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

Remove 'prolounged sound mark' from the end of katakana words

Browse Source
This commit is contained in:
Jean-Francois Dockes 2017-04-21 12:13:10 +02:00
parent adaf7c77f9
commit fca5e19504
1 changed files with 25 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
src/rcldb/termproc.h
View File

@ -23,6 +23,7 @@
#include "textsplit.h" #include "textsplit.h"
#include "stoplist.h" #include "stoplist.h"
#include "smallut.h" #include "smallut.h"
#include "utf8iter.h"
namespace Rcl { namespace Rcl {
@ -128,15 +129,17 @@ public:
{ {
m_totalterms++; m_totalterms++;
string otrm; string otrm;
if (!unacmaybefold(itrm, otrm, "UTF-8", UNACOP_UNACFOLD)) { if (!unacmaybefold(itrm, otrm, "UTF-8", UNACOP_UNACFOLD)) {
LOGDEB("splitter::takeword: unac [" << (itrm) << "] failed\n" ); LOGDEB("splitter::takeword: unac [" << itrm << "] failed\n");
m_unacerrors++; m_unacerrors++;
// We don't generate a fatal error because of a bad term, // We don't generate a fatal error because of a bad term,
// but one has to put the limit somewhere // but one has to put the limit somewhere
if (m_unacerrors > 500 && if (m_unacerrors > 500 &&
(double(m_totalterms) / double(m_unacerrors)) < 2.0) { (double(m_totalterms) / double(m_unacerrors)) < 2.0) {
// More than 1 error for every other term // More than 1 error for every other term
LOGERR("splitter::takeword: too many unac errors " << (m_unacerrors) << "/" << (m_totalterms) << "\n" ); LOGERR("splitter::takeword: too many unac errors " <<
m_unacerrors << "/" << m_totalterms << "\n");
return false; return false;
} }
return true; return true;
@ -150,6 +153,26 @@ public:
return true; return true;
} }
// We should have a Japanese stemmer to handle this, but for
// experimenting, let's do it here: remove 'prolounged sound
// mark' and its halfwidth variant from the end of terms.
if ((unsigned int)otrm[0] > 127) {
Utf8Iter it(otrm);
if (TextSplit::isKATAKANA(*it)) {
Utf8Iter itprev = it;
while (*it != (unsigned int)-1) {
itprev = it;
it++;
}
if (*itprev == 0x30fc || *itprev == 0xff70) {
otrm = otrm.substr(0, itprev.getBpos());
}
}
}
if (otrm.empty()) {
return true;
}
// It may also occur that unac introduces spaces in the string // It may also occur that unac introduces spaces in the string
// (when removing isolated accents, may happen for Greek // (when removing isolated accents, may happen for Greek
// for example). This is a pathological situation. We // for example). This is a pathological situation. We