fix: optimistic word sort

Author: NiuGuohui

.nvmrc

@@ -1 +1 @@
-v20.19.0
+v20

cpp/JSMdict.cpp

@@ -95,20 +95,19 @@ namespace JSMdict {
         return result;
     }
 
-    int levenshteinDistance(const std::string &s1, const std::string &s2) {
-        size_t len1 = s1.size(), len2 = s2.size();
-        std::vector<std::vector<int>> dp(len1 + 1, std::vector<int>(len2 + 1));
-
-        for (size_t i = 0; i <= len1; ++i) dp[i][0] = i;
-        for (size_t j = 0; j <= len2; ++j) dp[0][j] = j;
-
-        for (size_t i = 1; i <= len1; ++i) {
-            for (size_t j = 1; j <= len2; ++j) {
-                dp[i][j] = std::min({dp[i - 1][j] + 1, dp[i][j - 1] + 1, dp[i - 1][j - 1] + (s1[i - 1] != s2[j - 1])});
-            }
-        }
-
-        return dp[len1][len2];
+    int bigramJaccard(std::string_view a, std::string_view b) {
+        auto bigrams = [](std::string_view s, auto &out) {
+            for (size_t i = 0; i + 1 < s.size(); ++i)
+                out.insert((unsigned char) std::tolower(s[i]) << 8 |
+                           (unsigned char) std::tolower(s[i + 1]));
+        };
+        std::unordered_set<unsigned> A, B;
+        bigrams(a, A);
+        bigrams(b, B);
+        if (A.empty() || B.empty()) return 0;
+        size_t comm = 0;
+        for (auto v: A) if (B.count(v)) ++comm;
+        return static_cast<int>(std::round(float(comm) / float(A.size() + B.size() - comm) * 100));
     }
 
     int computeScore(const std::string &candidateOriginal, const std::string &queryOriginal) {
@@ -118,28 +117,25 @@ namespace JSMdict {
         if (candidate == query) return 100;
         if (candidate.starts_with(query)) return 80;
         if (candidate.find(query) != std::string::npos) return 50;
-        return std::max(0, 100 - levenshteinDistance(candidate, query) * 10);
+        return bigramJaccard(candidate, query);
     }
 
     std::vector<JSVariant>
     sortKeyListByKeywordRelevance(std::vector<mdict::key_list_item *> &items, const std::string &query) {
         auto compare = [&query](const mdict::key_list_item *a, const mdict::key_list_item *b) {
             int scoreA = computeScore(a->key_word, query);
             int scoreB = computeScore(b->key_word, query);
-            return scoreA > scoreB || (scoreA == scoreB && a->record_start < b->record_start);
+            if (a->key_word.size() != b->key_word.size()) return a->key_word.size() < b->key_word.size();
+            return a->record_start < b->record_start;
         };
-        if (items.size() >= 10) {
-            std::partial_sort(items.begin(), items.begin() + 10, items.end(), compare);
-            items.resize(10);
-        } else {
-            std::sort(items.begin(), items.begin(), compare);
-        }
+        size_t k = std::min(size_t(10), items.size());
+        std::partial_sort(items.begin(), items.begin() + k, items.end(), compare);
+        items.resize(k);
         auto list = std::vector<JSVariant>();
         for (auto item: items) list.push_back(JSVariant(item->key_word));
         return list;
     }
 
-
     jsi::Value JSMdict::keyList(jsi::Runtime &runtime, std::string query) {
         checkInitialized(runtime);
         return Promise::createPromise(
@@ -159,4 +155,3 @@ namespace JSMdict {
                 });
     }
 }
-

package.json

@@ -1,6 +1,6 @@
 {
   "name": "react-native-mdict",
-  "version": "0.3.1",
+  "version": "0.3.2",
   "description": "A mdict(*.mdx/mdd) file parser.",
   "main": "./lib/module/index.js",
   "types": "./lib/typescript/src/index.d.ts",