Add C++ HelpSearch and call from XMLHelp. Fix string conversion bug.

7 files changed, 131 insertions, 18 deletions

diff --git a/l10ntools/inc/l10ntools/HelpSearch.hxx b/l10ntools/inc/l10ntools/HelpSearch.hxx
new file mode 100644
index 000000000000..4885b5698222
--- /dev/null
+++ b/l10ntools/inc/l10ntools/HelpSearch.hxx
@@ -0,0 +1,36 @@
+#ifndef HELPSEARCH_HXX
+#define HELPSEARCH_HXX
+
+#include <l10ntools/dllapi.h>
+
+#include <CLucene/StdHeader.h>
+#include <CLucene.h>
+
+#include <rtl/ustring.hxx>
+#include <vector>
+
+class L10N_DLLPUBLIC HelpSearch {
+	private:
+		rtl::OUString d_lang;
+		rtl::OUString d_indexDir;
+
+	public:
+
+	/**
+	 * @param lang Help files language.
+	 * @param indexDir The directory where the index files are stored.
+	 */
+	HelpSearch(rtl::OUString const &lang, rtl::OUString const &indexDir);
+
+	/**
+	 * Query the index for a certain query string.
+	 * @param queryStr The query.
+	 * @param captionOnly Set to true to search in the caption, not the content.
+	 * @param rDocuments Vector to write the paths of the found documents.
+	 * @param rScores Vector to write the scores to.
+	 */
+	bool query(rtl::OUString const &queryStr, bool captionOnly,
+		std::vector<rtl::OUString> &rDocuments, std::vector<float> &rScores);
+};
+
+#endif
diff --git a/l10ntools/prj/d.lst b/l10ntools/prj/d.lst
index 44cf5f001e14..e9329dc93855 100644
--- a/l10ntools/prj/d.lst
+++ b/l10ntools/prj/d.lst
@@ -48,6 +48,7 @@ mkdir: %_DEST%\bin\help\com\sun\star\help
 ..\inc\l10ntools\directory.hxx %_DEST%\inc\l10ntools\directory.hxx
 ..\inc\l10ntools\file.hxx %_DEST%\inc\l10ntools\file.hxx
 ..\inc\l10ntools\HelpIndexer.hxx %_DEST%\inc\l10ntools\HelpIndexer.hxx
+..\inc\l10ntools\HelpSearch.hxx %_DEST%\inc\l10ntools\HelpSearch.hxx
 ..\source\filter\merge\FCFGMerge.cfg  %_DEST%\inc\l10ntools\FCFGMerge.cfg
 
 ..\%__SRC%\lib\transex.lib %_DEST%\lib\transex.lib
diff --git a/l10ntools/source/help/HelpIndexer.cxx b/l10ntools/source/help/HelpIndexer.cxx
index b54814a41895..793348b2b2fa 100644
--- a/l10ntools/source/help/HelpIndexer.cxx
+++ b/l10ntools/source/help/HelpIndexer.cxx
@@ -1,4 +1,5 @@
 #include <l10ntools/HelpIndexer.hxx>
+#include "LuceneHelper.hxx"
 
 #define TODO
 
@@ -100,22 +101,6 @@ bool HelpIndexer::scanForFiles(rtl::OUString const & path) {
 	return true;
 }
 
-std::vector<TCHAR> OUStringToTCHARVec(rtl::OUString const &rStr)
-{
-    //UTF-16
-    if (sizeof(wchar_t) == sizeof(sal_Unicode))
-        return std::vector<TCHAR>(rStr.getStr(), rStr.getStr() + rStr.getLength());
-
-    //UTF-32
-    std::vector<TCHAR> aRet;
-    for (sal_Int32 nStrIndex = 0; nStrIndex < rStr.getLength();)
-    {
-        const sal_uInt32 nCode = rStr.iterateCodePoints(&nStrIndex);
-        aRet.push_back(nCode);
-    }
-    return aRet;
-}
-
 bool HelpIndexer::helpDocument(rtl::OUString const & fileName, Document *doc) {
 	// Add the help path as an indexed, untokenized field.
 	rtl::OUString path = rtl::OUString(RTL_CONSTASCII_USTRINGPARAM("#HLP#")) + d_module + rtl::OUString(RTL_CONSTASCII_USTRINGPARAM("/")) + fileName;
diff --git a/l10ntools/source/help/HelpSearch.cxx b/l10ntools/source/help/HelpSearch.cxx
new file mode 100644
index 000000000000..f50c44eb7cbd
--- /dev/null
+++ b/l10ntools/source/help/HelpSearch.cxx
@@ -0,0 +1,40 @@
+#include <l10ntools/HelpSearch.hxx>
+#include "LuceneHelper.hxx"
+
+#include <iostream>
+
+HelpSearch::HelpSearch(rtl::OUString const &lang, rtl::OUString const &indexDir) :
+d_lang(lang), d_indexDir(indexDir) {}
+
+bool HelpSearch::query(rtl::OUString const &queryStr, bool captionOnly,
+		std::vector<rtl::OUString> &rDocuments, std::vector<float> &rScores) {
+	rtl::OString pathStr;
+	d_indexDir.convertToString(&pathStr, RTL_TEXTENCODING_ASCII_US, 0);
+	lucene::index::IndexReader *reader = lucene::index::IndexReader::open(pathStr.getStr());
+	lucene::search::IndexSearcher searcher(reader);
+
+	TCHAR captionField[] = L"caption";
+	TCHAR contentField[] = L"content";
+	TCHAR *field = captionOnly ? captionField : contentField;
+
+	bool isWildcard = queryStr[queryStr.getLength() - 1] == L'*';
+	std::vector<TCHAR> aQueryStr(OUStringToTCHARVec(queryStr));
+	lucene::search::Query *aQuery = (isWildcard ?
+		(lucene::search::Query*)new lucene::search::WildcardQuery(new lucene::index::Term(field, &aQueryStr[0])) :
+		(lucene::search::Query*)new lucene::search::TermQuery(new lucene::index::Term(field, &aQueryStr[0])));
+	// FIXME: who is responsible for the Term*?
+
+	lucene::search::Hits *hits = searcher.search(aQuery);
+	for (unsigned i = 0; i < hits->length(); ++i) {
+		lucene::document::Document &doc = hits->doc(i); // Document* belongs to Hits.
+		wchar_t const *path = doc.get(L"path");
+		rDocuments.push_back(TCHARArrayToOUString(path != 0 ? path : L""));
+		rScores.push_back(hits->score(i));
+	}
+
+	delete hits;
+	delete aQuery;
+
+	reader->close();
+	return true;
+}
diff --git a/l10ntools/source/help/LuceneHelper.cxx b/l10ntools/source/help/LuceneHelper.cxx
new file mode 100644
index 000000000000..a88542f93009
--- /dev/null
+++ b/l10ntools/source/help/LuceneHelper.cxx
@@ -0,0 +1,33 @@
+#include "LuceneHelper.hxx"
+
+std::vector<TCHAR> OUStringToTCHARVec(rtl::OUString const &rStr)
+{
+    //UTF-16
+    if (sizeof(TCHAR) == sizeof(sal_Unicode))
+        return std::vector<TCHAR>(rStr.getStr(), rStr.getStr() + rStr.getLength() + 1);
+
+    //UTF-32
+    std::vector<TCHAR> aRet;
+    for (sal_Int32 nStrIndex = 0; nStrIndex < rStr.getLength() + 1; )
+    {
+        const sal_uInt32 nCode = rStr.iterateCodePoints(&nStrIndex);
+        aRet.push_back(nCode);
+    }
+    return aRet;
+}
+
+inline unsigned tstrlen(TCHAR const *str) {
+	unsigned i;
+	for (i = 0; str[i] != 0; ++i) {}
+	return i;
+}
+
+rtl::OUString TCHARArrayToOUString(TCHAR const *str)
+{
+	// UTF-16
+	if (sizeof(TCHAR) == sizeof(sal_Unicode))
+		return rtl::OUString((sal_Unicode*) str);
+
+	// UTF-32
+	return rtl::OUString((char*) str, tstrlen(str), RTL_TEXTENCODING_UCS4);
+}
diff --git a/l10ntools/source/help/LuceneHelper.hxx b/l10ntools/source/help/LuceneHelper.hxx
new file mode 100644
index 000000000000..7591b8ca0760
--- /dev/null
+++ b/l10ntools/source/help/LuceneHelper.hxx
@@ -0,0 +1,13 @@
+#ifndef LUCENEHELPER_HXX
+#define LUCENEHELPER_HXX
+
+#include <CLucene/StdHeader.h>
+#include <CLucene.h>
+
+#include <rtl/ustring.hxx>
+#include <vector>
+
+std::vector<TCHAR> OUStringToTCHARVec(rtl::OUString const &rStr);
+rtl::OUString TCHARArrayToOUString(TCHAR const *str);
+
+#endif
diff --git a/l10ntools/source/help/makefile.mk b/l10ntools/source/help/makefile.mk
index 2ae32329d546..a466e2c9fc24 100644
--- a/l10ntools/source/help/makefile.mk
+++ b/l10ntools/source/help/makefile.mk
@@ -56,12 +56,16 @@ OBJFILES=\
         $(OBJ)$/HelpLinker.obj \
         $(OBJ)$/HelpCompiler.obj \
         $(OBJ)$/HelpIndexer.obj \
-        $(OBJ)$/HelpIndexer_main.obj
+        $(OBJ)$/HelpIndexer_main.obj \
+	$(OBJ)$/HelpSearch.obj \
+	$(OBJ)$/LuceneHelper.obj
 
 SLOFILES=\
         $(SLO)$/HelpLinker.obj \
         $(SLO)$/HelpCompiler.obj \
-        $(SLO)$/HelpIndexer.obj
+	$(SLO)$/LuceneHelper.obj \
+        $(SLO)$/HelpIndexer.obj \
+	$(SLO)$/HelpSearch.obj
 
 .IF "$(OS)" == "MACOSX" && "$(CPU)" == "P" && "$(COM)" == "GCC"
 # There appears to be a GCC 4.0.1 optimization error causing _file:good() to
@@ -85,6 +89,7 @@ APP1STDLIBS+=$(SALLIB) $(BERKELEYLIB) $(XSLTLIB) $(EXPATASCII3RDLIB)
 
 APP2TARGET=HelpIndexer
 APP2OBJS=\
+      $(OBJ)$/LuceneHelper.obj \
       $(OBJ)$/HelpIndexer.obj \
       $(OBJ)$/HelpIndexer_main.obj
 APP2RPATH = NONE