
+        verifyTargetHitsMaxAdjustmentFactor(2.0);
+    }
+
+    private void verifyTargetHitsMaxAdjustmentFactor(Double targetHitsMaxAdjustmentFactor) throws ParseException {
+        var rp = createRankProfile(null, null, targetHitsMaxAdjustmentFactor);
+        var rankProfile = rp.getFirst();
+        var rawRankProfile = rp.getSecond();
+        if (targetHitsMaxAdjustmentFactor != null) {
+            assertEquals((double)targetHitsMaxAdjustmentFactor, rankProfile.getTargetHitsMaxAdjustmentFactor().getAsDouble(), 0.000001);
+            assertEquals(String.valueOf(targetHitsMaxAdjustmentFactor), findProperty(rawRankProfile.configProperties(), "vespa.matching.nns.target_hits_max_adjustment_factor").get());
+        } else {
+            assertTrue(rankProfile.getTargetHitsMaxAdjustmentFactor().isEmpty());
+            assertFalse(findProperty(rawRankProfile.configProperties(), "vespa.matching.nns.target_hits_max_adjustment_factor").isPresent());
+        }
+    }
+
+    private Pair<RankProfile, RawRankProfile> createRankProfile(Double postFilterThreshold,
+                                                                Double approximateThreshold,
+                                                                Double targetHitsMaxAdjustmentFactor) throws ParseException {
+        var rankProfileRegistry = new RankProfileRegistry();
+        var props = new TestProperties();
+        var queryProfileRegistry = new QueryProfileRegistry();
+        var builder = new ApplicationBuilder(rankProfileRegistry, queryProfileRegistry, props);
+        builder.addSchema(createSDWithRankProfile(postFilterThreshold, approximateThreshold, targetHitsMaxAdjustmentFactor));
+        builder.build(true);
+
+        var schema = builder.getSchema();
+        var rankProfile = rankProfileRegistry.get(schema, "my_profile");
+        var rawRankProfile = new RawRankProfile(rankProfile, new LargeRankingExpressions(new MockFileRegistry()), queryProfileRegistry,
+                new ImportedMlModels(), new AttributeFields(schema), props);
+        return new Pair<>(rankProfile, rawRankProfile);
+    }
+
+    private String createSDWithRankProfile(Double postFilterThreshold,
+                                           Double approximateThreshold,
+                                           Double targetHitsMaxAdjustmentFactor) {
         return joinLines(
                 "search test {",
                 "    document test {}",
                 "    rank-profile my_profile {",
-                (postFilterThreshold != null ? ("        post-filter-threshold: " + postFilterThreshold) : ""),
-                (approximateThreshold != null ? ("        approximate-threshold: " + approximateThreshold) : ""),
+                (postFilterThreshold != null ?           ("        post-filter-threshold: " + postFilterThreshold) : ""),
+                (approximateThreshold != null ?          ("        approximate-threshold: " + approximateThreshold) : ""),
+                (targetHitsMaxAdjustmentFactor != null ? ("        target-hits-max-adjustment-factor: " + targetHitsMaxAdjustmentFactor) : ""),
                 "    }",
                 "}");
     }