利用cre2进行分组模式匹配的实例

左手的ㄟ右手 2022-05-15 09:36 189阅读 0赞

# code #

#include <stdio.h>
    #include <stdlib.h>
    #include <string.h>
    #include <cre2.h>
    
    
    int init_ptn(const char *str_ptn, cre2_regexp_t **pPtn)
    {
        // 初始化cre2的正则表达式
        cre2_options_t* opt;
        opt = cre2_opt_new();
        //cre2_opt_set_posix_syntax(opt, 1);
        cre2_opt_set_perl_classes(opt, 1);
        cre2_opt_set_one_line(opt, 1);
        // cre2_opt_set_literal(opt, 1);
        cre2_opt_set_longest_match(opt, 1);
        *pPtn = cre2_new(str_ptn, strlen(str_ptn), opt);
        if (cre2_error_code(*pPtn))
        {
            fprintf(stderr, "cre2_new error\n");
            return -1;
        }
        return 0;
    }
    
    
    
    int main(int argv, char* argc[])
    {
        cre2_regexp_t * rex = NULL;
        const char *    pattern = "^(?P<SRC>.+?)(?P<CMD>修改为|修改成|改为|改成|换成|变成|替换成|替成|替换为|换为|更替为)(?P<DEST>.+)$";
        const char *    text    = "你好修改为不好";
        init_ptn(pattern, &rex);
    
        cre2_string_t input = {
          .data   = text,
          .length = strlen(text)
        };
        int nmatch = cre2_num_capturing_groups(rex) + 1;;
        printf("nmatch is %d\n", nmatch);
        cre2_string_t match[nmatch];
        memset(match, '\0', nmatch * sizeof(cre2_string_t));
        int e, src_index, cmd_index, dest_index;
        e = cre2_match(rex, text, strlen(text), 0, strlen(text), CRE2_UNANCHORED, match, nmatch);
        printf("e is %d\n", e);
        if (e != 1)
        {
            printf("match error\n");
        }
        src_index = cre2_find_named_capturing_groups(rex, "SRC");
        printf("src_index is %d\n", src_index);
        printf("src is %.*s\n", match[src_index].length, match[src_index].data);
    
        cmd_index = cre2_find_named_capturing_groups(rex, "CMD");
        printf("cmd_index is %d\n", cmd_index);
        printf("cmd is %.*s\n", match[cmd_index].length, match[cmd_index].data);
    
        dest_index = cre2_find_named_capturing_groups(rex, "DEST");
        printf("dest_index is %d\n", dest_index);
        printf("dest is %.*s\n", match[dest_index].length, match[dest_index].data);
        cre2_delete(rex);
        return 0;
    }

# res #

nmatch is 4
    e is 1
    src_index is 1
    src is 你好
    cmd_index is 2
    cmd is 修改为
    dest_index is 3
    dest is 不好