s434730/RandomSec
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Increased file upload size limit to 1GB.

Browse Source 
Fixed charset detector to be more robust in trying more than one charset.

git-svn-id: http://google-refine.googlecode.com/svn/trunk@326 7d457c2a-affb-35e4-300a-418c747d4874
This commit is contained in:
David Huynh 2010-03-19 19:53:11 +00:00
parent fd85be7816
commit ff0049307e
1 changed files with 42 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

82
src/main/java/com/metaweb/gridworks/commands/edit/CreateProjectCommand.java
View File

@ -1,12 +1,12 @@
package com.metaweb.gridworks.commands.edit; package com.metaweb.gridworks.commands.edit;
import java.io.ByteArrayInputStream; import java.io.BufferedInputStream;
import java.io.ByteArrayOutputStream;
import java.io.IOException; import java.io.IOException;
import java.io.InputStream; import java.io.InputStream;
import java.io.InputStreamReader; import java.io.InputStreamReader;
import java.io.Reader; import java.io.Reader;
import java.io.StringReader; import java.io.StringReader;
import java.io.UnsupportedEncodingException;
import java.net.URL; import java.net.URL;
import java.net.URLConnection; import java.net.URLConnection;
import java.util.Properties; import java.util.Properties;
@ -67,12 +67,7 @@ public class CreateProjectCommand extends Command {
Project project, Project project,
Properties options Properties options
) throws Exception { ) throws Exception {
MultipartParser parser = null; MultipartParser parser = new MultipartParser(request, 1024 * 1024 * 1024);
try {
parser = new MultipartParser(request, 20 * 1024 * 1024);
} catch (Exception e) {
// silent
}
if (parser != null) { if (parser != null) {
Part part = null; Part part = null;
@ -100,24 +95,56 @@ public class CreateProjectCommand extends Command {
if (part.isFile()) { if (part.isFile()) {
FilePart filePart = (FilePart) part; FilePart filePart = (FilePart) part;
BufferedInputStream inputStream = new BufferedInputStream(filePart.getInputStream());
Importer importer = guessImporter(options, null, filePart.getFileName()); Importer importer = guessImporter(options, null, filePart.getFileName());
if (importer.takesReader()) { if (importer.takesReader()) {
/*
* NOTE(SM): The ICU4J char detection code requires the input stream to support mark/reset.
* Unfortunately, not all ServletInputStream implementations are marking, so we need do
* this memory-expensive wrapping to make it work. It's far from ideal but I don't have
* a more efficient solution.
*/
byte[] bytes = new byte[1024 * 4];
{
inputStream.mark(bytes.length);
inputStream.read(bytes);
inputStream.reset();
}
CharsetDetector detector = new CharsetDetector(); CharsetDetector detector = new CharsetDetector();
detector.setDeclaredEncoding("utf8"); // the content on the web is encoded in UTF-8 so assume that detector.setDeclaredEncoding("utf8"); // the content on the web is encoded in UTF-8 so assume that
CharsetMatch charsetMatch = detector.setText(enforceMarking(filePart.getInputStream())).detect();
options.setProperty("encoding", charsetMatch.getName()); Reader reader = null;
options.setProperty("encoding_confidence", Integer.toString(charsetMatch.getConfidence())); CharsetMatch[] charsetMatches = detector.setText(bytes).detectAll();
Gridworks.log("Best encoding guess: " + charsetMatch.getName() + " [confidence: " + charsetMatch.getConfidence() + "]"); for (CharsetMatch charsetMatch : charsetMatches) {
Reader reader = charsetMatch.getReader(); try {
reader = new InputStreamReader(inputStream, charsetMatch.getName());
options.setProperty("encoding", charsetMatch.getName());
options.setProperty("encoding_confidence", Integer.toString(charsetMatch.getConfidence()));
Gridworks.log(
"Best encoding guess: " +
charsetMatch.getName() +
" [confidence: " + charsetMatch.getConfidence() + "]");
break;
} catch (UnsupportedEncodingException e) {
// silent
}
}
if (reader == null) {
reader = new InputStreamReader(inputStream); // all else has failed
}
try { try {
importer.read(charsetMatch.getReader(), project, options, skip, limit); importer.read(reader, project, options, skip, limit);
} finally { } finally {
reader.close(); reader.close();
} }
} else { } else {
InputStream inputStream = filePart.getInputStream();
try { try {
importer.read(inputStream, project, options, skip, limit); importer.read(inputStream, project, options, skip, limit);
} finally { } finally {
@ -224,29 +251,4 @@ public class CreateProjectCommand extends Command {
return new TsvCsvImporter(); return new TsvCsvImporter();
} }
/*
* NOTE(SM): The ICU4J char detection code requires the input stream to support mark/reset. Unfortunately, not
* all ServletInputStream implementations are marking, so we need do this memory-expensive wrapping to make
* it work. It's far from ideal but I don't have a more efficient solution.
*/
private static InputStream enforceMarking(InputStream input) throws IOException {
if (input.markSupported()) {
return input;
} else {
ByteArrayOutputStream output = new ByteArrayOutputStream(64 * 1024);
byte[] buffer = new byte[1024 * 4];
long count = 0;
int n = 0;
while (-1 != (n = input.read(buffer))) {
output.write(buffer, 0, n);
count += n;
}
input.close();
return new ByteArrayInputStream(output.toByteArray());
}
}
} }