dbcheck.cc

Go to the documentation of this file.

 
/* Copyright 2009 Richard Boulton
 * Copyright 2010,2015 Olly Betts
 *
 * This program 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 version 2 of the
 * License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301
 * USA
 */
 
#include <config.h>
 
#include "dbcheck.h"
 
#include "str.h"
#include "testsuite.h"
 
using namespace std;
 
string
positions_to_string(Xapian::PositionIterator & it,
                    const Xapian::PositionIterator & end,
                    Xapian::termcount * count)
{
    string result;
    bool need_comma = false;
    Xapian::termcount c = 0;
    while (it != end) {
        if (need_comma)
            result += ", ";
        result += str(*it);
        need_comma = true;
        ++it;
        ++c;
    }
    if (count) {
        *count = c;
    }
    return result;
}
 
string
postlist_to_string(const Xapian::Database & db, const string & tname)
{
    string result;
    bool need_comma = false;
 
    for (Xapian::PostingIterator p = db.postlist_begin(tname);
         p != db.postlist_end(tname);
         ++p) {
        if (need_comma)
            result += ", ";
 
        Xapian::PositionIterator it(p.positionlist_begin());
        string posrepr = positions_to_string(it, p.positionlist_end());
        if (!posrepr.empty()) {
            posrepr = ", pos=[" + posrepr + "]";
        }
 
        result += "(" + str(*p) +
                ", doclen=" + str(p.get_doclength()) +
                ", wdf=" + str(p.get_wdf()) +
                posrepr + ")";
        need_comma = true;
    }
    return result;
}
 
string
docterms_to_string(const Xapian::Database & db, Xapian::docid did)
{
    string result;
    bool need_comma = false;
 
    for (Xapian::TermIterator t = db.termlist_begin(did);
         t != db.termlist_end(did);
         ++t) {
        Xapian::PositionIterator it(t.positionlist_begin());
        string posrepr = positions_to_string(it, t.positionlist_end());
        if (!posrepr.empty()) {
            posrepr = ", pos=[" + posrepr + "]";
        }
        if (need_comma)
            result += ", ";
        result += "Term(" + *t + ", wdf=" + str(t.get_wdf()) + posrepr;
        result += ")";
        need_comma = true;
    }
    return result;
}
 
string
docstats_to_string(const Xapian::Database & db, Xapian::docid did)
{
    string result;
 
    result += "len=" + str(db.get_doclength(did));
 
    return result;
}
 
string
termstats_to_string(const Xapian::Database & db, const string & term)
{
    string result;
 
    result += "tf=" + str(db.get_termfreq(term));
    result += ",cf=" + str(db.get_collection_freq(term));
 
    return result;
}
 
void
dbcheck(const Xapian::Database & db,
        Xapian::doccount expected_doccount,
        Xapian::docid expected_lastdocid)
{
    TEST_EQUAL(db.get_doccount(), expected_doccount);
    TEST_EQUAL(db.get_lastdocid(), expected_lastdocid);
 
    // Note - may not be a very big type, but we're only expecting to use this
    // for small databases, so should be fine.
    unsigned long totlen = 0;
 
    // A map from term to a representation of the posting list for that term.
    // We build this up from the documents, and then check it against the
    // equivalent built up from the posting lists.
    map<string, string> posting_reprs;
    map<Xapian::valueno, string> value_reprs;
 
    Xapian::termcount doclen_lower_bound = Xapian::termcount(-1);
    Xapian::termcount doclen_upper_bound = 0;
 
    for (Xapian::PostingIterator dociter = db.postlist_begin(string());
         dociter != db.postlist_end(string());
         ++dociter) {
        Xapian::docid did = *dociter;
        TEST_EQUAL(dociter.get_wdf(), 1);
        Xapian::Document doc(db.get_document(did));
        Xapian::termcount doclen(db.get_doclength(did));
        Xapian::termcount unique_terms(db.get_unique_terms(did));
        if (doclen < doclen_lower_bound)
            doclen_lower_bound = doclen;
        if (doclen > doclen_upper_bound)
            doclen_upper_bound = doclen;
        totlen += doclen;
 
        Xapian::termcount found_termcount = 0;
        Xapian::termcount found_unique_terms = 0;
        Xapian::termcount wdf_sum = 0;
        Xapian::TermIterator t, t2;
        for (t = doc.termlist_begin(), t2 = db.termlist_begin(did);
             t != doc.termlist_end();
             ++t, ++t2) {
            TEST(t2 != db.termlist_end(did));
 
            ++found_termcount;
            auto wdf = t.get_wdf();
            if (wdf) ++found_unique_terms;
            wdf_sum += wdf;
 
            TEST_EQUAL(*t, *t2);
            TEST_EQUAL(t.get_wdf(), t2.get_wdf());
            TEST_EQUAL(db.get_termfreq(*t), t.get_termfreq());
            TEST_EQUAL(db.get_termfreq(*t), t2.get_termfreq());
 
            // Check the position lists are equal.
            Xapian::termcount tc1, tc2;
            Xapian::PositionIterator it1(t.positionlist_begin());
            string posrepr = positions_to_string(it1, t.positionlist_end(), &tc1);
            Xapian::PositionIterator it2(t2.positionlist_begin());
            string posrepr2 = positions_to_string(it2, t2.positionlist_end(), &tc2);
            TEST_EQUAL(posrepr, posrepr2);
            TEST_EQUAL(tc1, tc2);
            try {
                TEST_EQUAL(tc1, t.positionlist_count());
            } catch (const Xapian::UnimplementedError &) {
                // positionlist_count() isn't implemented for remote databases.
            }
 
            // Make a representation of the posting.
            if (!posrepr.empty()) {
                posrepr = ",[" + posrepr + "]";
            }
            string posting_repr = "(" + str(did) + "," +
                    str(t.get_wdf()) + "/" + str(doclen) +
                    posrepr + ")";
 
            // Append the representation to the list for the term.
            map<string, string>::iterator i = posting_reprs.find(*t);
            if (i == posting_reprs.end()) {
                posting_reprs[*t] = posting_repr;
            } else {
                i->second += "," + posting_repr;
            }
        }
 
        Xapian::termcount vcount = 0;
        for (Xapian::ValueIterator v = doc.values_begin();
             v != doc.values_end();
             ++v, ++vcount) {
            TEST((*v).size() != 0);
            string value_repr = "(" + str(did) + "," + *v + ")";
 
            // Append the values to the value lists.
            map<Xapian::valueno, string>::iterator i;
            i = value_reprs.find(v.get_valueno());
            if (i == value_reprs.end()) {
                value_reprs[v.get_valueno()] = value_repr;
            } else {
                i->second += "," + value_repr;
            }
        }
        TEST_EQUAL(vcount, doc.values_count());
        TEST(t2 == db.termlist_end(did));
        Xapian::termcount expected_termcount = doc.termlist_count();
        TEST_EQUAL(expected_termcount, found_termcount);
        // Ideally this would be equal, but currently we don't store the
        // unique_terms values but calculate them, and scanning the termlist
        // of each document would be slow, so instead get_unique_terms(did)
        // returns min(doclen, termcount) at present.
        TEST_REL(unique_terms, >=, found_unique_terms);
        TEST_REL(unique_terms, <=, found_termcount);
        TEST_REL(unique_terms, <=, doclen);
        TEST_EQUAL(doclen, wdf_sum);
    }
 
    TEST_REL(doclen_lower_bound, >=, db.get_doclength_lower_bound());
    TEST_REL(doclen_upper_bound, <=, db.get_doclength_upper_bound());
 
    Xapian::TermIterator t;
    map<string, string>::const_iterator i;
    for (t = db.allterms_begin(), i = posting_reprs.begin();
         t != db.allterms_end();
         ++t, ++i) {
        TEST(db.term_exists(*t));
        TEST(i != posting_reprs.end());
        TEST_EQUAL(i->first, *t);
 
        Xapian::doccount tf_count = 0;
        Xapian::termcount cf_count = 0;
        Xapian::termcount wdf_upper_bound = 0;
        string posting_repr;
        bool need_comma = false;
        for (Xapian::PostingIterator p = db.postlist_begin(*t);
             p != db.postlist_end(*t);
             ++p) {
            if (need_comma) {
                posting_repr += ",";
            }
 
            ++tf_count;
            cf_count += p.get_wdf();
 
            Xapian::PositionIterator it(p.positionlist_begin());
            string posrepr = positions_to_string(it, p.positionlist_end());
            if (!posrepr.empty()) {
                posrepr = ",[" + posrepr + "]";
            }
            posting_repr += "(" + str(*p) + "," +
                    str(p.get_wdf()) + "/" +
                    str(p.get_doclength()) + posrepr + ")";
            if (wdf_upper_bound < p.get_wdf())
                wdf_upper_bound = p.get_wdf();
            need_comma = true;
        }
 
        TEST_EQUAL(posting_repr, i->second);
        TEST_EQUAL(tf_count, t.get_termfreq());
        TEST_EQUAL(tf_count, db.get_termfreq(*t));
        TEST_EQUAL(cf_count, db.get_collection_freq(*t));
        TEST_REL(wdf_upper_bound, <=, db.get_wdf_upper_bound(*t));
    }
    TEST(i == posting_reprs.end());
 
    map<Xapian::valueno, string>::const_iterator j;
    for (j = value_reprs.begin(); j != value_reprs.end(); ++j) {
        string value_repr;
        string value_lower_bound;
        string value_upper_bound;
        bool first = true;
        for (Xapian::ValueIterator v = db.valuestream_begin(j->first);
             v != db.valuestream_end(j->first); ++v) {
            if (first) {
                value_lower_bound = *v;
                value_upper_bound = *v;
                first = false;
            } else {
                value_repr += ",";
                if (*v > value_upper_bound) {
                    value_upper_bound = *v;
                }
                if (*v < value_lower_bound) {
                    value_lower_bound = *v;
                }
            }
            value_repr += "(" + str(v.get_docid()) + "," + *v + ")";
        }
        TEST_EQUAL(value_repr, j->second);
        try {
            TEST_REL(value_upper_bound, <=, db.get_value_upper_bound(j->first));
            TEST_REL(value_lower_bound, >=, db.get_value_lower_bound(j->first));
        } catch (const Xapian::UnimplementedError &) {
            // Skip the checks if the methods to get the bounds aren't
            // implemented for this backend.
        }
    }
 
    if (expected_doccount == 0) {
        TEST_EQUAL(0, db.get_avlength());
    } else {
        TEST_EQUAL_DOUBLE(double(totlen) / expected_doccount,
                          db.get_avlength());
    }
}