static void createInvertedIndex()
{
using (SessionNoServer session = new SessionNoServer(s_systemDir))
{
session.BeginUpdate();
session.EnableAutoPageFlush = false; // so that threads don't stomb on each other
Console.WriteLine(DateTime.Now.ToString() + ", start creating inverted index");
ParallelOptions pOptions = new ParallelOptions();
pOptions.MaxDegreeOfParallelism = 2; // set to what is appropriate for your computer (cores & memory size)
//pOptions.MaxDegreeOfParallelism = 1; // appears to work best with only 16GB of memory
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
BTreeSet <Document> documentSet = indexRoot.repository.documentSet;
List <Database> dbs = session.OpenAllDatabases(true);
Parallel.ForEach <Database>(dbs, pOptions,
(Database db, ParallelLoopState loop) => // method invoked by the loop on each iteration
{
if (db.DatabaseNumber >= Document.PlaceInDatabase)
{
createDocumentInvertedIndex(session, db, documentSet);
}
});
session.Commit();
Console.WriteLine(DateTime.Now.ToString() + ", done creating inverted index");
}
}
public void createGlobalInvertedIndex(IndexRoot indexRoot)
{
Placement wordPlacement = new Placement(Lexicon.PlaceInDatabase, 2);
BTreeSetOidShort <Word> wordSet = indexRoot.lexicon.WordSet;
BTreeSet <Document> docSet = indexRoot.repository.documentSet;
Word existingWord = null;
foreach (Document doc in docSet)
{
if (doc.Indexed == false)
{
foreach (Word word in doc.WordSet)
{
WordHit wordHit = doc.WordHit[word];
if (wordSet.TryGetKey(word, ref existingWord))
{
existingWord.GlobalCount = existingWord.GlobalCount + (uint)wordHit.Count;
}
else
{
existingWord = new WordGlobal(word.aWord, session, (uint)wordHit.Count);
existingWord.Persist(wordPlacement, session);
wordSet.Add(existingWord);
}
existingWord.DocumentHit.AddFast(doc);
}
doc.Indexed = true;
}
}
}
public void textToWords(Document doc, IndexRoot indexRoot, string docTextString)
{
DocumentText docText = new DocumentText(docTextString, doc);
session.Persist(doc);
doc.Page.Database.Name = doc.Name;
session.Persist(docText);
indexRoot.Repository.DocumentSet.Add(doc);
doc.Content = docText;
docTextString = docTextString.ToLower();
string[] excludedWords = new string[] { "and", "the" };
char[] splitChars = new char[] { ' ', '\n', '(', '"', '!', ',', '(', ')', '\t' };
string[] words = docTextString.Split(splitChars, StringSplitOptions.RemoveEmptyEntries);
int i = 0;
string aWord;
char[] trimEndChars = new char[] { ';', '.', '"', ',', '\r', ':', ']', '!', '?', '+', '(', ')', '\'', '{', '}', '-', '`', '/', '=' };
char[] trimStartChars = new char[] { ';', '&', '-', '#', '*', '[', '.', '"', ',', '\r', ')', '(', '\'', '{', '}', '-', '`' };
foreach (string wordStr in words)
{
i++;
aWord = wordStr.TrimEnd(trimEndChars);
aWord = aWord.TrimStart(trimStartChars);
if (aWord.Length > 1 && excludedWords.Contains(aWord) == false)
{
indexRoot.Lexicon.PossiblyAddToken(aWord, doc);
}
}
}
public void textToWords(Document doc, IndexRoot indexRoot, string docTextString, Placement documentPlacement,
Placement documentTextPlacement, Placement wordPlacement, Placement wordHitPlacement)
{
DocumentText docText = new DocumentText(docTextString, doc);
Word word;
doc.Persist(documentPlacement, session);
doc.Page.Database.Name = doc.Name;
docText.Persist(documentTextPlacement, session);
indexRoot.repository.documentSet.Add(doc);
doc.Content = docText;
docTextString = docTextString.ToLower();
string[] excludedWords = new string[] { "and", "the" };
char[] splitChars = new char[] { ' ', '\n', '(', '"', '!', ',', '(', ')', '\t' };
string[] words = docTextString.Split(splitChars, StringSplitOptions.RemoveEmptyEntries);
UInt64 wordCt = 0;
int i = 0;
string aWord;
char[] trimEndChars = new char[] { ';', '.', '"', ',', '\r', ':', ']', '!', '?', '+', '(', ')', '\'', '{', '}', '-', '`', '/', '=' };
char[] trimStartChars = new char[] { ';', '&', '-', '#', '*', '[', '.', '"', ',', '\r', ')', '(', '\'', '{', '}', '-', '`' };
foreach (string wordStr in words)
{
i++;
aWord = wordStr.TrimEnd(trimEndChars);
aWord = aWord.TrimStart(trimStartChars);
word = new Word(aWord);
if (aWord.Length > 1 && excludedWords.Contains(aWord) == false)
{
createLocalInvertedIndex(doc, word, wordCt, wordPlacement, wordHitPlacement);
++wordCt;
}
}
}
private void button1_Click(object sender, RoutedEventArgs e)
{
session.BeginUpdate();
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
Placement docPlacement = new Placement(Document.PlaceInDatabase);
foreach (string str in listBoxPagesToAdd.Items)
{
Document doc = null;
try
{
if (str.Contains(".html") || str.Contains(".htm") || str.Contains("http") || str.Contains("aspx"))
{
doc = parseHtml(str, indexRoot);
}
else
{
doc = parseTextFile(str, indexRoot, docPlacement);
}
}
catch (WebException ex)
{
Console.WriteLine(ex.ToString());
}
}
createGlobalInvertedIndex(indexRoot);
listBoxPagesToAdd.Items.Clear();
List <Document> docs = indexRoot.repository.documentSet.ToList <Document>().Take(50).ToList <Document>();
inDbListBox.ItemsSource = docs;
session.Commit();
session.BeginRead();
updateDataGrids(indexRoot);
session.Commit();
}
public async Task <IndexRoot> Integrate(ICollection <VirtualFile> files)
{
Utils.Log($"Integrating {files.Count} files");
var allFiles = AllFiles.Concat(files).GroupBy(f => f.Name).Select(g => g.Last()).ToImmutableList();
var byFullPath = Task.Run(() => allFiles.SelectMany(f => f.ThisAndAllChildren)
.ToImmutableDictionary(f => f.FullPath));
var byHash = Task.Run(() => allFiles.SelectMany(f => f.ThisAndAllChildren)
.Where(f => f.Hash != null)
.ToGroupedImmutableDictionary(f => f.Hash));
var byName = Task.Run(() => allFiles.SelectMany(f => f.ThisAndAllChildren)
.ToGroupedImmutableDictionary(f => f.Name));
var byRootPath = Task.Run(() => allFiles.ToImmutableDictionary(f => f.Name));
var result = new IndexRoot(allFiles,
await byFullPath,
await byHash,
await byRootPath,
await byName);
Utils.Log($"Done integrating");
return(result);
}
private void CreateRoot(string name)
{
Root = new FileRecordSegmentHeader();
Root.CreateFileRecordHeader(Enums.MFTEntryFlags.FileNameIndexPresent, null);
StandardInformation stdInfo = new StandardInformation();
stdInfo.CreateStandInfoFile(FileAttributes.Normal);
Root.UsedEntrySize += stdInfo.RecordLength;
IndexRoot indRoot = new IndexRoot();
//Root.UsedEntrySize += (uint)indRoot.numberOfChildren * 8; // reference is 8 B
FileName fileName = new FileName();
fileName.CreateFileNameFile(name);
fileName.RealSize = fileName.AllocatedSize = 0;
Root.UsedEntrySize += fileName.RecordLength;
Root.attributes.Add(stdInfo);
Root.attributes.Add(fileName);
Root.attributes.Add(indRoot);
BytesOccupied += Root.UsedEntrySize;
}
public async Task <IndexRoot> Integrate(ICollection <VirtualFile> files)
{
Utils.Log($"Integrating {files.Count} files");
var allFiles = AllFiles.Concat(files)
.OrderByDescending(f => f.LastModified)
.GroupBy(f => f.FullPath).Select(g => g.Last())
.ToList();
var byFullPath = Task.Run(() => allFiles.SelectMany(f => f.ThisAndAllChildren)
.ToDictionary(f => f.FullPath));
var byHash = Task.Run(() => allFiles.SelectMany(f => f.ThisAndAllChildren)
.Where(f => f.Hash != Hash.Empty)
.ToLookup(f => f.Hash));
var byName = Task.Run(() => allFiles.SelectMany(f => f.ThisAndAllChildren)
.ToLookup(f => f.Name));
var byRootPath = Task.Run(() => allFiles.ToDictionary(f => f.AbsoluteName));
var result = new IndexRoot(allFiles,
await byFullPath,
await byHash,
await byRootPath,
await byName);
Utils.Log($"Done integrating");
return(result);
}
public Index(File file, string name, BiosParameterBlock bpb, UpperCase upCase)
{
_file = file;
_name = name;
_bpb = bpb;
_isFileIndex = name == "$I30";
_blockCache = new ObjectCache<long, IndexBlock>();
_root = _file.GetStream(AttributeType.IndexRoot, _name).GetContent<IndexRoot>();
_comparer = _root.GetCollator(upCase);
using (Stream s = _file.OpenStream(AttributeType.IndexRoot, _name, FileAccess.Read))
{
byte[] buffer = Utilities.ReadFully(s, (int)s.Length);
_rootNode = new IndexNode(WriteRootNodeToDisk, 0, this, true, buffer, IndexRoot.HeaderOffset);
// Give the attribute some room to breathe, so long as it doesn't squeeze others out
// BROKEN, BROKEN, BROKEN - how to figure this out? Query at the point of adding entries to the root node?
_rootNode.TotalSpaceAvailable += _file.MftRecordFreeSpace(AttributeType.IndexRoot, _name) - 100;
}
if (_file.StreamExists(AttributeType.IndexAllocation, _name))
{
_indexStream = _file.OpenStream(AttributeType.IndexAllocation, _name, FileAccess.ReadWrite);
}
if (_file.StreamExists(AttributeType.Bitmap, _name))
{
_indexBitmap = new Bitmap(_file.OpenStream(AttributeType.Bitmap, _name, FileAccess.ReadWrite), long.MaxValue);
}
}
private void MenuItem_Click(object sender, RoutedEventArgs e)
{
MenuItem menuItem = (MenuItem)e.Source;
Document myItem = (Document)menuItem.DataContext;
session.BeginUpdate();
try
{
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
int index;
if (indexRoot.repository.documentSet.TryGetKey(myItem, ref myItem))
{
index = myItem.Remove(indexRoot, session);
}
else
{
index = -1; // weird case - should not happen
}
inDbListBox.ItemsSource = indexRoot.repository.documentSet.ToList <Document>();
updateDataGrids(indexRoot, index);
session.Commit();
}
catch
{
session.Abort();
}
}
static void createTopLevelInvertedIndex()
{
Console.WriteLine(DateTime.Now.ToString() + ", start creating top level inverted index");
using (SessionNoServer session = new SessionNoServer(s_systemDir))
{
Placement wordPlacement = new Placement(Lexicon.PlaceInDatabase, 2, 1, 1000, 50000, true, false, UInt32.MaxValue, false);
session.BeginUpdate();
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
BTreeSetOidShort <Word> wordSet = indexRoot.lexicon.WordSet;
BTreeSet <Document> documentSet = indexRoot.repository.documentSet;
Word existingWord = null;
foreach (Document doc in documentSet)
{
foreach (Word word in doc.WordSet)
{
WordHit wordHit = doc.WordHit[word];
if (wordSet.TryGetKey(word, ref existingWord))
{
existingWord.GlobalCount = existingWord.GlobalCount + (uint)wordHit.Count;
}
else
{
existingWord = new WordGlobal(word.aWord, session, (uint)wordHit.Count);
existingWord.Persist(wordPlacement, session);
indexRoot.lexicon.WordSet.Add(existingWord);
}
existingWord.DocumentHit.AddFast(doc);
}
doc.Indexed = true;
}
session.Commit();
Console.WriteLine(DateTime.Now.ToString() + ", done creating top level inverted index");
}
}
public Document parseTextFile(string url, IndexRoot indexRoot)
{
Document doc = new Document(Path.GetFileName(url), indexRoot, session);
using (StreamReader reader = new StreamReader(url))
{
textToWords(doc, indexRoot, reader.ReadToEnd());
}
return(doc);
}
public Database(string dataDir, string indexDir)
{
Directory.CreateDirectory(dataDir);
var logPath = Path.Combine(dataDir, "log.dat");
var datafilePath = Path.Combine(dataDir, "data.dat");
_wal = new Wal(logPath);
_dataFile = new DataFile(datafilePath);
_index = new IndexRoot(indexDir);
Recover().GetAwaiter().GetResult();
}
public Document parseTextFile(string url, IndexRoot indexRoot, Placement docPlacement)
{
Document doc = new Document(Path.GetFileName(url), indexRoot, session);
Placement docTextPlacement = new Placement(docPlacement.TryDatabaseNumber, (ushort)(docPlacement.TryPageNumber + 1));
Placement wordPlacement = new Placement(docPlacement.TryDatabaseNumber, (ushort)(docPlacement.TryPageNumber + 2));
Placement wordHitPlacement = new Placement(docPlacement.TryDatabaseNumber, (ushort)(docPlacement.TryPageNumber + 10));
using (StreamReader reader = new StreamReader(url))
{
textToWords(doc, indexRoot, reader.ReadToEnd(), docPlacement, docTextPlacement, wordPlacement, wordHitPlacement);
}
return(doc);
}
static void outputSomeInfo(SessionNoServer session)
{
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
BTreeSetOidShort <Word> wordSet = indexRoot.lexicon.WordSet;
using (StreamWriter writer = new StreamWriter("Wikipedia.txt"))
{
writer.WriteLine("Number of words in Lexicon is: " + wordSet.Count);
foreach (Word word in wordSet)
{
writer.WriteLine(word.aWord + " " + word.DocumentHit.Count);
}
writer.Close();
}
}
private void updateWordTables_Click(object sender, RoutedEventArgs e)
{
if (session != null)
{
try
{
session.BeginRead();
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
updateDataGrids(indexRoot);
}
finally
{
session.Commit();
}
}
}
static void outputSomeInfo(SessionNoServer session)
{
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
var wordHits = indexRoot.Lexicon.TokenMap;
using (StreamWriter writer = new StreamWriter("Wikipedia.txt"))
{
writer.WriteLine("Number of words in Lexicon is: " + indexRoot.Lexicon.IdToValue.Count);
foreach (var p in wordHits)
{
var word = indexRoot.Lexicon.IdToValue[p.Key];
writer.WriteLine(word + " " + p.Value.Count);
}
writer.Close();
}
}
public Document parseHtml(string url, IndexRoot indexRoot)
{
Document doc = new Document(url, indexRoot, session);
using (WebClient client = new WebClient())
{
string html = client.DownloadString(url);
string pageBody = "";
HtmlDocument htmlDoc = new HtmlDocument();
htmlDoc.LoadHtml(html);
foreach (HtmlNode node in htmlDoc.DocumentNode.SelectNodes("//text()"))
{
pageBody += " " + node.InnerText;
}
textToWords(doc, indexRoot, pageBody);
}
return(doc);
}
static void CreateInvertedIndex()
{
using (SessionNoServer session = new SessionNoServer(s_systemDir, 5000, false, false, CacheEnum.No)) // turn of page and object caching))
{
session.BeginUpdate();
session.RegisterClass(typeof(Repository));
session.RegisterClass(typeof(IndexRoot));
session.RegisterClass(typeof(Lexicon <string>));
Console.WriteLine(DateTime.Now.ToString() + ", start creating inverted index");
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
BTreeSet <Document> documentSet = indexRoot.Repository.DocumentSet;
foreach (var doc in documentSet)
{
createDocumentInvertedIndex(indexRoot, doc);
}
session.Commit();
Console.WriteLine(DateTime.Now.ToString() + ", done creating inverted index");
}
}
public void verify(bool startTrans = true)
{
if (startTrans)
{
session.BeginRead();
}
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
int i = 0;
int j = 0;
int k = 0;
foreach (var pair in indexRoot.Lexicon.TokenMap)
{
i++;
foreach (Document doc in pair.Value)
{
j++;
if (doc == null)
{
throw new UnexpectedException("bad documentHit BTreeSet");
}
foreach (KeyValuePair <UInt32, UInt32> pair2 in doc.WordHit)
{
k++;
if (pair2.Value == 0)
{
throw new UnexpectedException("bad document WordHit");
}
if (indexRoot.Lexicon.IdToValue.Contains(pair2.Key) == false)
{
throw new UnexpectedException("missing lexicon word");
}
}
}
}
if (startTrans)
{
session.Commit();
}
}
public void verify(bool startTrans = true)
{
if (startTrans)
{
session.BeginRead();
}
IndexRoot indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
int i = 0;
int j = 0;
int k = 0;
foreach (Word word in indexRoot.lexicon.WordSet)
{
i++;
foreach (Document doc in word.DocumentHit)
{
j++;
if (doc == null)
{
throw new UnexpectedException("bad documentHit BTreeSet");
}
foreach (KeyValuePair <Word, WordHit> pair in doc.WordHit)
{
k++;
if (pair.Value == null || pair.Key == null)
{
throw new UnexpectedException("bad document WordHit");
}
if (indexRoot.lexicon.WordSet.Contains(pair.Key) == false)
{
throw new UnexpectedException("missing lexicon word");
}
}
}
}
if (startTrans)
{
session.Commit();
}
}
static void createDocumentInvertedIndex(IndexRoot indexRoot, Document doc)
{
if (!doc.Indexed)
{
DocumentText docText = doc.Content;
string text = docText.Text.ToLower();
MatchCollection tagMatches = Regex.Matches(text, "[a-z][a-z$]+");
if (++s_docCountIndexed % 50000 == 0)
{
Console.WriteLine(DateTime.Now.ToString() + ", done indexing article: " + s_docCountIndexed);
}
foreach (Match m in tagMatches)
{
indexRoot.Lexicon.PossiblyAddToken(m.Value, doc);
}
if (s_docCountIndexed % 1000 == 0)
{
Console.WriteLine(DateTime.Now.ToString() + ", done indexing article: " + s_docCountIndexed + " Database: " + doc.DatabaseNumber + " is completed.");
}
doc.Indexed = true;
}
}
public Document parseHtml(string url, IndexRoot indexRoot)
{
Document doc = new Document(url, indexRoot, session);
Placement docPlacement = new Placement(Document.PlaceInDatabase);
Placement docTextPlacement = new Placement(Document.PlaceInDatabase, 2);
Placement wordPlacement = new Placement(Document.PlaceInDatabase, 3);
Placement wordHitPlacement = new Placement(Document.PlaceInDatabase, 100);
using (WebClient client = new WebClient())
{
string html = client.DownloadString(url);
string pageBody = "";
HtmlDocument htmlDoc = new HtmlDocument();
htmlDoc.LoadHtml(html);
foreach (HtmlNode node in htmlDoc.DocumentNode.SelectNodes("//text()"))
{
pageBody += " " + node.InnerText;
}
textToWords(doc, indexRoot, pageBody, docPlacement, docTextPlacement, wordPlacement, wordHitPlacement);
}
return(doc);
}
//static void MessageWrite(string str)
//{
// Console.Write(">" + "{0} {1}\n>", str, message1);
// Console.ReadLine();
//}
static void CreateDirectory(FileRecordSegmentHeader parentFolder, FileRecordSegmentHeader childFolder, string name, NtfsFileSystem fs)
{
childFolder.CreateFileRecordHeader(Enums.MFTEntryFlags.FileNameIndexPresent, parentFolder);
StandardInformation stdInfo = new StandardInformation();
stdInfo.CreateStandInfoFile(FileAttributes.Normal);
childFolder.UsedEntrySize += stdInfo.RecordLength;
IndexRoot indRoot = new IndexRoot();
//childFolder.UsedEntrySize += indRoot.RecordLength;
FileName fileName = new FileName();
fileName.CreateFileNameFile(name);
fileName.RealSize = fileName.AllocatedSize = 0;
childFolder.UsedEntrySize += fileName.RecordLength;
if (childFolder.UsedEntrySize >= (fs.VolumeSize - fs.BytesOccupied + 8))
{
Console.WriteLine("Volume is full!");
Save(fs.VolName.Name, fs.Save());
Environment.Exit(0);
}
childFolder.attributes.Add(stdInfo);
childFolder.attributes.Add(fileName);
childFolder.attributes.Add(indRoot);
((IndexRoot)parentFolder.attributes.ElementAt(2)).Children.Add(childFolder);
((IndexRoot)parentFolder.attributes.ElementAt(2)).numberOfChildren++;
childFolder.parent = parentFolder;
fs.BytesOccupied += childFolder.UsedEntrySize + 8;
fs.DirectoryRecordCount++;
parentFolder.UsedFolderSize += childFolder.UsedEntrySize;
}
private Index(AttributeType attrType, AttributeCollationRule collationRule, File file, string name, BiosParameterBlock bpb, UpperCase upCase)
{
_file = file;
_name = name;
_bpb = bpb;
_isFileIndex = name == "$I30";
_blockCache = new ObjectCache<long, IndexBlock>();
_file.CreateStream(AttributeType.IndexRoot, _name);
_root = new IndexRoot()
{
AttributeType = (uint)attrType,
CollationRule = collationRule,
IndexAllocationSize = (uint)bpb.IndexBufferSize,
RawClustersPerIndexRecord = bpb.RawIndexBufferSize
};
_comparer = _root.GetCollator(upCase);
_rootNode = new IndexNode(WriteRootNodeToDisk, 0, this, true, 32);
}
static void ImportEntireWikipedia()
{
const ushort btreeNodeSize = 10000;
Console.WriteLine(DateTime.Now.ToString() + ", start importing Wikipedia text");
//System.Xml.Schema.XmlSchema docSchema;
//using (System.Xml.XmlTextReader schemaReader = new System.Xml.XmlTextReader("c:\\export-0_5.xsd"))
//{
// docSchema = System.Xml.Schema.XmlSchema.Read(schemaReader, ValidationCallBack);
// }
int docCount = 0;
using (SessionNoServer session = new SessionNoServer(s_systemDir, 5000, false, false, CacheEnum.No)) // turn of page and object caching
{
Console.WriteLine($"Running with databases in directory: {session.SystemDirectory}");
//GCSettings.LatencyMode = GCLatencyMode.Batch;// try to keep the WeakIOptimizedPersistableReference objects around longer
XmlComment xmlComment;
XmlElement xmlElement;
XmlEntity xmlEntity;
XmlText xmlText;
XmlWhitespace xmlWhitespace;
session.BeginUpdate();
// register all database schema classes used by the application in advance to avoid lock conflict later in parallel indexing
Database db = session.OpenDatabase(IndexRoot.PlaceInDatabase, false, false);
if (db != null)
{
outputSomeInfo(session);
session.Abort();
return;
}
//session.SetTraceDbActivity(Lexicon.PlaceInDatabase);
//session.SetTraceAllDbActivity();
XmlDocument xmlDocument = new XmlDocument("enwiki-latest-pages-articles.xml");
IndexRoot indexRoot = new IndexRoot(btreeNodeSize, session);
indexRoot.Persist(session, indexRoot, true);
UInt32 currentDocumentDatabaseNum = 0;
Document doc = null;
bool titleElement = false;
bool pageText = false;
using (FileStream fs = new FileStream(s_wikipediaXmlFile, FileMode.Open))
{
//using (GZipStream zipStream = new GZipStream(fs, CompressionMode.Decompress)) // if input was a .gz file
{
using (System.Xml.XmlTextReader textReader = new System.Xml.XmlTextReader(fs))
{
while (textReader.Read())
{
System.Xml.XmlNodeType nodeType = textReader.NodeType;
switch (nodeType)
{
case System.Xml.XmlNodeType.Attribute:
break;
case System.Xml.XmlNodeType.CDATA:
break;
case System.Xml.XmlNodeType.Comment:
xmlComment = new XmlComment(textReader.Value, xmlDocument);
break;
case System.Xml.XmlNodeType.Document:
break;
case System.Xml.XmlNodeType.DocumentFragment:
break;
case System.Xml.XmlNodeType.DocumentType:
break;
case System.Xml.XmlNodeType.Element:
xmlElement = new XmlElement(textReader.Prefix, textReader.LocalName, textReader.NamespaceURI, xmlDocument);
if (textReader.LocalName == "title")
{
titleElement = true;
}
else if (textReader.LocalName == "text")
{
pageText = true;
}
break;
case System.Xml.XmlNodeType.EndElement:
if (textReader.LocalName == "title" && doc != null)
{
titleElement = false;
}
else if (textReader.LocalName == "text" && doc != null)
{
pageText = false;
}
break;
case System.Xml.XmlNodeType.EndEntity:
break;
case System.Xml.XmlNodeType.Entity:
xmlEntity = new XmlEntity(textReader.LocalName, xmlDocument);
break;
case System.Xml.XmlNodeType.EntityReference:
break;
case System.Xml.XmlNodeType.None:
break;
case System.Xml.XmlNodeType.Notation:
break;
case System.Xml.XmlNodeType.ProcessingInstruction:
break;
case System.Xml.XmlNodeType.SignificantWhitespace:
break;
case System.Xml.XmlNodeType.Text:
xmlText = new XmlText(textReader.Value, xmlDocument);
if (titleElement)
{
doc = new Document(textReader.Value, indexRoot, session);
session.Persist(doc);
if (doc.DatabaseNumber != currentDocumentDatabaseNum)
{
if (currentDocumentDatabaseNum > 0)
{
session.FlushUpdates();
Console.WriteLine("Database: " + currentDocumentDatabaseNum + " is completed, done importing article " + docCount + " number of lines: " + textReader.LineNumber);
}
currentDocumentDatabaseNum = doc.DatabaseNumber;
}
//doc.Page.Database.Name = doc.Name;
}
else if (doc != null && pageText)
{
#if DEBUGx
Console.WriteLine(doc.Name + " line: " + textReader.LineNumber);
#endif
//if (textReader.LineNumber > 1000000)
//{
// session.Commit();
// return;
//}
DocumentText content = new DocumentText(textReader.Value, doc);
session.Persist(content, 10000);
doc.Content = content;
indexRoot.Repository.DocumentSet.AddFast(doc);
if (++docCount % 1000000 == 0)
{
//session.Commit(false); // skip recovery check, we do it in BeginUpdate which is enough
Console.WriteLine("Done importing article " + docCount + " number of lines: " + textReader.LineNumber);
//session.BeginUpdate();
}
}
break;
case System.Xml.XmlNodeType.Whitespace:
xmlWhitespace = new XmlWhitespace(textReader.Value, xmlDocument);
break;
case System.Xml.XmlNodeType.XmlDeclaration:
break;
}
;
}
Console.WriteLine("Finished importing article " + docCount + " number of lines: " + textReader.LineNumber);
}
}
}
session.Commit();
}
Console.WriteLine(DateTime.Now.ToString() + ", done importing Wikipedia text");
}
public FileRecord(byte[] rawBytes, int offset)
{
Offset = offset;
var sig = BitConverter.ToInt32(rawBytes, 0);
switch (sig)
{
case FileSig:
break;
case BaadSig:
_logger.Debug($"Bad signature at offset 0x{offset:X}");
IsBad = true;
return;
default:
//not initialized
_logger.Debug($"Uninitialized entry (no signature) at offset 0x{offset:X}");
IsUninitialized = true;
return;
}
_logger.Debug($"Processing FILE record at offset 0x{offset:X}");
Attributes = new List <Attribute>();
FixupOffset = BitConverter.ToInt16(rawBytes, 0x4);
FixupEntryCount = BitConverter.ToInt16(rawBytes, 0x6);
//to build fixup info, take FixupEntryCount x 2 bytes as each are 2 bytes long
var fixupTotalLength = FixupEntryCount * 2;
var fixupBuffer = new byte[fixupTotalLength];
Buffer.BlockCopy(rawBytes, FixupOffset, fixupBuffer, 0, fixupTotalLength);
//pull this early so we can check if its free in our fix up value messages
EntryFlags = (EntryFlag)BitConverter.ToInt16(rawBytes, 0x16);
FixupData = new FixupData(fixupBuffer);
FixupOk = true;
//fixup verification
var counter = 512;
foreach (var bytese in FixupData.FixupActual)
{
//adjust the offset to where we need to check
var fixupOffset = counter - 2;
var expected = BitConverter.ToInt16(rawBytes, fixupOffset);
if (expected != FixupData.FixupExpected && EntryFlags != 0x0)
{
FixupOk = false;
_logger.Warn(
$"Offset: 0x{Offset:X} Entry/seq: 0x{EntryNumber:X}/0x{SequenceNumber:X} Fixup values do not match at 0x{fixupOffset:X}. Expected: 0x{FixupData.FixupExpected:X2}, actual: 0x{expected:X2}");
}
//replace fixup expected with actual bytes. bytese has actual replacement values in it.
Buffer.BlockCopy(bytese, 0, rawBytes, fixupOffset, 2);
counter += 512;
}
LogSequenceNumber = BitConverter.ToInt64(rawBytes, 0x8);
SequenceNumber = BitConverter.ToUInt16(rawBytes, 0x10);
ReferenceCount = BitConverter.ToInt16(rawBytes, 0x12);
FirstAttributeOffset = BitConverter.ToInt16(rawBytes, 0x14);
ActualRecordSize = BitConverter.ToInt32(rawBytes, 0x18);
AllocatedRecordSize = BitConverter.ToInt32(rawBytes, 0x1c);
var entryBytes = new byte[8];
Buffer.BlockCopy(rawBytes, 0x20, entryBytes, 0, 8);
MftRecordToBaseRecord = new MftEntryInfo(entryBytes);
FirstAvailablAttribueId = BitConverter.ToInt16(rawBytes, 0x28);
EntryNumber = BitConverter.ToUInt32(rawBytes, 0x2c);
//start attribute processing at FirstAttributeOffset
var index = (int)FirstAttributeOffset;
while (index < ActualRecordSize)
{
var attrType = (AttributeType)BitConverter.ToInt32(rawBytes, index);
var attrSize = BitConverter.ToInt32(rawBytes, index + 4);
if (attrSize == 0 || attrType == AttributeType.EndOfAttributes)
{
index += 8; //skip -1 type and 0 size
if (index != ActualRecordSize)
{
_logger.Warn($"Slack space found in entry/seq: 0x{EntryNumber:X}/0x{SequenceNumber:X}");
}
//TODO process slack here?
break;
}
_logger.Debug(
$"Found Attribute Type {attrType.ToString()} at absolute offset: 0x{index + offset:X}");
_logger.Trace(
$"ActualRecordSize: 0x{ActualRecordSize:X}, size: 0x{attrSize:X}, index: 0x{index:X}");
var rawAttr = new byte[attrSize];
Buffer.BlockCopy(rawBytes, index, rawAttr, 0, attrSize);
switch (attrType)
{
case AttributeType.StandardInformation:
var si = new StandardInfo(rawAttr);
Attributes.Add(si);
break;
case AttributeType.FileName:
var fi = new FileName(rawAttr);
Attributes.Add(fi);
break;
case AttributeType.Data:
var d = new Data(rawAttr);
Attributes.Add(d);
break;
case AttributeType.IndexAllocation:
var ia = new IndexAllocation(rawAttr);
Attributes.Add(ia);
break;
case AttributeType.IndexRoot:
var ir = new IndexRoot(rawAttr);
Attributes.Add(ir);
break;
case AttributeType.Bitmap:
var bm = new Bitmap(rawAttr);
Attributes.Add(bm);
break;
case AttributeType.VolumeVersionObjectId:
var oi = new ObjectId_(rawAttr);
Attributes.Add(oi);
break;
case AttributeType.SecurityDescriptor:
var sd = new SecurityDescriptor(rawAttr);
Attributes.Add(sd);
break;
case AttributeType.VolumeName:
var vn = new VolumeName(rawAttr);
Attributes.Add(vn);
break;
case AttributeType.VolumeInformation:
var vi = new VolumeInformation(rawAttr);
Attributes.Add(vi);
break;
case AttributeType.LoggedUtilityStream:
var lus = new LoggedUtilityStream(rawAttr);
Attributes.Add(lus);
break;
case AttributeType.ReparsePoint:
try
{
var rp = new ReparsePoint(rawAttr);
Attributes.Add(rp);
}
catch (Exception)
{
var l = LogManager.GetLogger("ReparsePoint");
l.Error(
$"There was an error parsing a ReparsePoint in FILE record at offset 0x{Offset:X}. Please extract via --dd and --do and send to [email protected]");
}
break;
case AttributeType.AttributeList:
var al = new AttributeList(rawAttr);
Attributes.Add(al);
break;
case AttributeType.Ea:
var ea = new ExtendedAttribute(rawAttr);
Attributes.Add(ea);
break;
case AttributeType.EaInformation:
var eai = new ExtendedAttributeInformation(rawAttr);
Attributes.Add(eai);
break;
default:
throw new Exception($"Add me: {attrType} (0x{attrType:X})");
}
index += attrSize;
}
//rest is slack. handle here?
_logger.Trace($"Slack starts at 0x{index:X} Absolute offset: 0x{index + offset:X}");
}
private bool SelfCheckIndexNode(byte[] buffer, int offset, Bitmap bitmap, IndexRoot root, string fileName, string indexName)
{
bool ok = true;
IndexHeader header = new IndexHeader(buffer, offset);
IndexEntry lastEntry = null;
IComparer<byte[]> collator = root.GetCollator(_context.UpperCase);
int pos = (int)header.OffsetToFirstEntry;
while (pos < header.TotalSizeOfEntries)
{
IndexEntry entry = new IndexEntry(indexName == "$I30");
entry.Read(buffer, offset + pos);
pos += entry.Size;
if ((entry.Flags & IndexEntryFlags.Node) != 0)
{
long bitmapIdx = entry.ChildrenVirtualCluster / Utilities.Ceil(root.IndexAllocationSize, _context.BiosParameterBlock.SectorsPerCluster * _context.BiosParameterBlock.BytesPerSector);
if (!bitmap.IsPresent(bitmapIdx))
{
ReportError("Index entry {0} is non-leaf, but child vcn {1} is not in bitmap at index {2}", Index.EntryAsString(entry, fileName, indexName), entry.ChildrenVirtualCluster, bitmapIdx);
}
}
if ((entry.Flags & IndexEntryFlags.End) != 0)
{
if (pos != header.TotalSizeOfEntries)
{
ReportError("Found END index entry {0}, but not at end of node", Index.EntryAsString(entry, fileName, indexName));
ok = false;
}
}
if (lastEntry != null && collator.Compare(lastEntry.KeyBuffer, entry.KeyBuffer) >= 0)
{
ReportError("Found entries out of order {0} was before {1}", Index.EntryAsString(lastEntry, fileName, indexName), Index.EntryAsString(entry, fileName, indexName));
ok = false;
}
lastEntry = entry;
}
return ok;
}
public static AttributeRecord ReadSingleAttribute(byte[] data, int maxLength, int offset = 0)
{
Debug.Assert(data.Length - offset >= maxLength);
Debug.Assert(0 <= offset && offset <= data.Length);
AttributeTypeCode TypeCode = GetTypeCode(data, offset);
if (TypeCode == AttributeTypeCode.EndOfAttributes)
{
AttributeRecord tmpAR = new AttributeGeneric();
tmpAR.ReadARHeader(data, offset);
return(tmpAR);
}
AttributeRecord attRecord;
switch (TypeCode)
{
case AttributeTypeCode.STANDARD_INFORMATION:
attRecord = new StandardInformation();
break;
case AttributeTypeCode.ATTRIBUTE_LIST:
attRecord = new AttributeList();
break;
case AttributeTypeCode.FILE_NAME:
attRecord = new FileName();
break;
case AttributeTypeCode.OBJECT_ID:
attRecord = new ObjectId();
break;
// To complicated to quickly be implemented. Maybe one day. lol
// case AttributeTypeCode.SECURITY_DESCRIPTOR:
// attRecord = new SecurityDescriptor();
// break;
case AttributeTypeCode.VOLUME_NAME:
attRecord = new VolumeName();
break;
case AttributeTypeCode.VOLUME_INFORMATION:
attRecord = new VolumeInformation();
break;
case AttributeTypeCode.DATA:
attRecord = new Data();
break;
case AttributeTypeCode.INDEX_ROOT:
attRecord = new IndexRoot();
break;
// INDEX_ALLOCATION is stored as non resident and this project deals only with resident files
// case AttributeTypeCode.INDEX_ALLOCATION:
// attRecord = new IndexAllocation();
// break;
case AttributeTypeCode.BITMAP:
attRecord = new Bitmap();
break;
case AttributeTypeCode.EA_INFORMATION:
attRecord = new ExtenedAttributeInformation();
break;
case AttributeTypeCode.EA:
attRecord = new ExtenedAttributes();
break;
// PROPERTY_SET needs a pre NTFS 3.0 volume. This is probably obsolete!
// case AttributeTypeCode.PROPERTY_SET:
// attRecord = new PropertSet();
// break;
case AttributeTypeCode.LOGGED_UTILITY_STREAM:
attRecord = new LoggedUtilityStream();
break;
default: // ?? could be a problem
attRecord = new AttributeGeneric();
break;
}
attRecord.ReadARHeader(data, offset);
if (attRecord.FormCode == ResidentFileFlag.Resident)
{
attRecord.ResidentHeader = AttributeResidentHeader.ReadHeader(data, offset + 16);
int residentBodyOffset = offset + attRecord.ResidentHeader.ValueOffset;
int length = offset + attRecord.RecordLength - residentBodyOffset;
attRecord.ReadAttributeResident(data, length, residentBodyOffset);
}
else
{
throw new Exception("Could not read and process resident flag!\n");
}
return(attRecord);
}
public MainWindow()
{
const ushort btreeNodeSize = 5000;
GCSettings.LatencyMode = GCLatencyMode.Batch;// try to keep the WeakIOptimizedPersistableReference objects around longer
dataGridList = new List <DataGrid>();
dataTableList = new List <DataTable>();
InitializeComponent();
session = new SessionNoServer(s_systemDir);
Placement placerIndexRoot = new Placement(IndexRoot.PlaceInDatabase);
session.BeginUpdate();
Console.WriteLine("Running with databases in directory: " + session.SystemDirectory);
File.Copy(s_licenseDbFile, Path.Combine(session.SystemDirectory, "4.odb"), true);
IndexRoot indexRoot;
Database db = session.OpenDatabase(IndexRoot.PlaceInDatabase, false, false);
if (db == null)
{
session.NewDatabase(IndexRoot.PlaceInDatabase, 0, "IndexRoot");
session.NewDatabase(Lexicon.PlaceInDatabase, 0, "Lexicon");
session.NewDatabase(Document.PlaceInDatabase, 0, "Document");
session.NewDatabase(Repository.PlaceInDatabase, 0, "Repository");
session.NewDatabase(DocumentText.PlaceInDatabase, 0, "DocumentText");
session.NewDatabase(Word.PlaceInDatabase, 0, "Word");
indexRoot = new IndexRoot(btreeNodeSize, session);
if (Directory.Exists(s_booksDir))
{
string[] directoryTextFiles = Directory.GetFiles(s_booksDir, "*.txt");
foreach (string fileName in directoryTextFiles)
{
listBoxPagesToAdd.Items.Add(fileName);
}
}
else
{
wordMinCt.Text = 1.ToString();
listBoxPagesToAdd.Items.Add("http://www.VelocityDB.com/");
// other database products
listBoxPagesToAdd.Items.Add("https://foundationdb.com/");
listBoxPagesToAdd.Items.Add("http://www.oracle.com/us/products/database/index.html");
listBoxPagesToAdd.Items.Add("http://www-01.ibm.com/software/data/db2/");
listBoxPagesToAdd.Items.Add("http://www.versant.com/");
listBoxPagesToAdd.Items.Add("http://web.progress.com/en/objectstore/");
listBoxPagesToAdd.Items.Add("https://www.mongodb.org/");
listBoxPagesToAdd.Items.Add("http://cassandra.apache.org/");
listBoxPagesToAdd.Items.Add("http://www.sybase.com/");
listBoxPagesToAdd.Items.Add("http://www.mcobject.com/perst");
listBoxPagesToAdd.Items.Add("http://www.marklogic.com/what-is-marklogic/");
listBoxPagesToAdd.Items.Add("http://hamsterdb.com/");
listBoxPagesToAdd.Items.Add("http://www.firebirdsql.org/");
listBoxPagesToAdd.Items.Add("http://www.h2database.com/");
listBoxPagesToAdd.Items.Add("http://www.oracle.com/technology/products/berkeley-db");
listBoxPagesToAdd.Items.Add("http://www.scimore.com/");
listBoxPagesToAdd.Items.Add("http://www.stsdb.com/");
listBoxPagesToAdd.Items.Add("http://www.sqlite.org/about.html");
listBoxPagesToAdd.Items.Add("http://www.mysql.com/products/enterprise/techspec.html");
listBoxPagesToAdd.Items.Add("http://www.objectivity.com");
listBoxPagesToAdd.Items.Add("http://vistadb.net/");
listBoxPagesToAdd.Items.Add("http://www.google.com/search?q=object+database&sourceid=ie7&rls=com.microsoft:en-us:IE-SearchBox&ie=&oe=");
}
indexRoot.Persist(session, indexRoot);
}
else
{
indexRoot = (IndexRoot)session.Open(Oid.Encode(IndexRoot.PlaceInDatabase, 1, 1));
}
if (indexRoot.repository.documentSet.Count > 0)
{
List <Document> docs = indexRoot.repository.documentSet.ToList <Document>().Take(50).ToList <Document>();
inDbListBox.ItemsSource = docs;
}
updateDataGrids(indexRoot);
session.Commit();
//verify();
}
static void importEntireWikipedia()
{
const ushort btreeNodeSize = 10000;
Console.WriteLine(DateTime.Now.ToString() + ", start importing Wikipedia text");
//System.Xml.Schema.XmlSchema docSchema;
//using (System.Xml.XmlTextReader schemaReader = new System.Xml.XmlTextReader("c:\\export-0_5.xsd"))
//{
// docSchema = System.Xml.Schema.XmlSchema.Read(schemaReader, ValidationCallBack);
// }
int docCount = 0;
using (SessionNoServer session = new SessionNoServer(s_systemDir, 5000, false, false, CacheEnum.No)) // turn of page and object caching
{
Console.WriteLine("Running with databases in directory: " + session.SystemDirectory);
//GCSettings.LatencyMode = GCLatencyMode.Batch;// try to keep the WeakIOptimizedPersistableReference objects around longer
Placement documentPlacement = new Placement(Document.PlaceInDatabase, 1003, 1, 500, 1000, false, false, 1000, false);
Placement contentPlacement = new Placement(Document.PlaceInDatabase, 1, 1, 500, UInt16.MaxValue, false, false, 1, false);
XmlComment xmlComment;
XmlElement xmlElement;
XmlEntity xmlEntity;
XmlText xmlText;
XmlWhitespace xmlWhitespace;
session.BeginUpdate();
File.Copy(s_licenseDbFile, System.IO.Path.Combine(session.SystemDirectory, "4.odb"), true);
// register all database schema classes used by the application in advance to avoid lock conflict later in parallell indexing
session.RegisterClass(typeof(Repository));
session.RegisterClass(typeof(IndexRoot));
session.RegisterClass(typeof(Document));
session.RegisterClass(typeof(Lexicon));
session.RegisterClass(typeof(DocumentText));
session.RegisterClass(typeof(Word));
session.RegisterClass(typeof(WordGlobal));
session.RegisterClass(typeof(WordHit));
session.RegisterClass(typeof(BTreeSet <Document>));
session.RegisterClass(typeof(OidShort));
session.RegisterClass(typeof(BTreeMap <Word, WordHit>));
session.RegisterClass(typeof(HashCodeComparer <Word>));
session.RegisterClass(typeof(BTreeSetOidShort <Word>));
session.RegisterClass(typeof(BTreeMapOidShort <Word, WordHit>));
Database db = session.OpenDatabase(IndexRoot.PlaceInDatabase, false, false);
if (db != null)
{
outputSomeInfo(session);
session.Abort();
return;
}
session.NewDatabase(IndexRoot.PlaceInDatabase, 0, "IndexRoot");
session.NewDatabase(Lexicon.PlaceInDatabase, 0, "Lexicon");
session.NewDatabase(Repository.PlaceInDatabase, 0, "Repository");
for (UInt32 i = 40; i <= 186; i++)
{
session.NewDatabase(i, 512, "Document"); // pre allocate 146 Document databases presized to 512MB each
}
//session.SetTraceDbActivity(Lexicon.PlaceInDatabase);
//session.SetTraceAllDbActivity();
XmlDocument xmlDocument = new XmlDocument("enwiki-latest-pages-articles.xml");
IndexRoot indexRoot = new IndexRoot(btreeNodeSize, session);
indexRoot.Persist(session, indexRoot, true);
Document doc = null;
bool titleElement = false;
bool pageText = false;
UInt32 currentDocumentDatabaseNum = documentPlacement.StartDatabaseNumber;
using (FileStream fs = new FileStream(s_wikipediaXmlFile, FileMode.Open))
{
//using (GZipStream zipStream = new GZipStream(fs, CompressionMode.Decompress)) // if input was a .gz file
{
using (System.Xml.XmlTextReader textReader = new System.Xml.XmlTextReader(fs))
{
while (textReader.Read())
{
System.Xml.XmlNodeType nodeType = textReader.NodeType;
switch (nodeType)
{
case System.Xml.XmlNodeType.Attribute:
break;
case System.Xml.XmlNodeType.CDATA:
break;
case System.Xml.XmlNodeType.Comment:
xmlComment = new XmlComment(textReader.Value, xmlDocument);
break;
case System.Xml.XmlNodeType.Document:
break;
case System.Xml.XmlNodeType.DocumentFragment:
break;
case System.Xml.XmlNodeType.DocumentType:
break;
case System.Xml.XmlNodeType.Element:
xmlElement = new XmlElement(textReader.Prefix, textReader.LocalName, textReader.NamespaceURI, xmlDocument);
if (textReader.LocalName == "title")
{
titleElement = true;
}
else if (textReader.LocalName == "text")
{
pageText = true;
}
break;
case System.Xml.XmlNodeType.EndElement:
if (textReader.LocalName == "title" && doc != null)
{
titleElement = false;
}
else if (textReader.LocalName == "text" && doc != null)
{
pageText = false;
}
break;
case System.Xml.XmlNodeType.EndEntity:
break;
case System.Xml.XmlNodeType.Entity:
xmlEntity = new XmlEntity(textReader.LocalName, xmlDocument);
break;
case System.Xml.XmlNodeType.EntityReference:
break;
case System.Xml.XmlNodeType.None:
break;
case System.Xml.XmlNodeType.Notation:
break;
case System.Xml.XmlNodeType.ProcessingInstruction:
break;
case System.Xml.XmlNodeType.SignificantWhitespace:
break;
case System.Xml.XmlNodeType.Text:
xmlText = new XmlText(textReader.Value, xmlDocument);
if (titleElement)
{
doc = new Document(textReader.Value, indexRoot, session);
doc.Persist(documentPlacement, session, true);
if (doc.DatabaseNumber != currentDocumentDatabaseNum)
{
session.FlushUpdates(session.OpenDatabase(currentDocumentDatabaseNum));
Console.WriteLine("Database: " + currentDocumentDatabaseNum + " is completed, done importing article " + docCount + " number of lines: " + textReader.LineNumber);
currentDocumentDatabaseNum = doc.DatabaseNumber;
}
//doc.Page.Database.Name = doc.Name;
}
else if (doc != null && pageText)
{
#if DEBUGx
Console.WriteLine(doc.Name + " line: " + textReader.LineNumber);
#endif
//if (textReader.LineNumber > 1000000)
//{
// session.Commit();
// return;
//}
DocumentText content = new DocumentText(textReader.Value, doc);
if (doc.DatabaseNumber != contentPlacement.TryDatabaseNumber)
{
contentPlacement = new Placement(doc.DatabaseNumber, (ushort)contentPlacement.StartPageNumber, 1, contentPlacement.MaxObjectsPerPage, contentPlacement.MaxPagesPerDatabase, false, false, 1, false);
}
content.Persist(contentPlacement, session, false);
Debug.Assert(content.DatabaseNumber == doc.DatabaseNumber);
doc.Content = content;
indexRoot.repository.documentSet.AddFast(doc);
if (++docCount % 1000000 == 0)
{
//session.Commit(false); // skip recovery check, we do it in BeginUpdate which is enough
Console.WriteLine("Done importing article " + docCount + " number of lines: " + textReader.LineNumber);
//session.BeginUpdate();
}
}
break;
case System.Xml.XmlNodeType.Whitespace:
xmlWhitespace = new XmlWhitespace(textReader.Value, xmlDocument);
break;
case System.Xml.XmlNodeType.XmlDeclaration:
break;
}
;
}
Console.WriteLine("Finished importing article " + docCount + " number of lines: " + textReader.LineNumber);
}
}
}
session.Commit();
}
Console.WriteLine(DateTime.Now.ToString() + ", done importing Wikipedia text");
}
void updateDataGrids(IndexRoot indexRoot, int indexOfRemoved = -1)
{
if (indexRoot == null)
{
return;
}
if (indexRoot.lexicon.WordSet.Count == 0)
{
return;
}
stackPanel.IsEnabled = false;
bool aRefresh = stackPanel.Children.Count > 0;
if (indexOfRemoved >= 0 && aRefresh)
{
stackPanel.Children.RemoveAt(0);
}
else if (stackPanel.Children.Count > 0)
{
stackPanel.Children.Clear();
}
DataGrid dataGrid = new DataGrid();
dataGrid.AutoGenerateColumns = true;
dataGrid.MaxColumnWidth = 150;
dataGridList.Add(dataGrid);
DataTable table = new DataTable("Word Count");
DataColumn wordColumn = new DataColumn("Words (all pages)", Type.GetType("System.String"));
DataColumn countColumn = new DataColumn("Count", Type.GetType("System.UInt32"));
table.Columns.Add(wordColumn);
table.Columns.Add(countColumn);
DataRow newRow;
int pageIndex = 0;
int min = 3;
int.TryParse(wordMinCt.Text, out min);
foreach (Word word in indexRoot.lexicon.WordSet)
{
if (word.GlobalCount >= min)
{
newRow = table.NewRow();
newRow[0] = word.aWord;
newRow[1] = word.GlobalCount;
table.Rows.Add(newRow);
}
}
DataView dataView = new DataView(table);
dataView.Sort = "Count desc";
dataGrid.ItemsSource = dataView;
stackPanel.Children.Insert(pageIndex++, dataGrid);
if (indexOfRemoved >= 0 && aRefresh)
{
stackPanel.Children.RemoveAt(indexOfRemoved + 1);
}
else
{
List <Document> docs = indexRoot.repository.documentSet.ToList <Document>().ToList <Document>();
foreach (Document page in docs)
{
DataTable pageTable = new DataTable();
dataTableList.Add(pageTable);
string pageName = page.url.TrimEnd('/');
int index = pageName.IndexOf("//");
if (index >= 0)
{
pageName = pageName.Remove(0, index + 2);
}
index = pageName.IndexOf("www.");
if (index >= 0)
{
pageName = pageName.Remove(0, index + 4);
}
pageName = pageName.Replace('.', ' ');
pageName = pageName.Replace('/', ' ');
DataColumn wordColumnPage = new DataColumn(pageName, Type.GetType("System.String"));
DataColumn countColumnPage = new DataColumn("Count", Type.GetType("System.Int32"));
pageTable.Columns.Add(wordColumnPage);
pageTable.Columns.Add(countColumnPage);
foreach (KeyValuePair <Word, WordHit> pair in page.WordHit)
{
if ((int)pair.Value.Count >= min)
{
newRow = pageTable.NewRow();
string aString = pair.Key.aWord;
newRow.SetField <string>(wordColumnPage, aString);
newRow.SetField <int>(countColumnPage, (int)pair.Value.Count);
//wc.Add(new WordCount(aString, (uint) hit.wordPositionSet.Count));
pageTable.Rows.Add(newRow);
}
}
dataGrid = new DataGrid();
dataGrid.AutoGenerateColumns = true;
dataGrid.MaxColumnWidth = 150;
dataGridList.Add(dataGrid);
dataView = new DataView(pageTable);
dataView.Sort = "Count desc";
dataGrid.ItemsSource = dataView;
stackPanel.Children.Insert(pageIndex++, dataGrid);
}
}
stackPanel.IsEnabled = true;
}
public FileRecord(byte[] rawBytes, int offset)
{
Offset = offset;
var sig = BitConverter.ToInt32(rawBytes, 0);
if ((sig != _fileSig) && (sig != _baadSig) && (sig != 0x0))
{
Logger.Fatal($"Invalid signature! 0x{sig:X}");
return;
//throw new Exception("Invalid signature!");
}
if (sig == _baadSig)
{
Logger.Warn($"Bad signature at offset 0x{offset:X}");
return;
}
Attributes = new List <Attribute>();
FixupOffset = BitConverter.ToInt16(rawBytes, 2);
FixupEntryCount = BitConverter.ToInt16(rawBytes, 4);
LogSequenceNumber = BitConverter.ToInt64(rawBytes, 0x8);
SequenceNumber = BitConverter.ToInt16(rawBytes, 0x10);
ReferenceCount = BitConverter.ToInt16(rawBytes, 0x12);
FirstAttributeOffset = BitConverter.ToInt16(rawBytes, 0x14);
EntryFlags = (EntryFlag)BitConverter.ToInt16(rawBytes, 0x16);
Logger.Trace($"Entry flags: {EntryFlags}");
ActualRecordSize = BitConverter.ToInt32(rawBytes, 0x18);
AllocatedRecordSize = BitConverter.ToInt32(rawBytes, 0x1c);
var entryBytes = new byte[8];
Buffer.BlockCopy(rawBytes, 0x20, entryBytes, 0, 8);
MFTRecordToBaseRecord = new MftEntryInfo(entryBytes);
FirstAvailableAttribueId = BitConverter.ToInt16(rawBytes, 0x28);
EntryNumber = BitConverter.ToInt32(rawBytes, 0x2c);
var fixupExpectedBytes = new byte[2];
var fixupActual1 = new byte[2];
var fixupActual2 = new byte[2];
Buffer.BlockCopy(rawBytes, 0x30, fixupExpectedBytes, 0, 2);
Buffer.BlockCopy(rawBytes, 0x32, fixupActual1, 0, 2);
Buffer.BlockCopy(rawBytes, 0x34, fixupActual2, 0, 2);
//verify this record looks ok based on fixup bytes
//0x1FE and 0x3fe
var expectedFixupVal = BitConverter.ToInt16(fixupExpectedBytes, 0);
var x1FeValue = BitConverter.ToInt16(rawBytes, 0x1FE);
var x3FeValue = BitConverter.ToInt16(rawBytes, 0x3FE);
if ((x1FeValue != expectedFixupVal) &&
((EntryFlags & EntryFlag.FileRecordSegmentInUse) == EntryFlag.FileRecordSegmentInUse))
{
Logger.Warn(
$"FILE record at offset 0x{offset:X}! Fixup values do not match at 0x1FE. Expected: {expectedFixupVal}, actual: {x1FeValue}, EntryFlags: {EntryFlags}");
}
if ((x3FeValue != expectedFixupVal) &&
((EntryFlags & EntryFlag.FileRecordSegmentInUse) == EntryFlag.FileRecordSegmentInUse))
{
Logger.Warn(
$"FILE record at offset 0x{offset:X}! Fixup values do not match at 0x3FE. Expected: {expectedFixupVal}, actual: {x3FeValue}, EntryFlags: {EntryFlags}");
}
//header is done, replace fixup bytes with actual bytes
//0x1fe and 0x3fe should contain fixup bytes
Buffer.BlockCopy(fixupActual1, 0, rawBytes, 0x1fe, 2);
Buffer.BlockCopy(fixupActual2, 0, rawBytes, 0x3fe, 2);
//start attribute processing at FirstAttributeOffset
var index = (int)FirstAttributeOffset;
while (index < ActualRecordSize)
{
var attrType = BitConverter.ToInt32(rawBytes, index);
var attrSize = BitConverter.ToInt32(rawBytes, index + 4);
// Logger.Trace(
// $"ActualRecordSize: {ActualRecordSize} attrType: 0x{attrType:X}, size: {attrSize}, index: {index}, offset: 0x{offset:x}, i+o: 0x{index + offset:X}");
if ((attrSize == 0) || (attrType == -1))
{
index += 8; //skip -1 type and 0 size
if (EntryFlags == 0) //this is a free record
{
break;
}
continue;
}
var rawAttr = new byte[attrSize];
Buffer.BlockCopy(rawBytes, index, rawAttr, 0, attrSize);
switch ((AttributeType)attrType)
{
case AttributeType.StandardInformation:
var si = new StandardInfo(rawAttr);
Attributes.Add(si);
SILastAccessedOn = si.LastAccessedOn;
SICreatedOn = si.CreatedOn;
SIRecordModifiedOn = si.RecordModifiedOn;
SIContentModifiedOn = si.ContentModifiedOn;
break;
case AttributeType.FileName:
var fi = new FileName(rawAttr);
Attributes.Add(fi);
if ((fi.FileInfo.NameType & NameTypes.Windows) == NameTypes.Windows)
{
FName = fi.FileInfo.FileName;
}
//if (fi.FileInfo.LastAccessedOn.UtcDateTime != SILastAccessedOn.UtcDateTime)
//{
FNLastAccessedOn = fi.FileInfo.LastAccessedOn;
//}
//if (fi.FileInfo.CreatedOn.UtcDateTime != SICreatedOn.UtcDateTime)
//{
FNCreatedOn = fi.FileInfo.CreatedOn;
//}
//if (fi.FileInfo.RecordModifiedOn.UtcDateTime != SIRecordModifiedOn.UtcDateTime)
//{
FNRecordModifiedOn = fi.FileInfo.RecordModifiedOn;
//}
//if (fi.FileInfo.ContentModifiedOn.UtcDateTime != SIContentModifiedOn.UtcDateTime)
//{
FNContentModifiedOn = fi.FileInfo.ContentModifiedOn;
//}
break;
case AttributeType.Data:
var data = new Data(rawAttr);
Attributes.Add(data);
break;
case AttributeType.IndexAllocation:
var ia = new IndexAllocation(rawAttr);
Attributes.Add(ia);
break;
case AttributeType.IndexRoot:
var ir = new IndexRoot(rawAttr);
Attributes.Add(ir);
break;
case AttributeType.Bitmap:
var bm = new Bitmap(rawAttr);
Attributes.Add(bm);
break;
case AttributeType.VolumeVersionObjectId:
var oi = new ObjectId(rawAttr);
Attributes.Add(oi);
break;
case AttributeType.SecurityDescriptor:
var sd = new SecurityDescriptor(rawAttr);
Attributes.Add(sd);
break;
case AttributeType.VolumeName:
var vn = new VolumeName(rawAttr);
Attributes.Add(vn);
break;
case AttributeType.VolumeInformation:
var vi = new VolumeInformation(rawAttr);
Attributes.Add(vi);
break;
case AttributeType.LoggedUtilityStream:
var lus = new LoggedUtilityStream(rawAttr);
Attributes.Add(lus);
break;
case AttributeType.ReparsePoint:
var rp = new ReparsePoint(rawAttr);
Attributes.Add(rp);
break;
case AttributeType.AttributeList:
var al = new AttributeList(rawAttr);
Attributes.Add(al);
break;
case AttributeType.Ea:
//TODO Finish this
var ea = new ExtendedAttribute(rawAttr);
Attributes.Add(ea);
break;
case AttributeType.EaInformation:
var eai = new ExtendedAttributeInformation(rawAttr);
Attributes.Add(eai);
break;
default:
Logger.Warn($"Unhandled attribute type! Add me: {(AttributeType) attrType}");
throw new Exception($"Add me: {(AttributeType) attrType}");
break;
}
index += attrSize;
}
SlackStartOffset = index;
//rest is slack. handle here?
Logger.Trace($"Slack starts at {index} i+o: 0x{index + offset:X}");
}