Class: MultiCompress::Dictionary

Inherits:

Object

• Object
• MultiCompress::Dictionary

Defined in:

ext/multi_compress/multi_compress.c

Class Method Summary

• .load(*args) ⇒ Object

Instance Method Summary

• #algo ⇒ Object
• #bytes ⇒ Object
• #initialize(*args) ⇒ Object constructor
• #save(path) ⇒ Object
• #size ⇒ Object
• #zstd_id ⇒ Object

Constructor Details

#initialize(*args) ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3298

static VALUE dict_initialize(int argc, VALUE *argv, VALUE self) {
    VALUE raw, opts;
    scan_one_required_keywords(argc, argv, &raw, &opts);
    StringValue(raw);

    mc_opts_t parsed_opts;
    mc_parse_opts(opts, &parsed_opts);
    reject_algorithm_keyword(&parsed_opts);

    dictionary_t *d;
    TypedData_Get_Struct(self, dictionary_t, &dictionary_type, d);

    VALUE algo_sym = parsed_opts.algo;
    d->algo = NIL_P(algo_sym) ? ALGO_ZSTD : sym_to_algo(algo_sym);

    if (d->algo == ALGO_LZ4)
        rb_raise(eUnsupportedError, "LZ4 does not support dictionaries");

    d->size = RSTRING_LEN(raw);
    d->data = ALLOC_N(uint8_t, d->size);
    memcpy(d->data, RSTRING_PTR(raw), d->size);

    return self;
}

Class Method Details

.load(*args) ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3412

static VALUE dict_load(int argc, VALUE *argv, VALUE self) {
    VALUE path, opts;
    scan_one_required_keywords(argc, argv, &path, &opts);
    StringValue(path);
    raise_if_path_has_null_byte(path);

    mc_opts_t parsed_opts;
    mc_parse_opts(opts, &parsed_opts);
    reject_algorithm_keyword(&parsed_opts);

    VALUE algo_sym = parsed_opts.algo;
    compress_algo_t algo = NIL_P(algo_sym) ? ALGO_ZSTD : sym_to_algo(algo_sym);

    if (algo == ALGO_LZ4)
        rb_raise(eUnsupportedError, "LZ4 does not support dictionaries");

    const char *cpath = RSTRING_PTR(path);
    FILE *f = fopen(cpath, "rb");
    if (!f)
        rb_sys_fail(cpath);

    fseek(f, 0, SEEK_END);
    long file_size = ftell(f);
    fseek(f, 0, SEEK_SET);

    if (file_size <= 0) {
        fclose(f);
        rb_raise(eDataError, "dictionary file is empty: %s", cpath);
    }
    if ((unsigned long long)file_size > DICT_FILE_MAX_SIZE) {
        fclose(f);
        rb_raise(eDataError, "dictionary file too large (%ld bytes, max=%d)", file_size,
                 (int)DICT_FILE_MAX_SIZE);
    }

    uint8_t *buf = ALLOC_N(uint8_t, file_size);
    size_t read_bytes = fread(buf, 1, file_size, f);
    fclose(f);

    if ((long)read_bytes != file_size) {
        xfree(buf);
        rb_raise(eDataError, "failed to read dictionary: %s", cpath);
    }

    VALUE dict_obj = dict_alloc(cDictionary);
    dictionary_t *d;
    TypedData_Get_Struct(dict_obj, dictionary_t, &dictionary_type, d);
    d->algo = algo;
    d->data = buf;
    d->size = (size_t)file_size;
    return dict_obj;
}

Instance Method Details

#algo ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3482

static VALUE dict_algo(VALUE self) {
    dictionary_t *d;
    TypedData_Get_Struct(self, dictionary_t, &dictionary_type, d);
    return algo_to_sym(d->algo);
}

#bytes ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3488

static VALUE dict_bytes(VALUE self) {
    dictionary_t *d;
    TypedData_Get_Struct(self, dictionary_t, &dictionary_type, d);
    return rb_binary_str_new((const char *)d->data, (long)d->size);
}

#save(path) ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3465

static VALUE dict_save(VALUE self, VALUE path) {
    dictionary_t *d;
    TypedData_Get_Struct(self, dictionary_t, &dictionary_type, d);

    const char *cpath = StringValueCStr(path);
    FILE *f = fopen(cpath, "wb");
    if (!f)
        rb_sys_fail(cpath);

    size_t written = fwrite(d->data, 1, d->size, f);
    fclose(f);

    if (written != d->size)
        rb_raise(eError, "failed to write dictionary to %s", cpath);
    return path;
}

#size ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3502

static VALUE dict_size(VALUE self) {
    dictionary_t *d;
    TypedData_Get_Struct(self, dictionary_t, &dictionary_type, d);
    return SIZET2NUM(d->size);
}

#zstd_id ⇒ Object

# File 'ext/multi_compress/multi_compress.c', line 3494

static VALUE dict_zstd_id(VALUE self) {
    dictionary_t *d;
    TypedData_Get_Struct(self, dictionary_t, &dictionary_type, d);
    if (d->algo != ALGO_ZSTD)
        rb_raise(eUnsupportedError, "zstd_id is available only for zstd dictionaries");
    return UINT2NUM(ZSTD_getDictID_fromDict(d->data, d->size));
}