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 (NULL, 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_database_t *notmuch = query->notmuch;
178 const char *query_string = query->query_string;
179 notmuch_mset_messages_t *messages;
181 messages = talloc (query, notmuch_mset_messages_t);
182 if (unlikely (messages == NULL))
187 messages->base.is_of_list_type = FALSE;
188 messages->base.iterator = NULL;
189 messages->notmuch = notmuch;
190 new (&messages->iterator) Xapian::MSetIterator ();
191 new (&messages->iterator_end) Xapian::MSetIterator ();
193 talloc_set_destructor (messages, _notmuch_messages_destructor);
195 Xapian::Enquire enquire (*notmuch->xapian_db);
196 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
197 _find_prefix ("type"),
199 Xapian::Query string_query, final_query, exclude_query;
201 Xapian::MSetIterator iterator;
202 unsigned int flags = (Xapian::QueryParser::FLAG_BOOLEAN |
203 Xapian::QueryParser::FLAG_PHRASE |
204 Xapian::QueryParser::FLAG_LOVEHATE |
205 Xapian::QueryParser::FLAG_BOOLEAN_ANY_CASE |
206 Xapian::QueryParser::FLAG_WILDCARD |
207 Xapian::QueryParser::FLAG_PURE_NOT);
209 if (strcmp (query_string, "") == 0 ||
210 strcmp (query_string, "*") == 0)
212 final_query = mail_query;
214 string_query = notmuch->query_parser->
215 parse_query (query_string, flags);
216 final_query = Xapian::Query (Xapian::Query::OP_AND,
217 mail_query, string_query);
219 messages->base.excluded_doc_ids = NULL;
221 if ((query->omit_excluded != NOTMUCH_EXCLUDE_FALSE) && (query->exclude_terms)) {
222 exclude_query = _notmuch_exclude_tags (query, final_query);
224 if (query->omit_excluded == NOTMUCH_EXCLUDE_TRUE ||
225 query->omit_excluded == NOTMUCH_EXCLUDE_ALL)
227 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
228 final_query, exclude_query);
229 } else { /* NOTMUCH_EXCLUDE_FLAG */
230 exclude_query = Xapian::Query (Xapian::Query::OP_AND,
231 exclude_query, final_query);
233 enquire.set_weighting_scheme (Xapian::BoolWeight());
234 enquire.set_query (exclude_query);
236 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
238 GArray *excluded_doc_ids = g_array_new (FALSE, FALSE, sizeof (unsigned int));
240 for (iterator = mset.begin (); iterator != mset.end (); iterator++) {
241 unsigned int doc_id = *iterator;
242 g_array_append_val (excluded_doc_ids, doc_id);
244 messages->base.excluded_doc_ids = talloc (messages, _notmuch_doc_id_set);
245 _notmuch_doc_id_set_init (query, messages->base.excluded_doc_ids,
247 g_array_unref (excluded_doc_ids);
252 enquire.set_weighting_scheme (Xapian::BoolWeight());
254 switch (query->sort) {
255 case NOTMUCH_SORT_OLDEST_FIRST:
256 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, FALSE);
258 case NOTMUCH_SORT_NEWEST_FIRST:
259 enquire.set_sort_by_value (NOTMUCH_VALUE_TIMESTAMP, TRUE);
261 case NOTMUCH_SORT_MESSAGE_ID:
262 enquire.set_sort_by_value (NOTMUCH_VALUE_MESSAGE_ID, FALSE);
264 case NOTMUCH_SORT_UNSORTED:
268 if (_debug_query ()) {
269 fprintf (stderr, "Exclude query is:\n%s\n",
270 exclude_query.get_description ().c_str ());
271 fprintf (stderr, "Final query is:\n%s\n",
272 final_query.get_description ().c_str ());
275 enquire.set_query (final_query);
277 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
279 messages->iterator = mset.begin ();
280 messages->iterator_end = mset.end ();
282 return &messages->base;
284 } catch (const Xapian::Error &error) {
285 fprintf (stderr, "A Xapian exception occurred performing query: %s\n",
286 error.get_msg().c_str());
287 fprintf (stderr, "Query string was: %s\n", query->query_string);
288 notmuch->exception_reported = TRUE;
289 talloc_free (messages);
295 _notmuch_mset_messages_valid (notmuch_messages_t *messages)
297 notmuch_mset_messages_t *mset_messages;
299 mset_messages = (notmuch_mset_messages_t *) messages;
301 return (mset_messages->iterator != mset_messages->iterator_end);
305 _notmuch_mset_messages_get_doc_id (notmuch_messages_t *messages)
307 notmuch_mset_messages_t *mset_messages;
309 mset_messages = (notmuch_mset_messages_t *) messages;
311 if (! _notmuch_mset_messages_valid (&mset_messages->base))
314 return *mset_messages->iterator;
318 _notmuch_mset_messages_get (notmuch_messages_t *messages)
320 notmuch_message_t *message;
321 Xapian::docid doc_id;
322 notmuch_private_status_t status;
323 notmuch_mset_messages_t *mset_messages;
325 mset_messages = (notmuch_mset_messages_t *) messages;
327 if (! _notmuch_mset_messages_valid (&mset_messages->base))
330 doc_id = *mset_messages->iterator;
332 message = _notmuch_message_create (mset_messages,
333 mset_messages->notmuch, doc_id,
336 if (message == NULL &&
337 status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND)
339 INTERNAL_ERROR ("a messages iterator contains a non-existent document ID.\n");
342 if (messages->excluded_doc_ids &&
343 _notmuch_doc_id_set_contains (messages->excluded_doc_ids, doc_id))
344 notmuch_message_set_flag (message, NOTMUCH_MESSAGE_FLAG_EXCLUDED, TRUE);
350 _notmuch_mset_messages_move_to_next (notmuch_messages_t *messages)
352 notmuch_mset_messages_t *mset_messages;
354 mset_messages = (notmuch_mset_messages_t *) messages;
356 mset_messages->iterator++;
359 static notmuch_bool_t
360 _notmuch_doc_id_set_init (void *ctx,
361 notmuch_doc_id_set_t *doc_ids,
364 unsigned int max = 0;
365 unsigned char *bitmap;
367 for (unsigned int i = 0; i < arr->len; i++)
368 max = MAX(max, g_array_index (arr, unsigned int, i));
369 bitmap = talloc_zero_array (ctx, unsigned char, DOCIDSET_WORD(max) + 1);
374 doc_ids->bitmap = bitmap;
375 doc_ids->bound = max + 1;
377 for (unsigned int i = 0; i < arr->len; i++) {
378 unsigned int doc_id = g_array_index (arr, unsigned int, i);
379 bitmap[DOCIDSET_WORD(doc_id)] |= 1 << DOCIDSET_BIT(doc_id);
386 _notmuch_doc_id_set_contains (notmuch_doc_id_set_t *doc_ids,
389 if (doc_id >= doc_ids->bound)
391 return doc_ids->bitmap[DOCIDSET_WORD(doc_id)] & (1 << DOCIDSET_BIT(doc_id));
395 _notmuch_doc_id_set_remove (notmuch_doc_id_set_t *doc_ids,
398 if (doc_id < doc_ids->bound)
399 doc_ids->bitmap[DOCIDSET_WORD(doc_id)] &= ~(1 << DOCIDSET_BIT(doc_id));
402 /* Glib objects force use to use a talloc destructor as well, (but not
403 * nearly as ugly as the for messages due to C++ objects). At
404 * this point, I'd really like to have some talloc-friendly
405 * equivalents for the few pieces of glib that I'm using. */
407 _notmuch_threads_destructor (notmuch_threads_t *threads)
409 if (threads->doc_ids)
410 g_array_unref (threads->doc_ids);
416 notmuch_query_search_threads (notmuch_query_t *query)
418 notmuch_threads_t *threads;
419 notmuch_messages_t *messages;
421 threads = talloc (query, notmuch_threads_t);
424 threads->doc_ids = NULL;
425 talloc_set_destructor (threads, _notmuch_threads_destructor);
427 threads->query = query;
429 messages = notmuch_query_search_messages (query);
430 if (messages == NULL) {
431 talloc_free (threads);
435 threads->doc_ids = g_array_new (FALSE, FALSE, sizeof (unsigned int));
436 while (notmuch_messages_valid (messages)) {
437 unsigned int doc_id = _notmuch_mset_messages_get_doc_id (messages);
438 g_array_append_val (threads->doc_ids, doc_id);
439 notmuch_messages_move_to_next (messages);
441 threads->doc_id_pos = 0;
443 talloc_free (messages);
445 if (! _notmuch_doc_id_set_init (threads, &threads->match_set,
447 talloc_free (threads);
455 notmuch_query_destroy (notmuch_query_t *query)
461 notmuch_threads_valid (notmuch_threads_t *threads)
465 while (threads->doc_id_pos < threads->doc_ids->len) {
466 doc_id = g_array_index (threads->doc_ids, unsigned int,
467 threads->doc_id_pos);
468 if (_notmuch_doc_id_set_contains (&threads->match_set, doc_id))
471 threads->doc_id_pos++;
474 return threads->doc_id_pos < threads->doc_ids->len;
478 notmuch_threads_get (notmuch_threads_t *threads)
482 if (! notmuch_threads_valid (threads))
485 doc_id = g_array_index (threads->doc_ids, unsigned int,
486 threads->doc_id_pos);
487 return _notmuch_thread_create (threads->query,
488 threads->query->notmuch,
491 threads->query->exclude_terms,
492 threads->query->omit_excluded,
493 threads->query->sort);
497 notmuch_threads_move_to_next (notmuch_threads_t *threads)
499 threads->doc_id_pos++;
503 notmuch_threads_destroy (notmuch_threads_t *threads)
505 talloc_free (threads);
509 notmuch_query_count_messages (notmuch_query_t *query)
511 notmuch_database_t *notmuch = query->notmuch;
512 const char *query_string = query->query_string;
513 Xapian::doccount count = 0;
516 Xapian::Enquire enquire (*notmuch->xapian_db);
517 Xapian::Query mail_query (talloc_asprintf (query, "%s%s",
518 _find_prefix ("type"),
520 Xapian::Query string_query, final_query, exclude_query;
522 unsigned int flags = (Xapian::QueryParser::FLAG_BOOLEAN |
523 Xapian::QueryParser::FLAG_PHRASE |
524 Xapian::QueryParser::FLAG_LOVEHATE |
525 Xapian::QueryParser::FLAG_BOOLEAN_ANY_CASE |
526 Xapian::QueryParser::FLAG_WILDCARD |
527 Xapian::QueryParser::FLAG_PURE_NOT);
529 if (strcmp (query_string, "") == 0 ||
530 strcmp (query_string, "*") == 0)
532 final_query = mail_query;
534 string_query = notmuch->query_parser->
535 parse_query (query_string, flags);
536 final_query = Xapian::Query (Xapian::Query::OP_AND,
537 mail_query, string_query);
540 exclude_query = _notmuch_exclude_tags (query, final_query);
542 final_query = Xapian::Query (Xapian::Query::OP_AND_NOT,
543 final_query, exclude_query);
545 enquire.set_weighting_scheme(Xapian::BoolWeight());
546 enquire.set_docid_order(Xapian::Enquire::ASCENDING);
548 if (_debug_query ()) {
549 fprintf (stderr, "Exclude query is:\n%s\n",
550 exclude_query.get_description ().c_str ());
551 fprintf (stderr, "Final query is:\n%s\n",
552 final_query.get_description ().c_str ());
555 enquire.set_query (final_query);
557 mset = enquire.get_mset (0, notmuch->xapian_db->get_doccount ());
559 count = mset.get_matches_estimated();
561 } catch (const Xapian::Error &error) {
562 fprintf (stderr, "A Xapian exception occurred: %s\n",
563 error.get_msg().c_str());
564 fprintf (stderr, "Query string was: %s\n", query->query_string);
571 notmuch_query_count_threads (notmuch_query_t *query)
573 notmuch_messages_t *messages;
579 query->sort = NOTMUCH_SORT_UNSORTED;
580 messages = notmuch_query_search_messages (query);
582 if (messages == NULL)
585 hash = g_hash_table_new_full (g_str_hash, g_str_equal, NULL, NULL);
587 talloc_free (messages);
591 while (notmuch_messages_valid (messages)) {
592 notmuch_message_t *message = notmuch_messages_get (messages);
593 const char *thread_id = notmuch_message_get_thread_id (message);
594 char *thread_id_copy = talloc_strdup (messages, thread_id);
595 if (unlikely (thread_id_copy == NULL)) {
596 notmuch_message_destroy (message);
600 g_hash_table_insert (hash, thread_id_copy, NULL);
601 notmuch_message_destroy (message);
602 notmuch_messages_move_to_next (messages);
605 count = g_hash_table_size (hash);
608 g_hash_table_unref (hash);
609 talloc_free (messages);