Background
Our Eclipse RCP 3.6-based application lets people drag files in for storage/processing. This works fine when the files are dragged from a filesystem, but not when people drag items (messages or attachments) directly from Outlook.
This appears to be because Outlook wants to feed our application the files via a FileGroupDescriptorW and FileContents, but SWT only includes a FileTransfer type. (In a FileTransfer, only the file paths are passed, with the assumption that the receiver can locate and read them. The FileGroupDescriptorW/FileContents approach can supply files directly application-to-application without writing temporary files out to disk.)
We have tried to produce a ByteArrayTransfer subclass that could accept FileGroupDescriptorW and FileContents. Based on some examples on the Web, we were able to receive and parse the FileGroupDescriptorW, which (as the name implies) describes the files available for transfer. (See code sketch below.) But we have been unable to accept the FileContents.
This seems to be because Outlook offers the FileContents data only as TYMED_ISTREAM or TYMED_ISTORAGE, but SWT only understands how to exchange data as TYMED_HGLOBAL. Of those, it appears that TYMED_ISTORAGE would be preferable, since it's not clear how TYMED_ISTREAM could provide access to multiple files' contents.
(We also have some concerns about SWT's desire to pick and convert only a single TransferData type, given that we need to process two, but we think we could probably hack around that in Java somehow: it seems that all the TransferDatas are available at other points of the process.)
Questions
Are we on the right track here? Has anyone managed to accept FileContents in SWT yet? Is there any chance that we could process the TYMED_ISTORAGE data without leaving Java (even if by creating a fragment-based patch to, or a derived version of, SWT), or would we have to build some new native support code too?
Relevant code snippets
Sketch code that extracts file names:
// THIS IS NOT PRODUCTION-QUALITY CODE - FOR ILLUSTRATION ONLY
final Transfer transfer = new ByteArrayTransfer() {
private final String[] typeNames = new String[] { "FileGroupDescriptorW", "FileContents" };
private final int[] typeIds = new int[] { registerType(typeNames[0]), registerType(typeNames[1]) };
#Override
protected String[] getTypeNames() {
return typeNames;
}
#Override
protected int[] getTypeIds() {
return typeIds;
}
#Override
protected Object nativeToJava(TransferData transferData) {
if (!isSupportedType(transferData))
return null;
final byte[] buffer = (byte[]) super.nativeToJava(transferData);
if (buffer == null)
return null;
try {
final DataInputStream in = new DataInputStream(new ByteArrayInputStream(buffer));
long count = 0;
for (int i = 0; i < 4; i++) {
count += in.readUnsignedByte() << i;
}
for (int i = 0; i < count; i++) {
final byte[] filenameBytes = new byte[260 * 2];
in.skipBytes(72); // probable architecture assumption(s) - may be wrong outside standard 32-bit Win XP
in.read(filenameBytes);
final String fileNameIncludingTrailingNulls = new String(filenameBytes, "UTF-16LE");
int stringLength = fileNameIncludingTrailingNulls.indexOf('\0');
if (stringLength == -1)
stringLength = 260;
final String fileName = fileNameIncludingTrailingNulls.substring(0, stringLength);
System.out.println("File " + i + ": " + fileName);
}
in.close();
return buffer;
}
catch (final Exception e) {
return null;
}
}
};
In the debugger, we see that ByteArrayTransfer's isSupportedType() ultimately returns false for the FileContents because the following test is not passed (since its tymed is TYMED_ISTREAM | TYMED_ISTORAGE):
if (format.cfFormat == types[i] &&
(format.dwAspect & COM.DVASPECT_CONTENT) == COM.DVASPECT_CONTENT &&
(format.tymed & COM.TYMED_HGLOBAL) == COM.TYMED_HGLOBAL )
return true;
This excerpt from org.eclipse.swt.internal.ole.win32.COM leaves us feeling less hope for an easy solution:
public static final int TYMED_HGLOBAL = 1;
//public static final int TYMED_ISTORAGE = 8;
//public static final int TYMED_ISTREAM = 4;
Thanks.
even if
//public static final int TYMED_ISTREAM = 4;
Try below code.. it should work
package com.nagarro.jsag.poc.swtdrag;
imports ...
public class MyTransfer extends ByteArrayTransfer {
private static int BYTES_COUNT = 592;
private static int SKIP_BYTES = 72;
private final String[] typeNames = new String[] { "FileGroupDescriptorW", "FileContents" };
private final int[] typeIds = new int[] { registerType(typeNames[0]), registerType(typeNames[1]) };
#Override
protected String[] getTypeNames() {
return typeNames;
}
#Override
protected int[] getTypeIds() {
return typeIds;
}
#Override
protected Object nativeToJava(TransferData transferData) {
String[] result = null;
if (!isSupportedType(transferData) || transferData.pIDataObject == 0)
return null;
IDataObject data = new IDataObject(transferData.pIDataObject);
data.AddRef();
// Check for descriptor format type
try {
FORMATETC formatetcFD = transferData.formatetc;
STGMEDIUM stgmediumFD = new STGMEDIUM();
stgmediumFD.tymed = COM.TYMED_HGLOBAL;
transferData.result = data.GetData(formatetcFD, stgmediumFD);
if (transferData.result == COM.S_OK) {
// Check for contents format type
long hMem = stgmediumFD.unionField;
long fileDiscriptorPtr = OS.GlobalLock(hMem);
int[] fileCount = new int[1];
try {
OS.MoveMemory(fileCount, fileDiscriptorPtr, 4);
fileDiscriptorPtr += 4;
result = new String[fileCount[0]];
for (int i = 0; i < fileCount[0]; i++) {
String fileName = handleFile(fileDiscriptorPtr, data);
System.out.println("FileName : = " + fileName);
result[i] = fileName;
fileDiscriptorPtr += BYTES_COUNT;
}
} catch (Exception e) {
e.printStackTrace();
} finally {
OS.GlobalFree(hMem);
}
}
} finally {
data.Release();
}
return result;
}
private String handleFile(long fileDiscriptorPtr, IDataObject data) throws Exception {
// GetFileName
char[] fileNameChars = new char[OS.MAX_PATH];
byte[] fileNameBytes = new byte[OS.MAX_PATH];
COM.MoveMemory(fileNameBytes, fileDiscriptorPtr, BYTES_COUNT);
// Skip some bytes.
fileNameBytes = Arrays.copyOfRange(fileNameBytes, SKIP_BYTES, fileNameBytes.length);
String fileNameIncludingTrailingNulls = new String(fileNameBytes, "UTF-16LE");
fileNameChars = fileNameIncludingTrailingNulls.toCharArray();
StringBuilder builder = new StringBuilder(OS.MAX_PATH);
for (int i = 0; fileNameChars[i] != 0 && i < fileNameChars.length; i++) {
builder.append(fileNameChars[i]);
}
String name = builder.toString();
try {
File file = saveFileContent(name, data);
if (file != null) {
System.out.println("File Saved # " + file.getAbsolutePath());
;
}
} catch (IOException e) {
System.out.println("Count not save file content");
;
}
return name;
}
private File saveFileContent(String fileName, IDataObject data) throws IOException {
File file = null;
FORMATETC formatetc = new FORMATETC();
formatetc.cfFormat = typeIds[1];
formatetc.dwAspect = COM.DVASPECT_CONTENT;
formatetc.lindex = 0;
formatetc.tymed = 4; // content.
STGMEDIUM stgmedium = new STGMEDIUM();
stgmedium.tymed = 4;
if (data.GetData(formatetc, stgmedium) == COM.S_OK) {
file = new File(fileName);
IStream iStream = new IStream(stgmedium.unionField);
iStream.AddRef();
try (FileOutputStream outputStream = new FileOutputStream(file)) {
int increment = 1024 * 4;
long pv = COM.CoTaskMemAlloc(increment);
int[] pcbWritten = new int[1];
while (iStream.Read(pv, increment, pcbWritten) == COM.S_OK && pcbWritten[0] > 0) {
byte[] buffer = new byte[pcbWritten[0]];
OS.MoveMemory(buffer, pv, pcbWritten[0]);
outputStream.write(buffer);
}
COM.CoTaskMemFree(pv);
} finally {
iStream.Release();
}
return file;
} else {
return null;
}
}
}
Have you looked at https://bugs.eclipse.org/bugs/show_bug.cgi?id=132514 ?
Attached to this bugzilla entry is an patch (against an rather old version of SWT) that might be of interest.
I had the same problem and created a small library providing a Drag'n Drop Transfer Class for JAVA SWT. It can be found here:
https://github.com/HendrikHoetker/OutlookItemTransfer
Currently it supports dropping Mail Items from Outlook to your Java SWT application and will provide a list of OutlookItems with the Filename and a byte array of the file contents.
All is pure Java and in-memory (no temp files).
Usage in your SWT java application:
if (OutlookItemTransfer.getInstance().isSupportedType(event.currentDataType)) {
Object o = OutlookItemTransfer.getInstance().nativeToJava(event.currentDataType);
if (o != null && o instanceof OutlookMessage[]) {
OutlookMessage[] outlookMessages = (OutlookMessage[])o;
for (OutlookMessage msg: outlookMessages) {
//...
}
}
}
The OutlookItem will then provide two elements: filename as String and file contents as array of byte.
From here on, one could write it to a file or further process the byte array.
To your question above:
- What you find in the file descriptor is the filename of the outlook item and a pointer to an IDataObject
- the IDataObject can be parsed and will provide an IStorage object
- The IStorageObject will be then a root container providing further sub-IStorageObjects or IStreams similar to a filesystem (directory = IStorage, file = IStream
You find those elements in the following lines of code:
Get File Contents, see OutlookItemTransfer.java, method nativeToJava:
FORMATETC format = new FORMATETC();
format.cfFormat = getTypeIds()[1];
format.dwAspect = COM.DVASPECT_CONTENT;
format.lindex = <fileIndex>;
format.ptd = 0;
format.tymed = TYMED_ISTORAGE | TYMED_ISTREAM | COM.TYMED_HGLOBAL;
STGMEDIUM medium = new STGMEDIUM();
if (data.GetData(format, medium) == COM.S_OK) {
// medium.tymed will now contain TYMED_ISTORAGE
// in medium.unionfield you will find the root IStorage
}
Read the root IStorage, see CompoundStorage, method readOutlookStorage:
// open IStorage object
IStorage storage = new IStorage(pIStorage);
storage.AddRef();
// walk through the content of the IStorage object
long[] pEnumStorage = new long[1];
if (storage.EnumElements(0, 0, 0, pEnumStorage) == COM.S_OK) {
// get storage iterator
IEnumSTATSTG enumStorage = new IEnumSTATSTG(pEnumStorage[0]);
enumStorage.AddRef();
enumStorage.Reset();
// prepare statstg structure which tells about the object found by the iterator
long pSTATSTG = OS.GlobalAlloc(OS.GMEM_FIXED | OS.GMEM_ZEROINIT, STATSTG.sizeof);
int[] fetched = new int[1];
while (enumStorage.Next(1, pSTATSTG, fetched) == COM.S_OK && fetched[0] == 1) {
// get the description of the the object found
STATSTG statstg = new STATSTG();
COM.MoveMemory(statstg, pSTATSTG, STATSTG.sizeof);
// get the name of the object found
String name = readPWCSName(statstg);
// depending on type of object
switch (statstg.type) {
case COM.STGTY_STREAM: { // load an IStream (=File)
long[] pIStream = new long[1];
// get the pointer to the IStream
if (storage.OpenStream(name, 0, COM.STGM_DIRECT | COM.STGM_READ | COM.STGM_SHARE_EXCLUSIVE, 0, pIStream) == COM.S_OK) {
// load the IStream
}
}
case COM.STGTY_STORAGE: { // load an IStorage (=SubDirectory) - requires recursion to traverse the sub dies
}
}
}
}
// close the iterator
enumStorage.Release();
}
// close the IStorage object
storage.Release();
Related
Using OSMSharp I am having trouble to open a stream for a file (which I can provide on demand)
The error occurs in PBFReader (line 104)
using (var tmp = new LimitedStream(_stream, length))
{
header = _runtimeTypeModel.Deserialize(tmp, null, _blockHeaderType) as BlobHeader;
}
and states: "ProtoBuf.ProtoException: 'Invalid field in source data: 0'" which might mean different things as I have read in this SO question.
The file opens and is visualized with QGis so is not corrupt in my opinion.
Can it be that the contracts do not match? Is OsmSharp/core updated to the latest .proto files for OSM from here (although not sure if this is the real original source for the definition files).
And what might make more sense, can it be that the file I attached is generated for v2 of OSM PBF specification?
In the code at the line of the exception I see the following comment which makes me wonder:
// TODO: remove some of the v1 specific code.
// TODO: this means also to use the built-in capped streams.
// code borrowed from: http://stackoverflow.com/questions/4663298/protobuf-net-deserialize-open-street-maps
// I'm just being lazy and re-using something "close enough" here
// note that v2 has a big-endian option, but Fixed32 assumes little-endian - we
// actually need the other way around (network byte order):
// length = IntLittleEndianToBigEndian((uint)length);
BlobHeader header;
// again, v2 has capped-streams built in, but I'm deliberately
// limiting myself to v1 features
So this makes me wonder if OSM Sharp is (still) up-to-date.
My sandbox code looks like this:
using OsmSharp.Streams;
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using OsmSharp.Tags;
namespace OsmSharp
{
class Program
{
private const string Path = #"C:\Users\Bernoulli IT\Documents\Applications\Argaleo\Test\";
private const string FileNameAntarctica = "antarctica-latest.osm";
private const string FileNameOSPbf = "OSPbf";
private const Boolean useRegisterSource = false;
private static KeyValuePair<string, string> KeyValuePair = new KeyValuePair<string, string>("joep", "monita");
static void Main(string[] args)
{
Console.WriteLine("Hello World!");
//string fileName = $#"{Path}\{FileNameAntarctica}.pbf";
string fileName = $#"{Path}\{FileNameOSPbf}.pbf";
string newFileName = $"{fileName.Replace(".pbf", string.Empty)}-{Guid.NewGuid().ToString().Substring(0, 4)}.pbf";
Console.WriteLine("*** Complete");
string fileNameOutput = CompleteFlow(fileName, newFileName);
Console.WriteLine("");
Console.WriteLine("*** Display");
DisplayFlow(fileNameOutput);
Console.ReadLine();
}
private static string CompleteFlow(string fileName, string newFileName)
{
// 1. Open file and convert to bytes
byte[] fileBytes = FileToBytes(fileName);
// 2. Bytes to OSM stream source (pbf)
PBFOsmStreamSource osmStreamSource;
osmStreamSource = BytesToOsmStreamSource(fileBytes);
osmStreamSource.MoveNext();
if (osmStreamSource.Current() == null)
{
osmStreamSource = FileToOsmStreamSource(fileName);
osmStreamSource.MoveNext();
if (osmStreamSource.Current() == null)
{
throw new Exception("No current in stream.");
}
}
// 3. Add custom tag
AddTag(osmStreamSource);
// 4. OSM stream source to bytes
//byte[] osmStreamSourceBytes = OsmStreamSourceToBytes(osmStreamSource);
// 5. Bytes to file
//string fileNameOutput = BytesToFile(osmStreamSourceBytes, newFileName);
OsmStreamSourceToFile(osmStreamSource, newFileName);
Console.WriteLine(newFileName);
return newFileName;
}
private static void DisplayFlow(string fileName)
{
// 1. Open file and convert to bytes
byte[] fileBytes = FileToBytes(fileName);
// 2. Bytes to OSM stream source (pbf)
BytesToOsmStreamSource(fileBytes);
}
private static byte[] FileToBytes(string fileName)
{
Console.WriteLine(fileName);
return File.ReadAllBytes(fileName);
}
private static PBFOsmStreamSource BytesToOsmStreamSource(byte[] bytes)
{
MemoryStream memoryStream = new MemoryStream(bytes);
memoryStream.Position = 0;
PBFOsmStreamSource osmStreamSource = new PBFOsmStreamSource(memoryStream);
foreach (OsmGeo element in osmStreamSource.Where(osmGeo => osmGeo.Tags.Any(tag => tag.Key.StartsWith(KeyValuePair.Key))))
{
foreach (Tag elementTag in element.Tags.Where(tag => tag.Key.StartsWith(KeyValuePair.Key)))
{
Console.WriteLine("!!!!!!!!!!!!!! Tag found while reading !!!!!!!!!!!!!!!!!!".ToUpper());
}
}
return osmStreamSource;
}
private static PBFOsmStreamSource FileToOsmStreamSource(string fileName)
{
using (FileStream fileStream = new FileInfo(fileName).OpenRead())
{
PBFOsmStreamSource osmStreamSource = new PBFOsmStreamSource(fileStream);
return osmStreamSource;
}
}
private static void AddTag(PBFOsmStreamSource osmStreamSource)
{
osmStreamSource.Reset();
OsmGeo osmGeo = null;
while (osmGeo == null)
{
osmStreamSource.MoveNext();
osmGeo = osmStreamSource.Current();
if(osmGeo?.Tags == null)
{
osmGeo = null;
}
}
osmGeo.Tags.Add("joep", "monita");
Console.WriteLine($"{osmGeo.Tags.FirstOrDefault(tag => tag.Key.StartsWith(KeyValuePair.Key)).Key} - {osmGeo.Tags.FirstOrDefault(tag => tag.Key.StartsWith(KeyValuePair.Key)).Value}");
}
private static byte[] OsmStreamSourceToBytes(PBFOsmStreamSource osmStreamSource)
{
MemoryStream memoryStream = new MemoryStream();
PBFOsmStreamTarget target = new PBFOsmStreamTarget(memoryStream, true);
osmStreamSource.Reset();
target.Initialize();
UpdateTarget(osmStreamSource, target);
target.Flush();
target.Close();
return memoryStream.ToArray();
}
private static string BytesToFile(byte[] bytes, string fileName)
{
using (FileStream fs = new FileStream(fileName, FileMode.Create, FileAccess.Write))
{
fs.Write(bytes, 0, bytes.Length);
}
return fileName;
}
private static void OsmStreamSourceToFile(PBFOsmStreamSource osmStreamSource, string fileName)
{
using (FileStream fileStream = new FileInfo(fileName).OpenWrite())
{
PBFOsmStreamTarget target = new PBFOsmStreamTarget(fileStream, true);
osmStreamSource.Reset();
target.Initialize();
UpdateTarget(osmStreamSource, target);
target.Flush();
target.Close();
}
}
private static void UpdateTarget(OsmStreamSource osmStreamSource, OsmStreamTarget osmStreamTarget)
{
if (useRegisterSource)
{
osmStreamTarget.RegisterSource(osmStreamSource, osmGeo => true);
osmStreamTarget.Pull();
}
else
{
bool isFirst = true;
foreach (OsmGeo osmGeo in osmStreamSource)
{
Tag? tag = osmGeo.Tags?.FirstOrDefault(t => t.Key == KeyValuePair.Key);
switch (osmGeo.Type)
{
case OsmGeoType.Node:
if (isFirst)
{
for (int indexer = 0; indexer < 1; indexer++)
{
(osmGeo as Node).Tags.Add(new Tag(KeyValuePair.Key + Guid.NewGuid(), KeyValuePair.Value));
}
isFirst = false;
}
osmStreamTarget.AddNode(osmGeo as Node);
break;
case OsmGeoType.Way:
osmStreamTarget.AddWay(osmGeo as Way);
break;
case OsmGeoType.Relation:
osmStreamTarget.AddRelation(osmGeo as Relation);
break;
default:
throw new ArgumentOutOfRangeException();
}
}
}
}
}
}
Already I posted this question on the GITHube page of OSMSharp as is linked here. Any help would be very appreciated.
I'm working towards properly integrating the stanford segmenter within SOLR for chinese tokenization.
This plugin involves loading other jar files and model files. I've got it working in a crude manner by hardcoding the complete path for the files.
I'm looking for methods to create the plugin where the paths need not be hardcoded and also to have the plugin in conformance with the SOLR plugin architecture. Please let me know if there are any recommended sites or tutorials for this.
I've added my code below :
public class ChineseTokenizerFactory extends TokenizerFactory {
/** Creates a new WhitespaceTokenizerFactory */
public ChineseTokenizerFactory(Map<String,String> args) {
super(args);
assureMatchVersion();
if (!args.isEmpty()) {
throw new IllegalArgumentException("Unknown parameters: " + args);
}
}
#Override
public ChineseTokenizer create(AttributeFactory factory, Reader input) {
Reader processedStringReader = new ProcessedStringReader(input);
return new ChineseTokenizer(luceneMatchVersion, factory, processedStringReader);
}
}
public class ProcessedStringReader extends java.io.Reader {
private static final int BUFFER_SIZE = 1024 * 8;
//private static TextProcess m_textProcess = null;
private static final String basedir = "/home/praveen/PDS_Meetup/solr-4.9.0/custom_plugins/";
static Properties props = null;
static CRFClassifier<CoreLabel> segmenter = null;
private char[] m_inputData = null;
private int m_offset = 0;
private int m_length = 0;
public ProcessedStringReader(Reader input){
char[] arr = new char[BUFFER_SIZE];
StringBuffer buf = new StringBuffer();
int numChars;
if(segmenter == null)
{
segmenter = new CRFClassifier<CoreLabel>(getProperties());
segmenter.loadClassifierNoExceptions(basedir + "ctb.gz", getProperties());
}
try {
while ((numChars = input.read(arr, 0, arr.length)) > 0) {
buf.append(arr, 0, numChars);
}
} catch (IOException e) {
e.printStackTrace();
}
m_inputData = processText(buf.toString()).toCharArray();
m_offset = 0;
m_length = m_inputData.length;
}
#Override
public int read(char[] cbuf, int off, int len) throws IOException {
int charNumber = 0;
for(int i = m_offset + off;i<m_length && charNumber< len; i++){
cbuf[charNumber] = m_inputData[i];
m_offset ++;
charNumber++;
}
if(charNumber == 0){
return -1;
}
return charNumber;
}
#Override
public void close() throws IOException {
m_inputData = null;
m_offset = 0;
m_length = 0;
}
public String processText(String inputText)
{
List<String> segmented = segmenter.segmentString(inputText);
String output = "";
if(segmented.size() > 0)
{
output = segmented.get(0);
for(int i=1;i<segmented.size();i++)
{
output = output + " " +segmented.get(i);
}
}
System.out.println(output);
return output;
}
static Properties getProperties()
{
if (props == null) {
props = new Properties();
props.setProperty("sighanCorporaDict", basedir);
// props.setProperty("NormalizationTable", "data/norm.simp.utf8");
// props.setProperty("normTableEncoding", "UTF-8");
// below is needed because CTBSegDocumentIteratorFactory accesses it
props.setProperty("serDictionary",basedir+"dict-chris6.ser.gz");
props.setProperty("inputEncoding", "UTF-8");
props.setProperty("sighanPostProcessing", "true");
}
return props;
}
}
public final class ChineseTokenizer extends CharTokenizer {
public ChineseTokenizer(Version matchVersion, Reader in) {
super(matchVersion, in);
}
public ChineseTokenizer(Version matchVersion, AttributeFactory factory, Reader in) {
super(matchVersion, factory, in);
}
/** Collects only characters which do not satisfy
* {#link Character#isWhitespace(int)}.*/
#Override
protected boolean isTokenChar(int c) {
return !Character.isWhitespace(c);
}
}
You can pass the argument through the Factory's args parameter.
I was under the impression that it is now possible to copy AcroFields using PdfCopy. In the release notes for iText 5.4.4.0 this is listed as possible now. However, when I try to do so it appears all the annotations (I think I am using that term correctly, still fairly new to iText...) for the fields are stripped out. It looks like the fields are there (meaning I can see the blue boxes that indicate an editable field), but they are not editable. If I try to bring the PDF up in Acrobat I get a message saying that "there are no fields, would you like Acrobat to discover them?" and most are found and marked and fields properly (check boxes aren't, but the text fields are).
I assume there is an additional step somewhere along the lines to re-add the annotations to the PdfCopy object, but I do not see a way to get the annotations from the PdfReader. I also cannot seem to find any documentation on how to do this (since AcroFields were for so long not supported in PdfCopy most of what I find is along that vein).
Due to sensitivity I cannot provide a copy of the PDF's in question, but using an altered version of a test program used earlier you can see the issue with the following code. It should generate a table with some check boxes in the four right columns. If I use the exact same code with PdfCopyFields in the MergePdfs method instead of PdfCopy it works as expected. This code does not produce any text fields, but in my main project they are part of the original parent PDF that is used as a template.
(Sorry for the long example, it has been cherry picked from a much larger application. You will need a PDF with a field named "TableStartPosition" somewhere in it and update RunTest with the correct paths for your local machine to get this to work.)
Has the PdfCopy functionality not made it into iTextSharp yet? I am using version 5.4.5.0.
class Program
{
Stream _pdfTemplateStream;
MemoryStream _pdfResultStream;
PdfReader _pdfTemplateReader;
PdfStamper _pdfResultStamper;
static void Main(string[] args)
{
Program p = new Program();
try
{
p.RunTest();
}
catch (Exception f)
{
Console.WriteLine(f.Message);
Console.ReadLine();
}
}
internal void RunTest()
{
FileStream fs = File.OpenRead(#"C:\temp\a\RenameFieldTest\RenameFieldTest\Library\CoverPage.pdf");
_pdfTemplateStream = fs;
_pdfResultStream = new MemoryStream();
//PDFTemplateStream = new FileStream(_templatePath, FileMode.Open);
_pdfTemplateReader = new PdfReader(_pdfTemplateStream);
_pdfResultStamper = new PdfStamper(_pdfTemplateReader, _pdfResultStream);
#region setup objects
List<CustomCategory> Categories = new List<CustomCategory>();
CustomCategory c1 = new CustomCategory();
c1.CategorySizesInUse.Add(CustomCategory.AvailableSizes[1]);
c1.CategorySizesInUse.Add(CustomCategory.AvailableSizes[2]);
Categories.Add(c1);
CustomCategory c2 = new CustomCategory();
c2.CategorySizesInUse.Add(CustomCategory.AvailableSizes[0]);
c2.CategorySizesInUse.Add(CustomCategory.AvailableSizes[1]);
Categories.Add(c2);
List<CustomObject> Items = new List<CustomObject>();
CustomObject co1 = new CustomObject();
co1.Category = c1;
co1.Title = "Object 1";
Items.Add(co1);
CustomObject co2 = new CustomObject();
co2.Category = c2;
co2.Title = "Object 2";
Items.Add(co2);
#endregion
FillCoverPage(Items);
_pdfResultStamper.Close();
_pdfTemplateReader.Close();
List<MemoryStream> pdfStreams = new List<MemoryStream>();
pdfStreams.Add(new MemoryStream(_pdfResultStream.ToArray()));
MergePdfs(#"C:\temp\a\RenameFieldTest\RenameFieldTest\Library\Outfile.pdf", pdfStreams);
_pdfResultStream.Dispose();
_pdfTemplateStream.Dispose();
}
internal void FillCoverPage(List<CustomObject> Items)
{
//Before we start we need to figure out where to start adding the table
var fieldPositions = _pdfResultStamper.AcroFields.GetFieldPositions("TableStartPosition");
if (fieldPositions == null)
{ throw new Exception("Could not find the TableStartPosition field. Unable to determine point of origin for the table!"); }
_pdfResultStamper.AcroFields.RemoveField("TableStartPosition");
var fieldPosition = fieldPositions[0];
// Get the position of the field
var targetPosition = fieldPosition.position;
//First, get all the available card sizes
List<string> availableSizes = CustomCategory.AvailableSizes;
//Generate a table with the number of available card sizes + 1 for the device name
PdfPTable table = new PdfPTable(availableSizes.Count + 1);
float[] columnWidth = new float[availableSizes.Count + 1];
for (int y = 0; y < columnWidth.Length; y++)
{
if (y == 0)
{ columnWidth[y] = 320; }
else
{ columnWidth[y] = 120; }
}
table.SetTotalWidth(columnWidth);
table.WidthPercentage = 100;
PdfContentByte canvas;
List<PdfFormField> checkboxes = new List<PdfFormField>();
//Build the header row
table.Rows.Add(new PdfPRow(this.GetTableHeaderRow(availableSizes)));
//Insert the global check boxes
PdfPCell[] globalRow = new PdfPCell[availableSizes.Count + 1];
Phrase tPhrase = new Phrase("Select/Unselect All");
PdfPCell tCell = new PdfPCell();
tCell.BackgroundColor = BaseColor.LIGHT_GRAY;
tCell.AddElement(tPhrase);
globalRow[0] = tCell;
for (int x = 0; x < availableSizes.Count; x++)
{
tCell = new PdfPCell();
tCell.BackgroundColor = BaseColor.LIGHT_GRAY;
PdfFormField f = PdfFormField.CreateCheckBox(_pdfResultStamper.Writer);
string fieldName = string.Format("InkSaver.Global.chk{0}", availableSizes[x].Replace(".", ""));
//f.FieldName = fieldName;
string js = string.Format("hideAll(event.target, '{0}');", availableSizes[x].Replace(".", ""));
f.Action = PdfAction.JavaScript(js, _pdfResultStamper.Writer);
tCell.CellEvent = new ChildFieldEvent(_pdfResultStamper.Writer, f, fieldName);
globalRow[x + 1] = tCell;
checkboxes.Add(f);
}
table.Rows.Add(new PdfPRow(globalRow));
int status = 0;
int pageNum = 1;
for (int itemIndex = 0; itemIndex < Items.Count; itemIndex++)
{
tCell = new PdfPCell();
Phrase p = new Phrase(Items[itemIndex].Title);
tCell.AddElement(p);
tCell.HorizontalAlignment = Element.ALIGN_LEFT;
PdfPCell[] cells = new PdfPCell[availableSizes.Count + 1];
cells[0] = tCell;
for (int availCardSizeIndex = 0; availCardSizeIndex < availableSizes.Count; availCardSizeIndex++)
{
if (Items[itemIndex].Category.CategorySizesInUse.Contains(availableSizes[availCardSizeIndex]))
{
string str = availableSizes[availCardSizeIndex];
tCell = new PdfPCell();
tCell.PaddingLeft = 10f;
tCell.PaddingRight = 10f;
cells[availCardSizeIndex + 1] = tCell;
cells[availCardSizeIndex].HorizontalAlignment = Element.ALIGN_CENTER;
PdfFormField f = PdfFormField.CreateCheckBox(_pdfResultStamper.Writer);
string fieldName = string.Format("InkSaver.chk{0}.{1}", availableSizes[availCardSizeIndex].Replace(".", ""), itemIndex + 1);
//f.FieldName = fieldName; <-- This causes the checkbox to be double-named (i.e. InkSaver.Global.chk0.InkSaver.Global.chk0
string js = string.Format("hideCardSize(event.target, {0}, '{1}');", itemIndex + 1, availableSizes[availCardSizeIndex]);
f.Action = PdfAction.JavaScript(js, _pdfResultStamper.Writer);
tCell.CellEvent = new ChildFieldEvent(_pdfResultStamper.Writer, f, fieldName);
checkboxes.Add(f);
}
else
{
//Add a blank cell
tCell = new PdfPCell();
cells[availCardSizeIndex + 1] = tCell;
}
}
//Test if the column text will fit
table.Rows.Add(new PdfPRow(cells));
canvas = _pdfResultStamper.GetUnderContent(pageNum);
ColumnText ct2 = new ColumnText(canvas);
ct2.AddElement(new PdfPTable(table));
ct2.Alignment = Element.ALIGN_LEFT;
ct2.SetSimpleColumn(targetPosition.Left, 0, targetPosition.Right, targetPosition.Top, 0, 0);
status = ct2.Go(true);
if ((status != ColumnText.NO_MORE_TEXT) || (itemIndex == (Items.Count - 1)))
{
ColumnText ct3 = new ColumnText(canvas);
ct3.AddElement(table);
ct3.Alignment = Element.ALIGN_LEFT;
ct3.SetSimpleColumn(targetPosition.Left, 0, targetPosition.Right, targetPosition.Top, 0, 0);
ct3.Go();
foreach (PdfFormField f in checkboxes)
{
_pdfResultStamper.AddAnnotation(f, pageNum);
}
checkboxes.Clear();
if (itemIndex < (Items.Count - 1))
{
pageNum++;
_pdfResultStamper.InsertPage(pageNum, _pdfTemplateReader.GetPageSize(1));
table = new PdfPTable(availableSizes.Count + 1);
table.SetTotalWidth(columnWidth);
table.WidthPercentage = 100;
table.Rows.Add(new PdfPRow(this.GetTableHeaderRow(availableSizes)));
}
}
}
}
private PdfPCell[] GetTableHeaderRow(List<string> AvailableSizes)
{
PdfPCell[] sizeHeaders = new PdfPCell[AvailableSizes.Count + 1];
Phrase devName = new Phrase("Device Name");
PdfPCell deviceHeader = new PdfPCell(devName);
deviceHeader.HorizontalAlignment = Element.ALIGN_CENTER;
deviceHeader.BackgroundColor = BaseColor.GRAY;
sizeHeaders[0] = deviceHeader;
for (int x = 0; x < AvailableSizes.Count; x++)
{
PdfPCell hCell = new PdfPCell(new Phrase(AvailableSizes[x]));
hCell.HorizontalAlignment = Element.ALIGN_CENTER;
hCell.BackgroundColor = BaseColor.GRAY;
sizeHeaders[x + 1] = hCell;
}
return sizeHeaders;
}
public void MergePdfs(string filePath, List<MemoryStream> pdfStreams)
{
//Create output stream
FileStream outStream = new FileStream(filePath, FileMode.Create);
Document document = null;
if (pdfStreams.Count > 0)
{
try
{
int PageCounter = 0;
//Create Main reader
PdfReader reader = new PdfReader(pdfStreams[0]);
PageCounter = reader.NumberOfPages;//This is if we have multiple pages in the cover page, we need to adjust the offset.
//rename fields in the PDF. This is required because PDF's cannot have more than one field with the same name
RenameFields(reader, PageCounter++);
//Create Main Doc
document = new Document(reader.GetPageSizeWithRotation(1));
//Create main writer
PdfCopy Writer = new PdfCopy(document, outStream);
//PdfCopyFields Writer = new PdfCopyFields(outStream);
//Open document for writing
document.Open();
////Add pages
Writer.AddDocument(reader);
//For each additional pdf after first combine them into main document
foreach (var PdfStream in pdfStreams.Skip(1))
{
PdfReader reader2 = new PdfReader(PdfStream);
//rename PDF fields
RenameFields(reader2, PageCounter++);
// Add content
Writer.AddDocument(reader);
}
//Writer.AddJavaScript(PostProcessing.GetSuperscriptJavaScript());
Writer.Close();
}
catch (Exception ex)
{
Console.WriteLine(ex.ToString());
}
finally
{
if (document != null)
document.Close();
foreach (var Strm in pdfStreams)
{
try { if (null != Strm) Strm.Dispose(); }
catch { }
}
//pdfStamper.Close();
outStream.Close();
}
}
}
private void RenameFields(PdfReader reader, int PageNum)
{
int tempPageNum = 1;
//rename all fields
foreach (string field in reader.AcroFields.Fields.Keys)
{
if (((reader.AcroFields.GetFieldType(field) == 1) || (reader.AcroFields.GetFieldType(field) == 2)) && (field.StartsWith("InkSaver")))
{
//This is a InkSaver button, set the name so its subclassed
string classPath;
if (reader.AcroFields.GetFieldType(field) == 2)
{
classPath = field.Substring(0, field.LastIndexOf("."));
if (field.StartsWith("InkSaver.chk"))
{
int a = field.LastIndexOf(".");
string sub = field.Substring(a + 1, (field.Length - a - 1));
int pageNum = int.Parse(sub);
int realPageNum = pageNum + tempPageNum;//PostProcessing.Instance.CoverPageLength;
PageNum = realPageNum;
}
}
else
{
classPath = field.Substring(0, field.LastIndexOf("."));
}
string newID = classPath + ".page" + PageNum.ToString();
bool ret = reader.AcroFields.RenameField(field, newID);
}
else
{
reader.AcroFields.RenameField(field, field + "_" + PageNum.ToString());// field + Guid.NewGuid().ToString("N"));
}
}
}
}
public class ChildFieldEvent : IPdfPCellEvent
{
protected PdfWriter writer;
protected PdfFormField parent;
protected string checkBoxName;
internal ChildFieldEvent(PdfWriter writer, PdfFormField parent, string CheckBoxName)
{
this.writer = writer;
this.parent = parent;
this.checkBoxName = CheckBoxName;
}
public void CellLayout(PdfPCell cell, Rectangle rect, PdfContentByte[] cb)
{
createCheckboxField(rect);
}
private void createCheckboxField(Rectangle rect)
{
RadioCheckField bt = new RadioCheckField(this.writer, rect, this.checkBoxName, "Yes");
bt.CheckType = RadioCheckField.TYPE_SQUARE;
bt.Checked = true;
this.parent.AddKid(bt.CheckField);
}
}
internal class CustomCategory
{
internal static List<string> AvailableSizes
{
get
{
List<string> retVal = new List<string>();
retVal.Add("1");
retVal.Add("2");
retVal.Add("3");
retVal.Add("4");
return retVal;
}
}
internal CustomCategory()
{
CategorySizesInUse = new List<string>();
}
internal List<string> CategorySizesInUse { get; set; }
}
internal class CustomObject
{
internal string Title { get; set; }
internal CustomCategory Category { get;set; }
}
Please take a look at the MergeForms example. Your example is too long for me to read, but at first sight, I'm missing the following line:
copy.setMergeFields();
By the way, in MergeForms2, the fields are also renamed before the form is merged.
I am running my code on TREC documents and right now implementing scoring scheme to get number of relevant documents. However now i want to implement boolean retrieval, I am trying to use HitCollector.
below is my code..
public class BatchSearch {
private BatchSearch() {}
/** Simple command-line based search demo. */
public static void main(String[] args) throws Exception {
String usage =
"Usage:\tjava BatchSearch [-index dir] [-simfn similarity] [-field f] [-queries file]";
if (args.length > 0 && ("-h".equals(args[0]) || "-help".equals(args[0]))) {
System.out.println(usage);
System.out.println("Supported similarity functions:\ndefault: DefaultSimilary (tfidf)\n");
System.exit(0);
}
String index = "index";
String field = "contents";
String queries = null;
String simstring = "default";
for(int i = 0;i < args.length;i++) {
if ("-index".equals(args[i])) {
index = args[i+1];
i++;
} else if ("-field".equals(args[i])) {
field = args[i+1];
i++;
} else if ("-queries".equals(args[i])) {
queries = args[i+1];
i++;
} else if ("-simfn".equals(args[i])) {
simstring = args[i+1];
i++;
}
}
Similarity simfn = null;
if ("default".equals(simstring)) {
simfn = new DefaultSimilarity();
} else if ("bm25".equals(simstring)) {
simfn = new BM25Similarity();
} else if ("dfr".equals(simstring)) {
simfn = new DFRSimilarity(new BasicModelP(), new AfterEffectL(), new NormalizationH2());
} else if ("lm".equals(simstring)) {
simfn = new LMDirichletSimilarity();
}
if (simfn == null) {
System.out.println(usage);
System.out.println("Supported similarity functions:\ndefault: DefaultSimilary (tfidf)");
System.out.println("bm25: BM25Similarity (standard parameters)");
System.out.println("dfr: Divergence from Randomness model (PL2 variant)");
System.out.println("lm: Language model, Dirichlet smoothing");
System.exit(0);
}
IndexReader reader = DirectoryReader.open(FSDirectory.open(new File(index)));
IndexSearcher searcher = new IndexSearcher(reader);
searcher.setSimilarity(simfn);
Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_41);
BufferedReader in = null;
if (queries != null) {
in = new BufferedReader(new InputStreamReader(new FileInputStream(queries), "UTF-8"));
} else {
in = new BufferedReader(new InputStreamReader(new FileInputStream("queries"), "UTF-8"));
}
QueryParser parser = new QueryParser(Version.LUCENE_41, field, analyzer);
while (true) {
String line = in.readLine();
if (line == null || line.length() == -1) {
break;
}
line = line.trim();
if (line.length() == 0) {
break;
}
String[] pair = line.split(" ", 2);
Query query = parser.parse(pair[1]);
doBatchSearch(in, searcher, pair[0], query, simstring);
}
reader.close();
}
/**
* This function performs a top-1000 search for the query as a basic TREC run.
*/
public static void doBatchSearch(BufferedReader in, IndexSearcher searcher, String qid, Query query, String runtag)
throws IOException {
// Collect enough docs to show 5 pages
TopDocs results = searcher.search(query, 1000);
ScoreDoc[] hits = results.scoreDocs;
HashMap<String, String> seen = new HashMap<String, String>(1000);
int numTotalHits = results.totalHits;
int start = 0;
int end = Math.min(numTotalHits, 1000);
for (int i = start; i < end; i++) {
Document doc = searcher.doc(hits[i].doc);
String docno = doc.get("docno");
// There are duplicate document numbers in the FR collection, so only output a given
// docno once.
if (seen.containsKey(docno)) {
continue;
}
seen.put(docno, docno);
System.out.println(qid+" Q0 "+docno+" "+i+" "+hits[i].score+" "+runtag);
}
}
}
The scoring is done in doBatchSearch and now i want to implement HitCollector here.
Please advise me How do I convert text files to .arff format(weka)
because i wan to do data clustering for 1000 txt file.
regards
There are some converters implemented in WEKA, just find the right format or make little changes to your data (using awk, sed...).
Here is the API pages related to this topic: http://weka.sourceforge.net/doc.stable/weka/core/converters/package-summary.html
For exapmle here is how to convert from CSV to ARFF:
java weka.core.converters.CSVLoader filename.csv > filename.arff
Here is the code you can use
package text.Classification;
import java.io.*;
import weka.core.*;
public class TextDirectoryToArff {
public Instances createDataset(String directoryPath) throws Exception {
FastVector atts;
FastVector attVals;
atts = new FastVector();
atts.addElement(new Attribute("contents", (FastVector) null));
String[] s = { "class1", "class2", "class3" };
attVals = new FastVector();
for (String p : s)
attVals.addElement(p);
atts.addElement(new Attribute("class", attVals));
Instances data = new Instances("MyRelation", atts, 0);
System.out.println(data);
InputStreamReader is = null;
File dir = new File(directoryPath);
String[] files = dir.list();
for (int i = 0; i < files.length; i++) {
if (files[i].endsWith(".txt")) {
double[] newInst = new double[2];
File txt = new File(directoryPath + File.separator + files[i]);
is = new InputStreamReader(new FileInputStream(txt));
StringBuffer txtStr = new StringBuffer();
int c;
while ((c = is.read()) != -1) {
txtStr.append((char) c);
}
newInst[0] = data.attribute(0).addStringValue(txtStr.toString());
int j=i%(s.length-1);
newInst[1] = attVals.indexOf(s[j]);
data.add(new Instance(1.0, newInst));
}
}
return data;
}
public static void main(String[] args) {
TextDirectoryToArff tdta = new TextDirectoryToArff();
try {
Instances dataset = tdta.createDataset("/home/asadul/Desktop/Downloads/text_example/class5");
PrintWriter fileWriter = new PrintWriter("/home/asadul/Desktop/Downloads/text_example/abc.arff", "UTF-8");
fileWriter.println(dataset);
fileWriter.close();
} catch (Exception e) {
System.err.println(e.getMessage());
e.printStackTrace();
}
}
}