Xapian: Internal Source Documentation: xapian-core: matcher/andmaybepostlist.cc Source File

00001 /* andmaybepostlist.cc: Merged postlist; items from one list, weights from both
00002  *
00003  * Copyright 1999,2000,2001 BrightStation PLC
00004  * Copyright 2002 Ananova Ltd
00005  * Copyright 2003,2004,2005,2008,2009 Olly Betts
00006  * Copyright 2009 Lemur Consulting Ltd
00007  *
00008  * This program is free software; you can redistribute it and/or
00009  * modify it under the terms of the GNU General Public License as
00010  * published by the Free Software Foundation; either version 2 of the
00011  * License, or (at your option) any later version.
00012  *
00013  * This program is distributed in the hope that it will be useful,
00014  * but WITHOUT ANY WARRANTY; without even the implied warranty of
00015  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00016  * GNU General Public License for more details.
00017  *
00018  * You should have received a copy of the GNU General Public License
00019  * along with this program; if not, write to the Free Software
00020  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301
00021  * USA
00022  */
00023 
00024 #include <config.h>
00025 #include "andmaybepostlist.h"
00026 
00027 #include "debuglog.h"
00028 #include "multiandpostlist.h"
00029 #include "omassert.h"
00030 
00031 PostList *
00032 AndMaybePostList::process_next_or_skip_to(Xapian::weight w_min, PostList *ret)
00033 {
00034     LOGCALL(MATCH, PostList *, "AndMaybePostList::process_next_or_skip_to", w_min | ret);
00035     handle_prune(l, ret);
00036     if (l->at_end()) {
00037         // once l is over, so is the AND MAYBE
00038         lhead = 0;
00039         RETURN(NULL);
00040     }
00041 
00042     lhead = l->get_docid();
00043     if (lhead <= rhead) RETURN(NULL);
00044 
00045     bool valid;
00046     check_handling_prune(r, lhead, w_min - lmax, matcher, valid);
00047     if (r->at_end()) {
00048         PostList *tmp = l;
00049         l = NULL;
00050         RETURN(tmp);
00051     }
00052     if (valid) {
00053         rhead = r->get_docid();
00054     } else {
00055         rhead = 0;
00056     }
00057     RETURN(NULL);
00058 }
00059 
00060 PostList *
00061 AndMaybePostList::sync_rhs(Xapian::weight w_min)
00062 {
00063     LOGCALL(MATCH, PostList *, "AndMaybePostList::sync_rhs", w_min);
00064     bool valid;
00065     check_handling_prune(r, lhead, w_min - lmax, matcher, valid);
00066     if (r->at_end()) {
00067         PostList *tmp = l;
00068         l = NULL;
00069         RETURN(tmp);
00070     }
00071     if (valid) {
00072         rhead = r->get_docid();
00073     } else {
00074         rhead = 0;
00075     }
00076     RETURN(NULL);
00077 }
00078 
00079 PostList *
00080 AndMaybePostList::next(Xapian::weight w_min)
00081 {
00082     LOGCALL(MATCH, PostList *, "AndMaybePostList::next", w_min);
00083     if (w_min > lmax) {
00084         // we can replace the AND MAYBE with an AND
00085         PostList *ret;
00086         LOGLINE(MATCH, "AND MAYBE -> AND");
00087         ret = new MultiAndPostList(l, r, lmax, rmax, matcher, dbsize, true);
00088         l = r = NULL;
00089         skip_to_handling_prune(ret, std::max(lhead, rhead) + 1, w_min, matcher);
00090         RETURN(ret);
00091     }
00092     RETURN(process_next_or_skip_to(w_min, l->next(w_min - rmax)));
00093 }
00094 
00095 PostList *
00096 AndMaybePostList::skip_to(Xapian::docid did, Xapian::weight w_min)
00097 {
00098     LOGCALL(MATCH, PostList *, "AndMaybePostList::skip_to", did | w_min);
00099     if (w_min > lmax) {
00100         // we can replace the AND MAYBE with an AND
00101         PostList *ret;
00102         LOGLINE(MATCH, "AND MAYBE -> AND (in skip_to)");
00103         ret = new MultiAndPostList(l, r, lmax, rmax, matcher, dbsize, true);
00104         did = std::max(did, std::max(lhead, rhead));
00105         l = r = NULL;
00106         skip_to_handling_prune(ret, did, w_min, matcher);
00107         RETURN(ret);
00108     }
00109 
00110     // exit if we're already past the skip point (or at it)
00111     if (did <= lhead) RETURN(NULL);
00112 
00113     RETURN(process_next_or_skip_to(w_min, l->skip_to(did, w_min - rmax)));
00114 }
00115 
00116 Xapian::doccount
00117 AndMaybePostList::get_termfreq_max() const
00118 {
00119     LOGCALL(MATCH, Xapian::doccount, "AndMaybePostList::get_termfreq_max", NO_ARGS);
00120     // Termfreq is exactly that of left hand branch.
00121     RETURN(l->get_termfreq_max());
00122 }
00123 
00124 Xapian::doccount
00125 AndMaybePostList::get_termfreq_min() const
00126 {
00127     LOGCALL(MATCH, Xapian::doccount, "AndMaybePostList::get_termfreq_min", NO_ARGS);
00128     // Termfreq is exactly that of left hand branch.
00129     RETURN(l->get_termfreq_min());
00130 }
00131 
00132 Xapian::doccount
00133 AndMaybePostList::get_termfreq_est() const
00134 {
00135     LOGCALL(MATCH, Xapian::doccount, "AndMaybePostList::get_termfreq_est", NO_ARGS);
00136     // Termfreq is exactly that of left hand branch.
00137     RETURN(l->get_termfreq_est());
00138 }
00139 
00140 TermFreqs
00141 AndMaybePostList::get_termfreq_est_using_stats(
00142         const Xapian::Weight::Internal & stats) const
00143 {
00144     LOGCALL(MATCH, TermFreqs, "AndMaybePostList::get_termfreq_est_using_stats", stats);
00145     // Termfreq is exactly that of left hand branch.
00146     RETURN(l->get_termfreq_est_using_stats(stats));
00147 }
00148 
00149 Xapian::docid
00150 AndMaybePostList::get_docid() const
00151 {
00152     LOGCALL(MATCH, Xapian::docid, "AndMaybePostList::get_docid", NO_ARGS);
00153     Assert(lhead != 0); // check we've started
00154     RETURN(lhead);
00155 }
00156 
00157 // only called if we are doing a probabilistic AND MAYBE
00158 Xapian::weight
00159 AndMaybePostList::get_weight() const
00160 {
00161     LOGCALL(MATCH, Xapian::weight, "AndMaybePostList::get_weight", NO_ARGS);
00162     Assert(lhead != 0); // check we've started
00163     if (lhead == rhead) RETURN(l->get_weight() + r->get_weight());
00164     RETURN(l->get_weight());
00165 }
00166 
00167 // only called if we are doing a probabilistic operation
00168 Xapian::weight
00169 AndMaybePostList::get_maxweight() const
00170 {
00171     LOGCALL(MATCH, Xapian::weight, "AndMaybePostList::get_maxweight", NO_ARGS);
00172     RETURN(lmax + rmax);
00173 }
00174 
00175 Xapian::weight
00176 AndMaybePostList::recalc_maxweight()
00177 {
00178     LOGCALL(MATCH, Xapian::weight, "AndMaybePostList::recalc_maxweight", NO_ARGS);
00179     lmax = l->recalc_maxweight();
00180     rmax = r->recalc_maxweight();
00181     RETURN(AndMaybePostList::get_maxweight());
00182 }
00183 
00184 bool
00185 AndMaybePostList::at_end() const
00186 {
00187     LOGCALL(MATCH, bool, "AndMaybePostList::at_end", NO_ARGS);
00188     RETURN(lhead == 0);
00189 }
00190 
00191 std::string
00192 AndMaybePostList::get_description() const
00193 {
00194     return "(" + l->get_description() + " AndMaybe " + r->get_description() +
00195            ")";
00196 }
00197 
00198 Xapian::termcount
00199 AndMaybePostList::get_doclength() const
00200 {
00201     LOGCALL(MATCH, Xapian::termcount, "AndMaybePostList::get_doclength", NO_ARGS);
00202     Assert(lhead != 0); // check we've started
00203     if (lhead == rhead) AssertEq(l->get_doclength(), r->get_doclength());
00204     RETURN(l->get_doclength());
00205 }
00206 
00207 Xapian::termcount
00208 AndMaybePostList::get_wdf() const
00209 {
00210     LOGCALL(MATCH, Xapian::termcount, "AndMaybePostList::get_wdf", NO_ARGS);
00211     if (lhead == rhead) RETURN(l->get_wdf() + r->get_wdf());
00212     RETURN(l->get_wdf());
00213 }
00214 
00215 Xapian::termcount
00216 AndMaybePostList::count_matching_subqs() const
00217 {
00218     LOGCALL(MATCH, Xapian::termcount, "AndMaybePostList::count_matching_subqs", NO_ARGS);
00219     if (lhead == rhead)
00220         RETURN(l->count_matching_subqs() + r->count_matching_subqs());
00221     RETURN(l->count_matching_subqs());
00222 }