Chunker.java

Go to the documentation of this file.

/*
 * Autopsy Forensic Browser
 *
 * Copyright 2011-2018 Basis Technology Corp.
 * Contact: carrier <at> sleuthkit <dot> org
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package org.sleuthkit.autopsy.keywordsearch;

import java.io.IOException;
import java.io.PushbackReader;
import java.io.Reader;
import java.nio.charset.Charset;
import java.nio.charset.StandardCharsets;
import java.text.Normalizer;
import java.util.Iterator;
import java.util.NoSuchElementException;
import javax.annotation.concurrent.NotThreadSafe;
import org.sleuthkit.autopsy.coreutils.TextUtil;
import org.sleuthkit.autopsy.keywordsearch.Chunker.Chunk;

@NotThreadSafe
class Chunker implements Iterator<Chunk>, Iterable<Chunk> {

    //local references to standard encodings
    private static final Charset UTF_16 = StandardCharsets.UTF_16;
    private static final Charset UTF_8 = StandardCharsets.UTF_8;

    //Chunking algorithm paramaters-------------------------------------//
    private static final int MAX_TOTAL_CHUNK_SIZE = 32760; //bytes
    private static final int MINIMUM_BASE_CHUNK_SIZE = 30 * 1024; //bytes
    private static final int MAXIMUM_BASE_CHUNK_SIZE = 31 * 1024; //bytes
    private static final int WHITE_SPACE_BUFFER_SIZE = 512; //bytes
    private static final int READ_CHARS_BUFFER_SIZE = 512; //chars
    private static final int MAX_CHAR_SIZE_INCREASE_IN_BYTES = 10; //bytes


    private final PushbackReader reader;
    private final char[] tempChunkBuf = new char[READ_CHARS_BUFFER_SIZE];

    private int chunkSizeBytes = 0;

    private int lowerCasedChunkSizeBytes = 0;
    private boolean endOfReaderReached = false;
    private Exception ex;

    Chunker(Reader reader) {
        //Using MAX_TOTAL_CHUNK_SIZE is safe but probably overkill.
        this.reader = new PushbackReader(reader, MAX_TOTAL_CHUNK_SIZE);
    }

    @Override
    public Iterator<Chunk> iterator() {
        return this;
    }

    boolean hasException() {
        return ex != null;
    }

    public Exception getException() {
        return ex;
    }

    @Override
    public boolean hasNext() {
        return (ex == null)
                && (endOfReaderReached == false);
    }

    private static StringBuilder sanitizeToUTF8(StringBuilder sb) {
        final int length = sb.length();
        for (int i = 0; i < length; i++) {
            if (TextUtil.isValidSolrUTF8(sb.charAt(i)) == false) {
                sb.replace(i, i + 1, "^");
            }
        }
        return sb;
    }

    private static StringBuilder replaceInvalidUTF16(String s) {
        /* encode the string to UTF-16 which does the replcement, see
         * Charset.encode(), then decode back to a StringBuilder. */
        return new StringBuilder(UTF_16.decode(UTF_16.encode(s)));
    }

    private static StringBuilder sanitize(String s) {
        String normStr = Normalizer.normalize(s, Normalizer.Form.NFKC);
        return sanitizeToUTF8(replaceInvalidUTF16(normStr));

    }

    @Override
    public Chunk next() {
        if (hasNext() == false) {
            throw new NoSuchElementException("There are no more chunks.");
        }
        //reset state for the next chunk

        chunkSizeBytes = 0;
        lowerCasedChunkSizeBytes = 0;
        int baseChunkSizeChars = 0;
        StringBuilder currentChunk = new StringBuilder();
        StringBuilder currentWindow = new StringBuilder();
        StringBuilder lowerCasedChunk = new StringBuilder();

        try {
            readBaseChunk(currentChunk, lowerCasedChunk);
            baseChunkSizeChars = currentChunk.length(); //save the base chunk length
            readWindow(currentWindow, lowerCasedChunk);
            //add the window text to the current chunk.
            currentChunk.append(currentWindow);
            if (endOfReaderReached) {
                /* if we have reached the end of the content,we won't make
                 * another overlapping chunk, so the length of the base chunk
                 * can be extended to the end. */
                baseChunkSizeChars = currentChunk.length();
            } else {
                /* otherwise we will make another chunk, so unread the window */
                reader.unread(currentWindow.toString().toCharArray());
            }
        } catch (Exception ioEx) {
            /* Save the exception, which will cause hasNext() to return false,
             * and break any chunking loop in client code. */
            ex = ioEx;
        }

        //sanitize the text and return a Chunk object, that includes the base chunk length.
        return new Chunk(currentChunk, baseChunkSizeChars, lowerCasedChunk);
    }

    private void readBaseChunk(StringBuilder currentChunk, StringBuilder lowerCasedChunk) throws IOException {
        //read the chunk until the minimum base chunk size
        readHelper(MINIMUM_BASE_CHUNK_SIZE, currentChunk, lowerCasedChunk);

        //keep reading until the maximum base chunk size or white space is reached.
        readToWhiteSpaceHelper(MAXIMUM_BASE_CHUNK_SIZE, currentChunk, lowerCasedChunk);
    }

    private void readWindow(StringBuilder currentChunk, StringBuilder lowerCasedChunk) throws IOException {
        //read the window, leaving some room to look for white space to break at.
        readHelper(MAX_TOTAL_CHUNK_SIZE - WHITE_SPACE_BUFFER_SIZE, currentChunk, lowerCasedChunk);

        //keep reading until the max chunk size, or until whitespace is reached.
        readToWhiteSpaceHelper(MAX_TOTAL_CHUNK_SIZE, currentChunk, lowerCasedChunk);
    }

    private void readHelper(int maxBytes, StringBuilder currentSegment, StringBuilder currentLowerCasedSegment) throws IOException {
        int charsRead = 0;
        //read chars up to maxBytes, or the end of the reader.
        while ((chunkSizeBytes < maxBytes) && (lowerCasedChunkSizeBytes < maxBytes)
                && (endOfReaderReached == false)) {
            charsRead = reader.read(tempChunkBuf, 0, READ_CHARS_BUFFER_SIZE);
            if (-1 == charsRead) {
                //this is the last chunk
                endOfReaderReached = true;
                return;
            } else {
                //if the last char might be part of a surroate pair, unread it.
                final char lastChar = tempChunkBuf[charsRead - 1];
                if (Character.isHighSurrogate(lastChar)) {
                    charsRead--;
                    reader.unread(lastChar);
                }

                //cleanup any invalid utf-16 sequences
                StringBuilder chunkSegment = sanitize(new String(tempChunkBuf, 0, charsRead));

                //get the length in utf8 bytes of the read chars
                int segmentSize = chunkSegment.toString().getBytes(UTF_8).length;

                // lower case the string and get it's size. NOTE: lower casing can 
                // change the size of the string!
                String lowerCasedSegment = chunkSegment.toString().toLowerCase();
                int lowerCasedSegmentSize = lowerCasedSegment.getBytes(UTF_8).length;

                //if it will not put us past maxBytes
                if ((chunkSizeBytes + segmentSize < maxBytes) && (lowerCasedChunkSizeBytes + lowerCasedSegmentSize < maxBytes)) {
                    //add it to the chunk
                    currentSegment.append(chunkSegment);
                    chunkSizeBytes += segmentSize;

                    currentLowerCasedSegment.append(lowerCasedSegment);
                    lowerCasedChunkSizeBytes += lowerCasedSegmentSize;
                } else {
                    //unread it, and break out of read loop.
                    reader.unread(tempChunkBuf, 0, charsRead);
                    return;
                }
            }
        }
    }

    private void readToWhiteSpaceHelper(int maxBytes, StringBuilder currentChunk, StringBuilder lowerCasedChunk) throws IOException {
        int charsRead = 0;
        boolean whitespaceFound = false;
        //read 1 char at a time up to maxBytes, whitespaceFound, or we reach the end of the reader.
        while ((chunkSizeBytes < maxBytes - MAX_CHAR_SIZE_INCREASE_IN_BYTES) 
                && (lowerCasedChunkSizeBytes < maxBytes - MAX_CHAR_SIZE_INCREASE_IN_BYTES)
                && (whitespaceFound == false)
                && (endOfReaderReached == false)) {
            charsRead = reader.read(tempChunkBuf, 0, 1);
            if (-1 == charsRead) {
                //this is the last chunk
                endOfReaderReached = true;
                return;
            } else {
                //if the last charcter might be part of a surroate pair, read another char
                final char ch = tempChunkBuf[0];
                String chunkSegment;
                if (Character.isHighSurrogate(ch)) {
                    //read another char into the buffer.
                    charsRead = reader.read(tempChunkBuf, 1, 1);
                    if (charsRead == -1) {
                        //this is the last chunk, so just drop the unpaired surrogate
                        endOfReaderReached = true;
                        return;
                    } else {
                        //try to use the pair together.
                        chunkSegment = new String(tempChunkBuf, 0, 2);
                    }
                } else {
                    //one char
                    chunkSegment = new String(tempChunkBuf, 0, 1);
                }

                //cleanup any invalid utf-16 sequences
                StringBuilder sanitizedChunkSegment = sanitize(chunkSegment);
                //check for whitespace.
                whitespaceFound = Character.isWhitespace(sanitizedChunkSegment.codePointAt(0));
                //add read chars to the chunk and update the length.
                currentChunk.append(sanitizedChunkSegment);
                chunkSizeBytes += sanitizedChunkSegment.toString().getBytes(UTF_8).length;

                // lower case the string and get it's size. NOTE: lower casing can 
                // change the size of the string.
                String lowerCasedSegment = sanitizedChunkSegment.toString().toLowerCase();
                lowerCasedChunk.append(lowerCasedSegment);
                lowerCasedChunkSizeBytes += lowerCasedSegment.getBytes(UTF_8).length;
            }
        }
    }

    static class Chunk {

        private final StringBuilder sb;
        private final int baseChunkSizeChars;
        private final StringBuilder lowerCasedChunk;

        Chunk(StringBuilder sb, int baseChunkSizeChars, StringBuilder lowerCasedChunk) {
            this.sb = sb;
            this.baseChunkSizeChars = baseChunkSizeChars;
            this.lowerCasedChunk = lowerCasedChunk;
        }

        @Override
        public String toString() {
            return sb.toString();
        }

        public String geLowerCasedChunk() {
            return lowerCasedChunk.toString();
        }

        int getBaseChunkLength() {
            return baseChunkSizeChars;
        }
    }
}