/*
* Copyright 2002-2016 Drew Noakes
*
* 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.
*
* More information about this project is available at:
*
* https://drewnoakes.com/code/exif/
* https://github.com/drewnoakes/metadata-extractor
*/
package com.drew.imaging.tiff;
import com.drew.lang.RandomAccessReader;
import com.drew.lang.Rational;
import com.drew.lang.annotations.NotNull;
import java.io.IOException;
import java.util.HashSet;
import java.util.Set;
/**
* Processes TIFF-formatted data, calling into client code via that {@link TiffHandler} interface.
*
* @author Drew Noakes https://drewnoakes.com
*/
public class TiffReader
{
/**
* Processes a TIFF data sequence.
*
* @param reader the {@link RandomAccessReader} from which the data should be read
* @param handler the {@link TiffHandler} that will coordinate processing and accept read values
* @param tiffHeaderOffset the offset within <code>reader</code> at which the TIFF header starts
* @throws TiffProcessingException if an error occurred during the processing of TIFF data that could not be
* ignored or recovered from
* @throws IOException an error occurred while accessing the required data
*/
public void processTiff(@NotNull final RandomAccessReader reader,
@NotNull final TiffHandler handler,
final int tiffHeaderOffset) throws TiffProcessingException, IOException
{
// This must be either "MM" or "II".
short byteOrderIdentifier = reader.getInt16(tiffHeaderOffset);
if (byteOrderIdentifier == 0x4d4d) { // "MM"
reader.setMotorolaByteOrder(true);
} else if (byteOrderIdentifier == 0x4949) { // "II"
reader.setMotorolaByteOrder(false);
} else {
throw new TiffProcessingException("Unclear distinction between Motorola/Intel byte ordering: " + byteOrderIdentifier);
}
// Check the next two values for correctness.
final int tiffMarker = reader.getUInt16(2 + tiffHeaderOffset);
handler.setTiffMarker(tiffMarker);
int firstIfdOffset = reader.getInt32(4 + tiffHeaderOffset) + tiffHeaderOffset;
// David Ekholm sent a digital camera image that has this problem
// TODO getLength should be avoided as it causes RandomAccessStreamReader to read to the end of the stream
if (firstIfdOffset >= reader.getLength() - 1) {
handler.warn("First IFD offset is beyond the end of the TIFF data segment -- trying default offset");
// First directory normally starts immediately after the offset bytes, so try that
firstIfdOffset = tiffHeaderOffset + 2 + 2 + 4;
}
Set<Integer> processedIfdOffsets = new HashSet<Integer>();
processIfd(handler, reader, processedIfdOffsets, firstIfdOffset, tiffHeaderOffset);
handler.completed(reader, tiffHeaderOffset);
}
/**
* Processes a TIFF IFD.
*
* IFD Header:
* <ul>
* <li><b>2 bytes</b> number of tags</li>
* </ul>
* Tag structure:
* <ul>
* <li><b>2 bytes</b> tag type</li>
* <li><b>2 bytes</b> format code (values 1 to 12, inclusive)</li>
* <li><b>4 bytes</b> component count</li>
* <li><b>4 bytes</b> inline value, or offset pointer if too large to fit in four bytes</li>
* </ul>
*
*
* @param handler the {@link com.drew.imaging.tiff.TiffHandler} that will coordinate processing and accept read values
* @param reader the {@link com.drew.lang.RandomAccessReader} from which the data should be read
* @param processedIfdOffsets the set of visited IFD offsets, to avoid revisiting the same IFD in an endless loop
* @param ifdOffset the offset within <code>reader</code> at which the IFD data starts
* @param tiffHeaderOffset the offset within <code>reader</code> at which the TIFF header starts
* @throws IOException an error occurred while accessing the required data
*/
public static void processIfd(@NotNull final TiffHandler handler,
@NotNull final RandomAccessReader reader,
@NotNull final Set<Integer> processedIfdOffsets,
final int ifdOffset,
final int tiffHeaderOffset) throws IOException
{
Boolean resetByteOrder = null;
try {
// check for directories we've already visited to avoid stack overflows when recursive/cyclic directory structures exist
if (processedIfdOffsets.contains(Integer.valueOf(ifdOffset))) {
return;
}
// remember that we've visited this directory so that we don't visit it again later
processedIfdOffsets.add(ifdOffset);
if (ifdOffset >= reader.getLength() || ifdOffset < 0) {
handler.error("Ignored IFD marked to start outside data segment");
return;
}
// First two bytes in the IFD are the number of tags in this directory
int dirTagCount = reader.getUInt16(ifdOffset);
// Some software modifies the byte order of the file, but misses some IFDs (such as makernotes).
// The entire test image repository doesn't contain a single IFD with more than 255 entries.
// Here we detect switched bytes that suggest this problem, and temporarily swap the byte order.
// This was discussed in GitHub issue #136.
if (dirTagCount > 0xFF && (dirTagCount & 0xFF) == 0) {
resetByteOrder = reader.isMotorolaByteOrder();
dirTagCount >>= 8;
reader.setMotorolaByteOrder(!reader.isMotorolaByteOrder());
}
int dirLength = (2 + (12 * dirTagCount) + 4);
if (dirLength + ifdOffset > reader.getLength()) {
handler.error("Illegally sized IFD");
return;
}
//
// Handle each tag in this directory
//
int invalidTiffFormatCodeCount = 0;
for (int tagNumber = 0; tagNumber < dirTagCount; tagNumber++) {
final int tagOffset = calculateTagOffset(ifdOffset, tagNumber);
// 2 bytes for the tag id
final int tagId = reader.getUInt16(tagOffset);
// 2 bytes for the format code
final int formatCode = reader.getUInt16(tagOffset + 2);
final TiffDataFormat format = TiffDataFormat.fromTiffFormatCode(formatCode);
// 4 bytes dictate the number of components in this tag's data
final long componentCount = reader.getUInt32(tagOffset + 4);
final long byteCount;
if (format == null) {
Long byteCountOverride = handler.tryCustomProcessFormat(tagId, formatCode, componentCount);
if (byteCountOverride == null) {
// This error suggests that we are processing at an incorrect index and will generate
// rubbish until we go out of bounds (which may be a while). Exit now.
handler.error(String.format("Invalid TIFF tag format code %d for tag 0x%04X", formatCode, tagId));
// TODO specify threshold as a parameter, or provide some other external control over this behaviour
if (++invalidTiffFormatCodeCount > 5) {
handler.error("Stopping processing as too many errors seen in TIFF IFD");
return;
}
continue;
}
byteCount = byteCountOverride;
} else {
byteCount = componentCount * format.getComponentSizeBytes();
}
final long tagValueOffset;
if (byteCount > 4) {
// If it's bigger than 4 bytes, the dir entry contains an offset.
final long offsetVal = reader.getUInt32(tagOffset + 8);
if (offsetVal + byteCount > reader.getLength()) {
// Bogus pointer offset and / or byteCount value
handler.error("Illegal TIFF tag pointer offset");
continue;
}
tagValueOffset = tiffHeaderOffset + offsetVal;
} else {
// 4 bytes or less and value is in the dir entry itself.
tagValueOffset = tagOffset + 8;
}
if (tagValueOffset < 0 || tagValueOffset > reader.getLength()) {
handler.error("Illegal TIFF tag pointer offset");
continue;
}
// Check that this tag isn't going to allocate outside the bounds of the data array.
// This addresses an uncommon OutOfMemoryError.
if (byteCount < 0 || tagValueOffset + byteCount > reader.getLength()) {
handler.error("Illegal number of bytes for TIFF tag data: " + byteCount);
continue;
}
// Some tags point to one or more additional IFDs to process
boolean isIfdPointer = false;
if (byteCount == 4 * componentCount) {
for (int i = 0; i < componentCount; i++) {
if (handler.tryEnterSubIfd(tagId)) {
isIfdPointer = true;
int subDirOffset = tiffHeaderOffset + reader.getInt32((int) (tagValueOffset + i * 4));
processIfd(handler, reader, processedIfdOffsets, subDirOffset, tiffHeaderOffset);
}
}
}
// If it wasn't an IFD pointer, allow custom tag processing to occur
if (!isIfdPointer && !handler.customProcessTag((int) tagValueOffset, processedIfdOffsets, tiffHeaderOffset, reader, tagId, (int) byteCount)) {
// If no custom processing occurred, process the tag in the standard fashion
processTag(handler, tagId, (int) tagValueOffset, (int) componentCount, formatCode, reader);
}
}
// at the end of each IFD is an optional link to the next IFD
final int finalTagOffset = calculateTagOffset(ifdOffset, dirTagCount);
int nextIfdOffset = reader.getInt32(finalTagOffset);
if (nextIfdOffset != 0) {
nextIfdOffset += tiffHeaderOffset;
if (nextIfdOffset >= reader.getLength()) {
// Last 4 bytes of IFD reference another IFD with an address that is out of bounds
// Note this could have been caused by jhead 1.3 cropping too much
return;
} else if (nextIfdOffset < ifdOffset) {
// TODO is this a valid restriction?
// Last 4 bytes of IFD reference another IFD with an address that is before the start of this directory
return;
}
if (handler.hasFollowerIfd()) {
processIfd(handler, reader, processedIfdOffsets, nextIfdOffset, tiffHeaderOffset);
}
}
} finally {
handler.endingIFD();
if (resetByteOrder != null)
reader.setMotorolaByteOrder(resetByteOrder);
}
}
private static void processTag(@NotNull final TiffHandler handler,
final int tagId,
final int tagValueOffset,
final int componentCount,
final int formatCode,
@NotNull final RandomAccessReader reader) throws IOException
{
switch (formatCode) {
case TiffDataFormat.CODE_UNDEFINED:
// this includes exif user comments
handler.setByteArray(tagId, reader.getBytes(tagValueOffset, componentCount));
break;
case TiffDataFormat.CODE_STRING:
handler.setString(tagId, reader.getNullTerminatedString(tagValueOffset, componentCount));
break;
case TiffDataFormat.CODE_RATIONAL_S:
if (componentCount == 1) {
handler.setRational(tagId, new Rational(reader.getInt32(tagValueOffset), reader.getInt32(tagValueOffset + 4)));
} else if (componentCount > 1) {
Rational[] array = new Rational[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = new Rational(reader.getInt32(tagValueOffset + (8 * i)), reader.getInt32(tagValueOffset + 4 + (8 * i)));
handler.setRationalArray(tagId, array);
}
break;
case TiffDataFormat.CODE_RATIONAL_U:
if (componentCount == 1) {
handler.setRational(tagId, new Rational(reader.getUInt32(tagValueOffset), reader.getUInt32(tagValueOffset + 4)));
} else if (componentCount > 1) {
Rational[] array = new Rational[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = new Rational(reader.getUInt32(tagValueOffset + (8 * i)), reader.getUInt32(tagValueOffset + 4 + (8 * i)));
handler.setRationalArray(tagId, array);
}
break;
case TiffDataFormat.CODE_SINGLE:
if (componentCount == 1) {
handler.setFloat(tagId, reader.getFloat32(tagValueOffset));
} else {
float[] array = new float[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getFloat32(tagValueOffset + (i * 4));
handler.setFloatArray(tagId, array);
}
break;
case TiffDataFormat.CODE_DOUBLE:
if (componentCount == 1) {
handler.setDouble(tagId, reader.getDouble64(tagValueOffset));
} else {
double[] array = new double[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getDouble64(tagValueOffset + (i * 4));
handler.setDoubleArray(tagId, array);
}
break;
case TiffDataFormat.CODE_INT8_S:
if (componentCount == 1) {
handler.setInt8s(tagId, reader.getInt8(tagValueOffset));
} else {
byte[] array = new byte[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getInt8(tagValueOffset + i);
handler.setInt8sArray(tagId, array);
}
break;
case TiffDataFormat.CODE_INT8_U:
if (componentCount == 1) {
handler.setInt8u(tagId, reader.getUInt8(tagValueOffset));
} else {
short[] array = new short[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getUInt8(tagValueOffset + i);
handler.setInt8uArray(tagId, array);
}
break;
case TiffDataFormat.CODE_INT16_S:
if (componentCount == 1) {
handler.setInt16s(tagId, (int)reader.getInt16(tagValueOffset));
} else {
short[] array = new short[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getInt16(tagValueOffset + (i * 2));
handler.setInt16sArray(tagId, array);
}
break;
case TiffDataFormat.CODE_INT16_U:
if (componentCount == 1) {
handler.setInt16u(tagId, reader.getUInt16(tagValueOffset));
} else {
int[] array = new int[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getUInt16(tagValueOffset + (i * 2));
handler.setInt16uArray(tagId, array);
}
break;
case TiffDataFormat.CODE_INT32_S:
// NOTE 'long' in this case means 32 bit, not 64
if (componentCount == 1) {
handler.setInt32s(tagId, reader.getInt32(tagValueOffset));
} else {
int[] array = new int[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getInt32(tagValueOffset + (i * 4));
handler.setInt32sArray(tagId, array);
}
break;
case TiffDataFormat.CODE_INT32_U:
// NOTE 'long' in this case means 32 bit, not 64
if (componentCount == 1) {
handler.setInt32u(tagId, reader.getUInt32(tagValueOffset));
} else {
long[] array = new long[componentCount];
for (int i = 0; i < componentCount; i++)
array[i] = reader.getUInt32(tagValueOffset + (i * 4));
handler.setInt32uArray(tagId, array);
}
break;
default:
handler.error(String.format("Invalid TIFF tag format code %d for tag 0x%04X", formatCode, tagId));
}
}
/**
* Determine the offset of a given tag within the specified IFD.
*
* @param ifdStartOffset the offset at which the IFD starts
* @param entryNumber the zero-based entry number
*/
private static int calculateTagOffset(int ifdStartOffset, int entryNumber)
{
// Add 2 bytes for the tag count.
// Each entry is 12 bytes.
return ifdStartOffset + 2 + (12 * entryNumber);
}
}