From c712cc0ca72a331d290ea0764847127c4a3dc38b Mon Sep 17 00:00:00 2001
From: Kishore Nallan <kishorenc@gmail.com>
Date: Fri, 21 May 2021 19:01:39 +0530
Subject: [PATCH] Ensure floating point precision is correct when faceted.

---
 include/string_utils.h            |  2 ++
 src/collection.cpp                |  4 ++--
 src/index.cpp                     |  6 +++---
 src/string_utils.cpp              |  6 ++++++
 test/collection_faceting_test.cpp | 36 +++++++++++++++++++++++++++++++
 5 files changed, 49 insertions(+), 5 deletions(-)

diff --git a/include/string_utils.h b/include/string_utils.h
index c1d7f201..552a2b6b 100644
--- a/include/string_utils.h
+++ b/include/string_utils.h
@@ -294,4 +294,6 @@ struct StringUtils {
     }
 
     static std::map<std::string, std::string> parse_query_string(const std::string& query);
+
+    static std::string float_to_str(float value);
 };
\ No newline at end of file
diff --git a/src/collection.cpp b/src/collection.cpp
index f72e17d7..2e443153 100644
--- a/src/collection.cpp
+++ b/src/collection.cpp
@@ -1370,11 +1370,11 @@ bool Collection::facet_value_to_string(const facet &a_facet, const facet_count_t
         value = std::to_string(raw_val);
     } else if(facet_schema.at(a_facet.field_name).type == field_types::FLOAT) {
         float raw_val = document[a_facet.field_name].get<float>();
-        value = std::to_string(raw_val);
+        value = StringUtils::float_to_str(raw_val);
         value.erase ( value.find_last_not_of('0') + 1, std::string::npos ); // remove trailing zeros
     } else if(facet_schema.at(a_facet.field_name).type == field_types::FLOAT_ARRAY) {
         float raw_val = document[a_facet.field_name][facet_count.array_pos].get<float>();
-        value = std::to_string(raw_val);
+        value = StringUtils::float_to_str(raw_val);
         value.erase ( value.find_last_not_of('0') + 1, std::string::npos );  // remove trailing zeros
     } else if(facet_schema.at(a_facet.field_name).type == field_types::BOOL) {
         value = std::to_string(document[a_facet.field_name].get<bool>());
diff --git a/src/index.cpp b/src/index.cpp
index aa4908ab..bfeed79c 100644
--- a/src/index.cpp
+++ b/src/index.cpp
@@ -163,7 +163,7 @@ Option<uint32_t> Index::index_in_memory(const nlohmann::json &document, uint32_t
                     }
                 } else if(field_pair.second.type == field_types::FLOAT_ARRAY) {
                     for(float value: document[field_name]){
-                        strings.push_back(std::to_string(value));
+                        strings.push_back(StringUtils::float_to_str(value));
                     }
                 } else if(field_pair.second.type == field_types::BOOL_ARRAY) {
                     for(bool value: document[field_name]){
@@ -179,7 +179,7 @@ Option<uint32_t> Index::index_in_memory(const nlohmann::json &document, uint32_t
                 } else if(field_pair.second.type == field_types::INT64) {
                     text = std::to_string(document[field_name].get<int64_t>());
                 } else if(field_pair.second.type == field_types::FLOAT) {
-                    text = std::to_string(document[field_name].get<float>());
+                    text = StringUtils::float_to_str(document[field_name].get<float>());
                 } else if(field_pair.second.type == field_types::BOOL) {
                     text = std::to_string(document[field_name].get<bool>());
                 }
@@ -2608,7 +2608,7 @@ Option<uint32_t> Index::coerce_string(const DIRTY_VALUES& dirty_values, const st
     }
 
     else if(item.is_number_float()) {
-        item = std::to_string((float)item);
+        item = StringUtils::float_to_str((float)item);
     }
 
     else if(item.is_boolean()) {
diff --git a/src/string_utils.cpp b/src/string_utils.cpp
index 717405f4..617b1b00 100644
--- a/src/string_utils.cpp
+++ b/src/string_utils.cpp
@@ -180,6 +180,12 @@ void StringUtils::split_to_values(const std::string& vals_str, std::vector<std::
     }
 }
 
+std::string StringUtils::float_to_str(float value) {
+    std::ostringstream os;
+    os << value;
+    return os.str();
+}
+
 /*size_t StringUtils::unicode_length(const std::string& bytes) {
     std::wstring_convert<std::codecvt_utf8<char32_t>, char32_t> utf8conv;
     return utf8conv.from_bytes(bytes).size();
diff --git a/test/collection_faceting_test.cpp b/test/collection_faceting_test.cpp
index bfbcb1ca..3aba9e64 100644
--- a/test/collection_faceting_test.cpp
+++ b/test/collection_faceting_test.cpp
@@ -404,6 +404,42 @@ TEST_F(CollectionFacetingTest, FacetCountsBool) {
     collectionManager.drop_collection("coll1");
 }
 
+TEST_F(CollectionFacetingTest, FacetCountsFloatPrecision) {
+    Collection *coll1;
+
+    std::vector<field> fields = {field("title", field_types::STRING, false),
+                                 field("points", field_types::FLOAT, true)};
+
+    std::vector<sort_by> sort_fields = {sort_by("points", "DESC")};
+
+    coll1 = collectionManager.get_collection("coll1").get();
+    if (coll1 == nullptr) {
+        coll1 = collectionManager.create_collection("coll1", 4, fields, "points").get();
+    }
+
+    nlohmann::json doc;
+    doc["id"] = "100";
+    doc["title"] = "Ford Mustang";
+    doc["points"] = 113.4;
+
+    coll1->add(doc.dump());
+
+    std::vector<std::string> facets = {"points"};
+
+    nlohmann::json results = coll1->search("*", {"title"}, "", facets, sort_fields, {0}, 10, 1,
+                                           token_ordering::FREQUENCY, true).get();
+
+    ASSERT_EQ(1, results["facet_counts"].size());
+    ASSERT_EQ(1, results["facet_counts"][0]["counts"].size());
+
+    ASSERT_STREQ("points", results["facet_counts"][0]["field_name"].get<std::string>().c_str());
+    ASSERT_EQ(1, (int) results["facet_counts"][0]["counts"][0]["count"]);
+    ASSERT_STREQ("113.4", results["facet_counts"][0]["counts"][0]["value"].get<std::string>().c_str());
+    ASSERT_STREQ("113.4",results["facet_counts"][0]["counts"][0]["highlighted"].get<std::string>().c_str());
+
+    collectionManager.drop_collection("coll1");
+}
+
 TEST_F(CollectionFacetingTest, FacetCountsHighlighting) {
     Collection *coll1;