[Opentrep-svn] SF.net SVN: opentrep:[160] trunk/opentrep
Status: Beta
Brought to you by:
denis_arnaud
Menu
▾
▴
[Opentrep-svn] SF.net SVN: opentrep:[160] trunk/opentrep
|
From: <den...@us...> - 2009-07-23 05:13:33
|
Revision: 160
http://opentrep.svn.sourceforge.net/opentrep/?rev=160&view=rev
Author: denis_arnaud
Date: 2009-07-23 05:13:30 +0000 (Thu, 23 Jul 2009)
Log Message:
-----------
[Dev] Second attempt to fix a matching bug.
Modified Paths:
--------------
trunk/opentrep/db/data/ref_place_names.csv
trunk/opentrep/opentrep/bom/ResultHolder.cpp
trunk/opentrep/opentrep/bom/ResultHolder.hpp
trunk/opentrep/opentrep/bom/StringMatcher.cpp
trunk/opentrep/opentrep/bom/StringMatcher.hpp
Modified: trunk/opentrep/db/data/ref_place_names.csv
===================================================================
--- trunk/opentrep/db/data/ref_place_names.csv 2009-07-23 04:35:16 UTC (rev 159)
+++ trunk/opentrep/db/data/ref_place_names.csv 2009-07-23 05:13:30 UTC (rev 160)
@@ -3550,7 +3550,7 @@
en,cdd,cauquira,cauquira,cauquira/hn:cauquira airport
en,cde,caledonia,caledonia,caledonia/pa
en,cdf,cortina d'ampezzo,cortina d'ampez,cortina d'ampezzo/it:fiames
-en,cdg,paris cdg,paris cdg,paris/fr:charles de gaulle,cdg,cdg
+en,cdg,paris cdg,paris cdg,paris/fr:charles de gaulle,cdg,cdg,charles de gaulle
en,cdh,camden,camden,camden/ar/us:harrell fld
en,cdi,cachoeiro,cachoeiro,cachoeiro de i/es/br:cachoeiro
en,cdj,conceicao do arag,conceicao do ar,conceicao do arag/pa/br
Modified: trunk/opentrep/opentrep/bom/ResultHolder.cpp
===================================================================
--- trunk/opentrep/opentrep/bom/ResultHolder.cpp 2009-07-23 04:35:16 UTC (rev 159)
+++ trunk/opentrep/opentrep/bom/ResultHolder.cpp 2009-07-23 05:13:30 UTC (rev 160)
@@ -68,30 +68,53 @@
}
// //////////////////////////////////////////////////////////////////////
- std::string ResultHolder::
- searchString (Xapian::MSet& ioMatchingSet,
- TravelQuery_T& ioPartialQueryString,
- NbOfErrors_T& ioMaxEditDistance,
- bool ioHasReachedMaximalAllowableEditDistance,
- Document& ioMatchingDocument) {
+ std::string ResultHolder::searchString (Xapian::MSet& ioMatchingSet,
+ TravelQuery_T& ioPartialQueryString,
+ Document& ioMatchingDocument) {
std::string oMatchedString;
// Catch any Xapian::Error exceptions thrown
try {
+ /**
+ The query string must first be checked, without allowing any
+ spelling errors, but by removing the furthest right word at
+ every step.
+ <br>If no match is found, the maximal allowable edit
+ distance/error becomes 1, and the process (trying to match
+ the whole sentence, then by removing the furthest right word,
+ etc.) is re-performed.
+ <br>If no match is found, the maximal allowable edit
+ distance/error becomes 2.
+ <br>And so on until the maximum of the edit distance/error
+ becomes greater than the maximal allowable distance/error.
+ reached.
+ */
+ NbOfErrors_T lMaxEditDistance = 0;
+ bool hasReachedMaximalAllowableEditDistance = false;
bool shouldStop = false;
while (shouldStop == false) {
- // Retrieve the list of documents matching the query string
+
+ // DEBUG
+ OPENTREP_LOG_DEBUG ("--------");
+ OPENTREP_LOG_DEBUG ("Current query string: `"
+ << ioPartialQueryString
+ << "', with a maximal edit distance of "
+ << lMaxEditDistance << ".");
+
+ // Retrieve the list of Xapian documents matching the query string
oMatchedString =
StringMatcher::searchString (ioMatchingSet, ioPartialQueryString,
- ioMaxEditDistance,
- ioHasReachedMaximalAllowableEditDistance,
+ lMaxEditDistance,
+ hasReachedMaximalAllowableEditDistance,
_database);
// DEBUG
- OPENTREP_LOG_DEBUG ("Current initial query string: `"
+ OPENTREP_LOG_DEBUG ("---- Current query string: `"
<< ioPartialQueryString << "' --- Kept query: `"
- << oMatchedString << "' for "
+ << oMatchedString
+ << "', with a maximal edit distance of "
+ << lMaxEditDistance << ", for "
<< ioMatchingSet.size() << " matches.");
if (ioMatchingSet.empty() == false) {
@@ -106,22 +129,15 @@
break;
}
- // Since the query, as is, yield no match, the furthest right
- // word must be removed from the query string.
- StringMatcher::removeFurthestRightWord (ioPartialQueryString);
-
+ // Allow for one more spelling error
+ ++lMaxEditDistance;
+
/**
- Stop when the resulting string gets empty.
-
- <br>Note that whether maximal allowable edit distance/error
- has been reached is not checked at that stage. That
- algorithm is performed independently for each level of
- maximal allowable edit distance/error. Only the caller
- (below) retriggers this process by changing the level of
- maximal allowable edit distance/error, until that latter be
- reached.
+ Stop when it is no longer necessary to increase the maximal
+ allowable edit distance, as it is already greater than the
+ maximum of the calculated edit distance.
*/
- if (ioPartialQueryString.empty() == true) {
+ if (hasReachedMaximalAllowableEditDistance == true) {
shouldStop = true;
}
}
@@ -141,39 +157,26 @@
// Catch any Xapian::Error exceptions thrown
try {
- bool shouldStop = false;
- NbOfErrors_T lMaxEditDistance = 0;
-
/**
- The query string must first be checked, without allowing any
- spelling errors, but by removing the furthest right word at
- every step.
- <br>If no match is found, the maximal allowable edit
- distance/error becomes 1, and the process (trying to match
- the whole sentence, then by removing the furthest right word,
- etc.) is re-performed.
- <br>If no match is found, the maximal allowable edit
- distance/error becomes 2.
- <br>And so on until the maximum of the edit distance/error
- becomes greater than the maximal allowable distance/error.
- reached.
+ A copy of the query is made, as that copy will be altered by
+ the below process, whereas a clean copy needs to be reprocessed
+ for each level of maximal edit distance/error.
+ <br>However, in case of match, the modifications on the query
+ string (lPartialQueryString) must be replicated on the
+ original one (ioPartialQueryString).
*/
+ TravelQuery_T lPartialQueryString (ioPartialQueryString);
+
+ bool shouldStop = false;
while (shouldStop == false) {
- /**
- A copy of the query is made, as that copy will be altered by
- the below process, whereas a clean copy needs to be reprocessed
- for each level of maximal edit distance/error.
- <br>However, at the end, the modifications on the query
- string (lPartialQueryString) must be replicated on the
- original one (ioPartialQueryString).
- */
- TravelQuery_T lPartialQueryString (ioPartialQueryString);
+ // DEBUG
+ OPENTREP_LOG_DEBUG ("----------------");
+ OPENTREP_LOG_DEBUG ("Current query string: `"
+ << lPartialQueryString << "'");
+
Xapian::MSet lMatchingSet;
- bool hasReachedMaximalAllowableEditDistance = false;
oMatchedString = searchString (lMatchingSet, lPartialQueryString,
- lMaxEditDistance,
- hasReachedMaximalAllowableEditDistance,
ioMatchingDocument);
if (oMatchedString.empty() == false) {
@@ -189,17 +192,28 @@
break;
}
- // Allow for one more spelling error
- ++lMaxEditDistance;
+ // Since the query, as is, yields no match, the furthest right
+ // word must be removed from the query string.
+ StringMatcher::removeFurthestRightWord (lPartialQueryString);
+
+ /**
+ Stop when the resulting string gets empty.
- /**
- Stop when it is no longer necessary to increase the maximal
- allowable edit distance, as it is already greater than the
- maximum of the calculated edit distance.
+ <br>Note that whether maximal allowable edit distance/error
+ has been reached is not checked at that stage. That
+ algorithm is performed independently for each level of
+ maximal allowable edit distance/error. Only the caller
+ (below) retriggers this process by changing the level of
+ maximal allowable edit distance/error, until that latter be
+ reached.
*/
- if (hasReachedMaximalAllowableEditDistance == true) {
- ioPartialQueryString = lPartialQueryString;
+ if (lPartialQueryString.empty() == true) {
shouldStop = true;
+
+ // DEBUG
+ OPENTREP_LOG_DEBUG ("----------------");
+ OPENTREP_LOG_DEBUG ("Still no match for current query string: `"
+ << ioPartialQueryString << "'");
}
}
@@ -221,7 +235,7 @@
bool shouldStop = false;
while (shouldStop == false) {
// DEBUG
- OPENTREP_LOG_DEBUG ("---------------------")
+ OPENTREP_LOG_DEBUG ("+++++++++++++++++++++")
OPENTREP_LOG_DEBUG ("Remaining part of the query string: `"
<< lRemainingQueryString << "'");
Modified: trunk/opentrep/opentrep/bom/ResultHolder.hpp
===================================================================
--- trunk/opentrep/opentrep/bom/ResultHolder.hpp 2009-07-23 04:35:16 UTC (rev 159)
+++ trunk/opentrep/opentrep/bom/ResultHolder.hpp 2009-07-23 05:13:30 UTC (rev 160)
@@ -48,16 +48,10 @@
/** Retrieve the document best matching the query string.
@param Xapian::MSet& The Xapian matching set. It can be empty.
@param TravelQuery_T& The partial query string.
- @param NbOfErrors_T& The maximal allowable edit distance/error.
- @param bool Whether or not the maximal allowable edit distance/error
- has become greater than the maximum of the edit distance/errors
- calculated on the phrase.
@param MatchingDocument_T& The best matching Xapian document (if found).
@return bool Whether such a best matching document has been found. */
std::string searchString (Xapian::MSet& ioMatchingSet,
TravelQuery_T& ioPartialQueryString,
- NbOfErrors_T& ioMaxEditDistance,
- bool ioHasReachedMaximalAllowableEditDistance,
Document& ioMatchingDocument);
/** Retrieve the document best matching the query string.
Modified: trunk/opentrep/opentrep/bom/StringMatcher.cpp
===================================================================
--- trunk/opentrep/opentrep/bom/StringMatcher.cpp 2009-07-23 04:35:16 UTC (rev 159)
+++ trunk/opentrep/opentrep/bom/StringMatcher.cpp 2009-07-23 05:13:30 UTC (rev 160)
@@ -92,7 +92,7 @@
searchString (Xapian::MSet& ioMatchingSet,
const std::string& iSearchString,
NbOfErrors_T& ioMaxEditDistance,
- bool ioHasReachedMaximalAllowableEditDistance,
+ bool& ioHasReachedMaximalAllowableEditDistance,
const Xapian::Database& ioDatabase) {
NbOfErrors_T lMaxEditDistance = std::numeric_limits<EditDistance_T>::min();
@@ -176,9 +176,11 @@
int nbMatches = ioMatchingSet.size();
// DEBUG
+ /*
OPENTREP_LOG_DEBUG ("Original query `" << lOriginalQueryString
<< "', i.e., `" << lOriginalQuery.get_description()
<< "' => " << nbMatches << " results found");
+ */
/**
When no match is found, we search on the corrected phrase/string
@@ -241,10 +243,12 @@
nbMatches = ioMatchingSet.size();
// DEBUG
+ /*
OPENTREP_LOG_DEBUG ("Corrected query `" << lCorrectedQueryString
<< "', i.e., `" << lCorrectedQuery.get_description()
<< "' => " << nbMatches
<< " results found on corrected string");
+ */
if (nbMatches != 0) {
/**
@@ -300,6 +304,7 @@
nbMatches = ioMatchingSet.size();
// DEBUG
+ /*
OPENTREP_LOG_DEBUG ("Query corrected as a full sentence `"
<< lFullWordCorrectedString
<< "' with an allowable maximal edit distance of "
@@ -308,6 +313,7 @@
<< ", i.e., `"<< lFullQueryCorrected.get_description()
<< "' => " << nbMatches
<< " results found on corrected full string");
+ */
if (nbMatches != 0) {
oMatchedString = lFullWordCorrectedString;
@@ -329,7 +335,7 @@
of the calculated edit distance, it becomes useless to go on
increasing the maximal allowable edit distance.
*/
- if (lMaxEditDistance <= ioMaxEditDistance) {
+ if (ioMaxEditDistance >= lMaxEditDistance) {
ioHasReachedMaximalAllowableEditDistance = true;
}
Modified: trunk/opentrep/opentrep/bom/StringMatcher.hpp
===================================================================
--- trunk/opentrep/opentrep/bom/StringMatcher.hpp 2009-07-23 04:35:16 UTC (rev 159)
+++ trunk/opentrep/opentrep/bom/StringMatcher.hpp 2009-07-23 05:13:30 UTC (rev 160)
@@ -29,7 +29,7 @@
@param Xapian::MSet& The Xapian matching set. It can be empty.
@param const std::string& The query string.
@param NbOfErrors_T& The maximal allowable edit distance/error.
- @param bool Whether or not the maximal allowable edit distance/error
+ @param bool& Whether or not the maximal allowable edit distance/error
has become greater than the maximum of the edit distance/errors
calculated on the phrase.
@param const Xapian::Database& The Xapian index/database.
@@ -38,7 +38,7 @@
static std::string searchString (Xapian::MSet&,
const std::string& iSearchString,
NbOfErrors_T& ioMaxEditDistance,
- bool ioHasReachedMaximalAllowableEditDistance,
+ bool& ioHasReachedMaximalAllowableEditDistance,
const Xapian::Database&);
/** Extract the best matching Xapian document.
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
