diff options
author | Gert van Valkenhoef <g.h.m.van.valkenhoef@rug.nl> | 2012-02-19 18:34:44 +0100 |
---|---|---|
committer | Caolán McNamara <caolanm@redhat.com> | 2012-02-23 10:31:18 +0000 |
commit | c4e3596820d0d71a86bd2f88a428b34d2c587074 (patch) | |
tree | 9741818786bc63e16bf5c221c12826fe221e298c /l10ntools | |
parent | 71231b9ea4a16e3582566e12a619b4deddef87c5 (diff) |
HelpIndexer and HelpSearch: remove Java
Diffstat (limited to 'l10ntools')
-rw-r--r-- | l10ntools/source/help/HelpFileDocument.java | 86 | ||||
-rw-r--r-- | l10ntools/source/help/HelpIndexerTool.java | 390 |
2 files changed, 0 insertions, 476 deletions
diff --git a/l10ntools/source/help/HelpFileDocument.java b/l10ntools/source/help/HelpFileDocument.java deleted file mode 100644 index 15e62d6c6c52..000000000000 --- a/l10ntools/source/help/HelpFileDocument.java +++ /dev/null @@ -1,86 +0,0 @@ -/************************************************************************* - * - * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. - * - * Copyright 2000, 2010 Oracle and/or its affiliates. - * - * OpenOffice.org - a multi-platform office productivity suite - * - * This file is part of OpenOffice.org. - * - * OpenOffice.org is free software: you can redistribute it and/or modify - * it under the terms of the GNU Lesser General Public License version 3 - * only, as published by the Free Software Foundation. - * - * OpenOffice.org is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU Lesser General Public License version 3 for more details - * (a copy is included in the LICENSE file that accompanied this code). - * - * You should have received a copy of the GNU Lesser General Public License - * version 3 along with OpenOffice.org. If not, see - * <http://www.openoffice.org/license.html> - * for a copy of the LGPLv3 License. - * - ************************************************************************/ - -package com.sun.star.help; - -import java.io.File; -import java.io.Reader; -import java.io.FileInputStream; -import java.io.InputStreamReader; -//import java.io.FileReader; -import java.io.StringReader; - -import org.apache.lucene.document.Document; -import org.apache.lucene.document.Field; - -/** Lucene Document for help files */ -public class HelpFileDocument -{ - /** Creates reader for UTF-8 files - */ - private static Reader getReaderForFile( File aFile ) - throws java.io.FileNotFoundException, java.io.UnsupportedEncodingException { - Reader aReader; - if( aFile != null ) { - FileInputStream fis = new FileInputStream( aFile ); - aReader = new InputStreamReader( fis, "UTF-8" ); - } - else { - aReader = new StringReader( "" ); - } - return aReader; - } - - /** Makes a document for a File. - */ - public static Document Document( String aModule, File aCaptionFile, File aContentFile ) - throws java.io.FileNotFoundException, java.io.UnsupportedEncodingException { - Document doc = new Document(); - - // Add the path of the file as a field named "path". Use a field that is - // indexed (i.e. searchable), but don't tokenize the field into words. - File aFile = aCaptionFile != null ? aCaptionFile : aContentFile; - if( aFile != null ) - { - String aPath = "#HLP#" + aModule + "/" + aFile.getName(); - doc.add(new Field("path", aPath, Field.Store.YES, Field.Index.UN_TOKENIZED)); - } - - // Add the caption of the file to a field named "caption". Specify a Reader, - // so that the text of the file is tokenized and indexed, but not stored. - doc.add( new Field( "caption", getReaderForFile( aCaptionFile ) ) ); - - // Add the contents of the file to a field named "content". Specify a Reader, - // so that the text of the file is tokenized and indexed, but not stored. - doc.add( new Field( "content", getReaderForFile( aContentFile ) ) ); - - // return the document - return doc; - } - - private HelpFileDocument() {} -} diff --git a/l10ntools/source/help/HelpIndexerTool.java b/l10ntools/source/help/HelpIndexerTool.java deleted file mode 100644 index 8e42d55d2768..000000000000 --- a/l10ntools/source/help/HelpIndexerTool.java +++ /dev/null @@ -1,390 +0,0 @@ -/************************************************************************* - * - * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. - * - * Copyright 2000, 2010 Oracle and/or its affiliates. - * - * OpenOffice.org - a multi-platform office productivity suite - * - * This file is part of OpenOffice.org. - * - * OpenOffice.org is free software: you can redistribute it and/or modify - * it under the terms of the GNU Lesser General Public License version 3 - * only, as published by the Free Software Foundation. - * - * OpenOffice.org is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU Lesser General Public License version 3 for more details - * (a copy is included in the LICENSE file that accompanied this code). - * - * You should have received a copy of the GNU Lesser General Public License - * version 3 along with OpenOffice.org. If not, see - * <http://www.openoffice.org/license.html> - * for a copy of the LGPLv3 License. - * - ************************************************************************/ - -package com.sun.star.help; - -import java.io.FileInputStream; -import java.io.FileOutputStream; -import java.util.Arrays; -import java.util.HashSet; -import java.util.List; -import java.util.zip.ZipEntry; -import java.util.zip.ZipOutputStream; -import java.util.zip.CRC32; -import org.apache.lucene.analysis.standard.StandardAnalyzer; -import org.apache.lucene.analysis.cjk.CJKAnalyzer; -import org.apache.lucene.analysis.Analyzer; -import org.apache.lucene.index.IndexWriter; - -import java.io.File; -import java.io.FileNotFoundException; -import java.io.IOException; -import java.util.Date; - - -/** - When this tool is used with long path names on Windows, that is paths which start - with \\?\, then the caller must make sure that the path is unique. This is achieved - by removing '.' and '..' from the path. Paths which are created by - osl_getSystemPathFromFileURL fulfill this requirement. This is necessary because - lucene is patched to not use File.getCanonicalPath. See long_path.patch in the lucene - module. - */ -public class HelpIndexerTool -{ - public HelpIndexerTool() - { - } - - - /** - * @param args the command line arguments - */ - public static void main( String[] args ) - { - boolean bExtensionMode = false; - mainImpl( args, bExtensionMode ); - } - - public static void mainImpl( String[] args, boolean bExtensionMode ) - { - String aDirToZipStr = ""; - String aSrcDirStr = ""; - String aLanguageStr = ""; - String aModule = ""; - String aTargetZipFileStr = ""; - String aCfsName = ""; - String aSegmentName = ""; - - // Scan arguments - //If this tool is invoked in the build process for extensions help, - //then -extension must be set. - boolean bExtension = false; - boolean bLang = false; - boolean bMod = false; - boolean bZipDir = false; - boolean bSrcDir = false; - boolean bOutput = false; - boolean bCfsName = false; - boolean bSegmentName = false; - - int nArgCount = args.length; - for( int i = 0 ; i < nArgCount ; i++ ) - { - if( "-extension".equals(args[i]) ) - { - bExtension = true; - } - else if( "-lang".equals(args[i]) ) - { - if( i + 1 < nArgCount ) - { - aLanguageStr = args[i + 1]; - bLang = true; - } - i++; - } - else if( "-mod".equals(args[i]) ) - { - if( i + 1 < nArgCount ) - { - aModule = args[i + 1]; - bMod = true; - } - i++; - } - else if( "-zipdir".equals(args[i]) ) - { - if( i + 1 < nArgCount ) - { - aDirToZipStr = args[i + 1]; - bZipDir = true; - } - i++; - } - else if( "-srcdir".equals(args[i]) ) - { - if( i + 1 < nArgCount ) - { - aSrcDirStr = args[i + 1]; - bSrcDir = true; - } - i++; - } - else if( "-o".equals(args[i]) ) - { - if( i + 1 < nArgCount ) - { - aTargetZipFileStr = args[i + 1]; - bOutput = true; - } - i++; - } - else if( "-checkcfsandsegname".equals(args[i]) ) - { - if( i + 1 < nArgCount ) - { - aCfsName = args[i + 1] + ".cfs"; - bCfsName = true; - } - i++; - if( i + 1 < nArgCount ) - { - aSegmentName = "segments" + args[i + 1]; - bSegmentName = true; - } - i++; - if (!(bCfsName && bSegmentName)) - { - System.out.println("Usage: HelpIndexer -checkcfsandsegname _0 _3 (2 arguments needed)"); - System.exit( -1 ); - } - } - } - - if( !bLang || !bMod || !bZipDir || (!bOutput && !bExtensionMode && !bExtension) ) - { - if( bExtensionMode ) - return; - - System.out.println("Usage: HelpIndexer -lang ISOLangCode -mod HelpModule -zipdir TempZipDir -o OutputZipFile"); - System.out.println("Usage: HelpIndexer -extension -lang ISOLangCode -mod HelpModule -zipdir PathToLangDir"); - System.exit( -1 ); - } - - String aIndexDirName = aModule + ".idxl"; - File aIndexDir = new File( aDirToZipStr + File.separator + aIndexDirName ); - if( !bSrcDir ) - aSrcDirStr = aDirToZipStr; - File aCaptionFilesDir = new File( aSrcDirStr + File.separator + "caption" ); - File aContentFilesDir = new File( aSrcDirStr + File.separator + "content" ); - - try - { - Analyzer analyzer = ( aLanguageStr.equals("ja") - || aLanguageStr.equals("ko") - || aLanguageStr.equals("zh-CN") - || aLanguageStr.equals("zh-TW") ) ? (Analyzer)new CJKAnalyzer() : (Analyzer)new StandardAnalyzer(); - IndexWriter writer = new IndexWriter( aIndexDir, analyzer, true ); - int nRet = indexDocs( writer, aModule, bExtensionMode, aCaptionFilesDir, aContentFilesDir ); - if( nRet != -1 ) - writer.optimize(); - writer.close(); - - boolean bCfsFileOk = true; - boolean bSegmentFileOk = true; - if( bCfsName && bSegmentName && !bExtensionMode && nRet != -1 ) - { - String aCompleteCfsFileName = aDirToZipStr + File.separator + aIndexDirName + File.separator + aCfsName; - String aCompleteSegmentFileName = aDirToZipStr + File.separator + aIndexDirName + File.separator + aSegmentName; - File aCfsFile = new File( aCompleteCfsFileName ); - File aSegmentFile = new File( aCompleteSegmentFileName ); - bCfsFileOk = aCfsFile.exists(); - bSegmentFileOk = aSegmentFile.exists(); - System.out.println( "Checking cfs file " + aCfsName+ ": " + (bCfsFileOk ? "Found" : "Not found") ); - System.out.println( "Checking segment file " + aSegmentName+ ": " + (bSegmentFileOk ? "Found" : "Not found") ); - } - - if( bExtensionMode || bExtension) - { - if( !bSrcDir ) - { - deleteRecursively( aCaptionFilesDir ); - deleteRecursively( aContentFilesDir ); - } - } - else - { - if( nRet == -1 ) - deleteRecursively( aIndexDir ); - - File aDirToZipFile = new File( aDirToZipStr ); - createZipFile( aDirToZipFile, aTargetZipFileStr ); - deleteRecursively( aDirToZipFile ); - } - - if( !bCfsFileOk ) - { - System.out.println( "cfs file check failed, terminating..." ); - System.exit( -1 ); - } - - if( !bSegmentFileOk ) - { - System.out.println( "segment file check failed, terminating..." ); - System.exit( -1 ); - } - } - catch (IOException e) - { - if( bExtensionMode ) - return; - - System.out.println(" caught a " + e.getClass() + - "\n with message: " + e.getMessage()); - System.exit( -1 ); - } - } - - private static int indexDocs(IndexWriter writer, String aModule, boolean bExtensionMode, - File aCaptionFilesDir, File aContentFilesDir) throws IOException - { - if( !aCaptionFilesDir.canRead() || !aCaptionFilesDir.isDirectory() ) - { - if( !bExtensionMode ) - System.out.println( "Not found: " + aCaptionFilesDir ); - return -1; - } - if( !aContentFilesDir.canRead() || !aContentFilesDir.isDirectory() ) - { - if( !bExtensionMode ) - System.out.println( "Not found: " + aContentFilesDir ); - return -1; - } - - String[] aCaptionFiles = aCaptionFilesDir.list(); - List aCaptionFilesList = Arrays.asList( aCaptionFiles ); - HashSet aCaptionFilesHashSet = new HashSet( aCaptionFilesList ); - - String[] aContentFiles = aContentFilesDir.list(); - List aContentFilesList = Arrays.asList( aContentFiles ); - HashSet aContentFilesHashSet = new HashSet( aContentFilesList ); - - // Loop over caption files and find corresponding content file - int nCaptionFilesLen = aCaptionFiles.length; - for( int i = 0 ; i < nCaptionFilesLen ; i++ ) - { - String aCaptionFileStr = aCaptionFiles[i]; - File aCaptionFile = new File( aCaptionFilesDir, aCaptionFileStr ); - File aContentFile = null; - if( aContentFilesHashSet.contains( aCaptionFileStr ) ) - aContentFile = new File( aContentFilesDir, aCaptionFileStr ); - writer.addDocument( HelpFileDocument.Document( aModule, aCaptionFile, aContentFile ) ); - } - - // Loop over content files to find remaining files not mapped to caption files - int nContentFilesLen = aContentFiles.length; - for( int i = 0 ; i < nContentFilesLen ; i++ ) - { - String aContentFileStr = aContentFiles[i]; - if( !aCaptionFilesHashSet.contains( aContentFileStr ) ) - { - // Not already handled in caption files loop - File aCaptionFile = null; - File aContentFile = new File( aContentFilesDir, aContentFileStr ); - writer.addDocument( HelpFileDocument.Document( aModule, aCaptionFile, aContentFile ) ); - } - } - return 0; - } - - public static void createZipFile( File aDirToZip, String aTargetZipFileStr ) - throws FileNotFoundException, IOException - { - FileOutputStream fos = new FileOutputStream( aTargetZipFileStr ); - ZipOutputStream zos = new ZipOutputStream( fos ); - - File[] aChildrenFiles = aDirToZip.listFiles(); - int nFileCount = aChildrenFiles.length; - for( int i = 0 ; i < nFileCount ; i++ ) - addToZipRecursively( zos, aChildrenFiles[i], null ); - - zos.close(); - } - - public static void addToZipRecursively( ZipOutputStream zos, File aFile, String aBasePath ) - throws FileNotFoundException, IOException - { - if( aFile.isDirectory() ) - { - String aDirName = aFile.getName(); - if( aDirName.equalsIgnoreCase( "caption" ) || aDirName.equalsIgnoreCase( "content" ) ) - return; - - File[] aChildrenFiles = aFile.listFiles(); - String aNewBasePath = ""; - if( aBasePath != null ) - aNewBasePath += aBasePath + File.separator; - aNewBasePath += aDirName; - - int nFileCount = aChildrenFiles.length; - for( int i = 0 ; i < nFileCount ; i++ ) - addToZipRecursively( zos, aChildrenFiles[i], aNewBasePath ); - - return; - } - - // No directory - // read contents of file we are going to put in the zip - int fileLength = (int) aFile.length(); - FileInputStream fis = new FileInputStream( aFile ); - byte[] wholeFile = new byte[fileLength]; - int bytesRead = fis.read( wholeFile, 0, fileLength ); - fis.close(); - - String aFileName = aFile.getName(); - String aEntryName = ""; - if( aBasePath != null ) - aEntryName += aBasePath + "/"; - aEntryName += aFileName; - ZipEntry aZipEntry = new ZipEntry( aEntryName ); - aZipEntry.setTime( aFile.lastModified() ); - aZipEntry.setSize( fileLength ); - - int nMethod = ( aFileName.toLowerCase().endsWith( ".jar" ) ) - ? ZipEntry.STORED : ZipEntry.DEFLATED; - aZipEntry.setMethod( nMethod ); - - CRC32 tempCRC = new CRC32(); - tempCRC.update( wholeFile, 0, wholeFile.length ); - aZipEntry.setCrc( tempCRC.getValue() ); - - // write the contents into the zip element - zos.putNextEntry( aZipEntry ); - zos.write( wholeFile, 0, fileLength ); - zos.closeEntry(); - } - - static public boolean deleteRecursively( File aFile ) - { - if( aFile.isDirectory() ) - { - File[] aChildrenFiles = aFile.listFiles(); - int nFileCount = aChildrenFiles.length; - for( int i = 0 ; i < nFileCount ; i++ ) - { - File aChildrenFile = aChildrenFiles[i]; - boolean bSuccess = deleteRecursively( aChildrenFile ); - if( !bSuccess ) - return false; - } - } - - return aFile.delete(); - } -} - |