1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
|
/*
Copyright(C) 2020 Sutou Kouhei <kou@clear-code.com>
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <cstring>
#include "grn.h"
#include "grn_str.h"
#include "grn_token_column.h"
#include "grn_token_cursor.h"
#ifdef GRN_WITH_APACHE_ARROW
# include "grn_arrow.hpp"
# include <arrow/util/thread_pool.h>
# include <mutex>
#endif
namespace grn {
namespace token_column {
static uint32_t parallel_chunk_size = 1024;
static uint32_t parallel_table_size_threshold = parallel_chunk_size * 10;
class Builder
{
public:
Builder(grn_ctx *ctx,
grn_obj *column)
: ctx_(ctx),
column_(column),
with_weight_(grn_column_get_flags(ctx_, column_) &
GRN_OBJ_WITH_WEIGHT),
table_(grn_ctx_at(ctx, column->header.domain)),
lexicon_(grn_ctx_at(ctx, DB_OBJ(column)->range)),
source_(NULL) {
grn_id *source_ids = static_cast<grn_id *>(DB_OBJ(column)->source);
source_ = grn_ctx_at(ctx, source_ids[0]);
}
~Builder() {
grn_obj_unref(ctx_, source_);
grn_obj_unref(ctx_, lexicon_);
grn_obj_unref(ctx_, table_);
}
void
build() {
grn_obj_set_visibility(ctx_, column_, false);
if (use_parallel()) {
build_parallel();
} else {
build_sequential();
}
grn_obj_set_visibility(ctx_, column_, true);
}
private:
bool
use_parallel() {
#ifdef GRN_WITH_APACHE_ARROW
return grn_table_size(ctx_, table_) >= parallel_table_size_threshold;
#else
return false;
#endif
}
void
build_parallel() {
#ifdef GRN_WITH_APACHE_ARROW
std::mutex mutex;
auto pool = ::arrow::internal::GetCpuThreadPool();
uint32_t token_cursor_flags = GRN_TOKEN_CURSOR_PARALLEL;
grn_obj *db = grn_ctx_db(ctx_);
uint32_t chunk_size = parallel_chunk_size;
auto build_chunk = [&](std::vector<grn_id> ids) {
grn_ctx ctx;
grn_ctx_init(&ctx, 0);
grn_ctx_use(&ctx, db);
grn_obj tokens;
grn_obj value;
GRN_RECORD_INIT(&tokens, GRN_OBJ_VECTOR, DB_OBJ(lexicon_)->id);
if (with_weight_) {
tokens.header.flags |= GRN_OBJ_WITH_WEIGHT;
}
GRN_VOID_INIT(&value);
for (auto id : ids) {
GRN_BULK_REWIND(&tokens);
GRN_BULK_REWIND(&value);
grn_obj_get_value(&ctx, source_, id, &value);
if (GRN_TEXT_LEN(&value) > 0) {
auto token_cursor = grn_token_cursor_open(&ctx,
lexicon_,
GRN_BULK_HEAD(&value),
GRN_BULK_VSIZE(&value),
GRN_TOKEN_ADD,
token_cursor_flags);
if (token_cursor) {
grn_token_cursor_set_source_column(&ctx, token_cursor, source_);
grn_token_cursor_set_source_id(&ctx, token_cursor, id);
grn_token_cursor_set_index_column(&ctx, token_cursor, column_);
while (token_cursor->status == GRN_TOKEN_CURSOR_DOING) {
grn_id token_id = grn_token_cursor_next(&ctx, token_cursor);
if (token_id == GRN_ID_NIL) {
break;
}
if (with_weight_) {
grn_token *token =
grn_token_cursor_get_token(&ctx, token_cursor);
float weight = grn_token_get_weight(&ctx, token);
grn_uvector_add_element_record(&ctx,
&tokens,
token_id,
weight);
} else {
GRN_RECORD_PUT(&ctx, &tokens, token_id);
}
}
grn_token_cursor_close(&ctx, token_cursor);
}
}
grn_obj_set_value(&ctx, column_, id, &tokens, GRN_OBJ_SET);
if (ctx.rc != GRN_SUCCESS) {
break;
}
}
GRN_OBJ_FIN(&ctx, &value);
GRN_OBJ_FIN(&ctx, &tokens);
if (ctx.rc != GRN_SUCCESS) {
std::lock_guard<std::mutex> lock(mutex);
if (ctx_->rc == GRN_SUCCESS) {
ctx_->rc = ctx.rc;
ctx_->errlvl = ctx.errlvl;
ctx_->errfile = ctx.errfile;
ctx_->errline = ctx.errline;
ctx_->errfunc = ctx.errfunc;
grn_strcpy(ctx_->errbuf, GRN_CTX_MSGSIZE, ctx.errbuf);
}
}
grn_ctx_fin(&ctx);
return ::arrow::Status::OK();
};
std::vector<grn_id> ids;
std::vector<::arrow::Future<>> futures;
GRN_TABLE_EACH_BEGIN_FLAGS(ctx_, table_, cursor, id, GRN_CURSOR_BY_ID) {
ids.push_back(id);
if (ids.size() == chunk_size) {
auto future = pool->Submit(build_chunk, ids);
if (!grnarrow::check(ctx_,
future,
"[token-column][build][parallel] "
"failed to submit a job")) {
break;
}
futures.push_back(*future);
ids.clear();
}
} GRN_TABLE_EACH_END(ctx_, cursor);
if (ctx_->rc == GRN_SUCCESS && !ids.empty()) {
auto future = pool->Submit(build_chunk, ids);
if (grnarrow::check(ctx_,
future,
"[token-column][build][parallel] "
"failed to submit a job")) {
futures.push_back(*future);
}
}
auto status = ::arrow::Status::OK();
for (auto& future : futures) {
status &= future.status();
}
grnarrow::check(ctx_,
status,
"[token-column][build][parallel] "
"failed to complete a job");
#endif
}
void
build_sequential() {
grn_obj tokens;
grn_obj value;
GRN_RECORD_INIT(&tokens, GRN_OBJ_VECTOR, DB_OBJ(lexicon_)->id);
if (with_weight_) {
tokens.header.flags |= GRN_OBJ_WITH_WEIGHT;
}
GRN_VOID_INIT(&value);
const unsigned int token_flags = 0;
GRN_TABLE_EACH_BEGIN_FLAGS(ctx_, table_, cursor, id, GRN_CURSOR_BY_ID) {
GRN_BULK_REWIND(&tokens);
GRN_BULK_REWIND(&value);
grn_obj_get_value(ctx_, source_, id, &value);
if (GRN_BULK_VSIZE(&value) > 0) {
grn_token_cursor *token_cursor =
grn_token_cursor_open(ctx_,
lexicon_,
GRN_BULK_HEAD(&value),
GRN_BULK_VSIZE(&value),
GRN_TOKEN_ADD,
token_flags);
if (token_cursor) {
grn_token_cursor_set_source_column(ctx_, token_cursor, source_);
grn_token_cursor_set_source_id(ctx_, token_cursor, id);
grn_token_cursor_set_index_column(ctx_, token_cursor, column_);
while (token_cursor->status == GRN_TOKEN_CURSOR_DOING) {
grn_id token_id = grn_token_cursor_next(ctx_, token_cursor);
if (token_id == GRN_ID_NIL) {
break;
}
if (with_weight_) {
grn_token *token =
grn_token_cursor_get_token(ctx_, token_cursor);
float weight = grn_token_get_weight(ctx_, token);
grn_uvector_add_element_record(ctx_,
&tokens,
token_id,
weight);
} else {
GRN_RECORD_PUT(ctx_, &tokens, token_id);
}
}
grn_token_cursor_close(ctx_, token_cursor);
}
}
grn_obj_set_value(ctx_, column_, id, &tokens, GRN_OBJ_SET);
if (ctx_->rc != GRN_SUCCESS) {
break;
}
} GRN_TABLE_EACH_END(ctx_, cursor);
GRN_OBJ_FIN(ctx_, &value);
GRN_OBJ_FIN(ctx_, &tokens);
}
grn_ctx *ctx_;
grn_obj *column_;
bool with_weight_;
grn_obj *table_;
grn_obj *lexicon_;
grn_obj *source_;
};
}
}
extern "C" {
void
grn_token_column_init_from_env(void)
{
{
char grn_token_column_parallel_chunk_size_env[GRN_ENV_BUFFER_SIZE];
grn_getenv("GRN_TOKEN_COLUMN_PARALLEL_CHUNK_SIZE",
grn_token_column_parallel_chunk_size_env,
GRN_ENV_BUFFER_SIZE);
if (grn_token_column_parallel_chunk_size_env[0]) {
size_t env_len = strlen(grn_token_column_parallel_chunk_size_env);
uint32_t chunk_size =
grn_atoui(grn_token_column_parallel_chunk_size_env,
grn_token_column_parallel_chunk_size_env + env_len,
NULL);
if (chunk_size > 0) {
grn::token_column::parallel_chunk_size = chunk_size;
}
}
}
{
char grn_token_column_parallel_table_size_threshold_env[GRN_ENV_BUFFER_SIZE];
grn_getenv("GRN_TOKEN_COLUMN_PARALLEL_TABLE_SIZE_THRESHOLD",
grn_token_column_parallel_table_size_threshold_env,
GRN_ENV_BUFFER_SIZE);
if (grn_token_column_parallel_table_size_threshold_env[0]) {
size_t env_len =
strlen(grn_token_column_parallel_table_size_threshold_env);
uint32_t threshold =
grn_atoui(grn_token_column_parallel_table_size_threshold_env,
grn_token_column_parallel_table_size_threshold_env + env_len,
NULL);
if (threshold > 0) {
grn::token_column::parallel_table_size_threshold = threshold;
}
}
}
}
void
grn_token_column_update(grn_ctx *ctx,
grn_obj *column,
grn_id id,
int section,
grn_obj *old_value,
grn_obj *new_value)
{
grn_obj *lexicon = grn_ctx_at(ctx, DB_OBJ(column)->range);
grn_obj tokens;
GRN_RECORD_INIT(&tokens, GRN_OBJ_VECTOR, DB_OBJ(lexicon)->id);
if (GRN_TEXT_LEN(new_value) > 0) {
unsigned int token_flags = 0;
grn_token_cursor *token_cursor =
grn_token_cursor_open(ctx,
lexicon,
GRN_TEXT_VALUE(new_value),
GRN_TEXT_LEN(new_value),
GRN_TOKEN_ADD,
token_flags);
if (token_cursor) {
while (token_cursor->status == GRN_TOKEN_CURSOR_DOING) {
grn_id token_id = grn_token_cursor_next(ctx, token_cursor);
if (token_id == GRN_ID_NIL) {
break;
}
GRN_RECORD_PUT(ctx, &tokens, token_id);
}
grn_token_cursor_close(ctx, token_cursor);
}
}
grn_obj_set_value(ctx, column, id, &tokens, GRN_OBJ_SET);
GRN_OBJ_FIN(ctx, &tokens);
grn_obj_unref(ctx, lexicon);
}
void
grn_token_column_build(grn_ctx *ctx, grn_obj *column)
{
grn::token_column::Builder builder(ctx, column);
builder.build();
}
}
|
