Open Source Projects Europe

    return sterms;

    return sterms;

}

}

// Text splitter callback used to take note of the position of query terms 

// Text splitter callback used to take note of the position of query terms 

// inside the result text. This is then used to insert highlight tags. 

// inside the result text. This is then used to insert highlight tags. 

 public:

 public:

    // Out: begin and end byte positions of query terms/groups in text

    // Out: begin and end byte positions of query terms/groups in text

    vector<pair<int, int> > tboffs;  

    vector<pair<int, int> > tboffs;  

    :  m_wcount(0), m_groups(groups), m_slacks(slacks)

    :  m_wcount(0), m_groups(groups), m_slacks(slacks)

    {

    {

    for (vector<string>::const_iterator it = its.begin(); 

    for (vector<string>::const_iterator it = its.begin(); 

         it != its.end(); it++) {

         it != its.end(); it++) {

        m_terms.insert(*it);

        m_terms.insert(*it);

    // Callback called by the text-to-words breaker for each word

    // Callback called by the text-to-words breaker for each word

    virtual bool takeword(const std::string& term, int pos, int bts, int bte) {

    virtual bool takeword(const std::string& term, int pos, int bts, int bte) {

    string dumb;

    string dumb;

    if (!unacmaybefold(term, dumb, "UTF-8", true)) {

    if (!unacmaybefold(term, dumb, "UTF-8", true)) {

        LOGINFO(("PlainToRich::splitter::takeword: unac failed for [%s]\n",

        return true;

        return true;

    }

    }

    //LOGDEB2(("Input dumbbed term: '%s' %d %d %d\n", dumb.c_str(), 

    //LOGDEB2(("Input dumbbed term: '%s' %d %d %d\n", dumb.c_str(), 

    // pos, bts, bte));

    // pos, bts, bte));

    }

    }

    return false;

    return false;

}

}

// Check if there is a NEAR match for the group of terms

// Check if there is a NEAR match for the group of terms

{

{

        vecStringToString(terms).c_str()));

        vecStringToString(terms).c_str()));

    // The position lists we are going to work with. We extract them from the 

    // The position lists we are going to work with. We extract them from the 

    // (string->plist) map

    // (string->plist) map

    vector<vector<int>* > plists;

    vector<vector<int>* > plists;

    // stem-expanded: we don't know which matched)

    // stem-expanded: we don't know which matched)

    for (vector<string>::const_iterator it = terms.begin(); 

    for (vector<string>::const_iterator it = terms.begin(); 

     it != terms.end(); it++) {

     it != terms.end(); it++) {

    map<string, vector<int> >::iterator pl = m_plists.find(*it);

    map<string, vector<int> >::iterator pl = m_plists.find(*it);

    if (pl == m_plists.end()) {

    if (pl == m_plists.end()) {

            (*it).c_str()));

            (*it).c_str()));

        continue;

        continue;

    }

    }

    plists.push_back(&(pl->second));

    plists.push_back(&(pl->second));

    plistToTerm[&(pl->second)] = *it;

    plistToTerm[&(pl->second)] = *it;

    realgroup.push_back(*it);

    realgroup.push_back(*it);

    }

    }

         window, vecStringToString(realgroup).c_str()));

         window, vecStringToString(realgroup).c_str()));

    if (plists.size() < 2) {

    if (plists.size() < 2) {

    return false;

    return false;

    }

    }

    // Sort the positions lists so that the shorter is first

    // Sort the positions lists so that the shorter is first

    std::sort(plists.begin(), plists.end(), VecIntCmpShorter());

    std::sort(plists.begin(), plists.end(), VecIntCmpShorter());

     it != plists[0]->end(); it++) {

     it != plists[0]->end(); it++) {

    int pos = *it;

    int pos = *it;

    int sta = int(10E9), sto = 0;

    int sta = int(10E9), sto = 0;

    LOGDEB0(("MatchGroup: Testing at pos %d\n", pos));

    LOGDEB0(("MatchGroup: Testing at pos %d\n", pos));

    if (do_proximity_test(window, plists, 1, pos, pos, &sta, &sto)) {

    if (do_proximity_test(window, plists, 1, pos, pos, &sta, &sto)) {

             sta, sto)); 

             sta, sto)); 

        // Maybe extend the window by 1st term position, this was not

        // Maybe extend the window by 1st term position, this was not

        // done by do_prox..

        // done by do_prox..

        SETMINMAX(pos, sta, sto);

        SETMINMAX(pos, sta, sto);

        // Translate the position window into a byte offset window

        // Translate the position window into a byte offset window

        int bs = 0;

        int bs = 0;

        map<int, pair<int, int> >::iterator i1 =  m_gpostobytes.find(sta);

        map<int, pair<int, int> >::iterator i1 =  m_gpostobytes.find(sta);

        map<int, pair<int, int> >::iterator i2 =  m_gpostobytes.find(sto);

        map<int, pair<int, int> >::iterator i2 =  m_gpostobytes.find(sto);

        if (i1 != m_gpostobytes.end() && i2 != m_gpostobytes.end()) {

        if (i1 != m_gpostobytes.end() && i2 != m_gpostobytes.end()) {

            i1->second.first, i2->second.second));

            i1->second.first, i2->second.second));

        tboffs.push_back(pair<int, int>(i1->second.first, 

        tboffs.push_back(pair<int, int>(i1->second.first, 

                        i2->second.second));

                        i2->second.second));

        bs = i1->second.first;

        bs = i1->second.first;

        } else {

        } else {

    return a.second > b.second;

    return a.second > b.second;

    }

    }

};

};

// Do the phrase match thing, then merge the highlight lists

// Do the phrase match thing, then merge the highlight lists

{

{

    vector<vector<string> >::const_iterator vit = m_groups.begin();

    vector<vector<string> >::const_iterator vit = m_groups.begin();

    vector<int>::const_iterator sit = m_slacks.begin();

    vector<int>::const_iterator sit = m_slacks.begin();

    for (; vit != m_groups.end() && sit != m_slacks.end(); vit++, sit++) {

    for (; vit != m_groups.end() && sit != m_slacks.end(); vit++, sit++) {

    matchGroup(*vit, *sit + (*vit).size());

    matchGroup(*vit, *sit + (*vit).size());

    }

    }

    // Compute the positions for the query terms.  We use the text

    // Compute the positions for the query terms.  We use the text

    // splitter to break the text into words, and compare the words to

    // splitter to break the text into words, and compare the words to

    // the search terms,

    // the search terms,

    // Note: the splitter returns the term locations in byte, not

    // Note: the splitter returns the term locations in byte, not

    // character, offsets.

    // character, offsets.

    splitter.text_to_words(in);

    splitter.text_to_words(in);

    LOGDEB0(("plaintorich: split done %d mS\n", chron.millis()));

    LOGDEB0(("plaintorich: split done %d mS\n", chron.millis()));

    // Compute the positions for NEAR and PHRASE groups.

    // Compute the positions for NEAR and PHRASE groups.

    out.clear();

    out.clear();

    out.push_back("");

    out.push_back("");

    list<string>::iterator olit = out.begin();

    list<string>::iterator olit = out.begin();

    *olit = header();

    *olit = header();

    // Iterator for the list of input term positions. We use it to

    // Iterator for the list of input term positions. We use it to

    // output highlight tags and to compute term positions in the

    // output highlight tags and to compute term positions in the

    // output text

    // output text

#if 0

#if 0

    LOGDEB2(("plaintorich: region: %d %d\n", it->first, it->second));

    LOGDEB2(("plaintorich: region: %d %d\n", it->first, it->second));

    }

    }

#endif

#endif

    // Input character iterator

    // Input character iterator

            *olit += endMatch();

            *olit += endMatch();

            *olit += endAnchor();

            *olit += endAnchor();

        }

        }

        // Skip all highlight areas that would overlap this one

        // Skip all highlight areas that would overlap this one

        int crend = tPosIt->second;

        int crend = tPosIt->second;

            tPosIt++;

            tPosIt++;

                inrcltag = 0;

                inrcltag = 0;

        }

        }

    }

    }