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 Xapian::Query xapian_query;
34 std::set<std::string> terms;
37 typedef struct _notmuch_mset_messages {
38 notmuch_messages_t base;
39 notmuch_database_t *notmuch;
40 Xapian::MSetIterator iterator;
41 Xapian::MSetIterator iterator_end;
42 } notmuch_mset_messages_t;
44 struct _notmuch_doc_id_set {
45 unsigned char *bitmap;
49 #define DOCIDSET_WORD(bit) ((bit) / CHAR_BIT)
50 #define DOCIDSET_BIT(bit) ((bit) % CHAR_BIT)
52 struct _notmuch_threads {
53 notmuch_query_t *query;
55 /* The ordered list of doc ids matched by the query. */
57 /* Our iterator's current position in doc_ids. */
58 unsigned int doc_id_pos;
59 /* The set of matched docid's that have not been assigned to a
60 * thread. Initially, this contains every docid in doc_ids. */
61 notmuch_doc_id_set_t match_set;
64 /* We need this in the message functions so forward declare. */
66 _notmuch_doc_id_set_init (void *ctx,
67 notmuch_doc_id_set_t *doc_ids,
73 char *env = getenv ("NOTMUCH_DEBUG_QUERY");
74 return (env && strcmp (env, "") != 0);
77 /* Explicit destructor call for placement new */
79 _notmuch_query_destructor (notmuch_query_t *query) {
80 query->xapian_query.~Query();
81 query->terms.~set<std::string>();
86 notmuch_query_create (notmuch_database_t *notmuch,
87 const char *query_string)
89 notmuch_query_t *query;
92 fprintf (stderr, "Query string is:\n%s\n", query_string);
94 query = talloc (notmuch, notmuch_query_t);
95 if (unlikely (query == NULL))
98 new (&query->xapian_query) Xapian::Query ();
99 new (&query->terms) std::set<std::string> ();
100 query->parsed = false;
102 talloc_set_destructor (query, _notmuch_query_destructor);
104 query->notmuch = notmuch;
106 query->query_string = talloc_strdup (query, query_string);
108 query->sort = NOTMUCH_SORT_NEWEST_FIRST;
110 query->exclude_terms = _notmuch_string_list_create (query);
112 query->omit_excluded = NOTMUCH_EXCLUDE_TRUE;
117 static notmuch_status_t
118 _notmuch_query_ensure_parsed (notmuch_query_t *query)
121 return NOTMUCH_STATUS_SUCCESS;
124 query->xapian_query =
125 query->notmuch->query_parser->
126 parse_query (query->query_string, NOTMUCH_QUERY_PARSER_FLAGS);
128 /* Xapian doesn't support skip_to on terms from a query since
129 * they are unordered, so cache a copy of all terms in
130 * something searchable.
133 for (Xapian::TermIterator t = query->xapian_query.get_terms_begin ();
134 t != query->xapian_query.get_terms_end (); ++t)
135 query->terms.insert (*t);
137 query->parsed = true;
139 } catch (const Xapian::Error &error) {
140 if (!query->notmuch->exception_reported) {
141 _notmuch_database_log (query->notmuch,
142 "A Xapian exception occurred parsing query: %s\n",
143 error.get_msg ().c_str ());
144 _notmuch_database_log_append (query->notmuch,
145 "Query string was: %s\n",
146 query->query_string);
147 query->notmuch->exception_reported = true;
150 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
152 return NOTMUCH_STATUS_SUCCESS;
156 notmuch_query_get_query_string (const notmuch_query_t *query)
158 return query->query_string;
162 notmuch_query_set_omit_excluded (notmuch_query_t *query,
163 notmuch_exclude_t omit_excluded)
165 query->omit_excluded = omit_excluded;
169 notmuch_query_set_sort (notmuch_query_t *query, notmuch_sort_t sort)
175 notmuch_query_get_sort (const notmuch_query_t *query)
181 notmuch_query_add_tag_exclude (notmuch_query_t *query, const char *tag)
183 notmuch_status_t status;
186 status = _notmuch_query_ensure_parsed (query);
190 term = talloc_asprintf (query, "%s%s", _find_prefix ("tag"), tag);
191 if (query->terms.count(term) != 0)
192 return NOTMUCH_STATUS_IGNORED;
194 _notmuch_string_list_append (query->exclude_terms, term);
195 return NOTMUCH_STATUS_SUCCESS;
198 /* We end up having to call the destructors explicitly because we had
199 * to use "placement new" in order to initialize C++ objects within a
200 * block that we allocated with talloc. So C++ is making talloc
201 * slightly less simple to use, (we wouldn't need
202 * talloc_set_destructor at all otherwise).
205 _notmuch_messages_destructor (notmuch_mset_messages_t *messages)
207 messages->iterator.~MSetIterator ();
208 messages->iterator_end.~MSetIterator ();
213 /* Return a query that matches messages with the excluded tags
214 * registered with query. The caller of this function has to combine the returned
215 * query appropriately.*/
217 _notmuch_exclude_tags (notmuch_query_t *query)
219 Xapian::Query exclude_query = Xapian::Query::MatchNothing;
221 for (notmuch_string_node_t *term = query->exclude_terms->head; term;
223 exclude_query = Xapian::Query (Xapian::Query::OP_OR,
224 exclude_query, Xapian::Query (term->string));
226 return exclude_query;
231 notmuch_query_search_messages_st (notmuch_query_t *query,
232 notmuch_messages_t **out)
234 return notmuch_query_search_messages (query, out);
238 notmuch_query_search_messages (notmuch_query_t *query,
239 notmuch_messages_t **out)
241 return _notmuch_query_search_documents (query, "mail", out);
245 _notmuch_query_search_documents (notmuch_query_t *query,
247 notmuch_messages_t **out)
249 notmuch_database_t *notmuch = query->notmuch;
250 const char *query_string = query->query_string;
251 notmuch_mset_messages_t *messages;
252 notmuch_status_t status;
254 status = _notmuch_query_ensure_parsed (query);
258 messages = talloc (query, notmuch_mset_messages_t);
259 if (unlikely (messages == NULL))
260 return NOTMUCH_STATUS_OUT_OF_MEMORY;
264 messages->base.is_of_list_type = false;
265 messages->base.iterator = NULL;
266 messages->notmuch = notmuch;
267 new (&messages->iterator) Xapian::MSetIterator ();
268 new (&messages->iterator_end) Xapian::MSetIterator ();
270 talloc_set_destructor (messages, _notmuch_messages_destructor);
272 Xapian::Enquire enquire (*notmuch->xapian_db);
273 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
274 _find_prefix ("type"),
276 Xapian::Query final_query, exclude_query;
278 Xapian::MSetIterator iterator;
280 if (strcmp (query_string, "") == 0 ||
281 strcmp (query_string, "*") == 0)
283 final_query = mail_query;
285 final_query = Xapian::Query (Xapian::Query::OP_AND,
286 mail_query, query->xapian_query);
288 messages->base.excluded_doc_ids = NULL;
290 if ((query->omit_excluded != NOTMUCH_EXCLUDE_FALSE) && (query->exclude_terms)) {
291 exclude_query = _notmuch_exclude_tags (query);
293 if (query->omit_excluded == NOTMUCH_EXCLUDE_TRUE ||
294 query->omit_excluded == NOTMUCH_EXCLUDE_ALL)
296 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
297 final_query, exclude_query);
298 } else { /* NOTMUCH_EXCLUDE_FLAG */
299 exclude_query = Xapian::Query (Xapian::Query::OP_AND,
300 exclude_query, final_query);
302 enquire.set_weighting_scheme (Xapian::BoolWeight());
303 enquire.set_query (exclude_query);
305 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
307 GArray *excluded_doc_ids = g_array_new (false, false, sizeof (unsigned int));
309 for (iterator = mset.begin (); iterator != mset.end (); iterator++) {
310 unsigned int doc_id = *iterator;
311 g_array_append_val (excluded_doc_ids, doc_id);
313 messages->base.excluded_doc_ids = talloc (messages, _notmuch_doc_id_set);
314 _notmuch_doc_id_set_init (query, messages->base.excluded_doc_ids,
316 g_array_unref (excluded_doc_ids);
321 enquire.set_weighting_scheme (Xapian::BoolWeight());
323 switch (query->sort) {
324 case NOTMUCH_SORT_OLDEST_FIRST:
325 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, false);
327 case NOTMUCH_SORT_NEWEST_FIRST:
328 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, true);
330 case NOTMUCH_SORT_MESSAGE_ID:
331 enquire.set_sort_by_value (NOTMUCH_VALUE_MESSAGE_ID, false);
333 case NOTMUCH_SORT_UNSORTED:
337 if (_debug_query ()) {
338 fprintf (stderr, "Exclude query is:\n%s\n",
339 exclude_query.get_description ().c_str ());
340 fprintf (stderr, "Final query is:\n%s\n",
341 final_query.get_description ().c_str ());
344 enquire.set_query (final_query);
346 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
348 messages->iterator = mset.begin ();
349 messages->iterator_end = mset.end ();
351 *out = &messages->base;
352 return NOTMUCH_STATUS_SUCCESS;
354 } catch (const Xapian::Error &error) {
355 _notmuch_database_log (notmuch,
356 "A Xapian exception occurred performing query: %s\n",
357 error.get_msg().c_str());
358 _notmuch_database_log_append (notmuch,
359 "Query string was: %s\n",
360 query->query_string);
362 notmuch->exception_reported = true;
363 talloc_free (messages);
364 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
369 _notmuch_mset_messages_valid (notmuch_messages_t *messages)
371 notmuch_mset_messages_t *mset_messages;
373 mset_messages = (notmuch_mset_messages_t *) messages;
375 return (mset_messages->iterator != mset_messages->iterator_end);
379 _notmuch_mset_messages_get_doc_id (notmuch_messages_t *messages)
381 notmuch_mset_messages_t *mset_messages;
383 mset_messages = (notmuch_mset_messages_t *) messages;
385 if (! _notmuch_mset_messages_valid (&mset_messages->base))
388 return *mset_messages->iterator;
392 _notmuch_mset_messages_get (notmuch_messages_t *messages)
394 notmuch_message_t *message;
395 Xapian::docid doc_id;
396 notmuch_private_status_t status;
397 notmuch_mset_messages_t *mset_messages;
399 mset_messages = (notmuch_mset_messages_t *) messages;
401 if (! _notmuch_mset_messages_valid (&mset_messages->base))
404 doc_id = *mset_messages->iterator;
406 message = _notmuch_message_create (mset_messages,
407 mset_messages->notmuch, doc_id,
410 if (message == NULL &&
411 status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
413 INTERNAL_ERROR ("a messages iterator contains a non-existent document ID.\n");
416 if (messages->excluded_doc_ids &&
417 _notmuch_doc_id_set_contains (messages->excluded_doc_ids, doc_id))
418 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, true);
424 _notmuch_mset_messages_move_to_next (notmuch_messages_t *messages)
426 notmuch_mset_messages_t *mset_messages;
428 mset_messages = (notmuch_mset_messages_t *) messages;
430 mset_messages->iterator++;
434 _notmuch_doc_id_set_init (void *ctx,
435 notmuch_doc_id_set_t *doc_ids,
438 unsigned int max = 0;
439 unsigned char *bitmap;
441 for (unsigned int i = 0; i < arr->len; i++)
442 max = MAX(max, g_array_index (arr, unsigned int, i));
443 bitmap = talloc_zero_array (ctx, unsigned char, DOCIDSET_WORD(max) + 1);
448 doc_ids->bitmap = bitmap;
449 doc_ids->bound = max + 1;
451 for (unsigned int i = 0; i < arr->len; i++) {
452 unsigned int doc_id = g_array_index (arr, unsigned int, i);
453 bitmap[DOCIDSET_WORD(doc_id)] |= 1 << DOCIDSET_BIT(doc_id);
460 _notmuch_doc_id_set_contains (notmuch_doc_id_set_t *doc_ids,
463 if (doc_id >= doc_ids->bound)
465 return doc_ids->bitmap[DOCIDSET_WORD(doc_id)] & (1 << DOCIDSET_BIT(doc_id));
469 _notmuch_doc_id_set_remove (notmuch_doc_id_set_t *doc_ids,
472 if (doc_id < doc_ids->bound)
473 doc_ids->bitmap[DOCIDSET_WORD(doc_id)] &= ~(1 << DOCIDSET_BIT(doc_id));
476 /* Glib objects force use to use a talloc destructor as well, (but not
477 * nearly as ugly as the for messages due to C++ objects). At
478 * this point, I'd really like to have some talloc-friendly
479 * equivalents for the few pieces of glib that I'm using. */
481 _notmuch_threads_destructor (notmuch_threads_t *threads)
483 if (threads->doc_ids)
484 g_array_unref (threads->doc_ids);
490 notmuch_query_search_threads_st (notmuch_query_t *query, notmuch_threads_t **out)
492 return notmuch_query_search_threads(query, out);
496 notmuch_query_search_threads (notmuch_query_t *query,
497 notmuch_threads_t **out)
499 notmuch_threads_t *threads;
500 notmuch_messages_t *messages;
501 notmuch_status_t status;
503 threads = talloc (query, notmuch_threads_t);
505 return NOTMUCH_STATUS_OUT_OF_MEMORY;
506 threads->doc_ids = NULL;
507 talloc_set_destructor (threads, _notmuch_threads_destructor);
509 threads->query = query;
511 status = notmuch_query_search_messages (query, &messages);
513 talloc_free (threads);
517 threads->doc_ids = g_array_new (false, false, sizeof (unsigned int));
518 while (notmuch_messages_valid (messages)) {
519 unsigned int doc_id = _notmuch_mset_messages_get_doc_id (messages);
520 g_array_append_val (threads->doc_ids, doc_id);
521 notmuch_messages_move_to_next (messages);
523 threads->doc_id_pos = 0;
525 talloc_free (messages);
527 if (! _notmuch_doc_id_set_init (threads, &threads->match_set,
529 talloc_free (threads);
530 return NOTMUCH_STATUS_OUT_OF_MEMORY;
534 return NOTMUCH_STATUS_SUCCESS;
538 notmuch_query_destroy (notmuch_query_t *query)
544 notmuch_threads_valid (notmuch_threads_t *threads)
551 while (threads->doc_id_pos < threads->doc_ids->len) {
552 doc_id = g_array_index (threads->doc_ids, unsigned int,
553 threads->doc_id_pos);
554 if (_notmuch_doc_id_set_contains (&threads->match_set, doc_id))
557 threads->doc_id_pos++;
560 return threads->doc_id_pos < threads->doc_ids->len;
564 notmuch_threads_get (notmuch_threads_t *threads)
568 if (! notmuch_threads_valid (threads))
571 doc_id = g_array_index (threads->doc_ids, unsigned int,
572 threads->doc_id_pos);
573 return _notmuch_thread_create (threads->query,
574 threads->query->notmuch,
577 threads->query->exclude_terms,
578 threads->query->omit_excluded,
579 threads->query->sort);
583 notmuch_threads_move_to_next (notmuch_threads_t *threads)
585 threads->doc_id_pos++;
589 notmuch_threads_destroy (notmuch_threads_t *threads)
591 talloc_free (threads);
595 notmuch_query_count_messages_st (notmuch_query_t *query, unsigned *count_out)
597 return notmuch_query_count_messages (query, count_out);
601 notmuch_query_count_messages (notmuch_query_t *query, unsigned *count_out)
603 return _notmuch_query_count_documents (query, "mail", count_out);
607 _notmuch_query_count_documents (notmuch_query_t *query, const char *type, unsigned *count_out)
609 notmuch_database_t *notmuch = query->notmuch;
610 const char *query_string = query->query_string;
611 Xapian::doccount count = 0;
612 notmuch_status_t status;
614 status = _notmuch_query_ensure_parsed (query);
619 Xapian::Enquire enquire (*notmuch->xapian_db);
620 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
621 _find_prefix ("type"),
623 Xapian::Query final_query, exclude_query;
626 if (strcmp (query_string, "") == 0 ||
627 strcmp (query_string, "*") == 0)
629 final_query = mail_query;
631 final_query = Xapian::Query (Xapian::Query::OP_AND,
632 mail_query, query->xapian_query);
635 exclude_query = _notmuch_exclude_tags (query);
637 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
638 final_query, exclude_query);
640 enquire.set_weighting_scheme(Xapian::BoolWeight());
641 enquire.set_docid_order(Xapian::Enquire::ASCENDING);
643 if (_debug_query ()) {
644 fprintf (stderr, "Exclude query is:\n%s\n",
645 exclude_query.get_description ().c_str ());
646 fprintf (stderr, "Final query is:\n%s\n",
647 final_query.get_description ().c_str ());
650 enquire.set_query (final_query);
653 * Set the checkatleast parameter to the number of documents
654 * in the database to make get_matches_estimated() exact.
655 * Set the max parameter to 0 to avoid fetching documents we will discard.
657 mset = enquire.get_mset (0, 0,
658 notmuch->xapian_db->get_doccount ());
660 count = mset.get_matches_estimated();
662 } catch (const Xapian::Error &error) {
663 _notmuch_database_log (notmuch,
664 "A Xapian exception occurred performing query: %s\n",
665 error.get_msg().c_str());
666 _notmuch_database_log_append (notmuch,
667 "Query string was: %s\n",
668 query->query_string);
669 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
673 return NOTMUCH_STATUS_SUCCESS;
677 notmuch_query_count_threads_st (notmuch_query_t *query, unsigned *count)
679 return notmuch_query_count_threads (query, count);
683 notmuch_query_count_threads (notmuch_query_t *query, unsigned *count)
685 notmuch_messages_t *messages;
688 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
691 query->sort = NOTMUCH_SORT_UNSORTED;
692 ret = notmuch_query_search_messages (query, &messages);
696 if (messages == NULL)
697 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
699 hash = g_hash_table_new_full (g_str_hash, g_str_equal, NULL, NULL);
701 talloc_free (messages);
702 return NOTMUCH_STATUS_OUT_OF_MEMORY;
705 while (notmuch_messages_valid (messages)) {
706 notmuch_message_t *message = notmuch_messages_get (messages);
707 const char *thread_id = notmuch_message_get_thread_id (message);
708 char *thread_id_copy = talloc_strdup (messages, thread_id);
709 if (unlikely (thread_id_copy == NULL)) {
710 notmuch_message_destroy (message);
711 ret = NOTMUCH_STATUS_OUT_OF_MEMORY;
714 g_hash_table_insert (hash, thread_id_copy, NULL);
715 notmuch_message_destroy (message);
716 notmuch_messages_move_to_next (messages);
719 *count = g_hash_table_size (hash);
722 g_hash_table_unref (hash);
723 talloc_free (messages);
729 notmuch_query_get_database (const notmuch_query_t *query)
731 return query->notmuch;