void
test_remove_tokenized_delimiter(gconstpointer data)
{
grn_obj *string;
grn_obj *normalizer = NULL;
const gchar *expected;
const gchar *input;
const gchar *normalized;
unsigned int length_in_bytes;
int flags = GRN_STRING_REMOVE_TOKENIZED_DELIMITER;
GRN_CTX_SET_ENCODING(&context, GRN_ENC_UTF8);
input = gcut_data_get_string(data, "input");
flags |= gcut_data_get_int(data, "flags");
if (flags & GRN_OBJ_KEY_NORMALIZE) {
normalizer = GRN_NORMALIZER_AUTO;
}
string = grn_string_open(&context, input, strlen(input), normalizer, flags);
grn_string_get_normalized(&context, string,
&normalized, &length_in_bytes, NULL);
normalized = cut_take_strndup(normalized, length_in_bytes);
grn_obj_unlink(&context, string);
expected = gcut_data_get_string(data, "expected");
cut_assert_equal_string(expected, normalized);
}
void
test_normalize(gconstpointer data)
{
const gchar *utf8_expected, *encoded_expected;
const gchar *utf8_input, *encoded_input;
grn_str *string;
const gchar *normalized_text;
guint normalized_text_len;
int flags;
grn_encoding encoding;
encoding = gcut_data_get_int(data, "encoding");
GRN_CTX_SET_ENCODING(&context, encoding);
flags = GRN_STR_NORMALIZE | GRN_STR_WITH_CHECKS | GRN_STR_WITH_CTYPES;
utf8_input = gcut_data_get_string(data, "input");
encoded_input = convert_encoding(utf8_input, encoding);
string = grn_str_open(&context, encoded_input, strlen(encoded_input), flags);
normalized_text = cut_take_strndup(string->norm, string->norm_blen);
normalized_text_len = string->norm_blen;
grn_test_assert(grn_str_close(&context, string));
utf8_expected = gcut_data_get_string(data, "expected");
encoded_expected = convert_encoding(utf8_expected, encoding);
cut_assert_equal_string(encoded_expected, normalized_text);
cut_assert_equal_int(strlen(encoded_expected), normalized_text_len);
}
static gboolean
run_test(const gchar **test_case_names, const grn_test_data *data)
{
const gchar *type_name, *table_name;
gchar *path;
grn_test_assert(grn_ctx_init(context, GRN_CTX_USE_QL));
GRN_CTX_SET_ENCODING(context, GRN_ENC_UTF8);
type_name = "name";
type = grn_type_create(context, type_name, strlen(type_name),
GRN_OBJ_KEY_UINT, sizeof(grn_id));
path = g_build_filename(base_dir, "table", NULL);
g_setenv(GRN_TEST_ENV_TABLE_PATH, path, TRUE);
table_name = cut_take_printf("%s: performance-read-write", data->type_name);
g_setenv(GRN_TEST_ENV_TABLE_TYPE, data->type_name, TRUE);
table = grn_table_create(context,
table_name, strlen(table_name),
path, GRN_OBJ_PERSISTENT | data->flags,
type, NULL);
g_free(path);
cut_assert_not_null(table);
return run(test_case_names, data);
}
void
test_normalize_broken(gconstpointer data)
{
grn_obj *string;
const gchar *input, *encoded_input;
const gchar *normalized_text;
grn_encoding input_encoding, context_encoding;
gint input_length;
guint normalized_text_length, normalized_text_n_characters;
int flags = GRN_STRING_WITH_CHECKS | GRN_STRING_WITH_TYPES;
context_encoding = gcut_data_get_int(data, "context-encoding");
GRN_CTX_SET_ENCODING(&context, context_encoding);
input = gcut_data_get_string(data, "input");
input_encoding = gcut_data_get_int(data, "input-encoding");
input_length = gcut_data_get_int(data, "input-length");
encoded_input = convert_encoding(input, input_encoding);
if (input_length < 0) {
input_length = strlen(encoded_input);
}
string = grn_string_open(&context, encoded_input, input_length,
GRN_NORMALIZER_AUTO, flags);
grn_string_get_normalized(&context, string,
&normalized_text,
&normalized_text_length,
&normalized_text_n_characters);
normalized_text = cut_take_strndup(normalized_text, normalized_text_length);
grn_obj_unlink(&context, string);
cut_assert_equal_string("", normalized_text);
cut_assert_equal_int(0, normalized_text_length);
cut_assert_equal_int(0, normalized_text_n_characters);
}
grn_obj *FieldNormalizer::normalize(const char *string,
unsigned int string_length) {
MRN_DBUG_ENTER_METHOD();
grn_obj *normalizer = find_grn_normalizer();
int flags = 0;
grn_encoding original_encoding = GRN_CTX_GET_ENCODING(ctx_);
encoding::set(ctx_, field_->charset());
grn_obj *grn_string = grn_string_open(ctx_, string, string_length,
normalizer, flags);
GRN_CTX_SET_ENCODING(ctx_, original_encoding);
DBUG_RETURN(grn_string);
}
/*
* call-seq:
* context.encoding=(encoding)
*
* コンテキストが使うエンコーディングを設定する。エンコーディ
* ングの指定のしかたはGroonga::Encodingを参照。
*/
static VALUE
rb_grn_context_set_encoding (VALUE self, VALUE rb_encoding)
{
grn_ctx *context;
grn_encoding encoding;
context = SELF(self);
encoding = RVAL2GRNENCODING(rb_encoding, NULL);
GRN_CTX_SET_ENCODING(context, encoding);
return rb_encoding;
}
void
cut_setup(void)
{
gchar *table_path, *vgram_path;
const gchar *type_name, *table_name;
cut_set_fixture_data_dir(grn_test_get_base_dir(),
"fixtures",
"inverted-index",
NULL);
logger = setup_grn_logger();
expected_messages = NULL;
record_ids = NULL;
remove_tmp_directory();
g_mkdir_with_parents(tmp_directory, 0700);
path = g_build_filename(tmp_directory, "inverted-index", NULL);
context = g_new0(grn_ctx, 1);
grn_test_assert(grn_ctx_init(context, GRN_CTX_USE_QL));
GRN_CTX_SET_ENCODING(context, GRN_ENC_UTF8);
db = grn_db_create(context, NULL, NULL);
grn_ctx_use(context, db);
type_name = "name";
type = grn_type_create(context, type_name, strlen(type_name),
GRN_OBJ_KEY_VAR_SIZE, TYPE_SIZE);
table_name = "lexicon";
table_path = g_build_filename(tmp_directory, "lexicon-table", NULL);
lexicon = grn_table_create(context,
table_name, strlen(table_name),
table_path,
GRN_OBJ_PERSISTENT | GRN_OBJ_TABLE_PAT_KEY,
type, NULL);
grn_obj_set_info(context, lexicon, GRN_INFO_DEFAULT_TOKENIZER,
grn_ctx_at(context, GRN_DB_BIGRAM));
g_free(table_path);
vgram_path = g_build_filename(tmp_directory, "vgram", NULL);
/*
vgram = grn_vgram_create(vgram_path);
*/
g_free(vgram_path);
inverted_index = NULL;
}
static grn_obj *
grn_table_factory_make(grn_table_factory *factory)
{
grn_obj *value_type = grn_ctx_get(factory->context, VALUE_TYPE_NAME, strlen(VALUE_TYPE_NAME));
if (!value_type) {
value_type = grn_type_create(factory->context, VALUE_TYPE_NAME, strlen(VALUE_TYPE_NAME),
0, factory->value_size);
}
GRN_CTX_SET_ENCODING(factory->context, factory->encoding);
return grn_table_create(factory->context,
factory->name, factory->name_size,
factory->path, factory->flags,
factory->key_type, value_type);
}
static grn_obj *
open_snip(void)
{
if (snip) {
grn_obj_close(&context, (grn_obj *)snip);
}
GRN_CTX_SET_ENCODING(&context, default_encoding);
snip = grn_snip_open(&context, default_flags,
default_width, default_max_results,
default_open_tag, default_open_tag_len,
default_close_tag, default_close_tag_len,
default_mapping);
return snip;
}
static void
bench_normal_temporary(gpointer user_data)
{
BenchmarkData *data = user_data;
grn_obj *table;
grn_obj *value_type = grn_ctx_get(data->context, VALUE_TYPE_NAME, strlen(VALUE_TYPE_NAME));
if (!value_type) {
value_type = grn_type_create(data->context, VALUE_TYPE_NAME, strlen(VALUE_TYPE_NAME),
0, data->value_size);
}
GRN_CTX_SET_ENCODING(data->context, data->encoding);
table = grn_table_create(data->context,
data->name, data->name_size,
NULL, data->flags & ~GRN_OBJ_PERSISTENT,
data->key_type, value_type);
grn_obj_close(data->context, table);
}
/*
* call-seq:
* Groonga::Context.new(options=nil)
*
* コンテキストを作成する。_options_に指定可能な値は以下の通
* り。
*
* [+:encoding+]
* エンコーディングを指定する。エンコーディングの指定方法
* はGroonga::Encodingを参照。
*/
static VALUE
rb_grn_context_initialize (int argc, VALUE *argv, VALUE self)
{
RbGrnContext *rb_grn_context;
grn_ctx *context;
int flags = 0;
VALUE options, default_options;
VALUE rb_encoding;
rb_scan_args(argc, argv, "01", &options);
default_options = rb_grn_context_s_get_default_options(rb_obj_class(self));
if (NIL_P(default_options))
default_options = rb_hash_new();
if (NIL_P(options))
options = rb_hash_new();
options = rb_funcall(default_options, rb_intern("merge"), 1, options);
rb_grn_scan_options(options,
"encoding", &rb_encoding,
NULL);
rb_grn_context = ALLOC(RbGrnContext);
DATA_PTR(self) = rb_grn_context;
rb_grn_context->self = self;
context = rb_grn_context->context = grn_ctx_open(flags);
rb_grn_context_check(context, self);
GRN_CTX_USER_DATA(context)->ptr = rb_grn_context;
if (!NIL_P(rb_encoding)) {
grn_encoding encoding;
encoding = RVAL2GRNENCODING(rb_encoding, NULL);
GRN_CTX_SET_ENCODING(context, encoding);
}
debug("context new: %p\n", context);
return Qnil;
}
void
test_charlen_broken(gconstpointer data)
{
const gchar *input, *encoded_input, *encoded_input_end;
grn_encoding encoding;
gint input_length;
encoding = gcut_data_get_int(data, "encoding");
GRN_CTX_SET_ENCODING(&context, encoding);
input = gcut_data_get_string(data, "input");
input_length = gcut_data_get_int(data, "input-length");
encoded_input = convert_encoding(input, encoding);
if (input_length < 0) {
input_length = strlen(encoded_input);
}
encoded_input_end = encoded_input + input_length;
cut_assert_equal_uint(0, grn_charlen(&context,
encoded_input,
encoded_input_end));
}
void
test_normalize_without_database(void)
{
grn_obj *string;
const char *input = "Groonga";
int flags = 0;
grn_obj_close(&context, database);
database = NULL;
GRN_CTX_SET_ENCODING(&context, GRN_ENC_UTF8);
string = grn_string_open(&context,
input,
strlen(input),
GRN_NORMALIZER_AUTO,
flags);
cut_assert_null(string);
grn_test_assert_error(GRN_INVALID_ARGUMENT,
"[string][open] "
"NormalizerAuto normalizer isn't available",
&context);
}
grn_obj *FieldNormalizer::normalize(const char *string,
unsigned int string_length) {
MRN_DBUG_ENTER_METHOD();
grn_obj normalizer;
GRN_TEXT_INIT(&normalizer, 0);
find_grn_normalizer(&normalizer);
int flags = 0;
grn_encoding original_encoding = GRN_CTX_GET_ENCODING(ctx_);
encoding::set_raw(ctx_, field_->charset());
grn_obj *grn_string;
if (GRN_TEXT_VALUE(&normalizer)[GRN_TEXT_LEN(&normalizer) - 1] == ')') {
if (!lexicon_) {
lexicon_ = grn_table_create(ctx_,
NULL, 0,
NULL,
GRN_OBJ_TABLE_PAT_KEY,
grn_ctx_at(ctx_, GRN_DB_SHORT_TEXT),
NULL);
}
grn_obj_set_info(ctx_, lexicon_, GRN_INFO_NORMALIZER, &normalizer);
grn_string = grn_string_open(ctx_,
string,
string_length,
lexicon_,
flags);
} else {
grn_string = grn_string_open(ctx_,
string,
string_length,
grn_ctx_get(ctx_,
GRN_TEXT_VALUE(&normalizer),
GRN_TEXT_LEN(&normalizer)),
flags);
}
GRN_OBJ_FIN(ctx_, &normalizer);
GRN_CTX_SET_ENCODING(ctx_, original_encoding);
DBUG_RETURN(grn_string);
}
void
test_normalize_broken(gconstpointer data)
{
grn_str *string;
const gchar *input, *encoded_input;
grn_encoding input_encoding, context_encoding;
gint input_length;
int flags = GRN_STR_NORMALIZE | GRN_STR_WITH_CHECKS | GRN_STR_WITH_CTYPES;
context_encoding = gcut_data_get_int(data, "context-encoding");
GRN_CTX_SET_ENCODING(&context, context_encoding);
input = gcut_data_get_string(data, "input");
input_encoding = gcut_data_get_int(data, "input-encoding");
input_length = gcut_data_get_int(data, "input-length");
encoded_input = convert_encoding(input, input_encoding);
if (input_length < 0) {
input_length = strlen(encoded_input);
}
string = grn_str_open(&context, encoded_input, input_length, flags);
cut_assert_equal_string("", string->norm);
cut_assert_equal_int(0, string->norm_blen);
grn_test_assert(grn_str_close(&context, string));
}
void
test_normalize(gconstpointer data)
{
const gchar *utf8_expected, *encoded_expected;
const gchar *utf8_input, *encoded_input;
grn_obj *string;
const gchar *normalized_text;
guint normalized_text_length;
guint normalized_text_n_characters;
int flags;
grn_encoding encoding;
encoding = gcut_data_get_int(data, "encoding");
GRN_CTX_SET_ENCODING(&context, encoding);
flags = GRN_STRING_WITH_CHECKS | GRN_STRING_WITH_TYPES;
utf8_input = gcut_data_get_string(data, "input");
encoded_input = convert_encoding(utf8_input, encoding);
string = grn_string_open(&context,
encoded_input,
strlen(encoded_input),
GRN_NORMALIZER_AUTO,
flags);
grn_string_get_normalized(&context, string,
&normalized_text,
&normalized_text_length,
&normalized_text_n_characters);
normalized_text = cut_take_strndup(normalized_text, normalized_text_length);
grn_obj_unlink(&context, string);
utf8_expected = gcut_data_get_string(data, "expected");
encoded_expected = convert_encoding(utf8_expected, encoding);
cut_assert_equal_string(encoded_expected, normalized_text);
cut_assert_equal_uint(strlen(encoded_expected), normalized_text_length);
cut_assert_equal_uint(g_utf8_strlen(utf8_expected, -1),
normalized_text_n_characters);
}
int set(grn_ctx *ctx, const CHARSET_INFO *charset) {
MRN_DBUG_ENTER_FUNCTION();
if (!charset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_NONE);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_utf8->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_UTF8);
DBUG_RETURN(0);
}
if (mrn_charset_utf8mb4 && charset->cset == mrn_charset_utf8mb4->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_UTF8);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_cp932->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_SJIS);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_eucjpms->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_EUC_JP);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_latin1_1->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_LATIN1);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_latin1_2->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_LATIN1);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_koi8r->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_KOI8R);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_binary->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_NONE);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_ascii->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_UTF8);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_sjis->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_SJIS);
DBUG_RETURN(0);
}
if (charset->cset == mrn_charset_ujis->cset)
{
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_EUC_JP);
DBUG_RETURN(0);
}
GRN_CTX_SET_ENCODING(ctx, GRN_ENC_NONE);
my_printf_error(ER_MRN_CHARSET_NOT_SUPPORT_NUM,
ER_MRN_CHARSET_NOT_SUPPORT_STR,
MYF(0), charset->name, charset->csname);
DBUG_RETURN(ER_MRN_CHARSET_NOT_SUPPORT_NUM);
}
