From 6091cc0b805bdd7b083612f05fe6098dcdbe6c37 Mon Sep 17 00:00:00 2001
From: seb <sebgreenbus@gmail.com>
Date: Fri, 18 Apr 2025 09:16:26 +0200
Subject: [PATCH] Increase default capacity in SearchIndexWrapper and enhance
 similarity calculation in calculate_similarity function to boost similarity
 score when all query words are found. Update MAX_WORDS and MAX_STRING_LEN
 definitions for improved handling.

---
 similarity_search.c        | 25 +++++++++++++++++++++----
 similarity_search.h        |  6 +++---
 similarity_search_addon.cc |  2 +-
 3 files changed, 25 insertions(+), 8 deletions(-)

diff --git a/similarity_search.c b/similarity_search.c
index 9cb8b33..0c1e0ee 100644
--- a/similarity_search.c
+++ b/similarity_search.c
@@ -70,19 +70,33 @@ float calculate_similarity(const char *query, const char *target, float cutoff)
         return 0.0;
     }
     
-    // Count matches
+    // Count matches and track which query words were found
     int matches = 0;
+    int query_words_found = 0;
+    int found_query_words[MAX_WORDS] = {0}; // Track which query words were found
+    
     for (int i = 0; i < query_word_count; i++) {
         for (int j = 0; j < target_word_count; j++) {
             if (str_case_cmp(query_words[i], target_words[j]) == 0) {
                 matches++;
+                if (!found_query_words[i]) {
+                    found_query_words[i] = 1;
+                    query_words_found++;
+                }
                 break;
             }
         }
     }
     
-    // Calculate Jaccard similarity (intersection over union)
-    float similarity = (float)matches / (query_word_count + target_word_count - matches);
+    // Calculate base similarity (intersection over union)
+    float base_similarity = (float)matches / (query_word_count + target_word_count - matches);
+    
+    // If all query words were found, boost the similarity
+    float similarity = base_similarity;
+    if (query_words_found == query_word_count) {
+        // If all query words were found, similarity should be at least 0.8
+        similarity = base_similarity > 0.8f ? base_similarity : 0.8f;
+    }
     
     free_words(query_words, query_word_count);
     free_words(target_words, target_word_count);
@@ -247,7 +261,10 @@ SearchResult* search_index(SearchIndex* index, const char* query, float cutoff,
     }
     
     // Copy results to final array
-    memcpy(results, temp_results, *num_results * sizeof(SearchResult));
+    for (int i = 0; i < *num_results; i++) {
+        results[i].string = temp_results[i].string;
+        results[i].similarity = temp_results[i].similarity;
+    }
     free(temp_results);
     
     return results;
diff --git a/similarity_search.h b/similarity_search.h
index 613dd99..c93405b 100644
--- a/similarity_search.h
+++ b/similarity_search.h
@@ -5,8 +5,8 @@
 extern "C" {
 #endif
 
-#define MAX_STRING_LEN 100
-#define MAX_WORDS 20
+#define MAX_STRING_LEN 1000
+#define MAX_WORDS 100
 
 // Public API
 
@@ -19,7 +19,7 @@ typedef struct {
 
 // Structure to hold a search result
 typedef struct {
-    const char *string;
+    char *string;
     float similarity;
 } SearchResult;
 
diff --git a/similarity_search_addon.cc b/similarity_search_addon.cc
index f065134..40aa2f4 100644
--- a/similarity_search_addon.cc
+++ b/similarity_search_addon.cc
@@ -41,7 +41,7 @@ SearchIndexWrapper::SearchIndexWrapper(const Napi::CallbackInfo& info)
   Napi::Env env = info.Env();
   Napi::HandleScope scope(env);
   
-  int capacity = 500; // Default capacity
+  int capacity = 10000; // Increased default capacity from 500 to 10000
   if (info.Length() > 0 && info[0].IsNumber()) {
     capacity = info[0].As<Napi::Number>().Int32Value();
   }