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 https://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;
33 notmuch_query_syntax_t syntax;
34 Xapian::Query xapian_query;
35 std::set<std::string> terms;
38 typedef struct _notmuch_mset_messages {
39 notmuch_messages_t base;
40 notmuch_database_t *notmuch;
41 Xapian::MSetIterator iterator;
42 Xapian::MSetIterator iterator_end;
43 } notmuch_mset_messages_t;
45 struct _notmuch_doc_id_set {
46 unsigned char *bitmap;
50 #define DOCIDSET_WORD(bit) ((bit) / CHAR_BIT)
51 #define DOCIDSET_BIT(bit) ((bit) % CHAR_BIT)
53 struct _notmuch_threads {
54 notmuch_query_t *query;
56 /* The ordered list of doc ids matched by the query. */
58 /* Our iterator's current position in doc_ids. */
59 unsigned int doc_id_pos;
60 /* The set of matched docid's that have not been assigned to a
61 * thread. Initially, this contains every docid in doc_ids. */
62 notmuch_doc_id_set_t match_set;
65 /* We need this in the message functions so forward declare. */
67 _notmuch_doc_id_set_init (void *ctx,
68 notmuch_doc_id_set_t *doc_ids,
74 char *env = getenv ("NOTMUCH_DEBUG_QUERY");
76 return (env && strcmp (env, "") != 0);
79 /* Explicit destructor call for placement new */
81 _notmuch_query_destructor (notmuch_query_t *query)
83 query->xapian_query.~Query();
84 query->terms.~set<std::string>();
88 static notmuch_query_t *
89 _notmuch_query_constructor (notmuch_database_t *notmuch,
90 const char *query_string)
92 notmuch_query_t *query;
95 fprintf (stderr, "Query string is:\n%s\n", query_string);
97 query = talloc (notmuch, notmuch_query_t);
98 if (unlikely (query == NULL))
101 new (&query->xapian_query) Xapian::Query ();
102 new (&query->terms) std::set<std::string> ();
103 query->parsed = false;
105 talloc_set_destructor (query, _notmuch_query_destructor);
107 query->notmuch = notmuch;
110 query->query_string = talloc_strdup (query, query_string);
112 query->query_string = NULL;
114 query->sort = NOTMUCH_SORT_NEWEST_FIRST;
116 query->exclude_terms = _notmuch_string_list_create (query);
118 query->omit_excluded = NOTMUCH_EXCLUDE_TRUE;
124 notmuch_query_create (notmuch_database_t *notmuch,
125 const char *query_string)
128 notmuch_query_t *query;
129 notmuch_status_t status;
131 status = notmuch_query_create_with_syntax (notmuch, query_string,
132 NOTMUCH_QUERY_SYNTAX_XAPIAN,
141 notmuch_query_create_with_syntax (notmuch_database_t *notmuch,
142 const char *query_string,
143 notmuch_query_syntax_t syntax,
144 notmuch_query_t **output)
147 notmuch_query_t *query;
150 return NOTMUCH_STATUS_NULL_POINTER;
152 query = _notmuch_query_constructor (notmuch, query_string);
154 return NOTMUCH_STATUS_OUT_OF_MEMORY;
156 if (syntax == NOTMUCH_QUERY_SYNTAX_SEXP && ! HAVE_SFSEXP) {
157 _notmuch_database_log (notmuch, "sexp query parser not available");
158 return NOTMUCH_STATUS_ILLEGAL_ARGUMENT;
161 query->syntax = syntax;
165 return NOTMUCH_STATUS_SUCCESS;
169 _notmuch_query_cache_terms (notmuch_query_t *query)
171 /* Xapian doesn't support skip_to on terms from a query since
172 * they are unordered, so cache a copy of all terms in
173 * something searchable.
176 for (Xapian::TermIterator t = query->xapian_query.get_terms_begin ();
177 t != query->xapian_query.get_terms_end (); ++t)
178 query->terms.insert (*t);
181 static notmuch_status_t
182 _notmuch_query_ensure_parsed_xapian (notmuch_query_t *query)
185 query->xapian_query =
186 query->notmuch->query_parser->
187 parse_query (query->query_string, NOTMUCH_QUERY_PARSER_FLAGS);
189 _notmuch_query_cache_terms (query);
190 query->parsed = true;
192 } catch (const Xapian::Error &error) {
193 if (! query->notmuch->exception_reported) {
194 _notmuch_database_log (query->notmuch,
195 "A Xapian exception occurred parsing query: %s\n",
196 error.get_msg ().c_str ());
197 _notmuch_database_log_append (query->notmuch,
198 "Query string was: %s\n",
199 query->query_string);
200 query->notmuch->exception_reported = true;
203 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
205 return NOTMUCH_STATUS_SUCCESS;
208 static notmuch_status_t
209 _notmuch_query_ensure_parsed_sexpr (notmuch_query_t *query)
211 notmuch_status_t status;
214 return NOTMUCH_STATUS_SUCCESS;
216 status = _notmuch_sexp_string_to_xapian_query (query->notmuch, query->query_string,
217 query->xapian_query);
221 _notmuch_query_cache_terms (query);
222 return NOTMUCH_STATUS_SUCCESS;
225 static notmuch_status_t
226 _notmuch_query_ensure_parsed (notmuch_query_t *query)
229 return NOTMUCH_STATUS_SUCCESS;
232 if (query->syntax == NOTMUCH_QUERY_SYNTAX_SEXP)
233 return _notmuch_query_ensure_parsed_sexpr (query);
236 return _notmuch_query_ensure_parsed_xapian (query);
240 notmuch_query_get_query_string (const notmuch_query_t *query)
242 return query->query_string;
246 notmuch_query_set_omit_excluded (notmuch_query_t *query,
247 notmuch_exclude_t omit_excluded)
249 query->omit_excluded = omit_excluded;
253 notmuch_query_set_sort (notmuch_query_t *query, notmuch_sort_t sort)
259 notmuch_query_get_sort (const notmuch_query_t *query)
265 notmuch_query_add_tag_exclude (notmuch_query_t *query, const char *tag)
267 notmuch_status_t status;
270 status = _notmuch_query_ensure_parsed (query);
274 term = talloc_asprintf (query, "%s%s", _find_prefix ("tag"), tag);
275 if (query->terms.count (term) != 0)
276 return NOTMUCH_STATUS_IGNORED;
278 _notmuch_string_list_append (query->exclude_terms, term);
279 return NOTMUCH_STATUS_SUCCESS;
282 /* We end up having to call the destructors explicitly because we had
283 * to use "placement new" in order to initialize C++ objects within a
284 * block that we allocated with talloc. So C++ is making talloc
285 * slightly less simple to use, (we wouldn't need
286 * talloc_set_destructor at all otherwise).
289 _notmuch_messages_destructor (notmuch_mset_messages_t *messages)
291 messages->iterator.~MSetIterator ();
292 messages->iterator_end.~MSetIterator ();
297 /* Return a query that matches messages with the excluded tags
298 * registered with query. The caller of this function has to combine the returned
299 * query appropriately.*/
301 _notmuch_exclude_tags (notmuch_query_t *query)
303 Xapian::Query exclude_query = Xapian::Query::MatchNothing;
305 for (notmuch_string_node_t *term = query->exclude_terms->head; term;
307 exclude_query = Xapian::Query (Xapian::Query::OP_OR,
308 exclude_query, Xapian::Query (term->string));
310 return exclude_query;
315 notmuch_query_search_messages_st (notmuch_query_t *query,
316 notmuch_messages_t **out)
318 return notmuch_query_search_messages (query, out);
322 notmuch_query_search_messages (notmuch_query_t *query,
323 notmuch_messages_t **out)
325 return _notmuch_query_search_documents (query, "mail", out);
329 _notmuch_query_search_documents (notmuch_query_t *query,
331 notmuch_messages_t **out)
333 notmuch_database_t *notmuch = query->notmuch;
334 const char *query_string = query->query_string;
335 notmuch_mset_messages_t *messages;
336 notmuch_status_t status;
338 status = _notmuch_query_ensure_parsed (query);
342 messages = talloc (query, notmuch_mset_messages_t);
343 if (unlikely (messages == NULL))
344 return NOTMUCH_STATUS_OUT_OF_MEMORY;
348 messages->base.is_of_list_type = false;
349 messages->base.iterator = NULL;
350 messages->notmuch = notmuch;
351 new (&messages->iterator) Xapian::MSetIterator ();
352 new (&messages->iterator_end) Xapian::MSetIterator ();
354 talloc_set_destructor (messages, _notmuch_messages_destructor);
356 Xapian::Enquire enquire (*notmuch->xapian_db);
357 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
358 _find_prefix ("type"),
360 Xapian::Query final_query, exclude_query;
362 Xapian::MSetIterator iterator;
364 if (strcmp (query_string, "") == 0 ||
365 strcmp (query_string, "*") == 0) {
366 final_query = mail_query;
368 final_query = Xapian::Query (Xapian::Query::OP_AND,
369 mail_query, query->xapian_query);
371 messages->base.excluded_doc_ids = NULL;
373 if ((query->omit_excluded != NOTMUCH_EXCLUDE_FALSE) && (query->exclude_terms)) {
374 exclude_query = _notmuch_exclude_tags (query);
376 if (query->omit_excluded == NOTMUCH_EXCLUDE_TRUE ||
377 query->omit_excluded == NOTMUCH_EXCLUDE_ALL) {
378 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
379 final_query, exclude_query);
380 } else { /* NOTMUCH_EXCLUDE_FLAG */
381 exclude_query = Xapian::Query (Xapian::Query::OP_AND,
382 exclude_query, final_query);
384 enquire.set_weighting_scheme (Xapian::BoolWeight ());
385 enquire.set_query (exclude_query);
387 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
389 GArray *excluded_doc_ids = g_array_new (false, false, sizeof (unsigned int));
391 for (iterator = mset.begin (); iterator != mset.end (); iterator++) {
392 unsigned int doc_id = *iterator;
393 g_array_append_val (excluded_doc_ids, doc_id);
395 messages->base.excluded_doc_ids = talloc (messages, _notmuch_doc_id_set);
396 _notmuch_doc_id_set_init (query, messages->base.excluded_doc_ids,
398 g_array_unref (excluded_doc_ids);
403 enquire.set_weighting_scheme (Xapian::BoolWeight ());
405 switch (query->sort) {
406 case NOTMUCH_SORT_OLDEST_FIRST:
407 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, false);
409 case NOTMUCH_SORT_NEWEST_FIRST:
410 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, true);
412 case NOTMUCH_SORT_MESSAGE_ID:
413 enquire.set_sort_by_value (NOTMUCH_VALUE_MESSAGE_ID, false);
415 case NOTMUCH_SORT_UNSORTED:
419 if (_debug_query ()) {
420 fprintf (stderr, "Exclude query is:\n%s\n",
421 exclude_query.get_description ().c_str ());
422 fprintf (stderr, "Final query is:\n%s\n",
423 final_query.get_description ().c_str ());
426 enquire.set_query (final_query);
428 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
430 messages->iterator = mset.begin ();
431 messages->iterator_end = mset.end ();
433 *out = &messages->base;
434 return NOTMUCH_STATUS_SUCCESS;
436 } catch (const Xapian::Error &error) {
437 _notmuch_database_log (notmuch,
438 "A Xapian exception occurred performing query: %s\n",
439 error.get_msg ().c_str ());
440 _notmuch_database_log_append (notmuch,
441 "Query string was: %s\n",
442 query->query_string);
444 notmuch->exception_reported = true;
445 talloc_free (messages);
446 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
451 _notmuch_mset_messages_valid (notmuch_messages_t *messages)
453 notmuch_mset_messages_t *mset_messages;
455 mset_messages = (notmuch_mset_messages_t *) messages;
457 return (mset_messages->iterator != mset_messages->iterator_end);
461 _notmuch_mset_messages_get_doc_id (notmuch_messages_t *messages)
463 notmuch_mset_messages_t *mset_messages;
465 mset_messages = (notmuch_mset_messages_t *) messages;
467 if (! _notmuch_mset_messages_valid (&mset_messages->base))
470 return *mset_messages->iterator;
474 _notmuch_mset_messages_get (notmuch_messages_t *messages)
476 notmuch_message_t *message;
477 Xapian::docid doc_id;
478 notmuch_private_status_t status;
479 notmuch_mset_messages_t *mset_messages;
481 mset_messages = (notmuch_mset_messages_t *) messages;
483 if (! _notmuch_mset_messages_valid (&mset_messages->base))
486 doc_id = *mset_messages->iterator;
488 message = _notmuch_message_create (mset_messages,
489 mset_messages->notmuch, doc_id,
492 if (message == NULL &&
493 status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) {
494 INTERNAL_ERROR ("a messages iterator contains a non-existent document ID.\n");
497 if (messages->excluded_doc_ids &&
498 _notmuch_doc_id_set_contains (messages->excluded_doc_ids, doc_id))
499 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, true);
505 _notmuch_mset_messages_move_to_next (notmuch_messages_t *messages)
507 notmuch_mset_messages_t *mset_messages;
509 mset_messages = (notmuch_mset_messages_t *) messages;
511 mset_messages->iterator++;
515 _notmuch_doc_id_set_init (void *ctx,
516 notmuch_doc_id_set_t *doc_ids,
519 unsigned int max = 0;
520 unsigned char *bitmap;
522 for (unsigned int i = 0; i < arr->len; i++)
523 max = MAX (max, g_array_index (arr, unsigned int, i));
524 bitmap = talloc_zero_array (ctx, unsigned char, DOCIDSET_WORD (max) + 1);
529 doc_ids->bitmap = bitmap;
530 doc_ids->bound = max + 1;
532 for (unsigned int i = 0; i < arr->len; i++) {
533 unsigned int doc_id = g_array_index (arr, unsigned int, i);
534 bitmap[DOCIDSET_WORD (doc_id)] |= 1 << DOCIDSET_BIT (doc_id);
541 _notmuch_doc_id_set_contains (notmuch_doc_id_set_t *doc_ids,
544 if (doc_id >= doc_ids->bound)
546 return doc_ids->bitmap[DOCIDSET_WORD (doc_id)] & (1 << DOCIDSET_BIT (doc_id));
550 _notmuch_doc_id_set_remove (notmuch_doc_id_set_t *doc_ids,
553 if (doc_id < doc_ids->bound)
554 doc_ids->bitmap[DOCIDSET_WORD (doc_id)] &= ~(1 << DOCIDSET_BIT (doc_id));
557 /* Glib objects force use to use a talloc destructor as well, (but not
558 * nearly as ugly as the for messages due to C++ objects). At
559 * this point, I'd really like to have some talloc-friendly
560 * equivalents for the few pieces of glib that I'm using. */
562 _notmuch_threads_destructor (notmuch_threads_t *threads)
564 if (threads->doc_ids)
565 g_array_unref (threads->doc_ids);
571 notmuch_query_search_threads_st (notmuch_query_t *query, notmuch_threads_t **out)
573 return notmuch_query_search_threads (query, out);
577 notmuch_query_search_threads (notmuch_query_t *query,
578 notmuch_threads_t **out)
580 notmuch_threads_t *threads;
581 notmuch_messages_t *messages;
582 notmuch_status_t status;
584 threads = talloc (query, notmuch_threads_t);
586 return NOTMUCH_STATUS_OUT_OF_MEMORY;
587 threads->doc_ids = NULL;
588 talloc_set_destructor (threads, _notmuch_threads_destructor);
590 threads->query = query;
592 status = notmuch_query_search_messages (query, &messages);
594 talloc_free (threads);
598 threads->doc_ids = g_array_new (false, false, sizeof (unsigned int));
599 while (notmuch_messages_valid (messages)) {
600 unsigned int doc_id = _notmuch_mset_messages_get_doc_id (messages);
601 g_array_append_val (threads->doc_ids, doc_id);
602 notmuch_messages_move_to_next (messages);
604 threads->doc_id_pos = 0;
606 talloc_free (messages);
608 if (! _notmuch_doc_id_set_init (threads, &threads->match_set,
610 talloc_free (threads);
611 return NOTMUCH_STATUS_OUT_OF_MEMORY;
615 return NOTMUCH_STATUS_SUCCESS;
619 notmuch_query_destroy (notmuch_query_t *query)
625 notmuch_threads_valid (notmuch_threads_t *threads)
632 while (threads->doc_id_pos < threads->doc_ids->len) {
633 doc_id = g_array_index (threads->doc_ids, unsigned int,
634 threads->doc_id_pos);
635 if (_notmuch_doc_id_set_contains (&threads->match_set, doc_id))
638 threads->doc_id_pos++;
641 return threads->doc_id_pos < threads->doc_ids->len;
645 notmuch_threads_get (notmuch_threads_t *threads)
649 if (! notmuch_threads_valid (threads))
652 doc_id = g_array_index (threads->doc_ids, unsigned int,
653 threads->doc_id_pos);
654 return _notmuch_thread_create (threads->query,
655 threads->query->notmuch,
658 threads->query->exclude_terms,
659 threads->query->omit_excluded,
660 threads->query->sort);
664 notmuch_threads_move_to_next (notmuch_threads_t *threads)
666 threads->doc_id_pos++;
670 notmuch_threads_destroy (notmuch_threads_t *threads)
672 talloc_free (threads);
676 notmuch_query_count_messages_st (notmuch_query_t *query, unsigned *count_out)
678 return notmuch_query_count_messages (query, count_out);
682 notmuch_query_count_messages (notmuch_query_t *query, unsigned *count_out)
684 return _notmuch_query_count_documents (query, "mail", count_out);
688 _notmuch_query_count_documents (notmuch_query_t *query, const char *type, unsigned *count_out)
690 notmuch_database_t *notmuch = query->notmuch;
691 const char *query_string = query->query_string;
692 Xapian::doccount count = 0;
693 notmuch_status_t status;
695 status = _notmuch_query_ensure_parsed (query);
700 Xapian::Enquire enquire (*notmuch->xapian_db);
701 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
702 _find_prefix ("type"),
704 Xapian::Query final_query, exclude_query;
707 if (strcmp (query_string, "") == 0 ||
708 strcmp (query_string, "*") == 0) {
709 final_query = mail_query;
711 final_query = Xapian::Query (Xapian::Query::OP_AND,
712 mail_query, query->xapian_query);
715 exclude_query = _notmuch_exclude_tags (query);
717 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
718 final_query, exclude_query);
720 enquire.set_weighting_scheme (Xapian::BoolWeight ());
721 enquire.set_docid_order (Xapian::Enquire::ASCENDING);
723 if (_debug_query ()) {
724 fprintf (stderr, "Exclude query is:\n%s\n",
725 exclude_query.get_description ().c_str ());
726 fprintf (stderr, "Final query is:\n%s\n",
727 final_query.get_description ().c_str ());
730 enquire.set_query (final_query);
733 * Set the checkatleast parameter to the number of documents
734 * in the database to make get_matches_estimated() exact.
735 * Set the max parameter to 1 to avoid fetching documents we will discard.
737 mset = enquire.get_mset (0, 1,
738 notmuch->xapian_db->get_doccount ());
740 count = mset.get_matches_estimated ();
742 } catch (const Xapian::Error &error) {
743 _notmuch_database_log (notmuch,
744 "A Xapian exception occurred performing query: %s\n",
745 error.get_msg ().c_str ());
746 _notmuch_database_log_append (notmuch,
747 "Query string was: %s\n",
748 query->query_string);
749 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
753 return NOTMUCH_STATUS_SUCCESS;
757 notmuch_query_count_threads_st (notmuch_query_t *query, unsigned *count)
759 return notmuch_query_count_threads (query, count);
763 notmuch_query_count_threads (notmuch_query_t *query, unsigned *count)
765 notmuch_messages_t *messages;
768 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
771 query->sort = NOTMUCH_SORT_UNSORTED;
772 ret = notmuch_query_search_messages (query, &messages);
776 if (messages == NULL)
777 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
779 hash = g_hash_table_new_full (g_str_hash, g_str_equal, NULL, NULL);
781 talloc_free (messages);
782 return NOTMUCH_STATUS_OUT_OF_MEMORY;
785 while (notmuch_messages_valid (messages)) {
786 notmuch_message_t *message = notmuch_messages_get (messages);
787 const char *thread_id = notmuch_message_get_thread_id (message);
788 char *thread_id_copy = talloc_strdup (messages, thread_id);
789 if (unlikely (thread_id_copy == NULL)) {
790 notmuch_message_destroy (message);
791 ret = NOTMUCH_STATUS_OUT_OF_MEMORY;
794 g_hash_table_insert (hash, thread_id_copy, NULL);
795 notmuch_message_destroy (message);
796 notmuch_messages_move_to_next (messages);
799 *count = g_hash_table_size (hash);
802 g_hash_table_unref (hash);
803 talloc_free (messages);
809 notmuch_query_get_database (const notmuch_query_t *query)
811 return query->notmuch;