Merge pull request #32656 from vespa-engine/havardpe/allow-at-least-o…

…ne-of-each-type allow at least one of each category
vespa-engine · Oct 24, 2024 · e8530ee · e8530ee
2 parents f5473ef + 6e5f359
commit e8530ee
Show file tree

Hide file tree

Showing 5 changed files with 8 additions and 6 deletions.
diff --git a/searchcore/src/tests/proton/matching/match_phase_limiter/match_phase_limiter_test.cpp b/searchcore/src/tests/proton/matching/match_phase_limiter/match_phase_limiter_test.cpp
@@ -174,7 +174,9 @@ TEST("require that max group size is calculated correctly") {
     for (size_t min_groups: std::vector<size_t>({0, 1, 2, 3, 4, 10, 500})) {
         for (size_t wanted_hits: std::vector<size_t>({0, 3, 321, 921})) {
             MatchPhaseLimitCalculator calc(100, min_groups, 0.2);
-            if (min_groups == 0) {
+            if (wanted_hits <= min_groups) {
+                EXPECT_EQUAL(size_t(1), calc.max_group_size(wanted_hits));
+            } else if (min_groups == 0) {
                 EXPECT_EQUAL(wanted_hits, calc.max_group_size(wanted_hits));
             } else {
                 EXPECT_EQUAL((wanted_hits / min_groups), calc.max_group_size(wanted_hits));

diff --git a/searchcore/src/vespa/searchcore/proton/matching/match_phase_limit_calculator.h b/searchcore/src/vespa/searchcore/proton/matching/match_phase_limit_calculator.h
@@ -39,8 +39,8 @@ class MatchPhaseLimitCalculator
     size_t estimated_hits(double hit_rate, size_t num_docs) const noexcept {
         return (size_t) (hit_rate * num_docs);
     }
-    size_t max_group_size(size_t wanted_num_docs_in) const noexcept {
-        return (wanted_num_docs_in / _min_groups);
+    size_t max_group_size(size_t wanted_num_docs) const noexcept {
+        return std::max(size_t(1), wanted_num_docs / _min_groups);
     }
 };
 

diff --git a/searchcore/src/vespa/searchcore/proton/matching/match_tools.cpp b/searchcore/src/vespa/searchcore/proton/matching/match_tools.cpp
@@ -267,7 +267,7 @@ MatchToolsFactory::createDiversifier(uint32_t heapSize) const
         Issue::report("Skipping diversity due to no %s attribute.", _diversityParams.attribute.c_str());
         return {};
     }
-    size_t max_per_group = heapSize/_diversityParams.min_groups;
+    size_t max_per_group = std::max(size_t(1), size_t(heapSize / _diversityParams.min_groups));
     return DiversityFilter::create(*attr, heapSize, max_per_group, _diversityParams.min_groups,
                                    _diversityParams.cutoff_strategy == DiversityParams::CutoffStrategy::STRICT);
 }

diff --git a/searchlib/src/vespa/searchlib/attribute/diversity.cpp b/searchlib/src/vespa/searchlib/attribute/diversity.cpp
@@ -101,7 +101,7 @@ DiversityFilterT<Fetcher>::accepted(uint32_t docId) {
 
 std::unique_ptr<DiversityFilter>
 DiversityFilter::create(const IAttributeVector &diversity_attr, size_t wanted_hits,
-                        size_t max_per_group,size_t cutoff_max_groups, bool cutoff_strict)
+                        size_t max_per_group, size_t cutoff_max_groups, bool cutoff_strict)
 {
     if (diversity_attr.hasEnum()) { // must handle enum first
         FetchEnumFast fastEnum(diversity_attr);

diff --git a/searchlib/src/vespa/searchlib/attribute/diversity.h b/searchlib/src/vespa/searchlib/attribute/diversity.h
@@ -65,7 +65,7 @@ class DiversityFilter : public queryeval::IDiversifier {
     size_t getMaxTotal() const { return _max_total; }
     static std::unique_ptr<DiversityFilter>
     create(const IAttributeVector &diversity_attr, size_t wanted_hits,
-           size_t max_per_group,size_t cutoff_max_groups, bool cutoff_strict);
+           size_t max_per_group, size_t cutoff_max_groups, bool cutoff_strict);
 protected:
     size_t _max_total;
 };