tracker r1772 - in branches/indexer-split: . src/trackerd
- From: mr svn gnome org
- To: svn-commits-list gnome org
- Subject: tracker r1772 - in branches/indexer-split: . src/trackerd
- Date: Thu, 26 Jun 2008 11:38:22 +0000 (UTC)
Author: mr
Date: Thu Jun 26 11:38:21 2008
New Revision: 1772
URL: http://svn.gnome.org/viewvc/tracker?rev=1772&view=rev
Log:
* src/trackerd/tracker-daemon.c:
* src/trackerd/tracker-db.c:
* src/trackerd/tracker-indexer.[ch]:
* src/trackerd/tracker-main.[ch]:
* src/trackerd/tracker-monitor.c:
* src/trackerd/tracker-query-tree.[ch]:
* src/trackerd/tracker-search.[ch]: Refactored the TrackerIndexer
code and API so we now use GObject and do proper type checking in
the query tree code. The indexer API has been cut down to cater
JUST the daemons needs here. All the indexing code is in the
indexer instead now.
Modified:
branches/indexer-split/ChangeLog
branches/indexer-split/src/trackerd/tracker-daemon.c
branches/indexer-split/src/trackerd/tracker-db.c
branches/indexer-split/src/trackerd/tracker-indexer.c
branches/indexer-split/src/trackerd/tracker-indexer.h
branches/indexer-split/src/trackerd/tracker-main.c
branches/indexer-split/src/trackerd/tracker-main.h
branches/indexer-split/src/trackerd/tracker-monitor.c
branches/indexer-split/src/trackerd/tracker-query-tree.c
branches/indexer-split/src/trackerd/tracker-query-tree.h
branches/indexer-split/src/trackerd/tracker-search.c
branches/indexer-split/src/trackerd/tracker-search.h
Modified: branches/indexer-split/src/trackerd/tracker-daemon.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-daemon.c (original)
+++ branches/indexer-split/src/trackerd/tracker-daemon.c Thu Jun 26 11:38:21 2008
@@ -22,6 +22,7 @@
#include "config.h"
#include <stdio.h>
+#include <stdlib.h>
#include <string.h>
#include <libtracker-common/tracker-log.h>
Modified: branches/indexer-split/src/trackerd/tracker-db.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-db.c (original)
+++ branches/indexer-split/src/trackerd/tracker-db.c Thu Jun 26 11:38:21 2008
@@ -24,6 +24,7 @@
#include "config.h"
#include <string.h>
+#include <stdlib.h>
#include <fcntl.h>
#include <zlib.h>
@@ -908,14 +909,16 @@
/* Delete duds */
if (duds) {
- GSList *words, *w;
- Indexer *indexer;
+ TrackerIndexer *indexer;
+ GSList *words, *w;
words = tracker_query_tree_get_words (tree);
indexer = tracker_query_tree_get_indexer (tree);
for (w = words; w; w = w->next) {
- tracker_remove_dud_hits (indexer, (const gchar *) w->data, duds);
+ tracker_indexer_remove_dud_hits (indexer,
+ (const gchar *) w->data,
+ duds);
}
g_slist_free (words);
Modified: branches/indexer-split/src/trackerd/tracker-indexer.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-indexer.c (original)
+++ branches/indexer-split/src/trackerd/tracker-indexer.c Thu Jun 26 11:38:21 2008
@@ -1,6 +1,8 @@
-/* Tracker - indexer and metadata database engine
+/* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8 -*- */
+/*
* Copyright (C) 2006, Mr Jamie McCracken (jamiemcc gnome org)
- *
+ * Copyright (C) 2008, Nokia
+
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License as published by the Free Software Foundation; either
@@ -58,461 +60,540 @@
#include "tracker-status.h"
/* Size of free block pool of inverted index */
-#define INDEXFBP 32
-#define SCORE_MULTIPLIER 100000
#define MAX_HIT_BUFFER 480000
-#define MAX_HITS_FOR_WORD 30000
#define MAX_INDEX_FILE_SIZE 2000000000
-#define CREATE_INDEX \
- "CREATE TABLE HitIndex (Word Text not null " \
- "unique, HitCount Integer, HitArraySize Integer, HitArray Blob);"
-
-extern Tracker *tracker;
-
-static gint merge_count = 0;
-static gint merge_processed = 0;
-
-struct Indexer_ {
- DEPOT *word_index; /* file hashtable handle for the word -> {serviceID, ServiceTypeID, Score} */
- GMutex *word_mutex;
- char *name;
- gboolean main_index;
- gboolean needs_merge; /* should new stuff be added directly or merged later on from a new index */
+#define TRACKER_INDEXER_GET_PRIVATE(o) (G_TYPE_INSTANCE_GET_PRIVATE ((o), TRACKER_TYPE_INDEXER, TrackerIndexerPrivate))
+
+typedef struct TrackerIndexerPrivate TrackerIndexerPrivate;
+
+struct TrackerIndexerPrivate {
+ /* File hashtable handle for the word -> {serviceID,
+ * ServiceTypeID, Score}.
+ */
+ TrackerConfig *config;
+
+ DEPOT *word_index;
+ GMutex *word_mutex;
+
+ gchar *name;
};
-static inline gint16
-get_score (WordDetails *details)
+static void tracker_indexer_class_init (TrackerIndexerClass *class);
+static void tracker_indexer_init (TrackerIndexer *tree);
+static void tracker_indexer_finalize (GObject *object);
+static void tracker_indexer_set_property (GObject *object,
+ guint prop_id,
+ const GValue *value,
+ GParamSpec *pspec);
+static void tracker_indexer_get_property (GObject *object,
+ guint prop_id,
+ GValue *value,
+ GParamSpec *pspec);
+
+enum {
+ PROP_0,
+ PROP_NAME,
+ PROP_CONFIG
+};
+
+G_DEFINE_TYPE (TrackerIndexer, tracker_indexer, G_TYPE_OBJECT)
+
+static void
+tracker_indexer_class_init (TrackerIndexerClass *klass)
{
- unsigned char a[2];
+ GObjectClass *object_class = G_OBJECT_CLASS (klass);
- a[0] = (details->amalgamated >> 16) & 0xFF;
- a[1] = (details->amalgamated >> 8) & 0xFF;
+ object_class->finalize = tracker_indexer_finalize;
+ object_class->set_property = tracker_indexer_set_property;
+ object_class->get_property = tracker_indexer_get_property;
+
+ g_object_class_install_property (object_class,
+ PROP_NAME,
+ g_param_spec_string ("name",
+ "Name",
+ "Name",
+ NULL,
+ G_PARAM_READABLE));
+ g_object_class_install_property (object_class,
+ PROP_CONFIG,
+ g_param_spec_object ("config",
+ "Config",
+ "Config",
+ tracker_config_get_type (),
+ G_PARAM_READWRITE));
- return (gint16) (a[0] << 8) | (a[1]);
+ g_type_class_add_private (object_class, sizeof (TrackerIndexerPrivate));
}
-
-static inline guint8
-get_service_type (WordDetails *details)
+static void
+tracker_indexer_init (TrackerIndexer *indexer)
{
- return (details->amalgamated >> 24) & 0xFF;
-}
+ TrackerIndexerPrivate *priv;
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
+
+ priv->word_mutex = g_mutex_new ();
+}
-guint32
-tracker_indexer_calc_amalgamated (gint service, gint score)
+static void
+tracker_indexer_finalize (GObject *object)
{
- unsigned char a[4];
- gint16 score16;
- guint8 service_type;
+ TrackerIndexerPrivate *priv;
- if (score > 30000) {
- score16 = 30000;
- } else {
- score16 = (gint16) score;
- }
+ priv = TRACKER_INDEXER_GET_PRIVATE (object);
- service_type = (guint8) service;
+ g_free (priv->name);
- /* amalgamate and combine score and service_type into a single 32-bit int for compact storage */
- a[0] = service_type;
- a[1] = (score16 >> 8 ) & 0xFF ;
- a[2] = score16 & 0xFF ;
- a[3] = 0;
+ g_mutex_lock (priv->word_mutex);
- return (a[0] << 24) | (a[1] << 16) | (a[2] << 8) | a[3];
-}
+ if (!dpclose (priv->word_index)) {
+ g_message ("Index closure has failed, %s", dperrmsg (dpecode));
+ }
+ g_mutex_unlock (priv->word_mutex);
-static int
-get_preferred_bucket_count (Indexer *indexer)
-{
- gint result;
- gint bucket_ratio;
+ g_mutex_free (priv->word_mutex);
- bucket_ratio = tracker_config_get_bucket_ratio (tracker->config);
- result = dprnum (indexer->word_index);
+ if (priv->config) {
+ g_object_unref (priv->config);
+ }
- if (bucket_ratio < 1) {
- result /= 2;
- } else if (bucket_ratio > 3) {
- result *= 4;
- } else {
- result *= bucket_ratio;
- }
+ G_OBJECT_CLASS (tracker_indexer_parent_class)->finalize (object);
+}
- g_message ("Preferred bucket count is %d", result);
+static void
+tracker_indexer_set_property (GObject *object,
+ guint prop_id,
+ const GValue *value,
+ GParamSpec *pspec)
+{
+ switch (prop_id) {
+ case PROP_CONFIG:
+ tracker_indexer_set_config (TRACKER_INDEXER (object),
+ g_value_get_object (value));
+ break;
+ default:
+ G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
+ }
+}
- return result;
+static void
+tracker_indexer_get_property (GObject *object,
+ guint prop_id,
+ GValue *value,
+ GParamSpec *pspec)
+{
+ TrackerIndexerPrivate *priv;
+
+ priv = TRACKER_INDEXER_GET_PRIVATE (object);
+
+ switch (prop_id) {
+ case PROP_NAME:
+ g_value_set_string (value, priv->name);
+ break;
+ case PROP_CONFIG:
+ g_value_set_object (value, priv->config);
+ break;
+ default:
+ G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
+ }
}
-gboolean
-tracker_indexer_repair (const char *name)
+static inline gint16
+get_score (TrackerIndexerWordDetails *details)
{
- gchar *index_name;
- gboolean result = TRUE;
+ unsigned char a[2];
- index_name = g_build_filename (tracker_get_data_dir (),
- name,
- NULL);
- result = dprepair (index_name);
- g_free (index_name);
+ a[0] = (details->amalgamated >> 16) & 0xFF;
+ a[1] = (details->amalgamated >> 8) & 0xFF;
- return result;
+ return (gint16) (a[0] << 8) | (a[1]);
}
+static inline guint8
+get_service_type (TrackerIndexerWordDetails *details)
+{
+ return (details->amalgamated >> 24) & 0xFF;
+}
static inline DEPOT *
-open_index (const gchar *name)
+open_index (const gchar *name,
+ gint min_bucket_count,
+ gint max_bucket_count)
{
DEPOT *word_index = NULL;
- if (!name) return NULL;
+ if (!name) {
+ return NULL;
+ }
- g_message ("Opening index %s", name);
+ g_message ("Opening index:'%s'", name);
if (strstr (name, "tmp")) {
- word_index = dpopen (name, DP_OWRITER | DP_OCREAT | DP_ONOLCK,
- tracker_config_get_min_bucket_count (tracker->config));
- } else {
- word_index = dpopen (name, DP_OWRITER | DP_OCREAT | DP_ONOLCK,
- tracker_config_get_max_bucket_count (tracker->config));
+ word_index = dpopen (name,
+ DP_OWRITER | DP_OCREAT | DP_ONOLCK,
+ min_bucket_count);
+ } else {
+ word_index = dpopen (name,
+ DP_OWRITER | DP_OCREAT | DP_ONOLCK,
+ max_bucket_count);
}
if (!word_index) {
- g_critical ("%s index was not closed properly and caused error %s- attempting repair", name, dperrmsg (dpecode));
+ g_critical ("Index was not closed properly, index:'%s', %s",
+ name,
+ dperrmsg (dpecode));
+ g_message ("Attempting to repair...");
+
if (dprepair (name)) {
- word_index = dpopen (name, DP_OWRITER | DP_OCREAT | DP_ONOLCK,
- tracker_config_get_min_bucket_count (tracker->config));
+ word_index = dpopen (name,
+ DP_OWRITER | DP_OCREAT | DP_ONOLCK,
+ min_bucket_count);
} else {
- g_assert ("FATAL: index file is dead (suggest delete index file and restart trackerd)");
+ g_critical ("Index file is dead, it is suggested you remove "
+ "the indexe file:'%s' and restart trackerd",
+ name);
+ return NULL;
}
}
return word_index;
-
}
-
-static inline char *
-get_index_file (const char *name)
+static inline gchar *
+get_index_file (const gchar *name)
{
return g_build_filename (tracker_get_data_dir (), name, NULL);
}
-Indexer *
-tracker_indexer_open (const gchar *name, gboolean main_index)
+static inline gboolean
+has_word (TrackerIndexer *indexer,
+ const gchar *word)
{
- char *word_dir;
- DEPOT *word_index;
- Indexer *result;
-
- if (!name) return NULL;
-
- word_dir = get_index_file (name);
+ TrackerIndexerPrivate *priv;
+ gchar buffer[32];
+ gint count;
- word_index = open_index (word_dir);
-
- g_free (word_dir);
-
- result = g_new0 (Indexer, 1);
-
- result->main_index = main_index;
-
- result->needs_merge = FALSE;
-
- result->name = g_strdup (name);
-
- result->word_index = word_index;
-
- result->word_mutex = g_mutex_new ();
-
- dpsetalign (word_index , 8);
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
- /* re optimize database if bucket count < rec count */
+ g_mutex_lock (priv->word_mutex);
+ count = dpgetwb (priv->word_index, word, -1, 0, 32, buffer);
+ g_mutex_unlock (priv->word_mutex);
- int bucket_count, rec_count;
+ return count > 7;
+}
- bucket_count = dpbnum (result->word_index);
- rec_count = dprnum (result->word_index);
+static inline gint
+count_hit_size_for_word (TrackerIndexer *indexer,
+ const gchar *word)
+{
+ TrackerIndexerPrivate *priv;
+ gint tsiz;
- g_message ("Bucket count (max is %d) is %d and Record Count is %d",
- tracker_config_get_max_bucket_count (tracker->config),
- bucket_count, rec_count);
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
+
+ g_mutex_lock (priv->word_mutex);
+ tsiz = dpvsiz (priv->word_index, word, -1);
+ g_mutex_unlock (priv->word_mutex);
- return result;
+ return tsiz;
}
+/* int levenshtein ()
+ * Original license: GNU Lesser Public License
+ * from the Dixit project, (http://dixit.sourceforge.net/)
+ * Author: Octavian Procopiuc <oprocopiuc gmail com>
+ * Created: July 25, 2004
+ * Copied into tracker, by Edward Duffy
+ */
+static gint
+levenshtein (const gchar *source,
+ gchar *target,
+ gint maxdist)
+{
+ gchar n, m;
+ gint l;
+ gchar mincolval;
+ gchar matrix[51][51];
+ gchar j;
+ gchar i;
+ gchar cell;
-void
-tracker_indexer_close (Indexer *indexer)
-{
- g_return_if_fail (indexer);
+ l = strlen (source);
+ if (l > 50)
+ return -1;
+ n = l;
- g_mutex_lock (indexer->word_mutex);
+ l = strlen (target);
+ if (l > 50)
+ return -1;
+ m = l;
- if (!dpclose (indexer->word_index)) {
- g_message ("Index closure has failed due to %s", dperrmsg (dpecode));
- }
+ if (maxdist == 0)
+ maxdist = MAX(m, n);
+ if (n == 0)
+ return MIN(m, maxdist);
+ if (m == 0)
+ return MIN(n, maxdist);
- g_mutex_unlock (indexer->word_mutex);
- g_mutex_free (indexer->word_mutex);
- g_free (indexer->name);
- g_free (indexer);
-}
+ /* Store the min. value on each column, so that, if it
+ * reaches. maxdist, we break early.
+ */
+ for (j = 0; j <= m; j++)
+ matrix[0][(gint)j] = j;
+
+ for (i = 1; i <= n; i++) {
+ gchar s_i;
+ mincolval = MAX(m, i);
+ matrix[(gint)i][0] = i;
-void
-tracker_indexer_free (Indexer *indexer, gboolean remove_file)
-{
-
+ s_i = source[i-1];
- if (remove_file) {
- gchar *dbname;
+ for (j = 1; j <= m; j++) {
+ gchar t_j = target[j-1];
+ gchar cost = (s_i == t_j ? 0 : 1);
+ gchar above = matrix[i-1][(gint)j];
+ gchar left = matrix[(gint)i][j-1];
+ gchar diag = matrix[i-1][j-1];
- dbname = g_build_filename (tracker_get_data_dir (),
- indexer->name,
- NULL);
+ cell = MIN(above + 1, MIN(left + 1, diag + cost));
- g_return_if_fail (indexer);
+ /* Cover transposition, in addition to deletion,
+ * insertion and substitution. This step is taken from:
+ * Berghel, Hal ; Roach, David : "An Extension of Ukkonen's
+ * Enhanced Dynamic Programming ASM Algorithm"
+ * (http://www.acm.org/~hlb/publications/asm/asm.html)
+ */
+ if (i > 2 && j > 2) {
+ gchar trans = matrix[i-2][j-2] + 1;
- g_mutex_lock (indexer->word_mutex);
+ if (source[i-2] != t_j)
+ trans++;
+ if (s_i != target[j-2])
+ trans++;
+ if (cell > trans)
+ cell = trans;
+ }
- dpremove (dbname);
+ mincolval = MIN(mincolval, cell);
+ matrix[(gint)i][(gint)j] = cell;
+ }
- g_mutex_unlock (indexer->word_mutex);
+ if (mincolval >= maxdist)
+ break;
+ }
- g_free (dbname);
+ if (i == n + 1) {
+ return (gint) matrix[(gint)n][(gint)m];
} else {
- g_mutex_lock (indexer->word_mutex);
- dpclose (indexer->word_index);
- g_mutex_unlock (indexer->word_mutex);
- }
+ return maxdist;
+ }
+}
- g_mutex_free (indexer->word_mutex);
+static gint
+count_hits_for_word (TrackerIndexer *indexer,
+ const gchar *str) {
+
+ gint tsiz;
+ gint hits = 0;
- g_free (indexer->name);
+ tsiz = count_hit_size_for_word (indexer, str);
- g_free (indexer);
+ if (tsiz == -1 ||
+ tsiz % sizeof (TrackerIndexerWordDetails) != 0) {
+ return -1;
+ }
-
-}
+ hits = tsiz / sizeof (TrackerIndexerWordDetails);
-const gchar *
-tracker_indexer_get_name (Indexer *indexer)
-{
- g_return_val_if_fail (indexer != NULL, NULL);
-
- return dpname (indexer->word_index);
+ return hits;
}
+TrackerIndexer *
+tracker_indexer_new (TrackerIndexerType type,
+ TrackerConfig *config)
+{
+ TrackerIndexer *indexer;
+ TrackerIndexerPrivate *priv;
+ const gchar *name;
+ gchar *directory;
+ gint bucket_count;
+ gint rec_count;
+
+ g_return_val_if_fail (TRACKER_IS_CONFIG (config), NULL);
+
+ indexer = g_object_new (TRACKER_TYPE_INDEXER,
+ "config", config,
+ NULL);
+
+ switch (type) {
+ case TRACKER_INDEXER_TYPE_FILES:
+ name = TRACKER_INDEXER_FILE_INDEX_DB_FILENAME;
+ break;
+ case TRACKER_INDEXER_TYPE_EMAILS:
+ name = TRACKER_INDEXER_EMAIL_INDEX_DB_FILENAME;
+ break;
+ case TRACKER_INDEXER_TYPE_FILES_UPDATE:
+ name = TRACKER_INDEXER_FILE_UPDATE_INDEX_DB_FILENAME;
+ break;
+ }
-guint32
-tracker_indexer_size (Indexer *indexer)
-{
- return dpfsiz (indexer->word_index);
-}
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
+ priv->name = g_strdup (name);
-void
-tracker_indexer_sync (Indexer *indexer)
-{
- g_mutex_lock (indexer->word_mutex);
- dpsync (indexer->word_index);
- g_mutex_unlock (indexer->word_mutex);
+ directory = get_index_file (name);
+ priv->word_index = open_index (directory,
+ tracker_config_get_min_bucket_count (priv->config),
+ tracker_config_get_max_bucket_count (priv->config));
+ g_free (directory);
+
+ dpsetalign (priv->word_index, 8);
+
+ /* Re optimize database if bucket count < rec count */
+ bucket_count = dpbnum (priv->word_index);
+ rec_count = dprnum (priv->word_index);
+
+ g_message ("Bucket count (max is %d) is %d and record count is %d",
+ tracker_config_get_max_bucket_count (priv->config),
+ bucket_count,
+ rec_count);
+
+ return indexer;
}
-
-gboolean
-tracker_indexer_optimize (Indexer *indexer)
+void
+tracker_indexer_set_config (TrackerIndexer *object,
+ TrackerConfig *config)
{
-
- int num, b_count;
+ TrackerIndexerPrivate *priv;
- /* set bucket count to bucket_ratio times no. of recs divided by no. of divisions */
- num = CLAMP (get_preferred_bucket_count (indexer),
- tracker_config_get_min_bucket_count (tracker->config),
- tracker_config_get_max_bucket_count (tracker->config));
-
- b_count = num / tracker_config_get_divisions (tracker->config);
- g_message ("No. of buckets per division is %d", b_count);
-
- g_message ("Please wait while optimization of indexes takes place...");
- g_message ("Index has file size %d and bucket count of %d of which %d are used...",
- tracker_indexer_size (indexer),
- dpbnum (indexer->word_index),
- dpbusenum (indexer->word_index));
-
- g_mutex_lock (indexer->word_mutex);
+ g_return_if_fail (TRACKER_IS_INDEXER (object));
+ g_return_if_fail (TRACKER_IS_CONFIG (config));
+
+ priv = TRACKER_INDEXER_GET_PRIVATE (object);
- if (!dpoptimize (indexer->word_index, b_count)) {
+ if (config) {
+ g_object_ref (config);
+ }
- g_mutex_unlock (indexer->word_mutex);
- g_message ("Optimization has failed due to %s", dperrmsg (dpecode));
- return FALSE;
+ if (priv->config) {
+ g_object_unref (priv->config);
}
- g_mutex_unlock (indexer->word_mutex);
+ priv->config = config;
- g_message ("Index has been successfully optimized to file size %d and with bucket count of %d of which %d are used...",
- tracker_indexer_size (indexer),
- dpbnum (indexer->word_index),
- dpbusenum (indexer->word_index));
-
-
- return TRUE;
+ g_object_notify (G_OBJECT (object), "config");
}
-static inline gboolean
-has_word (Indexer *index, const char *word)
+guint32
+tracker_indexer_get_size (TrackerIndexer *indexer)
{
- char buffer [32];
+ TrackerIndexerPrivate *priv;
+ guint32 size;
+
+ g_return_val_if_fail (TRACKER_IS_INDEXER (indexer), 0);
- int count = dpgetwb (index->word_index, word, -1, 0, 32, buffer);
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
- return (count > 7);
+ g_mutex_lock (priv->word_mutex);
+ dpfsiz (priv->word_index);
+ g_mutex_unlock (priv->word_mutex);
+ return size;
}
+gboolean
+tracker_indexer_are_databases_too_big (void)
+{
+ gchar *filename;
+ const gchar *filename_const;
+ const gchar *data_dir;
+ gboolean too_big;
+ data_dir = tracker_get_data_dir ();
-void
-tracker_indexer_apply_changes (Indexer *dest, Indexer *src, gboolean update)
-{
- GObject *object;
- char *str;
- char buffer[MAX_HIT_BUFFER];
- int bytes;
- int sz = sizeof (WordDetails);
- int i = 0, interval;
- int buff_size = MAX_HITS_FOR_WORD * sz;
-
- g_message ("applying incremental changes to indexes");
-
- guint32 size = tracker_indexer_size (dest);
-
- if (size < (10 * 1024 * 1024)) {
- interval = 20000;
- } else if (size < (20 * 1024 * 1024)) {
- interval = 10000;
- } else if (size < (30 * 1024 * 1024)) {
- interval = 5000;
- } else if (size < (100 * 1024 * 1024)) {
- interval = 3000;
- } else {
- interval = 2000;
+ filename = g_build_filename (data_dir, TRACKER_INDEXER_FILE_INDEX_DB_FILENAME, NULL);
+ too_big = tracker_file_get_size (filename) > MAX_INDEX_FILE_SIZE;
+ g_free (filename);
+
+ if (too_big) {
+ g_critical ("File index database is too big, discontinuing indexing");
+ return TRUE;
}
-#ifdef HAVE_HAL
- /* halve the interval value as notebook hard drives are smaller */
- if (tracker_hal_get_battery_exists (tracker->hal)) {
- interval /= 2;
- }
-#endif /* HAVE_HAL */
-
- dpiterinit (src->word_index);
-
- tracker->in_merge = TRUE;
- merge_count = 1;
- merge_processed = 0;
-
- /* Signal progress */
- object = tracker_dbus_get_object (TRACKER_TYPE_DAEMON);
- g_signal_emit_by_name (object,
- "index-progress",
- "Merging",
- "",
- tracker->index_count,
- merge_processed,
- merge_count);
-
- while ((str = dpiternext (src->word_index, NULL))) {
-
- i++;
-
- if (i > 1 && (i % 200 == 0)) {
-#if 0
- /* FIXME-indexer-split: This has been commented out as
- * a result of removing the tracker-cache.[ch] which
- * is no longer used. This code is in a transitional
- * period.
- *
- * -Martyn
- */
- if (!tracker_cache_process_events (NULL, FALSE)) {
- return;
- }
-#endif
- }
+ filename = g_build_filename (data_dir, TRACKER_INDEXER_EMAIL_INDEX_DB_FILENAME, NULL);
+ too_big = tracker_file_get_size (filename) > MAX_INDEX_FILE_SIZE;
+ g_free (filename);
+
+ if (too_big) {
+ g_critical ("Email index database is too big, discontinuing indexing");
+ return TRUE;
+ }
- if (i > 1 && (i % interval == 0)) {
- if (!tracker_config_get_fast_merges (tracker->config)) {
- dpsync (dest->word_index);
- }
- }
-
- bytes = dpgetwb (src->word_index, str, -1, 0, buff_size, buffer);
+ filename_const = tracker_db_manager_get_file (TRACKER_DB_FILE_METADATA);
+ too_big = tracker_file_get_size (filename_const) > MAX_INDEX_FILE_SIZE;
+
+ if (too_big) {
+ g_critical ("File metadata database is too big, discontinuing indexing");
+ return TRUE;
+ }
- if (bytes < 1) continue;
+ filename_const = tracker_db_manager_get_file (TRACKER_DB_EMAIL_METADATA);
+ too_big = tracker_file_get_size (filename_const) > MAX_INDEX_FILE_SIZE;
+
+ if (too_big) {
+ g_critical ("Email metadata database is too big, discontinuing indexing");
+ return TRUE;
+ }
- if (bytes % sz != 0) {
- g_critical ("possible corruption found during application of changes to index with word %s (ignoring update for this word)", str);
- continue;
- }
+ return FALSE;
+}
- if (update) {
- tracker_indexer_update_word_chunk (dest, str, (WordDetails *) buffer, bytes / sz);
- } else {
- tracker_indexer_append_word_chunk (dest, str, (WordDetails *) buffer, bytes / sz);
- }
-
- dpout (src->word_index, str, -1);
+guint32
+tracker_indexer_calc_amalgamated (gint service,
+ gint score)
+{
+ unsigned char a[4];
+ gint16 score16;
+ guint8 service_type;
- g_free (str);
+ if (score > 30000) {
+ score16 = 30000;
+ } else {
+ score16 = (gint16) score;
}
-
- dpsync (dest->word_index);
- /* delete src and recreate if file update index */
+ service_type = (guint8) service;
- tracker_indexer_free (src, TRUE);
+ /* Amalgamate and combine score and service_type into a single
+ * 32-bit int for compact storage.
+ */
+ a[0] = service_type;
+ a[1] = (score16 >> 8) & 0xFF;
+ a[2] = score16 & 0xFF;
+ a[3] = 0;
- if (update) {
- tracker->file_update_index = tracker_indexer_open (TRACKER_INDEXER_FILE_UPDATE_INDEX_DB_FILENAME, FALSE);
- }
-
- tracker->in_merge = FALSE;
- merge_count = 1;
- merge_processed = 1;
-
- /* Signal progress */
- object = tracker_dbus_get_object (TRACKER_TYPE_DAEMON);
- g_signal_emit_by_name (object,
- "index-progress",
- "Merging",
- "",
- tracker->index_count,
- merge_processed,
- merge_count);
+ return (a[0] << 24) | (a[1] << 16) | (a[2] << 8) | a[3];
}
gboolean
-tracker_indexer_has_tmp_merge_files (IndexType type)
+tracker_indexer_has_tmp_merge_files (TrackerIndexerType type)
{
- GSList *files = NULL;
+ GSList *files = NULL;
gboolean result = FALSE;
-
- if (type == INDEX_TYPE_FILES) {
- files = tracker_process_files_get_files_with_prefix (tracker_get_data_dir (),
- "file-index.tmp.");
+ if (type == TRACKER_INDEXER_TYPE_FILES) {
+ files = tracker_process_files_get_files_with_prefix (tracker_get_data_dir (),
+ "file-index.tmp.");
} else {
- files = tracker_process_files_get_files_with_prefix (tracker_get_data_dir (),
- "email-index.tmp.");
+ files = tracker_process_files_get_files_with_prefix (tracker_get_data_dir (),
+ "email-index.tmp.");
}
- result = (files != NULL);
+ result = files != NULL;
if (result) {
g_slist_foreach (files, (GFunc) g_free, NULL);
@@ -520,925 +601,211 @@
}
return result;
-
}
+guint8
+tracker_indexer_word_details_get_service_type (TrackerIndexerWordDetails *details)
+{
+ g_return_val_if_fail (details != NULL, 0);
+ return (details->amalgamated >> 24) & 0xFF;
+}
-gboolean
-tracker_indexer_has_merge_files (IndexType type)
+gint16
+tracker_indexer_word_details_get_score (TrackerIndexerWordDetails *details)
{
- GSList *files = NULL;
- gboolean result = FALSE;
- gchar *final;
- const gchar *data_dir;
+ unsigned char a[2];
- data_dir = tracker_get_data_dir ();
+ g_return_val_if_fail (details != NULL, 0);
- if (type == INDEX_TYPE_FILES) {
- files = tracker_process_files_get_files_with_prefix (data_dir,
- "file-index.tmp.");
- final = g_build_filename (data_dir, "file-index-final", NULL);
- } else {
- files = tracker_process_files_get_files_with_prefix (data_dir,
- "email-index.tmp.");
- final = g_build_filename (data_dir, "email-index-final", NULL);
- }
+ a[0] = (details->amalgamated >> 16) & 0xFF;
+ a[1] = (details->amalgamated >> 8) & 0xFF;
- result = files != NULL;
+ return (gint16) (a[0] << 8) | (a[1]);
+}
- if (!result) {
- result = g_file_test (final, G_FILE_TEST_EXISTS);
- } else {
- g_slist_foreach (files, (GFunc) g_free, NULL);
- g_slist_free (files);
- }
+char *
+tracker_indexer_get_suggestion (TrackerIndexer *indexer,
+ const gchar *term,
+ gint maxdist)
+{
+ TrackerIndexerPrivate *priv;
+ gchar *str;
+ gint dist;
+ gchar *winner_str;
+ gint winner_dist;
+ gint hits;
+ GTimeVal start, current;
- g_free (final);
+ g_return_val_if_fail (TRACKER_IS_INDEXER (indexer), NULL);
+ g_return_val_if_fail (term != NULL, NULL);
- return result;
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
-}
+ winner_str = g_strdup (term);
+ winner_dist = G_MAXINT; /* Initialize to the worst case */
-static void
-move_index (Indexer *src_index, Indexer *dest_index, const char *fname)
-{
+ g_mutex_lock (priv->word_mutex);
+ dpiterinit (priv->word_index);
- if (!src_index || !dest_index) {
- g_critical ("cannot move indexes");
- return;
- }
+ g_get_current_time (&start);
- /* remove existing main index */
- g_mutex_lock (dest_index->word_mutex);
+ str = dpiternext (priv->word_index, NULL);
+ g_mutex_unlock (priv->word_mutex);
- dpclose (dest_index->word_index);
+ while (str != NULL) {
+ dist = levenshtein (term, str, 0);
- dpremove (fname);
-
- char *final_name = dpname (src_index->word_index);
-
- tracker_indexer_close (src_index);
-
- /* rename and reopen final index as main index */
-
- g_message ("renaming %s to %s", final_name, fname);
-
- rename (final_name, fname);
-
- dest_index->word_index = open_index (fname);
-
- if (!dest_index->word_index) {
- g_critical ("index creation failure for %s from %s", fname, final_name);
- }
-
- g_free (final_name);
-
- g_mutex_unlock (dest_index->word_mutex);
-
-}
-
-
-void
-tracker_indexer_merge_indexes (IndexType type)
-{
- GObject *object;
- GSList *lst;
- Indexer *final_index;
- GSList *file_list = NULL, *index_list = NULL;
- const gchar *prefix;
- const gchar *data_dir;
- gint i = 0, index_count, interval = 5000;
- gboolean final_exists;
- gchar *tmp;
-
- data_dir = tracker_get_data_dir ();
- object = tracker_dbus_get_object (TRACKER_TYPE_DAEMON);
-
- if (type == INDEX_TYPE_FILES) {
- g_return_if_fail (tracker->file_index);
-
- prefix = "file-index.tmp.";
- index_list = g_slist_prepend (index_list, tracker->file_index);
-
- tmp = g_build_filename (data_dir, "file-index-final", NULL);
- final_exists = g_file_test (tmp, G_FILE_TEST_EXISTS);
- g_free (tmp);
- } else {
- g_return_if_fail (tracker->email_index);
-
- prefix = "email-index.tmp.";
- index_list = g_slist_prepend (index_list, tracker->email_index);
-
- tmp = g_build_filename (data_dir, "email-index-final", NULL);
- final_exists = g_file_test (tmp, G_FILE_TEST_EXISTS);
- g_free (tmp);
- }
-
- file_list = tracker_process_files_get_files_with_prefix (data_dir, prefix);
-
- if (!file_list || !file_list->data) {
- g_slist_free (index_list);
- return;
- } else {
- GSList *file;
+ if (dist != -1 &&
+ dist < maxdist &&
+ dist < winner_dist) {
+ hits = count_hits_for_word (indexer, str);
- for (file = file_list; file; file = file->next) {
- if (file->data) {
- gchar *name = g_path_get_basename (file->data);
-
- if (name) {
- if (g_file_test (file->data, G_FILE_TEST_EXISTS)) {
- Indexer *tmp_index = tracker_indexer_open (name, FALSE);
-
- if (tmp_index) {
- index_list = g_slist_prepend (index_list, tmp_index);
- }
- }
+ if (hits < 0) {
+ g_free (winner_str);
+ g_free (str);
- g_free (name);
- }
+ return NULL;
+ } else if (hits > 0) {
+ g_free (winner_str);
+ winner_str = g_strdup (str);
+ winner_dist = dist;
+ } else {
+ g_message ("No hits for:'%s'!", str);
}
}
- g_slist_foreach (file_list, (GFunc) g_free, NULL);
- g_slist_free (file_list);
- }
-
- index_count = g_slist_length (index_list);
-
- if (index_count < 2) {
- g_slist_free (index_list);
- return;
- }
-
- g_message ("starting merge of %d indexes", index_count);
- tracker->in_merge = TRUE;
- merge_count = index_count;
- merge_processed = 0;
-
- /* Signal progress */
- g_signal_emit_by_name (object,
- "index-progress",
- "Merging",
- "",
- tracker->index_count,
- merge_processed,
- merge_count);
-
- if (index_count == 2 && !final_exists) {
- Indexer *index1 = index_list->data ;
- Indexer *index2 = index_list->next->data ;
-
- if (tracker_indexer_size (index1) * 3 < tracker_indexer_size (index2)) {
- tracker_indexer_apply_changes (index2, index1, FALSE);
- g_slist_free (index_list);
- goto end_of_merging;
- }
- }
-
- /* Signal state change */
- g_signal_emit_by_name (object,
- "index-state-change",
- tracker_status_get_as_string (),
- tracker->first_time_index,
- tracker->in_merge,
- tracker->pause_manual,
- tracker_should_pause_on_battery (),
- tracker->pause_io,
- tracker_config_get_enable_indexing (tracker->config));
-
- if (type == INDEX_TYPE_FILES) {
- final_index = tracker_indexer_open ("file-index-final", TRUE);
- } else {
- final_index = tracker_indexer_open ("email-index-final", TRUE);
- }
-
- if (!final_index) {
- g_slist_free (index_list);
- g_critical ("could not open final index - abandoning index merge");
- goto end_of_merging;
- }
-
- for (lst = index_list; lst && lst->data; lst = lst->next) {
- gchar *str;
- Indexer *index = lst->data;
-
- dpiterinit (index->word_index);
-
- while ((str = dpiternext (index->word_index, NULL))) {
- gchar buffer[MAX_HIT_BUFFER];
- gint offset;
- gint sz = sizeof (WordDetails);
- gint buff_size = MAX_HITS_FOR_WORD * sz;
-
- if (!has_word (final_index, str)) {
-
- i++;
-
- if (i > 101 && (i % 100 == 0)) {
-#if 0
- /* FIXME-indexer-split: This has been commented out as
- * a result of removing the tracker-cache.[ch] which
- * is no longer used. This code is in a transitional
- * period.
- *
- * -Martyn
- */
- if (!tracker_cache_process_events (NULL, FALSE)) {
- tracker_status_set_and_signal (TRACKER_STATUS_SHUTDOWN,
- tracker->first_time_index,
- tracker->in_merge,
- tracker->pause_manual,
- tracker_should_pause_on_battery (),
- tracker->pause_io,
- tracker_config_get_enable_indexing (tracker->config));
- return;
- }
-#endif
- }
-
- if (i > interval && (i % interval == 0)) {
-
- if (!tracker_config_get_fast_merges (tracker->config)) {
-
- dpsync (final_index->word_index);
-
- guint32 size = tracker_indexer_size (final_index);
-
- if (size < (10 * 1024 * 1024)) {
- interval = 10000;
- } else if (size < (20 * 1024 * 1024)) {
- interval = 6000;
- } else if (size < (50 * 1024 * 1024)) {
- interval = 6000;
- } else if (size < (100 * 1024 * 1024)) {
- interval = 4000;
- } else {
- interval = 3000;
- }
-
-#ifdef HAVE_HAL
- /* halve the interval value as notebook hard drives are smaller */
- if (tracker_hal_get_battery_exists (tracker->hal)) {
- interval /= 2;
- }
-#endif /* HAVE_HAL */
- }
- }
-
- offset = dpgetwb (index->word_index, str, -1, 0, buff_size, buffer);
-
- if (offset < 1) {
- continue;
- }
-
- if (offset % sz != 0) {
- g_critical ("possible corruption found during merge of word %s - purging word from index (it will not be searchable)", str);
- continue;
- }
-
- if (offset > 7 && offset < buff_size) {
-
- GSList *list;
-
- for (list = lst->next; list; list = list->next) {
- gchar tmp_buffer[MAX_HIT_BUFFER];
- Indexer *tmp_index = list->data;
-
- if (!tmp_index) {
- continue;
- }
-
- gint tmp_offset = dpgetwb (tmp_index->word_index, str, -1, 0, (buff_size - offset), tmp_buffer);
-
- if (tmp_offset > 0 && (tmp_offset % sz != 0)) {
- g_critical ("possible corruption found during merge of word %s - purging word from index", str);
- continue;
- }
-
- if (tmp_offset > 7 && (tmp_offset % sz == 0)) {
- memcpy (buffer + offset, tmp_buffer, tmp_offset);
- offset += tmp_offset;
- }
- }
+ g_free (str);
- dpput (final_index->word_index, str, -1, buffer, offset, DP_DOVER);
- }
- }
+ g_get_current_time (¤t);
- g_free (str);
+ if (current.tv_sec - start.tv_sec >= 2) { /* 2 second time out */
+ g_message ("Timeout in tracker_dbus_method_search_suggest");
+ break;
}
-
-
-
- /* dont free last entry as that is the main index */
- if (lst->next) {
-
- if (index != tracker->file_index && index != tracker->email_index) {
- GObject *object;
-
- tracker_indexer_free (index, TRUE);
- merge_processed++;
-
- /* Signal progress */
- object = tracker_dbus_get_object (TRACKER_TYPE_DAEMON);
- g_signal_emit_by_name (object,
- "index-progress",
- "Merging",
- "",
- tracker->index_count,
- merge_processed,
- merge_count);
- }
-
-
- } else {
- if (type == INDEX_TYPE_FILES) {
-
- char *fname = get_index_file (TRACKER_INDEXER_FILE_INDEX_DB_FILENAME);
- move_index (final_index, tracker->file_index, fname);
- g_free (fname);
-
- } else {
- char *fname = get_index_file (TRACKER_INDEXER_EMAIL_INDEX_DB_FILENAME);
- move_index (final_index, tracker->email_index, fname);
- g_free (fname);
- }
- }
+ g_mutex_lock (priv->word_mutex);
+ str = dpiternext (priv->word_index, NULL);
+ g_mutex_unlock (priv->word_mutex);
}
-
- g_slist_free (index_list);
-
-
- end_of_merging:
- tracker->in_merge = FALSE;
-
- /* Signal state change */
- g_signal_emit_by_name (object,
- "index-state-change",
- tracker_status_get_as_string (),
- tracker->first_time_index,
- tracker->in_merge,
- tracker->pause_manual,
- tracker_should_pause_on_battery (),
- tracker->pause_io,
- tracker_config_get_enable_indexing (tracker->config));
+ return winner_str;
}
+TrackerIndexerWordDetails *
+tracker_indexer_get_word_hits (TrackerIndexer *indexer,
+ const gchar *word,
+ guint *count)
+{
+ TrackerIndexerPrivate *priv;
+ TrackerIndexerWordDetails *details;
+ gint tsiz;
+ gchar *tmp;
+ g_return_val_if_fail (TRACKER_IS_INDEXER (indexer), NULL);
+ g_return_val_if_fail (word != NULL, NULL);
-/* indexing api */
-
-/* use for fast insertion of a word for multiple documents at a time */
-
-gboolean
-tracker_indexer_append_word_chunk (Indexer *indexer, const gchar *word, WordDetails *details, gint word_detail_count)
-{
- g_return_val_if_fail (indexer, FALSE);
- g_return_val_if_fail (indexer->word_index, FALSE);
- g_return_val_if_fail (word, FALSE);
- g_return_val_if_fail (details, FALSE);
- g_return_val_if_fail (word_detail_count > 0, FALSE);
-
- g_mutex_lock (indexer->word_mutex);
- if (!dpput (indexer->word_index, word, -1, (char *) details, (word_detail_count * sizeof (WordDetails)), DP_DCAT)) {
- g_mutex_unlock (indexer->word_mutex);
- return FALSE;
- }
- g_mutex_unlock (indexer->word_mutex);
-
- return TRUE;
-}
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
+ g_mutex_lock (priv->word_mutex);
-/* append individual word for a document */
+ details = NULL;
-gboolean
-tracker_indexer_append_word (Indexer *indexer, const gchar *word, guint32 id, gint service, gint score)
-{
- if (score < 1) {
- return FALSE;
+ if (count) {
+ *count = 0;
}
- g_return_val_if_fail (indexer, FALSE);
- g_return_val_if_fail (indexer->word_index, FALSE);
- g_return_val_if_fail (word, FALSE);
+ if ((tmp = dpget (priv->word_index, word, -1, 0, MAX_HIT_BUFFER, &tsiz)) != NULL) {
+ if (tsiz >= (gint) sizeof (TrackerIndexerWordDetails)) {
+ details = (TrackerIndexerWordDetails *) tmp;
- WordDetails pair;
-
- pair.id = id;
- pair.amalgamated = tracker_indexer_calc_amalgamated (service, score);
-
- return tracker_indexer_append_word_chunk (indexer, word, &pair, 1);
-}
-
-
-/* append lists of words for a document - returns no. of hits added */
-gint
-tracker_indexer_append_word_list (Indexer *indexer, const gchar *word, GSList *list)
-{
- WordDetails word_details[MAX_HITS_FOR_WORD], *wd;
- gint i;
- GSList *lst;
-
- g_return_val_if_fail (indexer, 0);
- g_return_val_if_fail (indexer->word_index, 0);
- g_return_val_if_fail (word, 0);
-
- i = 0;
-
- if (list) {
- for (lst = list; (lst && i < MAX_HITS_FOR_WORD); lst = lst->next) {
-
- if (lst->data) {
- wd = lst->data;
- word_details[i].id = wd->id;
- word_details[i].amalgamated = wd->amalgamated;
- i++;
- }
- }
- }
-
- if (i > 0) {
- tracker_indexer_append_word_chunk (indexer, word, word_details, i);
- }
-
- return i;
-}
-
-
-
-/* use for deletes or updates of multiple entities when they are not new */
-gboolean
-tracker_indexer_update_word_chunk (Indexer *indexer, const gchar *word, WordDetails *detail_chunk, gint word_detail_count)
-{
- int tsiz, j, i, score;
- char *tmp;
- WordDetails *word_details;
- gboolean write_back = FALSE;
- GSList *list = NULL;
-
- g_return_val_if_fail (indexer, FALSE);
- g_return_val_if_fail (indexer->word_index, FALSE);
- g_return_val_if_fail (word, FALSE);
- g_return_val_if_fail (detail_chunk, FALSE);
- g_return_val_if_fail (word_detail_count > 0, FALSE);
-
- /* check if existing record is there */
- gint hit_count = 0;
-
- g_mutex_lock (indexer->word_mutex);
-
- if ((tmp = dpget (indexer->word_index, word, -1, 0, MAX_HIT_BUFFER, &tsiz)) != NULL) {
-
-
- WordDetails *details = (WordDetails *) tmp;
- hit_count = tsiz / sizeof (WordDetails);
-
- details = (WordDetails *) tmp;
-
- for (j = 0; j < word_detail_count; j++) {
-
- word_details = &detail_chunk[j];
-
- gboolean edited = FALSE;
-
- for (i = 0; i < hit_count; i++) {
-
- if (details[i].id == word_details->id) {
-
- write_back = TRUE;
-
- /* NB the paramter score can be negative */
- score = get_score (&details[i]) + get_score (word_details);
- //g_print ("current score for %s is %d and new is %d and final is %d\n", word, get_score (&details[i]), get_score (word_details), score);
-
-
- /* check for deletion */
- if (score < 1) {
-
- //g_print ("deleting word hit %s\n", word);
-
- gint k;
-
- /* shift all subsequent records in array down one place */
- for (k = i + 1; k < hit_count; k++) {
- details[k - 1] = details[k];
- }
-
- hit_count--;
-
- } else {
- details[i].amalgamated = tracker_indexer_calc_amalgamated (get_service_type (&details[i]), score);
- }
-
- edited = TRUE;
- break;
- }
- }
-
- /* add hits that could not be updated directly here so they can be appended later */
- if (!edited) {
- list = g_slist_prepend (list, &detail_chunk[j]);
- g_debug ("could not update word hit %s - appending", word);
- }
- }
-
- /* write back if we have modded anything */
- if (write_back) {
- dpput (indexer->word_index, word, -1, (char *) details, (hit_count * sizeof (WordDetails)), DP_DOVER);
- }
-
- g_mutex_unlock (indexer->word_mutex);
-
- if (list) {
- tracker_indexer_append_word_list (indexer, word, list);
- g_slist_free (list);
- }
-
- return TRUE;
- }
-
- g_mutex_unlock (indexer->word_mutex);
-
- /* none of the updates can be applied if word does not exist so return them all to be appended later */
- return tracker_indexer_append_word_chunk (indexer, word, detail_chunk, word_detail_count);
-
-}
-
-
-/* use for deletes or updates of multiple entities when they are not new */
-gboolean
-tracker_indexer_update_word_list (Indexer *indexer, const gchar *word, GSList *update_list)
-{
- WordDetails word_details[MAX_HITS_FOR_WORD], *wd;
- gint i;
- GSList *lst;
-
- g_return_val_if_fail (indexer, 0);
- g_return_val_if_fail (indexer->word_index, 0);
- g_return_val_if_fail (word, 0);
-
- i = 0;
-
- if (update_list) {
- for (lst = update_list; (lst && i < MAX_HITS_FOR_WORD); lst = lst->next) {
-
- if (lst->data) {
- wd = lst->data;
- word_details[i].id = wd->id;
- word_details[i].amalgamated = wd->amalgamated;
- i++;
- }
- }
- }
-
- if (i > 0) {
- tracker_indexer_update_word_chunk (indexer, word, word_details, i);
- }
-
- return i;
-}
-
-WordDetails *
-tracker_indexer_get_word_hits (Indexer *indexer,
- const gchar *word,
- guint *count)
-{
- WordDetails *details;
- gint tsiz;
- gchar *tmp;
-
- g_mutex_lock (indexer->word_mutex);
-
- details = NULL;
- *count = 0;
-
- if ((tmp = dpget (indexer->word_index, word, -1, 0, MAX_HIT_BUFFER, &tsiz)) != NULL) {
- if (tsiz >= (int) sizeof (WordDetails)) {
- details = (WordDetails *) tmp;
- *count = tsiz / sizeof (WordDetails);
+ if (count) {
+ *count = tsiz / sizeof (TrackerIndexerWordDetails);
+ }
}
}
- g_mutex_unlock (indexer->word_mutex);
+ g_mutex_unlock (priv->word_mutex);
return details;
}
-/* use to delete dud hits for a word - dud_list is a list of TrackerSearchHit structs */
+/* Use to delete dud hits for a word - dud_list is a list of
+ * TrackerSearchHit structs.
+ */
gboolean
-tracker_remove_dud_hits (Indexer *indexer, const gchar *word, GSList *dud_list)
-{
- gint tsiz;
- char *tmp;
+tracker_indexer_remove_dud_hits (TrackerIndexer *indexer,
+ const gchar *word,
+ GSList *dud_list)
+{
+ TrackerIndexerPrivate *priv;
+ gchar *tmp;
+ gint tsiz;
g_return_val_if_fail (indexer, FALSE);
- g_return_val_if_fail (indexer->word_index, FALSE);
+ g_return_val_if_fail (priv->word_index, FALSE);
g_return_val_if_fail (word, FALSE);
g_return_val_if_fail (dud_list, FALSE);
-
- g_mutex_lock (indexer->word_mutex);
-
- /* check if existing record is there */
- if ((tmp = dpget (indexer->word_index, word, -1, 0, MAX_HIT_BUFFER, &tsiz)) != NULL) {
-
- if (tsiz >= (int) sizeof (WordDetails)) {
-
- WordDetails *details;
- int wi, i, pnum;
-
- details = (WordDetails *) tmp;
- pnum = tsiz / sizeof (WordDetails);
- wi = 0;
-
- for (i = 0; i < pnum; i++) {
-
- GSList *lst;
-
- for (lst = dud_list; lst; lst = lst->next) {
- TrackerSearchHit *hit = lst->data;
-
- if (hit) {
- if (details[i].id == hit->service_id) {
- int k;
-
- /* shift all subsequent records in array down one place */
- for (k = i + 1; k < pnum; k++) {
- details[k - 1] = details[k];
- }
-
- /* make size of array one size smaller */
- tsiz -= sizeof (WordDetails);
- pnum--;
-
- break;
- }
- }
- }
- }
-
- dpput (indexer->word_index, word, -1, (char *) details, tsiz, DP_DOVER);
-
- g_mutex_unlock (indexer->word_mutex);
+ priv = TRACKER_INDEXER_GET_PRIVATE (indexer);
- g_free (tmp);
-
- return TRUE;
- }
-
- g_free (tmp);
- }
-
- g_mutex_unlock (indexer->word_mutex);
+ /* Check if existing record is there */
+ g_mutex_lock (priv->word_mutex);
+ tmp = dpget (priv->word_index,
+ word,
+ -1,
+ 0,
+ MAX_HIT_BUFFER,
+ &tsiz);
+ g_mutex_unlock (priv->word_mutex);
- return FALSE;
-}
-
-static inline gint
-count_hit_size_for_word (Indexer *indexer, const gchar *word)
-{
- int tsiz;
-
- g_mutex_lock (indexer->word_mutex);
- tsiz = dpvsiz (indexer->word_index, word, -1);
- g_mutex_unlock (indexer->word_mutex);
-
- return tsiz;
-}
-
-guint8
-tracker_word_details_get_service_type (WordDetails *details)
-{
- return (details->amalgamated >> 24) & 0xFF;
-}
-
-gint16
-tracker_word_details_get_score (WordDetails *details)
-{
- unsigned char a[2];
-
- a[0] = (details->amalgamated >> 16) & 0xFF;
- a[1] = (details->amalgamated >> 8) & 0xFF;
-
- return (gint16) (a[0] << 8) | (a[1]);
-}
-
-/* int levenshtein ()
- * Original license: GNU Lesser Public License
- * from the Dixit project, (http://dixit.sourceforge.net/)
- * Author: Octavian Procopiuc <oprocopiuc gmail com>
- * Created: July 25, 2004
- * Copied into tracker, by Edward Duffy
- */
-
-static int
-levenshtein(const char *source, char *target, int maxdist)
-{
- char n, m;
- int l;
- l = strlen (source);
- if (l > 50)
- return -1;
- n = l;
-
- l = strlen (target);
- if (l > 50)
- return -1;
- m = l;
-
- if (maxdist == 0)
- maxdist = MAX(m, n);
- if (n == 0)
- return MIN(m, maxdist);
- if (m == 0)
- return MIN(n, maxdist);
-
- // Store the min. value on each column, so that, if it reaches
- // maxdist, we break early.
- char mincolval;
-
- char matrix[51][51];
-
- char j;
- char i;
- char cell;
-
- for (j = 0; j <= m; j++)
- matrix[0][(int)j] = j;
-
- for (i = 1; i <= n; i++) {
-
- mincolval = MAX(m, i);
- matrix[(int)i][0] = i;
-
- char s_i = source[i-1];
-
- for (j = 1; j <= m; j++) {
-
- char t_j = target[j-1];
-
- char cost = (s_i == t_j ? 0 : 1);
-
- char above = matrix[i-1][(int)j];
- char left = matrix[(int)i][j-1];
- char diag = matrix[i-1][j-1];
- cell = MIN(above + 1, MIN(left + 1, diag + cost));
-
- // Cover transposition, in addition to deletion,
- // insertion and substitution. This step is taken from:
- // Berghel, Hal ; Roach, David : "An Extension of Ukkonen's
- // Enhanced Dynamic Programming ASM Algorithm"
- // (http://www.acm.org/~hlb/publications/asm/asm.html)
-
- if (i > 2 && j > 2) {
- char trans = matrix[i-2][j-2] + 1;
- if (source[i-2] != t_j)
- trans++;
- if (s_i != target[j-2])
- trans++;
- if (cell > trans)
- cell = trans;
- }
-
- mincolval = MIN(mincolval, cell);
- matrix[(int)i][(int)j] = cell;
- }
-
- if (mincolval >= maxdist)
- break;
-
- }
-
- if (i == n + 1)
- return (int) matrix[(int)n][(int)m];
- else
- return maxdist;
-}
-
-static int
-count_hits_for_word (Indexer *indexer, const gchar *str) {
-
- gint tsiz, hits = 0;
-
- tsiz = count_hit_size_for_word (indexer, str);
-
- if (tsiz == -1 || tsiz % sizeof (WordDetails) != 0) {
- return -1;
+ if (!tmp) {
+ return FALSE;
}
- hits = tsiz / sizeof (WordDetails);
-
- return hits;
-}
-
-char *
-tracker_indexer_get_suggestion (Indexer *indexer, const gchar *term, gint maxdist)
-{
-
- gchar *str;
- gint dist;
- gchar *winner_str;
- gint winner_dist;
- gint hits;
- GTimeVal start, current;
-
- winner_str = g_strdup (term);
- winner_dist = G_MAXINT; /* Initialize to the worst case */
-
- dpiterinit (indexer->word_index);
-
- g_get_current_time (&start);
-
- str = dpiternext (indexer->word_index, NULL);
-
- while (str != NULL) {
-
- dist = levenshtein (term, str, 0);
-
- if (dist != -1 && dist < maxdist && dist < winner_dist) {
-
- hits = count_hits_for_word (indexer, str);
-
- if (hits < 0) {
-
- g_free (winner_str);
- g_free (str);
- return NULL;
-
- } else if (hits > 0) {
-
- g_free (winner_str);
- winner_str = g_strdup (str);
- winner_dist = dist;
-
- } else {
- g_message ("No hits for %s!", str);
- }
- }
-
- g_free (str);
-
- g_get_current_time (¤t);
-
- if (current.tv_sec - start.tv_sec >= 2) { /* 2 second time out */
- g_message ("Timeout in tracker_dbus_method_search_suggest");
- break;
- }
-
- str = dpiternext (indexer->word_index, NULL);
- }
-
- return winner_str;
-}
-
-gboolean
-tracker_indexer_are_databases_too_big (void)
-{
- gchar *filename;
- const gchar *filename_const;
- const gchar *data_dir;
- gboolean too_big;
-
- data_dir = tracker_get_data_dir ();
-
- filename = g_build_filename (data_dir, TRACKER_INDEXER_FILE_INDEX_DB_FILENAME, NULL);
- too_big = tracker_file_get_size (filename) > MAX_INDEX_FILE_SIZE;
- g_free (filename);
-
- if (too_big) {
- g_critical ("File index database is too big, discontinuing indexing");
- return TRUE;
- }
-
- filename = g_build_filename (data_dir, TRACKER_INDEXER_EMAIL_INDEX_DB_FILENAME, NULL);
- too_big = tracker_file_get_size (filename) > MAX_INDEX_FILE_SIZE;
- g_free (filename);
-
- if (too_big) {
- g_critical ("Email index database is too big, discontinuing indexing");
- return TRUE;
- }
-
- filename_const = tracker_db_manager_get_file (TRACKER_DB_FILE_METADATA);
- too_big = tracker_file_get_size (filename_const) > MAX_INDEX_FILE_SIZE;
-
- if (too_big) {
- g_critical ("File metadata database is too big, discontinuing indexing");
- return TRUE;
- }
-
- filename_const = tracker_db_manager_get_file (TRACKER_DB_EMAIL_METADATA);
- too_big = tracker_file_get_size (filename_const) > MAX_INDEX_FILE_SIZE;
+ if (tsiz >= (int) sizeof (TrackerIndexerWordDetails)) {
+ TrackerIndexerWordDetails *details;
+ gint wi, i, pnum;
+
+ details = (TrackerIndexerWordDetails *) tmp;
+ pnum = tsiz / sizeof (TrackerIndexerWordDetails);
+ wi = 0;
+
+ for (i = 0; i < pnum; i++) {
+ GSList *lst;
+
+ for (lst = dud_list; lst; lst = lst->next) {
+ TrackerSearchHit *hit = lst->data;
+
+ if (hit) {
+ if (details[i].id == hit->service_id) {
+ gint k;
+
+ /* Shift all subsequent records in array down one place */
+ for (k = i + 1; k < pnum; k++) {
+ details[k - 1] = details[k];
+ }
+
+ /* Make size of array one size smaller */
+ tsiz -= sizeof (TrackerIndexerWordDetails);
+ pnum--;
+
+ break;
+ }
+ }
+ }
+ }
+
+ g_mutex_lock (priv->word_mutex);
+ dpput (priv->word_index, word, -1, (gchar *) details, tsiz, DP_DOVER);
+ g_mutex_unlock (priv->word_mutex);
+
+ g_free (tmp);
+
+ return TRUE;
+ }
- if (too_big) {
- g_critical ("Email metadata database is too big, discontinuing indexing");
- return TRUE;
- }
+ g_free (tmp);
return FALSE;
}
Modified: branches/indexer-split/src/trackerd/tracker-indexer.h
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-indexer.h (original)
+++ branches/indexer-split/src/trackerd/tracker-indexer.h Thu Jun 26 11:38:21 2008
@@ -1,7 +1,8 @@
/* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8 -*- */
-/*
+/*
* Copyright (C) 2006, Mr Jamie McCracken (jamiemcc gnome org)
- *
+ * Copyright (C) 2008, Nokia
+
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License as published by the Free Software Foundation; either
@@ -21,11 +22,9 @@
#ifndef __TRACKERD_INDEXER_H__
#define __TRACKERD_INDEXER_H__
-#include <stdlib.h>
-
#include <glib.h>
-#include <libtracker-db/tracker-db-interface.h>
+#include <libtracker-common/tracker-config.h>
#define TRACKER_INDEXER_FILE_INDEX_DB_FILENAME "file-index.db"
#define TRACKER_INDEXER_EMAIL_INDEX_DB_FILENAME "email-index.db"
@@ -33,74 +32,69 @@
G_BEGIN_DECLS
-typedef struct {
- guint32 id; /* Service ID number of the
- * document */
- int amalgamated; /* amalgamation of
- * service_type and score of
- * the word in the document's
- * metadata */
-} WordDetails;
+#define TRACKER_TYPE_INDEXER (tracker_indexer_get_type())
+#define TRACKER_INDEXER(o) (G_TYPE_CHECK_INSTANCE_CAST ((o), TRACKER_TYPE_INDEXER, TrackerIndexer))
+#define TRACKER_INDEXER_CLASS(c) (G_TYPE_CHECK_CLASS_CAST ((c), TRACKER_TYPE_INDEXER, TrackerIndexerClass))
+#define TRACKER_IS_INDEXER(o) (G_TYPE_CHECK_INSTANCE_TYPE ((o), TRACKER_TYPE_INDEXER))
+#define TRACKER_IS_INDEXER_CLASS(c) (G_TYPE_CHECK_CLASS_TYPE ((c), TRACKER_TYPE_INDEXER))
+#define TRACKER_INDEXER_GET_CLASS(o) (G_TYPE_INSTANCE_GET_CLASS ((o), TRACKER_TYPE_INDEXER, TrackerIndexerClass))
+
+typedef struct TrackerIndexer TrackerIndexer;
+typedef struct TrackerIndexerClass TrackerIndexerClass;
+typedef struct TrackerIndexerWordDetails TrackerIndexerWordDetails;
+
+struct TrackerIndexer {
+ GObject parent;
+};
+
+struct TrackerIndexerClass {
+ GObjectClass parent_class;
+};
+
+struct TrackerIndexerWordDetails {
+ /* Service ID number of the document */
+ guint32 id;
+
+ /* Amalgamation of service_type and score of the word in the
+ * document's metadata.
+ */
+ gint amalgamated;
+};
typedef enum {
- WordNormal,
- WordWildCard,
- WordExactPhrase
-} WordType;
-
-typedef struct {
- gchar *word;
- gint hit_count;
- gfloat idf;
- WordType word_type;
-} SearchWord;
-
-typedef struct Indexer_ Indexer;
-
-typedef enum {
- INDEX_TYPE_FILES,
- INDEX_TYPE_EMAILS,
- INDEX_TYPE_FILE_UPDATE
-} IndexType;
-
-guint32 tracker_indexer_calc_amalgamated (gint service, gint score);
-
-Indexer * tracker_indexer_open (const gchar *name, gboolean main_index);
-void tracker_indexer_close (Indexer *indexer);
-gboolean tracker_indexer_repair (const char *name);
-void tracker_indexer_free (Indexer *indexer, gboolean remove_file);
-gboolean tracker_indexer_has_merge_index (Indexer *indexer, gboolean update);
-
-const gchar * tracker_indexer_get_name (Indexer *indexer);
-guint32 tracker_indexer_size (Indexer *indexer);
-gboolean tracker_indexer_optimize (Indexer *indexer);
-void tracker_indexer_sync (Indexer *indexer);
-
-void tracker_indexer_apply_changes (Indexer *dest, Indexer *src, gboolean update);
-void tracker_indexer_merge_indexes (IndexType type);
-gboolean tracker_indexer_has_merge_files (IndexType type);
-gboolean tracker_indexer_has_tmp_merge_files (IndexType type);
-
-/* Indexing api */
-gboolean tracker_indexer_append_word (Indexer *indexer, const gchar *word, guint32 id, gint service, gint score);
-gboolean tracker_indexer_append_word_chunk (Indexer *indexer, const gchar *word, WordDetails *details, gint word_detail_count);
-gint tracker_indexer_append_word_list (Indexer *indexer, const gchar *word, GSList *list);
-
-gboolean tracker_indexer_update_word (Indexer *indexer, const gchar *word, guint32 id, gint service, gint score, gboolean remove_word);
-gboolean tracker_indexer_update_word_chunk (Indexer *indexer, const gchar *word, WordDetails *details, gint word_detail_count);
-gboolean tracker_indexer_update_word_list (Indexer *indexer, const gchar *word, GSList *update_list);
-
-WordDetails * tracker_indexer_get_word_hits (Indexer *indexer, const gchar *word, guint *count);
-
-gboolean tracker_remove_dud_hits (Indexer *indexer, const gchar *word, GSList *dud_list);
-
-char * tracker_indexer_get_suggestion (Indexer *indexer, const gchar *term, gint maxdist);
-
-gboolean tracker_indexer_are_databases_too_big (void);
-
-/* Word API */
-guint8 tracker_word_details_get_service_type (WordDetails *details);
-gint16 tracker_word_details_get_score (WordDetails *details);
+ TRACKER_INDEXER_TYPE_FILES,
+ TRACKER_INDEXER_TYPE_EMAILS,
+ TRACKER_INDEXER_TYPE_FILES_UPDATE
+} TrackerIndexerType;
+
+
+GType tracker_indexer_get_type (void);
+
+TrackerIndexer *tracker_indexer_new (TrackerIndexerType type,
+ TrackerConfig *config);
+void tracker_indexer_set_config (TrackerIndexer *object,
+ TrackerConfig *config);
+guint32 tracker_indexer_get_size (TrackerIndexer *indexer);
+
+gboolean tracker_indexer_are_databases_too_big (void);
+gboolean tracker_indexer_has_tmp_merge_files (TrackerIndexerType type);
+guint32 tracker_indexer_calc_amalgamated (gint service,
+ gint score);
+
+guint8 tracker_indexer_word_details_get_service_type (TrackerIndexerWordDetails *details);
+gint16 tracker_indexer_word_details_get_score (TrackerIndexerWordDetails *details);
+
+char * tracker_indexer_get_suggestion (TrackerIndexer *indexer,
+ const gchar *term,
+ gint maxdist);
+TrackerIndexerWordDetails *
+ tracker_indexer_get_word_hits (TrackerIndexer *indexer,
+ const gchar *word,
+ guint *count);
+
+gboolean tracker_indexer_remove_dud_hits (TrackerIndexer *indexer,
+ const gchar *word,
+ GSList *dud_list);
G_END_DECLS
Modified: branches/indexer-split/src/trackerd/tracker-main.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-main.c (original)
+++ branches/indexer-split/src/trackerd/tracker-main.c Thu Jun 26 11:38:21 2008
@@ -25,9 +25,10 @@
#include "config.h"
+#include <stdlib.h>
+#include <string.h>
#include <signal.h>
#include <locale.h>
-#include <string.h>
#include <unistd.h>
#include <fcntl.h>
@@ -483,9 +484,6 @@
static gboolean
initialize_databases (void)
{
- Indexer *index;
- gchar *final_index_name;
-
/*
* Create SQLite databases
*/
@@ -519,13 +517,22 @@
tracker_db_set_option_int ("InitialIndex", 1);
}
+ return TRUE;
+}
+
+static gboolean
+initialize_indexers (TrackerConfig *config)
+{
+ TrackerIndexer *indexer;
+ gchar *final_index_name;
+
/*
* Create index files
*/
final_index_name = g_build_filename (data_dir, "file-index-final", NULL);
if (g_file_test (final_index_name, G_FILE_TEST_EXISTS) &&
- !tracker_indexer_has_tmp_merge_files (INDEX_TYPE_FILES)) {
+ !tracker_indexer_has_tmp_merge_files (TRACKER_INDEXER_TYPE_FILES)) {
gchar *file_index_name;
file_index_name = g_build_filename (data_dir,
@@ -546,7 +553,7 @@
NULL);
if (g_file_test (final_index_name, G_FILE_TEST_EXISTS) &&
- !tracker_indexer_has_tmp_merge_files (INDEX_TYPE_EMAILS)) {
+ !tracker_indexer_has_tmp_merge_files (TRACKER_INDEXER_TYPE_EMAILS)) {
gchar *file_index_name;
file_index_name = g_build_filename (data_dir,
@@ -563,16 +570,26 @@
g_free (final_index_name);
/* Create indexers */
- index = tracker_indexer_open (TRACKER_INDEXER_FILE_INDEX_DB_FILENAME, TRUE);
- tracker->file_index = index;
+ indexer = tracker_indexer_new (TRACKER_INDEXER_TYPE_FILES, config);
+ if (!indexer) {
+ return FALSE;
+ }
- index = tracker_indexer_open (TRACKER_INDEXER_FILE_UPDATE_INDEX_DB_FILENAME, FALSE);
- tracker->file_update_index = index;
+ tracker->file_index = indexer;
- index = tracker_indexer_open (TRACKER_INDEXER_EMAIL_INDEX_DB_FILENAME, TRUE);
- tracker->email_index = index;
+ indexer = tracker_indexer_new (TRACKER_INDEXER_TYPE_FILES_UPDATE, config);
+ if (!indexer) {
+ return FALSE;
+ }
- /* db_con->word_index = tracker->file_index; */
+ tracker->file_update_index = indexer;
+
+ indexer = tracker_indexer_new (TRACKER_INDEXER_TYPE_EMAILS, config);
+ if (!indexer) {
+ return FALSE;
+ }
+
+ tracker->email_index = indexer;
return TRUE;
}
@@ -608,9 +625,17 @@
static void
shutdown_indexer (void)
{
- tracker_indexer_close (tracker->file_index);
- tracker_indexer_close (tracker->file_update_index);
- tracker_indexer_close (tracker->email_index);
+ if (tracker->file_index) {
+ g_object_unref (tracker->file_index);
+ }
+
+ if (tracker->file_update_index) {
+ g_object_unref (tracker->file_update_index);
+ }
+
+ if (tracker->email_index) {
+ g_object_unref (tracker->email_index);
+ }
}
static void
@@ -827,6 +852,10 @@
return EXIT_FAILURE;
}
+ if (!initialize_indexers (tracker->config)) {
+ return EXIT_FAILURE;
+ }
+
/* Set our status as running, if this is FALSE, threads stop
* doing what they do and shutdown.
*/
Modified: branches/indexer-split/src/trackerd/tracker-main.h
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-main.h (original)
+++ branches/indexer-split/src/trackerd/tracker-main.h Thu Jun 26 11:38:21 2008
@@ -50,6 +50,9 @@
TrackerLanguage *language;
TrackerCrawler *crawler;
+ TrackerIndexer *file_index;
+ TrackerIndexer *file_update_index;
+ TrackerIndexer *email_index;
gboolean is_running;
gboolean readonly;
@@ -64,9 +67,6 @@
gboolean pause_io;
/* Indexing options */
- Indexer *file_index;
- Indexer *file_update_index;
- Indexer *email_index;
/* Table of stop words that are to be ignored by the parser */
gboolean first_time_index;
Modified: branches/indexer-split/src/trackerd/tracker-monitor.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-monitor.c (original)
+++ branches/indexer-split/src/trackerd/tracker-monitor.c Thu Jun 26 11:38:21 2008
@@ -19,6 +19,7 @@
*/
#include <string.h>
+#include <stdlib.h>
#include <libtracker-common/tracker-dbus.h>
#include <libtracker-common/tracker-file-utils.h>
Modified: branches/indexer-split/src/trackerd/tracker-query-tree.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-query-tree.c (original)
+++ branches/indexer-split/src/trackerd/tracker-query-tree.c Thu Jun 26 11:38:21 2008
@@ -67,7 +67,7 @@
struct TrackerQueryTreePrivate {
gchar *query_str;
TreeNode *tree;
- Indexer *indexer;
+ TrackerIndexer *indexer;
TrackerConfig *config;
TrackerLanguage *language;
GArray *services;
@@ -125,10 +125,11 @@
G_PARAM_READWRITE));
g_object_class_install_property (object_class,
PROP_INDEXER,
- g_param_spec_pointer ("indexer",
- "Indexer",
- "Indexer",
- G_PARAM_READWRITE));
+ g_param_spec_object ("indexer",
+ "Indexer",
+ "Indexer",
+ tracker_indexer_get_type (),
+ G_PARAM_READWRITE));
g_object_class_install_property (object_class,
PROP_CONFIG,
g_param_spec_object ("config",
@@ -257,7 +258,7 @@
g_value_set_string (value, priv->query_str);
break;
case PROP_INDEXER:
- g_value_set_pointer (value, priv->indexer);
+ g_value_set_object (value, priv->indexer);
break;
case PROP_CONFIG:
g_value_set_object (value, priv->config);
@@ -275,13 +276,13 @@
TrackerQueryTree *
tracker_query_tree_new (const gchar *query_str,
- Indexer *indexer,
+ TrackerIndexer *indexer,
TrackerConfig *config,
TrackerLanguage *language,
GArray *services)
{
g_return_val_if_fail (query_str != NULL, NULL);
- g_return_val_if_fail (indexer != NULL, NULL);
+ g_return_val_if_fail (TRACKER_IS_INDEXER (indexer), NULL);
g_return_val_if_fail (TRACKER_IS_CONFIG (config), NULL);
g_return_val_if_fail (language != NULL, NULL);
@@ -482,20 +483,29 @@
void
tracker_query_tree_set_indexer (TrackerQueryTree *tree,
- Indexer *indexer)
+ TrackerIndexer *indexer)
{
TrackerQueryTreePrivate *priv;
g_return_if_fail (TRACKER_IS_QUERY_TREE (tree));
- g_return_if_fail (indexer != NULL);
+ g_return_if_fail (TRACKER_IS_INDEXER (indexer));
priv = TRACKER_QUERY_TREE_GET_PRIVATE (tree);
+
+ if (indexer) {
+ g_object_ref (indexer);
+ }
+
+ if (priv->indexer) {
+ g_object_unref (priv->indexer);
+ }
+
priv->indexer = indexer;
g_object_notify (G_OBJECT (tree), "indexer");
}
-Indexer *
+TrackerIndexer *
tracker_query_tree_get_indexer (TrackerQueryTree *tree)
{
TrackerQueryTreePrivate *priv;
@@ -641,10 +651,14 @@
}
static gint
-get_idf_score (WordDetails *details, float idf)
+get_idf_score (TrackerIndexerWordDetails *details,
+ gfloat idf)
{
- guint32 score = tracker_word_details_get_score (details);
- float f = idf * score * SCORE_MULTIPLIER;
+ guint32 score;
+ gfloat f;
+
+ score = tracker_indexer_word_details_get_score (details);
+ f = idf * score * SCORE_MULTIPLIER;
return (f > 1.0) ? lrintf (f) : 1;
}
@@ -676,8 +690,8 @@
const gchar *term)
{
TrackerQueryTreePrivate *priv;
+ TrackerIndexerWordDetails *details;
GHashTable *result;
- WordDetails *details;
guint count, i;
priv = TRACKER_QUERY_TREE_GET_PRIVATE (tree);
@@ -693,7 +707,7 @@
SearchHitData *data;
gint service;
- service = tracker_word_details_get_service_type (&details[i]);
+ service = tracker_indexer_word_details_get_service_type (&details[i]);
if (in_array (priv->services, service)) {
data = g_slice_new (SearchHitData);
Modified: branches/indexer-split/src/trackerd/tracker-query-tree.h
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-query-tree.h (original)
+++ branches/indexer-split/src/trackerd/tracker-query-tree.h Thu Jun 26 11:38:21 2008
@@ -65,16 +65,16 @@
GType tracker_query_tree_get_type (void);
TrackerQueryTree * tracker_query_tree_new (const gchar *query_str,
- Indexer *indexer,
+ TrackerIndexer *indexer,
TrackerConfig *config,
TrackerLanguage *language,
GArray *services);
G_CONST_RETURN gchar *tracker_query_tree_get_query (TrackerQueryTree *tree);
void tracker_query_tree_set_query (TrackerQueryTree *tree,
const gchar *query_str);
-Indexer * tracker_query_tree_get_indexer (TrackerQueryTree *tree);
+TrackerIndexer * tracker_query_tree_get_indexer (TrackerQueryTree *tree);
void tracker_query_tree_set_indexer (TrackerQueryTree *tree,
- Indexer *indexer);
+ TrackerIndexer *indexer);
TrackerConfig * tracker_query_tree_get_config (TrackerQueryTree *tree);
void tracker_query_tree_set_config (TrackerQueryTree *tree,
TrackerConfig *config);
Modified: branches/indexer-split/src/trackerd/tracker-search.c
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-search.c (original)
+++ branches/indexer-split/src/trackerd/tracker-search.c Thu Jun 26 11:38:21 2008
@@ -50,8 +50,8 @@
DBusGProxy *fd_proxy;
TrackerConfig *config;
TrackerLanguage *language;
- Indexer *file_index;
- Indexer *email_index;
+ TrackerIndexer *file_index;
+ TrackerIndexer *email_index;
} TrackerSearchPriv;
enum {
@@ -217,8 +217,8 @@
}
void
-tracker_search_set_file_index (TrackerSearch *object,
- Indexer *file_index)
+tracker_search_set_file_index (TrackerSearch *object,
+ TrackerIndexer *file_index)
{
TrackerSearchPriv *priv;
@@ -233,8 +233,8 @@
}
void
-tracker_search_set_email_index (TrackerSearch *object,
- Indexer *email_index)
+tracker_search_set_email_index (TrackerSearch *object,
+ TrackerIndexer *email_index)
{
TrackerSearchPriv *priv;
Modified: branches/indexer-split/src/trackerd/tracker-search.h
==============================================================================
--- branches/indexer-split/src/trackerd/tracker-search.h (original)
+++ branches/indexer-split/src/trackerd/tracker-search.h Thu Jun 26 11:38:21 2008
@@ -59,9 +59,9 @@
void tracker_search_set_language (TrackerSearch *object,
TrackerLanguage *language);
void tracker_search_set_file_index (TrackerSearch *object,
- Indexer *file_index);
+ TrackerIndexer *file_index);
void tracker_search_set_email_index (TrackerSearch *object,
- Indexer *email_index);
+ TrackerIndexer *email_index);
gboolean tracker_search_get_hit_count (TrackerSearch *object,
const gchar *service,
const gchar *search_text,
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]
