usaa24/sk1/compressor.c

#include "compressor.h"
#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
#include <string.h>

#define MAX_DICT_SIZE 4096

int read_file(const char* filename, unsigned char** data, size_t* size) {
    FILE* file = fopen(filename, "rb");
    if (!file) return -1;

    fseek(file, 0, SEEK_END);
    *size = ftell(file);
    fseek(file, 0, SEEK_SET);

    *data = malloc(*size);
    if (!*data) {
        fclose(file);
        return -1;
    }

    fread(*data, 1, *size, file);
    fclose(file);
    return 0;
}
int write_file(const char* filename, const unsigned char* data, size_t size) {
    FILE* file = fopen(filename, "wb");
    if (!file) return -1;

    fwrite(data, 1, size, file);
    fclose(file);
    return 0;
}

//RLE
int compress_1(const char* input_file_name, const char* output_file_name) {
    size_t size;
    unsigned char* data;
    if (read_file(input_file_name, &data, &size) != 0) return -1;
    FILE* output = fopen(output_file_name, "w");
    if (!output) {
        free(data);
        return -1;
    }
    for (size_t i = 0; i < size;) {
        unsigned char byte = data[i];
        size_t count = 1;
        while (i + count < size && data[i + count] == byte && count < 255) {
            count++;
        }
        fprintf(output, "%c%d", byte, (int)count);

        i += count;
    }

    fclose(output);
    free(data);
    return 0;
}
unsigned char* decompress_1_buffer(const unsigned char* data, size_t size, size_t* out_size) {
    unsigned char* decompressed = malloc(size * 255);
    if (!decompressed) return NULL;
    size_t write_idx = 0;
    for (size_t i = 0; i < size;) {
        unsigned char byte = data[i++];
        size_t count = 0;
        while (i < size && isdigit(data[i])) {
            count = count * 10 + (data[i++] - '0');
        }
        for (size_t j = 0; j < count; j++) {
            if (write_idx >= size * 255) {
                size_t new_size = write_idx * 2;
                unsigned char* new_buffer = realloc(decompressed, new_size);
                if (!new_buffer) {
                    free(decompressed);
                    return NULL;
                }
                decompressed = new_buffer;
            }
            decompressed[write_idx++] = byte;
        }
    }
    *out_size = write_idx;
    return decompressed;
}

int decompress_1(const char* input_file_name, const char* output_file_name) {
    size_t size;
    unsigned char* data;
    if (read_file(input_file_name, &data, &size) != 0) return -1;

    size_t out_size;
    unsigned char* decompressed = decompress_1_buffer(data, size, &out_size);
    if (!decompressed) {
        free(data);
        return -1;
    }
    int result = write_file(output_file_name, decompressed, out_size);
    free(data);
    free(decompressed);
    return result == 0 ? (int)out_size : -1;
}

// LZ78
int compress_2(const char* input_file_name, const char* output_file_name) {
    unsigned char* data;
    size_t size;
    if (read_file(input_file_name, &data, &size) != 0) {
        return -1;
    }
    FILE* output = fopen(output_file_name, "w");
    if (!output) {
        free(data);
        return -1;
    }
    unsigned char* dictionary[MAX_DICT_SIZE] = { NULL };
    size_t dict_size = 0;
    size_t current_index = 0;
    while (current_index < size) {
        size_t match_index = 0;
        size_t match_length = 0;
        for (size_t i = 0; i < dict_size; i++) {
            size_t len = 0;
            while (data[current_index + len] != '\0' &&
                   dictionary[i][len] != '\0' &&
                   data[current_index + len] == dictionary[i][len]) {
                len++;
            }
            if (len > match_length) {
                match_length = len;
                match_index = i + 1;
            }
        }
        if (current_index + match_length < size) {
            fprintf(output, "%zu%c", match_index, data[current_index + match_length]);
        }
        if (dict_size < MAX_DICT_SIZE) {
            dictionary[dict_size] = malloc(match_length + 2);
            if (!dictionary[dict_size]) {
                fclose(output);
                free(data);
                for (size_t i = 0; i < dict_size; i++) {
                    free(dictionary[i]);
                }
                return -1;
            }
            if (match_length > 0) {
                memcpy(dictionary[dict_size], &data[current_index], match_length);
            }
            dictionary[dict_size][match_length] = data[current_index + match_length];
            dictionary[dict_size][match_length + 1] = '\0';
            dict_size++;
        }
        current_index += match_length + 1;
    }

    fclose(output);
    free(data);
    for (size_t i = 0; i < dict_size; i++) {
        free(dictionary[i]);
    }
    return 0;
}
int decompress_2(const char* input_file_name, const char* output_file_name) {
    unsigned char* data;
    size_t size;
    if (read_file(input_file_name, &data, &size) != 0) {
        return -1;
    }
    FILE* output = fopen(output_file_name, "w");
    if (!output) {
        free(data);
        return -1;
    }
    unsigned char* dictionary[MAX_DICT_SIZE] = { NULL };
    size_t dict_size = 0;
    size_t current_index = 0;
    while (current_index < size) {
        unsigned short index = 0;
        unsigned char ch = '\0';
        while (current_index < size && isdigit(data[current_index])) {
            index = index * 10 + (data[current_index] - '0');
            current_index++;
        }
        if (current_index < size) {
            ch = data[current_index++];
        }
        if (index == 0) {
            fputc(ch, output);
        } else {
            fwrite(dictionary[index - 1], 1, strlen((char*)dictionary[index - 1]), output);
            fputc(ch, output);
        }
        if (dict_size < MAX_DICT_SIZE) {
            size_t length = (index > 0) ? strlen((char*)dictionary[index - 1]) : 0;
            dictionary[dict_size] = malloc(length + 2);
            if (!dictionary[dict_size]) {
                fclose(output);
                free(data);
                for (size_t i = 0; i < dict_size; i++) {
                    free(dictionary[i]);
                }
                return -1;
            }
            if (index > 0) {
                memcpy(dictionary[dict_size], dictionary[index - 1], length);
            }
            dictionary[dict_size][length] = ch;
            dictionary[dict_size][length + 1] = '\0';
            dict_size++;
        }
    }
    fclose(output);
    free(data);
    for (size_t i = 0; i < dict_size; i++) {
        free(dictionary[i]);
    }
    return 0;
}