handle empty strings more consistently

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: 2023-04-16 23:07:04.138210
+//  Generated: 2023-04-17 01:55:45.256062
 //  ----------------------------------------------------------
 //  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 (!P_len && !T_len) return 1.0;
+
     /* filter out based on the length difference between the two strings */
     if (!jaro_length_filter(P_len, T_len, score_cutoff)) return 0.0;
 
@@ -5248,6 +5250,8 @@ double jaro_similarity(const BlockPatternMatchVector& PM, Range<InputIt1> P, Ran
     int64_t P_len = P.size();
     int64_t T_len = T.size();
 
+    if (!P_len && !T_len) return 1.0;
+
     /* filter out based on the length difference between the two strings */
     if (!jaro_length_filter(P_len, T_len, score_cutoff)) return 0.0;
 

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 (!P_len && !T_len) return 1.0;
+
     /* filter out based on the length difference between the two strings */
     if (!jaro_length_filter(P_len, T_len, score_cutoff)) return 0.0;
 
@@ -388,6 +390,8 @@ double jaro_similarity(const BlockPatternMatchVector& PM, Range<InputIt1> P, Ran
     int64_t P_len = P.size();
     int64_t T_len = T.size();
 
+    if (!P_len && !T_len) return 1.0;
+
     /* filter out based on the length difference between the two strings */
     if (!jaro_length_filter(P_len, T_len, score_cutoff)) return 0.0;
 

rapidfuzz_reference/Jaro.hpp

@@ -20,7 +20,7 @@ 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 0;
+    if (!P_len || !T_len) return 1.0;
 
     std::vector<int> P_flag(P_len + 1);
     std::vector<int> T_flag(T_len + 1);

test/distance/tests-Jaro.cpp

@@ -56,10 +56,10 @@ double jaro_distance(const Sentence1& s1, const Sentence2& s2, double score_cuto
  */
 TEST_CASE("JaroWinklerTest")
 {
-    std::array<std::string, 19> names = {"james",    "robert",   "john",   "michael",   "william",
+    std::array<std::string, 20> names = {"james",    "robert",   "john",   "michael",   "william",
                                          "david",    "joseph",   "thomas", "charles",   "mary",
                                          "patricia", "jennifer", "linda",  "elizabeth", "barbara",
-                                         "susan",    "jessica",  "sarah",  "karen"};
+                                         "susan",    "jessica",  "sarah",  "karen", ""};
 
     SECTION("testFullResultWithScoreCutoff")
     {

test/distance/tests-JaroWinkler.cpp

@@ -60,10 +60,10 @@ double jaro_winkler_distance(const Sentence1& s1, const Sentence2& s2, double pr
  */
 TEST_CASE("JaroWinklerTest")
 {
-    std::array<std::string, 19> names = {"james",    "robert",   "john",   "michael",   "william",
+    std::array<std::string, 20> names = {"james",    "robert",   "john",   "michael",   "william",
                                          "david",    "joseph",   "thomas", "charles",   "mary",
                                          "patricia", "jennifer", "linda",  "elizabeth", "barbara",
-                                         "susan",    "jessica",  "sarah",  "karen"};
+                                         "susan",    "jessica",  "sarah",  "karen", ""};
 
     SECTION("testFullResultWithScoreCutoff")
     {