1 /* query.cc - Support for searching a notmuch database
3 * Copyright © 2009 Carl Worth
5 * This program is free software: you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation, either version 3 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program. If not, see http://www.gnu.org/licenses/ .
18 * Author: Carl Worth <cworth@cworth.org>
21 #include "notmuch-private.h"
22 #include "database-private.h"
24 #include <glib.h> /* GHashTable, GPtrArray */
26 struct _notmuch_query {
27 notmuch_database_t *notmuch;
28 const char *query_string;
30 notmuch_string_list_t *exclude_terms;
31 notmuch_exclude_t omit_excluded;
34 typedef struct _notmuch_mset_messages {
35 notmuch_messages_t base;
36 notmuch_database_t *notmuch;
37 Xapian::MSetIterator iterator;
38 Xapian::MSetIterator iterator_end;
39 } notmuch_mset_messages_t;
41 struct _notmuch_doc_id_set {
42 unsigned char *bitmap;
46 #define DOCIDSET_WORD(bit) ((bit) / CHAR_BIT)
47 #define DOCIDSET_BIT(bit) ((bit) % CHAR_BIT)
49 struct visible _notmuch_threads {
50 notmuch_query_t *query;
52 /* The ordered list of doc ids matched by the query. */
54 /* Our iterator's current position in doc_ids. */
55 unsigned int doc_id_pos;
56 /* The set of matched docid's that have not been assigned to a
57 * thread. Initially, this contains every docid in doc_ids. */
58 notmuch_doc_id_set_t match_set;
61 /* We need this in the message functions so forward declare. */
63 _notmuch_doc_id_set_init (void *ctx,
64 notmuch_doc_id_set_t *doc_ids,
70 char *env = getenv ("NOTMUCH_DEBUG_QUERY");
71 return (env && strcmp (env, "") != 0);
75 notmuch_query_create (notmuch_database_t *notmuch,
76 const char *query_string)
78 notmuch_query_t *query;
81 fprintf (stderr, "Query string is:\n%s\n", query_string);
83 query = talloc (notmuch, notmuch_query_t);
84 if (unlikely (query == NULL))
87 query->notmuch = notmuch;
89 query->query_string = talloc_strdup (query, query_string);
91 query->sort = NOTMUCH_SORT_NEWEST_FIRST;
93 query->exclude_terms = _notmuch_string_list_create (query);
95 query->omit_excluded = NOTMUCH_EXCLUDE_TRUE;
101 notmuch_query_get_query_string (notmuch_query_t *query)
103 return query->query_string;
107 notmuch_query_set_omit_excluded (notmuch_query_t *query,
108 notmuch_exclude_t omit_excluded)
110 query->omit_excluded = omit_excluded;
114 notmuch_query_set_sort (notmuch_query_t *query, notmuch_sort_t sort)
120 notmuch_query_get_sort (notmuch_query_t *query)
126 notmuch_query_add_tag_exclude (notmuch_query_t *query, const char *tag)
128 char *term = talloc_asprintf (query, "%s%s", _find_prefix ("tag"), tag);
129 _notmuch_string_list_append (query->exclude_terms, term);
132 /* We end up having to call the destructors explicitly because we had
133 * to use "placement new" in order to initialize C++ objects within a
134 * block that we allocated with talloc. So C++ is making talloc
135 * slightly less simple to use, (we wouldn't need
136 * talloc_set_destructor at all otherwise).
139 _notmuch_messages_destructor (notmuch_mset_messages_t *messages)
141 messages->iterator.~MSetIterator ();
142 messages->iterator_end.~MSetIterator ();
147 /* Return a query that matches messages with the excluded tags
148 * registered with query. Any tags that explicitly appear in xquery
149 * will not be excluded, and will be removed from the list of exclude
150 * tags. The caller of this function has to combine the returned
151 * query appropriately.*/
153 _notmuch_exclude_tags (notmuch_query_t *query, Xapian::Query xquery)
155 Xapian::Query exclude_query = Xapian::Query::MatchNothing;
157 for (notmuch_string_node_t *term = query->exclude_terms->head; term;
159 Xapian::TermIterator it = xquery.get_terms_begin ();
160 Xapian::TermIterator end = xquery.get_terms_end ();
161 for (; it != end; it++) {
162 if ((*it).compare (term->string) == 0)
166 exclude_query = Xapian::Query (Xapian::Query::OP_OR,
167 exclude_query, Xapian::Query (term->string));
169 term->string = talloc_strdup (query, "");
171 return exclude_query;
175 notmuch_query_search_messages (notmuch_query_t *query)
177 notmuch_status_t status;
178 notmuch_messages_t *messages;
179 status = notmuch_query_search_messages_st (query, &messages);
187 notmuch_query_search_messages_st (notmuch_query_t *query,
188 notmuch_messages_t **out)
190 notmuch_database_t *notmuch = query->notmuch;
191 const char *query_string = query->query_string;
192 notmuch_mset_messages_t *messages;
194 messages = talloc (query, notmuch_mset_messages_t);
195 if (unlikely (messages == NULL))
196 return NOTMUCH_STATUS_OUT_OF_MEMORY;
200 messages->base.is_of_list_type = FALSE;
201 messages->base.iterator = NULL;
202 messages->notmuch = notmuch;
203 new (&messages->iterator) Xapian::MSetIterator ();
204 new (&messages->iterator_end) Xapian::MSetIterator ();
206 talloc_set_destructor (messages, _notmuch_messages_destructor);
208 Xapian::Enquire enquire (*notmuch->xapian_db);
209 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
210 _find_prefix ("type"),
212 Xapian::Query string_query, final_query, exclude_query;
214 Xapian::MSetIterator iterator;
215 unsigned int flags = (Xapian::QueryParser::FLAG_BOOLEAN |
216 Xapian::QueryParser::FLAG_PHRASE |
217 Xapian::QueryParser::FLAG_LOVEHATE |
218 Xapian::QueryParser::FLAG_BOOLEAN_ANY_CASE |
219 Xapian::QueryParser::FLAG_WILDCARD |
220 Xapian::QueryParser::FLAG_PURE_NOT);
222 if (strcmp (query_string, "") == 0 ||
223 strcmp (query_string, "*") == 0)
225 final_query = mail_query;
227 string_query = notmuch->query_parser->
228 parse_query (query_string, flags);
229 final_query = Xapian::Query (Xapian::Query::OP_AND,
230 mail_query, string_query);
232 messages->base.excluded_doc_ids = NULL;
234 if ((query->omit_excluded != NOTMUCH_EXCLUDE_FALSE) && (query->exclude_terms)) {
235 exclude_query = _notmuch_exclude_tags (query, final_query);
237 if (query->omit_excluded == NOTMUCH_EXCLUDE_TRUE ||
238 query->omit_excluded == NOTMUCH_EXCLUDE_ALL)
240 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
241 final_query, exclude_query);
242 } else { /* NOTMUCH_EXCLUDE_FLAG */
243 exclude_query = Xapian::Query (Xapian::Query::OP_AND,
244 exclude_query, final_query);
246 enquire.set_weighting_scheme (Xapian::BoolWeight());
247 enquire.set_query (exclude_query);
249 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
251 GArray *excluded_doc_ids = g_array_new (FALSE, FALSE, sizeof (unsigned int));
253 for (iterator = mset.begin (); iterator != mset.end (); iterator++) {
254 unsigned int doc_id = *iterator;
255 g_array_append_val (excluded_doc_ids, doc_id);
257 messages->base.excluded_doc_ids = talloc (messages, _notmuch_doc_id_set);
258 _notmuch_doc_id_set_init (query, messages->base.excluded_doc_ids,
260 g_array_unref (excluded_doc_ids);
265 enquire.set_weighting_scheme (Xapian::BoolWeight());
267 switch (query->sort) {
268 case NOTMUCH_SORT_OLDEST_FIRST:
269 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, FALSE);
271 case NOTMUCH_SORT_NEWEST_FIRST:
272 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, TRUE);
274 case NOTMUCH_SORT_MESSAGE_ID:
275 enquire.set_sort_by_value (NOTMUCH_VALUE_MESSAGE_ID, FALSE);
277 case NOTMUCH_SORT_UNSORTED:
281 if (_debug_query ()) {
282 fprintf (stderr, "Exclude query is:\n%s\n",
283 exclude_query.get_description ().c_str ());
284 fprintf (stderr, "Final query is:\n%s\n",
285 final_query.get_description ().c_str ());
288 enquire.set_query (final_query);
290 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
292 messages->iterator = mset.begin ();
293 messages->iterator_end = mset.end ();
295 *out = &messages->base;
296 return NOTMUCH_STATUS_SUCCESS;
298 } catch (const Xapian::Error &error) {
299 _notmuch_database_log (notmuch,
300 "A Xapian exception occurred performing query: %s\n"
301 "Query string was: %s\n",
302 error.get_msg().c_str(),
303 query->query_string);
305 notmuch->exception_reported = TRUE;
306 talloc_free (messages);
307 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
312 _notmuch_mset_messages_valid (notmuch_messages_t *messages)
314 notmuch_mset_messages_t *mset_messages;
316 mset_messages = (notmuch_mset_messages_t *) messages;
318 return (mset_messages->iterator != mset_messages->iterator_end);
322 _notmuch_mset_messages_get_doc_id (notmuch_messages_t *messages)
324 notmuch_mset_messages_t *mset_messages;
326 mset_messages = (notmuch_mset_messages_t *) messages;
328 if (! _notmuch_mset_messages_valid (&mset_messages->base))
331 return *mset_messages->iterator;
335 _notmuch_mset_messages_get (notmuch_messages_t *messages)
337 notmuch_message_t *message;
338 Xapian::docid doc_id;
339 notmuch_private_status_t status;
340 notmuch_mset_messages_t *mset_messages;
342 mset_messages = (notmuch_mset_messages_t *) messages;
344 if (! _notmuch_mset_messages_valid (&mset_messages->base))
347 doc_id = *mset_messages->iterator;
349 message = _notmuch_message_create (mset_messages,
350 mset_messages->notmuch, doc_id,
353 if (message == NULL &&
354 status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
356 INTERNAL_ERROR ("a messages iterator contains a non-existent document ID.\n");
359 if (messages->excluded_doc_ids &&
360 _notmuch_doc_id_set_contains (messages->excluded_doc_ids, doc_id))
361 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, TRUE);
367 _notmuch_mset_messages_move_to_next (notmuch_messages_t *messages)
369 notmuch_mset_messages_t *mset_messages;
371 mset_messages = (notmuch_mset_messages_t *) messages;
373 mset_messages->iterator++;
376 static notmuch_bool_t
377 _notmuch_doc_id_set_init (void *ctx,
378 notmuch_doc_id_set_t *doc_ids,
381 unsigned int max = 0;
382 unsigned char *bitmap;
384 for (unsigned int i = 0; i < arr->len; i++)
385 max = MAX(max, g_array_index (arr, unsigned int, i));
386 bitmap = talloc_zero_array (ctx, unsigned char, DOCIDSET_WORD(max) + 1);
391 doc_ids->bitmap = bitmap;
392 doc_ids->bound = max + 1;
394 for (unsigned int i = 0; i < arr->len; i++) {
395 unsigned int doc_id = g_array_index (arr, unsigned int, i);
396 bitmap[DOCIDSET_WORD(doc_id)] |= 1 << DOCIDSET_BIT(doc_id);
403 _notmuch_doc_id_set_contains (notmuch_doc_id_set_t *doc_ids,
406 if (doc_id >= doc_ids->bound)
408 return doc_ids->bitmap[DOCIDSET_WORD(doc_id)] & (1 << DOCIDSET_BIT(doc_id));
412 _notmuch_doc_id_set_remove (notmuch_doc_id_set_t *doc_ids,
415 if (doc_id < doc_ids->bound)
416 doc_ids->bitmap[DOCIDSET_WORD(doc_id)] &= ~(1 << DOCIDSET_BIT(doc_id));
419 /* Glib objects force use to use a talloc destructor as well, (but not
420 * nearly as ugly as the for messages due to C++ objects). At
421 * this point, I'd really like to have some talloc-friendly
422 * equivalents for the few pieces of glib that I'm using. */
424 _notmuch_threads_destructor (notmuch_threads_t *threads)
426 if (threads->doc_ids)
427 g_array_unref (threads->doc_ids);
434 notmuch_query_search_threads (notmuch_query_t *query)
436 notmuch_status_t status;
437 notmuch_threads_t *threads;
438 status = notmuch_query_search_threads_st (query, &threads);
446 notmuch_query_search_threads_st (notmuch_query_t *query,
447 notmuch_threads_t **out)
449 notmuch_threads_t *threads;
450 notmuch_messages_t *messages;
451 notmuch_status_t status;
453 threads = talloc (query, notmuch_threads_t);
455 return NOTMUCH_STATUS_OUT_OF_MEMORY;
456 threads->doc_ids = NULL;
457 talloc_set_destructor (threads, _notmuch_threads_destructor);
459 threads->query = query;
461 status = notmuch_query_search_messages_st (query, &messages);
463 talloc_free (threads);
467 threads->doc_ids = g_array_new (FALSE, FALSE, sizeof (unsigned int));
468 while (notmuch_messages_valid (messages)) {
469 unsigned int doc_id = _notmuch_mset_messages_get_doc_id (messages);
470 g_array_append_val (threads->doc_ids, doc_id);
471 notmuch_messages_move_to_next (messages);
473 threads->doc_id_pos = 0;
475 talloc_free (messages);
477 if (! _notmuch_doc_id_set_init (threads, &threads->match_set,
479 talloc_free (threads);
480 return NOTMUCH_STATUS_OUT_OF_MEMORY;
484 return NOTMUCH_STATUS_SUCCESS;
488 notmuch_query_destroy (notmuch_query_t *query)
494 notmuch_threads_valid (notmuch_threads_t *threads)
501 while (threads->doc_id_pos < threads->doc_ids->len) {
502 doc_id = g_array_index (threads->doc_ids, unsigned int,
503 threads->doc_id_pos);
504 if (_notmuch_doc_id_set_contains (&threads->match_set, doc_id))
507 threads->doc_id_pos++;
510 return threads->doc_id_pos < threads->doc_ids->len;
514 notmuch_threads_get (notmuch_threads_t *threads)
518 if (! notmuch_threads_valid (threads))
521 doc_id = g_array_index (threads->doc_ids, unsigned int,
522 threads->doc_id_pos);
523 return _notmuch_thread_create (threads->query,
524 threads->query->notmuch,
527 threads->query->exclude_terms,
528 threads->query->omit_excluded,
529 threads->query->sort);
533 notmuch_threads_move_to_next (notmuch_threads_t *threads)
535 threads->doc_id_pos++;
539 notmuch_threads_destroy (notmuch_threads_t *threads)
541 talloc_free (threads);
545 notmuch_query_count_messages (notmuch_query_t *query)
547 notmuch_database_t *notmuch = query->notmuch;
548 const char *query_string = query->query_string;
549 Xapian::doccount count = 0;
552 Xapian::Enquire enquire (*notmuch->xapian_db);
553 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
554 _find_prefix ("type"),
556 Xapian::Query string_query, final_query, exclude_query;
558 unsigned int flags = (Xapian::QueryParser::FLAG_BOOLEAN |
559 Xapian::QueryParser::FLAG_PHRASE |
560 Xapian::QueryParser::FLAG_LOVEHATE |
561 Xapian::QueryParser::FLAG_BOOLEAN_ANY_CASE |
562 Xapian::QueryParser::FLAG_WILDCARD |
563 Xapian::QueryParser::FLAG_PURE_NOT);
565 if (strcmp (query_string, "") == 0 ||
566 strcmp (query_string, "*") == 0)
568 final_query = mail_query;
570 string_query = notmuch->query_parser->
571 parse_query (query_string, flags);
572 final_query = Xapian::Query (Xapian::Query::OP_AND,
573 mail_query, string_query);
576 exclude_query = _notmuch_exclude_tags (query, final_query);
578 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
579 final_query, exclude_query);
581 enquire.set_weighting_scheme(Xapian::BoolWeight());
582 enquire.set_docid_order(Xapian::Enquire::ASCENDING);
584 if (_debug_query ()) {
585 fprintf (stderr, "Exclude query is:\n%s\n",
586 exclude_query.get_description ().c_str ());
587 fprintf (stderr, "Final query is:\n%s\n",
588 final_query.get_description ().c_str ());
591 enquire.set_query (final_query);
594 * Set the checkatleast parameter to the number of documents
595 * in the database to make get_matches_estimated() exact.
597 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount (),
598 notmuch->xapian_db->get_doccount ());
600 count = mset.get_matches_estimated();
602 } catch (const Xapian::Error &error) {
603 _notmuch_database_log (notmuch,
604 "A Xapian exception occurred performing query: %s\n"
605 "Query string was: %s\n",
606 error.get_msg().c_str(),
607 query->query_string);
615 notmuch_query_count_threads (notmuch_query_t *query)
617 notmuch_messages_t *messages;
623 query->sort = NOTMUCH_SORT_UNSORTED;
624 messages = notmuch_query_search_messages (query);
626 if (messages == NULL)
629 hash = g_hash_table_new_full (g_str_hash, g_str_equal, NULL, NULL);
631 talloc_free (messages);
635 while (notmuch_messages_valid (messages)) {
636 notmuch_message_t *message = notmuch_messages_get (messages);
637 const char *thread_id = notmuch_message_get_thread_id (message);
638 char *thread_id_copy = talloc_strdup (messages, thread_id);
639 if (unlikely (thread_id_copy == NULL)) {
640 notmuch_message_destroy (message);
644 g_hash_table_insert (hash, thread_id_copy, NULL);
645 notmuch_message_destroy (message);
646 notmuch_messages_move_to_next (messages);
649 count = g_hash_table_size (hash);
652 g_hash_table_unref (hash);
653 talloc_free (messages);