Merge pull request #31127 from vespa-engine/geirst/query-cost-model-adjustments

Adjust down the estimate when its unknown and avoid WhiteListBlueprin…

4 files changed, 17 insertions, 9 deletions

diff --git a/searchcore/src/vespa/searchcore/proton/documentmetastore/lid_allocator.cpp b/searchcore/src/vespa/searchcore/proton/documentmetastore/lid_allocator.cpp
index 0c986422be6..758d1336399 100644
--- a/searchcore/src/vespa/searchcore/proton/documentmetastore/lid_allocator.cpp
+++ b/searchcore/src/vespa/searchcore/proton/documentmetastore/lid_allocator.cpp
@@ -210,7 +210,8 @@ private:
     }
     FlowStats calculate_flow_stats(uint32_t docid_limit) const override {
         double rel_est = abs_to_rel_est(_activeLids.size(), docid_limit);
-        return {rel_est, bitvector_cost(), bitvector_strict_cost(rel_est)};
+        double do_not_make_me_strict = 1000.0;
+        return {rel_est, bitvector_cost(), do_not_make_me_strict * bitvector_strict_cost(rel_est)};
     }
     SearchIterator::UP
     createLeafSearch(const TermFieldMatchDataArray &tfmda) const override
diff --git a/searchlib/src/vespa/searchlib/attribute/attribute_blueprint_factory.cpp b/searchlib/src/vespa/searchlib/attribute/attribute_blueprint_factory.cpp
index 5b17b491a20..70b86bf22a1 100644
--- a/searchlib/src/vespa/searchlib/attribute/attribute_blueprint_factory.cpp
+++ b/searchlib/src/vespa/searchlib/attribute/attribute_blueprint_factory.cpp
@@ -94,6 +94,7 @@ using search::queryeval::StrictHeapOrSearch;
 using search::queryeval::WeightedSetTermBlueprint;
 using search::queryeval::flow::btree_cost;
 using search::queryeval::flow::btree_strict_cost;
+using search::queryeval::flow::estimate_when_unknown;
 using search::queryeval::flow::get_num_indirections;
 using search::queryeval::flow::lookup_cost;
 using search::queryeval::flow::lookup_strict_cost;
@@ -150,10 +151,9 @@ public:
     search::queryeval::FlowStats calculate_flow_stats(uint32_t docid_limit) const override {
         if (_hit_estimate.is_unknown()) {
             // E.g. attributes without fast-search are not able to provide a hit estimate.
-            // In this case we just assume matching half of the document corpus.
             // In addition, matching is lookup based, and we are not able to skip documents efficiently when being strict.
             size_t indirections = get_num_indirections(_attr.getBasicType(), _attr.getCollectionType());
-            return {0.5, lookup_cost(indirections), lookup_strict_cost(indirections)};
+            return {estimate_when_unknown(), lookup_cost(indirections), lookup_strict_cost(indirections)};
         } else {
             double rel_est = abs_to_rel_est(_hit_estimate.est_hits(), docid_limit);
             return {rel_est, btree_cost(rel_est), btree_strict_cost(rel_est)};
diff --git a/searchlib/src/vespa/searchlib/queryeval/blueprint.cpp b/searchlib/src/vespa/searchlib/queryeval/blueprint.cpp
index 7334db4b716..cfa165be067 100644
--- a/searchlib/src/vespa/searchlib/queryeval/blueprint.cpp
+++ b/searchlib/src/vespa/searchlib/queryeval/blueprint.cpp
@@ -1,14 +1,15 @@
 // Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
 
 #include "blueprint.h"
-#include "leaf_blueprints.h"
+#include "andnotsearch.h"
+#include "andsearch.h"
 #include "emptysearch.h"
-#include "full_search.h"
 #include "field_spec.hpp"
-#include "andsearch.h"
-#include "orsearch.h"
-#include "andnotsearch.h"
+#include "flow_tuning.h"
+#include "full_search.h"
+#include "leaf_blueprints.h"
 #include "matching_elements_search.h"
+#include "orsearch.h"
 #include <vespa/searchlib/fef/termfieldmatchdataarray.h>
 #include <vespa/vespalib/objects/visit.hpp>
 #include <vespa/vespalib/objects/objectdumper.h>
@@ -238,7 +239,7 @@ Blueprint::default_flow_stats(uint32_t docid_limit, uint32_t abs_est, size_t chi
 FlowStats
 Blueprint::default_flow_stats(size_t child_cnt)
 {
-    return {0.5, 1.0 + child_cnt, 1.0 + child_cnt};
+    return {flow::estimate_when_unknown(), 1.0 + child_cnt, 1.0 + child_cnt};
 }
 
 std::unique_ptr<MatchingElementsSearch>
diff --git a/searchlib/src/vespa/searchlib/queryeval/flow_tuning.h b/searchlib/src/vespa/searchlib/queryeval/flow_tuning.h
index 22faa920bc0..5ed61ef9fc8 100644
--- a/searchlib/src/vespa/searchlib/queryeval/flow_tuning.h
+++ b/searchlib/src/vespa/searchlib/queryeval/flow_tuning.h
@@ -60,6 +60,12 @@ inline size_t get_num_indirections(const attribute::BasicType& basic_type,
     return res;
 }
 
+// Some blueprints are not able to provide a hit estimate (e.g. attributes without fast-search).
+// In such cases the following estimate is used instead. In most cases this is an overestimate.
+inline double estimate_when_unknown() {
+    return 0.1;
+}
+
 // Non-strict cost of lookup based matching in an attribute (not fast-search).
 // Test used: IteratorBenchmark::analyze_term_search_in_attributes_non_strict
 inline double lookup_cost(size_t num_indirections) {