tris/recoll
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

plaintorich: indent and log lines

Browse Source
This commit is contained in:
Jean-Francois Dockes 2017-12-26 12:42:53 +01:00
parent 51189fa49e
commit 3f8f31732d
2 changed files with 213 additions and 215 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

66
src/query/plaintorich.cpp
View File

@ -47,22 +47,20 @@ struct MatchEntry {
// match to the original user input. // match to the original user input.
size_t grpidx; size_t grpidx;
MatchEntry(int sta, int sto, size_t idx) MatchEntry(int sta, int sto, size_t idx)
: offs(sta, sto), grpidx(idx) : offs(sta, sto), grpidx(idx) {
{
} }
}; };
// Text splitter used to take note of the position of query terms // Text splitter 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.
class TextSplitPTR : public TextSplit { class TextSplitPTR : public TextSplit {
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<MatchEntry> tboffs; vector<MatchEntry> tboffs;
TextSplitPTR(const HighlightData& hdata) TextSplitPTR(const HighlightData& hdata)
: m_wcount(0), m_hdata(hdata) : m_wcount(0), m_hdata(hdata) {
{
// We separate single terms and groups and extract the group // We separate single terms and groups and extract the group
// terms for computing positions list before looking for group // terms for computing positions list before looking for group
// matches // matches
@ -86,17 +84,19 @@ class TextSplitPTR : public TextSplit {
string dumb = term; string dumb = term;
if (o_index_stripchars) { if (o_index_stripchars) {
if (!unacmaybefold(term, dumb, "UTF-8", UNACOP_UNACFOLD)) { if (!unacmaybefold(term, dumb, "UTF-8", UNACOP_UNACFOLD)) {
LOGINFO("PlainToRich::takeword: unac failed for [" << (term) << "]\n" ); LOGINFO("PlainToRich::takeword: unac failed for [" << term <<
"]\n");
return true; return true;
} }
} }
//LOGDEB2("Input dumbbed term: '" << (dumb) << "' " << (// pos) << " " << (bts) << " " << (bte) << "\n" ); LOGDEB2("Input dumbbed term: '" << dumb << "' " << pos << " " << bts
<< " " << bte << "\n");
// If this word is a search term, remember its byte-offset span. // If this word is a search term, remember its byte-offset span.
map<string, size_t>::const_iterator it = m_terms.find(dumb); map<string, size_t>::const_iterator it = m_terms.find(dumb);
if (it != m_terms.end()) { if (it != m_terms.end()) {
tboffs.push_back(MatchEntry(bts, bte, (*it).second)); tboffs.push_back(MatchEntry(bts, bte, it->second));
} }
// If word is part of a search group, update its positions list // If word is part of a search group, update its positions list
@ -104,7 +104,8 @@ class TextSplitPTR : public TextSplit {
// Term group (phrase/near) handling // Term group (phrase/near) handling
m_plists[dumb].push_back(pos); m_plists[dumb].push_back(pos);
m_gpostobytes[pos] = pair<int,int>(bts, bte); m_gpostobytes[pos] = pair<int,int>(bts, bte);
//LOGDEB2("Recorded bpos for " << (pos) << ": " << (bts) << " " << (bte) << "\n" ); LOGDEB2("Recorded bpos for " << pos << ": " << bts << " " <<
bte << "\n");
} }
// Check for cancellation request // Check for cancellation request
@ -139,9 +140,8 @@ private:
/** Sort by shorter comparison class */ /** Sort by shorter comparison class */
class VecIntCmpShorter { class VecIntCmpShorter {
public: public:
/** Return true if and only if a is strictly shorter than b. /** Return true if and only if a is strictly shorter than b. */
*/
bool operator()(const vector<int> *a, const vector<int> *b) { bool operator()(const vector<int> *a, const vector<int> *b) {
return a->size() < b->size(); return a->size() < b->size();
} }
@ -167,7 +167,8 @@ static bool do_proximity_test(int window, vector<vector<int>* >& plists,
unsigned int i, int min, int max, unsigned int i, int min, int max,
int *sp, int *ep, int minpos) int *sp, int *ep, int minpos)
{ {
LOGDEB1("do_prox_test: win " << (window) << " i " << (i) << " min " << (min) << " max " << (max) << " minpos " << (minpos) << "\n" ); LOGDEB1("do_prox_test: win " << window << " i " << i << " min " <<
min << " max " << max << " minpos " << minpos << "\n");
int tmp = max + 1 - window; int tmp = max + 1 - window;
if (tmp < minpos) if (tmp < minpos)
tmp = minpos; tmp = minpos;
@ -204,7 +205,8 @@ bool TextSplitPTR::matchGroup(unsigned int grpidx)
const vector<string>& terms = m_hdata.groups[grpidx]; const vector<string>& terms = m_hdata.groups[grpidx];
int window = int(m_hdata.groups[grpidx].size() + m_hdata.slacks[grpidx]); int window = int(m_hdata.groups[grpidx].size() + m_hdata.slacks[grpidx]);
LOGDEB1("TextSplitPTR::matchGroup:d " << (window) << ": " << (vecStringToString(terms)) << "\n" ); LOGDEB1("TextSplitPTR::matchGroup:d " << window << ": " <<
stringsToString(terms) << "\n");
// 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
@ -220,7 +222,8 @@ bool TextSplitPTR::matchGroup(unsigned int grpidx)
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()) {
LOGDEB1("TextSplitPTR::matchGroup: [" << ((*it)) << "] not found in m_plists\n" ); LOGDEB1("TextSplitPTR::matchGroup: [" << *it <<
"] not found in m_plists\n");
return false; return false;
} }
plists.push_back(&(pl->second)); plists.push_back(&(pl->second));
@ -229,7 +232,7 @@ bool TextSplitPTR::matchGroup(unsigned int grpidx)
// I think this can't actually happen, was useful when we used to // I think this can't actually happen, was useful when we used to
// prune the groups, but doesn't hurt. // prune the groups, but doesn't hurt.
if (plists.size() < 2) { if (plists.size() < 2) {
LOGDEB1("TextSplitPTR::matchGroup: no actual groups found\n" ); LOGDEB1("TextSplitPTR::matchGroup: no actual groups found\n");
return false; return false;
} }
// Sort the positions lists so that the shorter is first // Sort the positions lists so that the shorter is first
@ -240,10 +243,11 @@ bool TextSplitPTR::matchGroup(unsigned int grpidx)
it = plistToTerm.find(plists[0]); it = plistToTerm.find(plists[0]);
if (it == plistToTerm.end()) { if (it == plistToTerm.end()) {
// SuperWeird // SuperWeird
LOGERR("matchGroup: term for first list not found !?!\n" ); LOGERR("matchGroup: term for first list not found !?!\n");
return false; return false;
} }
LOGDEB1("matchGroup: walking the shortest plist. Term [" << (it->second) << "], len " << (plists[0]->size()) << "\n" ); LOGDEB1("matchGroup: walking the shortest plist. Term [" <<
it->second << "], len " << plists[0]->size() << "\n");
} }
// Minpos is the highest end of a found match. While looking for // Minpos is the highest end of a found match. While looking for
@ -256,9 +260,10 @@ bool TextSplitPTR::matchGroup(unsigned int grpidx)
it != plists[0]->end(); it++) { it != plists[0]->end(); it++) {
int pos = *it; int pos = *it;
int sta = INT_MAX, sto = 0; int sta = INT_MAX, sto = 0;
LOGDEB2("MatchGroup: Testing at pos " << (pos) << "\n" ); LOGDEB2("MatchGroup: Testing at pos " << pos << "\n");
if (do_proximity_test(window,plists, 1, pos, pos, &sta, &sto, minpos)) { if (do_proximity_test(window,plists, 1, pos, pos, &sta, &sto, minpos)) {
LOGDEB1("TextSplitPTR::matchGroup: MATCH termpos [" << (sta) << "," << (sto) << "]\n" ); LOGDEB1("TextSplitPTR::matchGroup: MATCH termpos [" << sta <<
"," << sto << "]\n");
// 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);
@ -267,14 +272,16 @@ bool TextSplitPTR::matchGroup(unsigned int grpidx)
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()) {
LOGDEB2("TextSplitPTR::matchGroup: pushing bpos " << (i1->second.first) << " " << (i2->second.second) << "\n" ); LOGDEB2("TextSplitPTR::matchGroup: pushing bpos " <<
i1->second.first << " " << i2->second.second << "\n");
tboffs.push_back(MatchEntry(i1->second.first, tboffs.push_back(MatchEntry(i1->second.first,
i2->second.second, grpidx)); i2->second.second, grpidx));
} else { } else {
LOGDEB0("matchGroup: no bpos found for " << (sta) << " or " << (sto) << "\n" ); LOGDEB0("matchGroup: no bpos found for " << sta << " or "
<< sto << "\n");
} }
} else { } else {
LOGDEB1("matchGroup: no group match found at this position\n" ); LOGDEB1("matchGroup: no group match found at this position\n");
} }
} }
@ -325,7 +332,7 @@ bool PlainToRich::plaintorich(const string& in,
{ {
Chrono chron; Chrono chron;
bool ret = true; bool ret = true;
LOGDEB1("plaintorichich: in: [" << (in) << "]\n" ); LOGDEB1("plaintorichich: in: [" << in << "]\n");
m_hdata = &hdata; m_hdata = &hdata;
// Compute the positions for the query terms. We use the text // Compute the positions for the query terms. We use the text
@ -335,10 +342,10 @@ bool PlainToRich::plaintorich(const string& in,
// 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);
LOGDEB2("plaintorich: split done " << (chron.millis()) << " mS\n" ); LOGDEB2("plaintorich: split done " << chron.millis() << " mS\n");
// Compute the positions for NEAR and PHRASE groups. // Compute the positions for NEAR and PHRASE groups.
splitter.matchGroups(); splitter.matchGroups();
LOGDEB2("plaintorich: group match done " << (chron.millis()) << " mS\n" ); LOGDEB2("plaintorich: group match done " << chron.millis() << " mS\n");
out.clear(); out.clear();
out.push_back(""); out.push_back("");
@ -351,7 +358,7 @@ bool PlainToRich::plaintorich(const string& in,
// a term match when we are actually looking for a group match // a term match when we are actually looking for a group match
// (the snippet generator does this...). // (the snippet generator does this...).
if (splitter.tboffs.empty()) { if (splitter.tboffs.empty()) {
LOGDEB1("plaintorich: no term matches\n" ); LOGDEB1("plaintorich: no term matches\n");
ret = false; ret = false;
} }
@ -364,7 +371,7 @@ bool PlainToRich::plaintorich(const string& in,
#if 0 #if 0
for (vector<pair<int, int> >::const_iterator it = splitter.tboffs.begin(); for (vector<pair<int, int> >::const_iterator it = splitter.tboffs.begin();
it != splitter.tboffs.end(); it++) { it != splitter.tboffs.end(); it++) {
LOGDEB2("plaintorich: region: " << (it->first) << " " << (it->second) << "\n" ); LOGDEB2("plaintorich: region: " << it->first << " "<<it->second<< "\n");
} }
#endif #endif
@ -524,7 +531,6 @@ bool PlainToRich::plaintorich(const string& in,
fclose(fp); fclose(fp);
} }
#endif #endif
LOGDEB2("plaintorich: done " << (chron.millis()) << " mS\n" ); LOGDEB2("plaintorich: done " << chron.millis() << " mS\n");
return ret; return ret;
} }

22
src/query/plaintorich.h
View File

@ -32,16 +32,12 @@
class PlainToRich { class PlainToRich {
public: public:
PlainToRich() PlainToRich()
: m_inputhtml(false), m_eolbr(false), m_hdata(0) : m_inputhtml(false), m_eolbr(false), m_hdata(0) {
{
} }
virtual ~PlainToRich() virtual ~PlainToRich() {}
{
}
void set_inputhtml(bool v) void set_inputhtml(bool v) {
{
m_inputhtml = v; m_inputhtml = v;
} }
@ -73,26 +69,22 @@ public:
/* Overridable output methods for headers, highlighting and marking tags */ /* Overridable output methods for headers, highlighting and marking tags */
virtual std::string header() virtual std::string header() {
{
return cstr_null; return cstr_null;
} }
/** Return match prefix (e.g.: <div class="match">). /** Return match prefix (e.g.: <div class="match">).
@param groupidx the index into hdata.groups */ @param groupidx the index into hdata.groups */
virtual std::string startMatch(unsigned int) virtual std::string startMatch(unsigned int) {
{
return cstr_null; return cstr_null;
} }
/** Return data for end of match area (e.g.: </div>). */ /** Return data for end of match area (e.g.: </div>). */
virtual std::string endMatch() virtual std::string endMatch() {
{
return cstr_null; return cstr_null;
} }
virtual std::string startChunk() virtual std::string startChunk() {
{
return cstr_null; return cstr_null;
} }