From 1f566ebe0f03ce969d9a6aedd4b4ffcd89bdb747 Mon Sep 17 00:00:00 2001
From: Ivan Leichenko <ivan.leichenko@student.tuke.sk>
Date: Sat, 18 Jan 2025 13:15:16 +0100
Subject: [PATCH] 1 v 1 sk1

---
 sk1/compressor.c | 85 ++++++++++++++++++++++++++++++++++++++----------
 sk1/main.c       |  2 +-
 2 files changed, 68 insertions(+), 19 deletions(-)

diff --git a/sk1/compressor.c b/sk1/compressor.c
index 309c6be..ece92d1 100644
--- a/sk1/compressor.c
+++ b/sk1/compressor.c
@@ -3,6 +3,7 @@
 #include <stdlib.h>
 #include <string.h>
 #include <stdint.h>
+#include <unistd.h>
 
 #define WINDOW_SIZE 4096  // Размер скользящего окна
 #define LOOKAHEAD_BUFFER_SIZE 15  // Размер буфера предпросмотра
@@ -39,7 +40,7 @@ void swap_nodes(Node** a, Node** b)
     *b = temp;
 }
 
-void heapify(MinHeap* heap, int idx) 
+void heapify(MinHeap* heap, int idx)
 {
     int smallest = idx;
     int left = 2 * idx + 1;
@@ -147,14 +148,18 @@ void free_tree(Node* root)
     }
 }
 
-int compress_2(const char* input_file_name, const char* output_file_name)
+int compress_2(const char* input_file_name, const char* output_file_name) 
 {
     FILE* input = fopen(input_file_name, "rb");
-    if (!input) return -1;
+    if (!input) 
+    {
+        return -1;
+    }
 
     int freq[MAX_TREE_NODES] = {0};
     unsigned char buffer;
 
+    // Читаем файл и подсчитываем частоты символов
     while (fread(&buffer, 1, 1, input))
     {
         freq[buffer]++;
@@ -175,7 +180,9 @@ int compress_2(const char* input_file_name, const char* output_file_name)
     char code[MAX_TREE_NODES];
     build_codes(heap->nodes[0], code, 0, codes);
 
+    // Вернемся в начало входного файла
     fseek(input, 0, SEEK_SET);
+
     FILE* output = fopen(output_file_name, "wb");
     if (!output)
     {
@@ -183,14 +190,17 @@ int compress_2(const char* input_file_name, const char* output_file_name)
         return -1;
     }
 
+    // Записываем частоты символов
     fwrite(freq, sizeof(freq), 1, output);
 
     unsigned char out_buffer = 0;
     int bit_count = 0;
 
+    // Записываем закодированные данные
     while (fread(&buffer, 1, 1, input))
     {
         char* symbol_code = codes[buffer];
+
         for (int i = 0; symbol_code[i] != '\0'; i++)
         {
             out_buffer <<= 1;
@@ -200,7 +210,7 @@ int compress_2(const char* input_file_name, const char* output_file_name)
             }
             bit_count++;
 
-            if (bit_count == 8)
+            if (bit_count == 8) 
             {
                 fwrite(&out_buffer, 1, 1, output);
                 bit_count = 0;
@@ -215,6 +225,13 @@ int compress_2(const char* input_file_name, const char* output_file_name)
         fwrite(&out_buffer, 1, 1, output);
     }
 
+    fseek(input, 0, SEEK_END);
+    long sizeIN = ftell(input);
+    fseek(input, 0, SEEK_SET);
+
+    // Записываем реальный размер входного файла для декомпрессии
+    fwrite(&sizeIN, sizeof(sizeIN), 1, output);
+
     fseek(output, 0, SEEK_END);
     int sizeOUT = ftell(output);
     fseek(output, 0, SEEK_SET);
@@ -230,7 +247,10 @@ int compress_2(const char* input_file_name, const char* output_file_name)
 int decompress_2(const char* input_file_name, const char* output_file_name)
 {
     FILE* input = fopen(input_file_name, "rb");
-    if (!input) return -1;
+    if (!input)
+    {
+        return -1;
+    }
 
     int freq[MAX_TREE_NODES];
     fread(freq, sizeof(freq), 1, input);
@@ -238,7 +258,7 @@ int decompress_2(const char* input_file_name, const char* output_file_name)
     MinHeap* heap = create_min_heap();
     for (int i = 0; i < MAX_TREE_NODES; i++)
     {
-        if (freq[i] > 0) 
+        if (freq[i] > 0)
         {
             insert_min_heap(heap, create_node((unsigned char)i, freq[i]));
         }
@@ -250,30 +270,55 @@ int decompress_2(const char* input_file_name, const char* output_file_name)
     if (!output)
     {
         fclose(input);
+        free_tree(heap->nodes[0]);
+        free(heap);
         return -1;
     }
 
     Node* root = heap->nodes[0];
     Node* current = root;
-    unsigned char buffer;
 
-    while (fread(&buffer, 1, 1, input))
+    // Перемещаемся к концу файла, чтобы прочитать оригинальный размер
+    fseek(input, -sizeof(long), SEEK_END);
+    long original_size;
+    fread(&original_size, sizeof(long), 1, input);
+
+    // Вернемся к началу сжатых данных
+    long data_end = ftell(input) - sizeof(long);
+    fseek(input, sizeof(freq), SEEK_SET);
+
+    long written_bytes = 0;
+    unsigned char buffer;
+    int bit_count = 0;
+
+    while (ftell(input) < data_end || (bit_count > 0 && written_bytes < original_size))
     {
-        for (int i = 7; i >= 0; i--)
+        if (bit_count == 0 && fread(&buffer, 1, 1, input) == 1)
         {
-            if ((buffer >> i) & 1)
-            {
-                current = current->right;
-            }
-            else
-            {
-                current = current->left;
-            }
+            bit_count = 8;
+        }
+
+        if (bit_count > 0)
+        {
+            int bit = (buffer >> (bit_count - 1)) & 1;
+            bit_count--;
+
+            current = bit ? current->right : current->left;
 
             if (!current->left && !current->right)
             {
-                fwrite(&current->symbol, 1, 1, output);
+                if (written_bytes < original_size)
+                {
+                    fwrite(&current->symbol, 1, 1, output);
+                    written_bytes++;
+                }
                 current = root;
+
+                // Остановимся, если записали все байты
+                if (written_bytes == original_size)
+                {
+                    break;
+                }
             }
         }
     }
@@ -290,6 +335,10 @@ int decompress_2(const char* input_file_name, const char* output_file_name)
     return sizeOUT;
 }
 
+
+
+
+
 // Функция для записи токена в файл в компактном формате
 void write_token(FILE *file, LZ77Token token)
 {
diff --git a/sk1/main.c b/sk1/main.c
index d77bef4..fe33b9b 100644
--- a/sk1/main.c
+++ b/sk1/main.c
@@ -50,4 +50,4 @@ int main(int argc, char* argv[])
     }
 
     return 0;
-}
\ No newline at end of file
+}