properly handle score_cutoff > 1.0

Author: maxbachmann

CHANGELOG.md

@@ -1,5 +1,9 @@
 ## Changelog
 
+### [1.11.2] - 2023-04-17
+#### Fixed
+- fix handling of `score_cutoff > 1.0` in `Jaro` and `JaroWinkler`
+
 ### [1.11.1] - 2023-04-16
 #### Fixed
 - fix division by zero in simd implementation of normalized string metrics, when comparing empty strings

CMakeLists.txt

@@ -25,7 +25,7 @@ if (CMAKE_BINARY_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
     message(FATAL_ERROR "Building in-source is not supported! Create a build dir and remove ${CMAKE_SOURCE_DIR}/CMakeCache.txt")
 endif()
 
-project(rapidfuzz LANGUAGES CXX VERSION 1.11.1)
+project(rapidfuzz LANGUAGES CXX VERSION 1.11.2)
 
 list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_LIST_DIR}/cmake")
 include(GNUInstallDirs)

extras/rapidfuzz_amalgamated.hpp

@@ -1,7 +1,7 @@
 //  Licensed under the MIT License <http://opensource.org/licenses/MIT>.
 //  SPDX-License-Identifier: MIT
 //  RapidFuzz v1.0.2
-//  Generated: 2023-04-17 01:55:45.256062
+//  Generated: 2023-04-17 13:47:21.759334
 //  ----------------------------------------------------------
 //  This file is an amalgamation of multiple different files.
 //  You probably shouldn't edit it directly.
@@ -5203,6 +5203,8 @@ double jaro_similarity(Range<InputIt1> P, Range<InputIt2> T, double score_cutoff
     int64_t P_len = P.size();
     int64_t T_len = T.size();
 
+    if (score_cutoff > 1.0) return 0.0;
+
     if (!P_len && !T_len) return 1.0;
 
     /* filter out based on the length difference between the two strings */
@@ -5250,6 +5252,8 @@ double jaro_similarity(const BlockPatternMatchVector& PM, Range<InputIt1> P, Ran
     int64_t P_len = P.size();
     int64_t T_len = T.size();
 
+    if (score_cutoff > 1.0) return 0.0;
+
     if (!P_len && !T_len) return 1.0;
 
     /* filter out based on the length difference between the two strings */

rapidfuzz/distance/Jaro_impl.hpp

@@ -343,6 +343,8 @@ double jaro_similarity(Range<InputIt1> P, Range<InputIt2> T, double score_cutoff
     int64_t P_len = P.size();
     int64_t T_len = T.size();
 
+    if (score_cutoff > 1.0) return 0.0;
+
     if (!P_len && !T_len) return 1.0;
 
     /* filter out based on the length difference between the two strings */
@@ -390,6 +392,8 @@ double jaro_similarity(const BlockPatternMatchVector& PM, Range<InputIt1> P, Ran
     int64_t P_len = P.size();
     int64_t T_len = T.size();
 
+    if (score_cutoff > 1.0) return 0.0;
+
     if (!P_len && !T_len) return 1.0;
 
     /* filter out based on the length difference between the two strings */

rapidfuzz_reference/Jaro.hpp

@@ -20,7 +20,9 @@ double jaro_similarity(InputIt1 P_first, InputIt1 P_last, InputIt2 T_first, Inpu
     size_t P_len = static_cast<size_t>(std::distance(P_first, P_last));
     size_t T_len = static_cast<size_t>(std::distance(T_first, T_last));
 
-    if (!P_len || !T_len) return 1.0;
+    if (score_cutoff > 1.0) return 0.0;
+
+    if (!P_len || !T_len) return double(!P_len && !T_len);
 
     std::vector<int> P_flag(P_len + 1);
     std::vector<int> T_flag(T_len + 1);

test/distance/tests-Jaro.cpp

@@ -16,8 +16,8 @@ double jaro_similarity(const Sentence1& s1, const Sentence2& s2, double score_cu
     rapidfuzz::CachedJaro scorer(s1);
     double res5 = scorer.similarity(s2, score_cutoff);
     double res6 = scorer.similarity(s2.begin(), s2.end(), score_cutoff);
-    double res7 = scorer.similarity(s2, score_cutoff);
-    double res8 = scorer.similarity(s2.begin(), s2.end(), score_cutoff);
+    double res7 = scorer.normalized_similarity(s2, score_cutoff);
+    double res8 = scorer.normalized_similarity(s2.begin(), s2.end(), score_cutoff);
     REQUIRE(res1 == Approx(res2));
     REQUIRE(res1 == Approx(res3));
     REQUIRE(res1 == Approx(res4));
@@ -39,8 +39,8 @@ double jaro_distance(const Sentence1& s1, const Sentence2& s2, double score_cuto
     rapidfuzz::CachedJaro scorer(s1);
     double res5 = scorer.distance(s2, score_cutoff);
     double res6 = scorer.distance(s2.begin(), s2.end(), score_cutoff);
-    double res7 = scorer.distance(s2, score_cutoff);
-    double res8 = scorer.distance(s2.begin(), s2.end(), score_cutoff);
+    double res7 = scorer.normalized_distance(s2, score_cutoff);
+    double res8 = scorer.normalized_distance(s2.begin(), s2.end(), score_cutoff);
     REQUIRE(res1 == Approx(res2));
     REQUIRE(res1 == Approx(res3));
     REQUIRE(res1 == Approx(res4));
@@ -63,7 +63,9 @@ TEST_CASE("JaroWinklerTest")
 
     SECTION("testFullResultWithScoreCutoff")
     {
-        for (double score_cutoff = 0.0; score_cutoff < 1.1; score_cutoff += 0.1)
+        auto score_cutoffs = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1};
+
+        for (double score_cutoff : score_cutoffs)
             for (const auto& name1 : names)
                 for (const auto& name2 : names) {
                     INFO("name1: " << name1 << ", name2: " << name2 << ", score_cutoff: " << score_cutoff);

test/distance/tests-JaroWinkler.cpp

@@ -18,8 +18,8 @@ double jaro_winkler_similarity(const Sentence1& s1, const Sentence2& s2, double
     rapidfuzz::CachedJaroWinkler scorer(s1, prefix_weight);
     double res5 = scorer.similarity(s2, score_cutoff);
     double res6 = scorer.similarity(s2.begin(), s2.end(), score_cutoff);
-    double res7 = scorer.similarity(s2, score_cutoff);
-    double res8 = scorer.similarity(s2.begin(), s2.end(), score_cutoff);
+    double res7 = scorer.normalized_similarity(s2, score_cutoff);
+    double res8 = scorer.normalized_similarity(s2.begin(), s2.end(), score_cutoff);
     REQUIRE(res1 == Approx(res2));
     REQUIRE(res1 == Approx(res3));
     REQUIRE(res1 == Approx(res4));
@@ -43,8 +43,8 @@ double jaro_winkler_distance(const Sentence1& s1, const Sentence2& s2, double pr
     rapidfuzz::CachedJaroWinkler scorer(s1, prefix_weight);
     double res5 = scorer.distance(s2, score_cutoff);
     double res6 = scorer.distance(s2.begin(), s2.end(), score_cutoff);
-    double res7 = scorer.distance(s2, score_cutoff);
-    double res8 = scorer.distance(s2.begin(), s2.end(), score_cutoff);
+    double res7 = scorer.normalized_distance(s2, score_cutoff);
+    double res8 = scorer.normalized_distance(s2.begin(), s2.end(), score_cutoff);
     REQUIRE(res1 == Approx(res2));
     REQUIRE(res1 == Approx(res3));
     REQUIRE(res1 == Approx(res4));
@@ -67,7 +67,9 @@ TEST_CASE("JaroWinklerTest")
 
     SECTION("testFullResultWithScoreCutoff")
     {
-        for (double score_cutoff = 0.0; score_cutoff < 1.1; score_cutoff += 0.1)
+        auto score_cutoffs = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1};
+
+        for (double score_cutoff : score_cutoffs)
             for (const auto& name1 : names)
                 for (const auto& name2 : names) {
                     INFO("name1: " << name1 << ", name2: " << name2 << ", score_cutoff: " << score_cutoff);