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;
32 notmuch_bool_t parsed;
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 visible _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);
187 /* The following is not ideal error handling, but to avoid
188 * breaking the ABI, we can live with it for now. In particular at
189 * least in the notmuch CLI, any syntax error in the query is
190 * caught in a later call to _notmuch_query_ensure_parsed with a
193 * TODO: add status return to this function.
198 term = talloc_asprintf (query, "%s%s", _find_prefix ("tag"), tag);
199 if (query->terms.count(term) != 0)
200 return; /* XXX report ignoring exclude? */
202 _notmuch_string_list_append (query->exclude_terms, term);
205 /* We end up having to call the destructors explicitly because we had
206 * to use "placement new" in order to initialize C++ objects within a
207 * block that we allocated with talloc. So C++ is making talloc
208 * slightly less simple to use, (we wouldn't need
209 * talloc_set_destructor at all otherwise).
212 _notmuch_messages_destructor (notmuch_mset_messages_t *messages)
214 messages->iterator.~MSetIterator ();
215 messages->iterator_end.~MSetIterator ();
220 /* Return a query that matches messages with the excluded tags
221 * registered with query. The caller of this function has to combine the returned
222 * query appropriately.*/
224 _notmuch_exclude_tags (notmuch_query_t *query)
226 Xapian::Query exclude_query = Xapian::Query::MatchNothing;
228 for (notmuch_string_node_t *term = query->exclude_terms->head; term;
230 exclude_query = Xapian::Query (Xapian::Query::OP_OR,
231 exclude_query, Xapian::Query (term->string));
233 return exclude_query;
237 notmuch_query_search_messages (notmuch_query_t *query)
239 notmuch_status_t status;
240 notmuch_messages_t *messages;
241 status = notmuch_query_search_messages_st (query, &messages);
249 notmuch_query_search_messages_st (notmuch_query_t *query,
250 notmuch_messages_t **out)
252 return _notmuch_query_search_documents (query, "mail", out);
256 _notmuch_query_search_documents (notmuch_query_t *query,
258 notmuch_messages_t **out)
260 notmuch_database_t *notmuch = query->notmuch;
261 const char *query_string = query->query_string;
262 notmuch_mset_messages_t *messages;
263 notmuch_status_t status;
265 status = _notmuch_query_ensure_parsed (query);
269 messages = talloc (query, notmuch_mset_messages_t);
270 if (unlikely (messages == NULL))
271 return NOTMUCH_STATUS_OUT_OF_MEMORY;
275 messages->base.is_of_list_type = FALSE;
276 messages->base.iterator = NULL;
277 messages->notmuch = notmuch;
278 new (&messages->iterator) Xapian::MSetIterator ();
279 new (&messages->iterator_end) Xapian::MSetIterator ();
281 talloc_set_destructor (messages, _notmuch_messages_destructor);
283 Xapian::Enquire enquire (*notmuch->xapian_db);
284 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
285 _find_prefix ("type"),
287 Xapian::Query final_query, exclude_query;
289 Xapian::MSetIterator iterator;
291 if (strcmp (query_string, "") == 0 ||
292 strcmp (query_string, "*") == 0)
294 final_query = mail_query;
296 final_query = Xapian::Query (Xapian::Query::OP_AND,
297 mail_query, query->xapian_query);
299 messages->base.excluded_doc_ids = NULL;
301 if ((query->omit_excluded != NOTMUCH_EXCLUDE_FALSE) && (query->exclude_terms)) {
302 exclude_query = _notmuch_exclude_tags (query);
304 if (query->omit_excluded == NOTMUCH_EXCLUDE_TRUE ||
305 query->omit_excluded == NOTMUCH_EXCLUDE_ALL)
307 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
308 final_query, exclude_query);
309 } else { /* NOTMUCH_EXCLUDE_FLAG */
310 exclude_query = Xapian::Query (Xapian::Query::OP_AND,
311 exclude_query, final_query);
313 enquire.set_weighting_scheme (Xapian::BoolWeight());
314 enquire.set_query (exclude_query);
316 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
318 GArray *excluded_doc_ids = g_array_new (FALSE, FALSE, sizeof (unsigned int));
320 for (iterator = mset.begin (); iterator != mset.end (); iterator++) {
321 unsigned int doc_id = *iterator;
322 g_array_append_val (excluded_doc_ids, doc_id);
324 messages->base.excluded_doc_ids = talloc (messages, _notmuch_doc_id_set);
325 _notmuch_doc_id_set_init (query, messages->base.excluded_doc_ids,
327 g_array_unref (excluded_doc_ids);
332 enquire.set_weighting_scheme (Xapian::BoolWeight());
334 switch (query->sort) {
335 case NOTMUCH_SORT_OLDEST_FIRST:
336 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, FALSE);
338 case NOTMUCH_SORT_NEWEST_FIRST:
339 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, TRUE);
341 case NOTMUCH_SORT_MESSAGE_ID:
342 enquire.set_sort_by_value (NOTMUCH_VALUE_MESSAGE_ID, FALSE);
344 case NOTMUCH_SORT_UNSORTED:
348 if (_debug_query ()) {
349 fprintf (stderr, "Exclude query is:\n%s\n",
350 exclude_query.get_description ().c_str ());
351 fprintf (stderr, "Final query is:\n%s\n",
352 final_query.get_description ().c_str ());
355 enquire.set_query (final_query);
357 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
359 messages->iterator = mset.begin ();
360 messages->iterator_end = mset.end ();
362 *out = &messages->base;
363 return NOTMUCH_STATUS_SUCCESS;
365 } catch (const Xapian::Error &error) {
366 _notmuch_database_log (notmuch,
367 "A Xapian exception occurred performing query: %s\n",
368 error.get_msg().c_str());
369 _notmuch_database_log_append (notmuch,
370 "Query string was: %s\n",
371 query->query_string);
373 notmuch->exception_reported = TRUE;
374 talloc_free (messages);
375 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
380 _notmuch_mset_messages_valid (notmuch_messages_t *messages)
382 notmuch_mset_messages_t *mset_messages;
384 mset_messages = (notmuch_mset_messages_t *) messages;
386 return (mset_messages->iterator != mset_messages->iterator_end);
390 _notmuch_mset_messages_get_doc_id (notmuch_messages_t *messages)
392 notmuch_mset_messages_t *mset_messages;
394 mset_messages = (notmuch_mset_messages_t *) messages;
396 if (! _notmuch_mset_messages_valid (&mset_messages->base))
399 return *mset_messages->iterator;
403 _notmuch_mset_messages_get (notmuch_messages_t *messages)
405 notmuch_message_t *message;
406 Xapian::docid doc_id;
407 notmuch_private_status_t status;
408 notmuch_mset_messages_t *mset_messages;
410 mset_messages = (notmuch_mset_messages_t *) messages;
412 if (! _notmuch_mset_messages_valid (&mset_messages->base))
415 doc_id = *mset_messages->iterator;
417 message = _notmuch_message_create (mset_messages,
418 mset_messages->notmuch, doc_id,
421 if (message == NULL &&
422 status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
424 INTERNAL_ERROR ("a messages iterator contains a non-existent document ID.\n");
427 if (messages->excluded_doc_ids &&
428 _notmuch_doc_id_set_contains (messages->excluded_doc_ids, doc_id))
429 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, TRUE);
435 _notmuch_mset_messages_move_to_next (notmuch_messages_t *messages)
437 notmuch_mset_messages_t *mset_messages;
439 mset_messages = (notmuch_mset_messages_t *) messages;
441 mset_messages->iterator++;
444 static notmuch_bool_t
445 _notmuch_doc_id_set_init (void *ctx,
446 notmuch_doc_id_set_t *doc_ids,
449 unsigned int max = 0;
450 unsigned char *bitmap;
452 for (unsigned int i = 0; i < arr->len; i++)
453 max = MAX(max, g_array_index (arr, unsigned int, i));
454 bitmap = talloc_zero_array (ctx, unsigned char, DOCIDSET_WORD(max) + 1);
459 doc_ids->bitmap = bitmap;
460 doc_ids->bound = max + 1;
462 for (unsigned int i = 0; i < arr->len; i++) {
463 unsigned int doc_id = g_array_index (arr, unsigned int, i);
464 bitmap[DOCIDSET_WORD(doc_id)] |= 1 << DOCIDSET_BIT(doc_id);
471 _notmuch_doc_id_set_contains (notmuch_doc_id_set_t *doc_ids,
474 if (doc_id >= doc_ids->bound)
476 return doc_ids->bitmap[DOCIDSET_WORD(doc_id)] & (1 << DOCIDSET_BIT(doc_id));
480 _notmuch_doc_id_set_remove (notmuch_doc_id_set_t *doc_ids,
483 if (doc_id < doc_ids->bound)
484 doc_ids->bitmap[DOCIDSET_WORD(doc_id)] &= ~(1 << DOCIDSET_BIT(doc_id));
487 /* Glib objects force use to use a talloc destructor as well, (but not
488 * nearly as ugly as the for messages due to C++ objects). At
489 * this point, I'd really like to have some talloc-friendly
490 * equivalents for the few pieces of glib that I'm using. */
492 _notmuch_threads_destructor (notmuch_threads_t *threads)
494 if (threads->doc_ids)
495 g_array_unref (threads->doc_ids);
501 notmuch_query_search_threads_st (notmuch_query_t *query, notmuch_threads_t **out)
503 return notmuch_query_search_threads(query, out);
507 notmuch_query_search_threads (notmuch_query_t *query,
508 notmuch_threads_t **out)
510 notmuch_threads_t *threads;
511 notmuch_messages_t *messages;
512 notmuch_status_t status;
514 threads = talloc (query, notmuch_threads_t);
516 return NOTMUCH_STATUS_OUT_OF_MEMORY;
517 threads->doc_ids = NULL;
518 talloc_set_destructor (threads, _notmuch_threads_destructor);
520 threads->query = query;
522 status = notmuch_query_search_messages_st (query, &messages);
524 talloc_free (threads);
528 threads->doc_ids = g_array_new (FALSE, FALSE, sizeof (unsigned int));
529 while (notmuch_messages_valid (messages)) {
530 unsigned int doc_id = _notmuch_mset_messages_get_doc_id (messages);
531 g_array_append_val (threads->doc_ids, doc_id);
532 notmuch_messages_move_to_next (messages);
534 threads->doc_id_pos = 0;
536 talloc_free (messages);
538 if (! _notmuch_doc_id_set_init (threads, &threads->match_set,
540 talloc_free (threads);
541 return NOTMUCH_STATUS_OUT_OF_MEMORY;
545 return NOTMUCH_STATUS_SUCCESS;
549 notmuch_query_destroy (notmuch_query_t *query)
555 notmuch_threads_valid (notmuch_threads_t *threads)
562 while (threads->doc_id_pos < threads->doc_ids->len) {
563 doc_id = g_array_index (threads->doc_ids, unsigned int,
564 threads->doc_id_pos);
565 if (_notmuch_doc_id_set_contains (&threads->match_set, doc_id))
568 threads->doc_id_pos++;
571 return threads->doc_id_pos < threads->doc_ids->len;
575 notmuch_threads_get (notmuch_threads_t *threads)
579 if (! notmuch_threads_valid (threads))
582 doc_id = g_array_index (threads->doc_ids, unsigned int,
583 threads->doc_id_pos);
584 return _notmuch_thread_create (threads->query,
585 threads->query->notmuch,
588 threads->query->exclude_terms,
589 threads->query->omit_excluded,
590 threads->query->sort);
594 notmuch_threads_move_to_next (notmuch_threads_t *threads)
596 threads->doc_id_pos++;
600 notmuch_threads_destroy (notmuch_threads_t *threads)
602 talloc_free (threads);
606 notmuch_query_count_messages (notmuch_query_t *query)
608 notmuch_status_t status;
611 status = notmuch_query_count_messages_st (query, &count);
612 return status ? 0 : count;
616 notmuch_query_count_messages_st (notmuch_query_t *query, unsigned *count_out)
618 return _notmuch_query_count_documents (query, "mail", count_out);
622 _notmuch_query_count_documents (notmuch_query_t *query, const char *type, unsigned *count_out)
624 notmuch_database_t *notmuch = query->notmuch;
625 const char *query_string = query->query_string;
626 Xapian::doccount count = 0;
627 notmuch_status_t status;
629 status = _notmuch_query_ensure_parsed (query);
634 Xapian::Enquire enquire (*notmuch->xapian_db);
635 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
636 _find_prefix ("type"),
638 Xapian::Query final_query, exclude_query;
641 if (strcmp (query_string, "") == 0 ||
642 strcmp (query_string, "*") == 0)
644 final_query = mail_query;
646 final_query = Xapian::Query (Xapian::Query::OP_AND,
647 mail_query, query->xapian_query);
650 exclude_query = _notmuch_exclude_tags (query);
652 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
653 final_query, exclude_query);
655 enquire.set_weighting_scheme(Xapian::BoolWeight());
656 enquire.set_docid_order(Xapian::Enquire::ASCENDING);
658 if (_debug_query ()) {
659 fprintf (stderr, "Exclude query is:\n%s\n",
660 exclude_query.get_description ().c_str ());
661 fprintf (stderr, "Final query is:\n%s\n",
662 final_query.get_description ().c_str ());
665 enquire.set_query (final_query);
668 * Set the checkatleast parameter to the number of documents
669 * in the database to make get_matches_estimated() exact.
670 * Set the max parameter to 0 to avoid fetching documents we will discard.
672 mset = enquire.get_mset (0, 0,
673 notmuch->xapian_db->get_doccount ());
675 count = mset.get_matches_estimated();
677 } catch (const Xapian::Error &error) {
678 _notmuch_database_log (notmuch,
679 "A Xapian exception occurred performing query: %s\n",
680 error.get_msg().c_str());
681 _notmuch_database_log_append (notmuch,
682 "Query string was: %s\n",
683 query->query_string);
684 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
688 return NOTMUCH_STATUS_SUCCESS;
692 notmuch_query_count_threads (notmuch_query_t *query)
694 notmuch_status_t status;
697 status = notmuch_query_count_threads_st (query, &count);
698 return status ? 0 : count;
702 notmuch_query_count_threads_st (notmuch_query_t *query, unsigned *count)
704 notmuch_messages_t *messages;
707 notmuch_status_t ret = NOTMUCH_STATUS_SUCCESS;
710 query->sort = NOTMUCH_SORT_UNSORTED;
711 ret = notmuch_query_search_messages_st (query, &messages);
715 if (messages == NULL)
716 return NOTMUCH_STATUS_XAPIAN_EXCEPTION;
718 hash = g_hash_table_new_full (g_str_hash, g_str_equal, NULL, NULL);
720 talloc_free (messages);
721 return NOTMUCH_STATUS_OUT_OF_MEMORY;
724 while (notmuch_messages_valid (messages)) {
725 notmuch_message_t *message = notmuch_messages_get (messages);
726 const char *thread_id = notmuch_message_get_thread_id (message);
727 char *thread_id_copy = talloc_strdup (messages, thread_id);
728 if (unlikely (thread_id_copy == NULL)) {
729 notmuch_message_destroy (message);
730 ret = NOTMUCH_STATUS_OUT_OF_MEMORY;
733 g_hash_table_insert (hash, thread_id_copy, NULL);
734 notmuch_message_destroy (message);
735 notmuch_messages_move_to_next (messages);
738 *count = g_hash_table_size (hash);
741 g_hash_table_unref (hash);
742 talloc_free (messages);
748 notmuch_query_get_database (const notmuch_query_t *query)
750 return query->notmuch;