1 | /*
|
---|
2 | * Copyright 2002-2012 Drew Noakes
|
---|
3 | *
|
---|
4 | * Licensed under the Apache License, Version 2.0 (the "License");
|
---|
5 | * you may not use this file except in compliance with the License.
|
---|
6 | * You may obtain a copy of the License at
|
---|
7 | *
|
---|
8 | * http://www.apache.org/licenses/LICENSE-2.0
|
---|
9 | *
|
---|
10 | * Unless required by applicable law or agreed to in writing, software
|
---|
11 | * distributed under the License is distributed on an "AS IS" BASIS,
|
---|
12 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
---|
13 | * See the License for the specific language governing permissions and
|
---|
14 | * limitations under the License.
|
---|
15 | *
|
---|
16 | * More information about this project is available at:
|
---|
17 | *
|
---|
18 | * http://drewnoakes.com/code/exif/
|
---|
19 | * http://code.google.com/p/metadata-extractor/
|
---|
20 | */
|
---|
21 | package com.drew.metadata.iptc;
|
---|
22 |
|
---|
23 | import com.drew.lang.BufferBoundsException;
|
---|
24 | import com.drew.lang.BufferReader;
|
---|
25 | import com.drew.lang.annotations.NotNull;
|
---|
26 | import com.drew.metadata.Directory;
|
---|
27 | import com.drew.metadata.Metadata;
|
---|
28 | import com.drew.metadata.MetadataReader;
|
---|
29 |
|
---|
30 | import java.util.Date;
|
---|
31 |
|
---|
32 | /**
|
---|
33 | * Decodes IPTC binary data, populating a <code>Metadata</code> object with tag values in an <code>IptcDirectory</code>.
|
---|
34 | *
|
---|
35 | * @author Drew Noakes http://drewnoakes.com
|
---|
36 | */
|
---|
37 | public class IptcReader implements MetadataReader
|
---|
38 | {
|
---|
39 | // TODO consider breaking the IPTC section up into multiple directories and providing segregation of each IPTC directory
|
---|
40 | /*
|
---|
41 | public static final int DIRECTORY_IPTC = 2;
|
---|
42 |
|
---|
43 | public static final int ENVELOPE_RECORD = 1;
|
---|
44 | public static final int APPLICATION_RECORD_2 = 2;
|
---|
45 | public static final int APPLICATION_RECORD_3 = 3;
|
---|
46 | public static final int APPLICATION_RECORD_4 = 4;
|
---|
47 | public static final int APPLICATION_RECORD_5 = 5;
|
---|
48 | public static final int APPLICATION_RECORD_6 = 6;
|
---|
49 | public static final int PRE_DATA_RECORD = 7;
|
---|
50 | public static final int DATA_RECORD = 8;
|
---|
51 | public static final int POST_DATA_RECORD = 9;
|
---|
52 | */
|
---|
53 |
|
---|
54 | /** Performs the IPTC data extraction, adding found values to the specified instance of <code>Metadata</code>. */
|
---|
55 | public void extract(@NotNull final BufferReader reader, @NotNull final Metadata metadata)
|
---|
56 | {
|
---|
57 | IptcDirectory directory = metadata.getOrCreateDirectory(IptcDirectory.class);
|
---|
58 |
|
---|
59 | int offset = 0;
|
---|
60 |
|
---|
61 | /*
|
---|
62 | // find start-of-segment marker (potentially need to skip some ASCII photoshop header info)
|
---|
63 | try {
|
---|
64 | while (offset < data.length - 1 && reader.getUInt16(offset) != 0x1c01 && reader.getUInt16(offset) != 0x1c02)
|
---|
65 | offset++;
|
---|
66 | } catch (BufferBoundsException e) {
|
---|
67 | directory.addError("Couldn't find start of IPTC data (invalid segment)");
|
---|
68 | return;
|
---|
69 | }
|
---|
70 | */
|
---|
71 |
|
---|
72 | // for each tag
|
---|
73 | while (offset < reader.getLength()) {
|
---|
74 |
|
---|
75 | // identifies start of a tag
|
---|
76 | short startByte;
|
---|
77 | try {
|
---|
78 | startByte = reader.getUInt8(offset);
|
---|
79 | } catch (BufferBoundsException e) {
|
---|
80 | directory.addError("Unable to read starting byte of IPTC tag");
|
---|
81 | break;
|
---|
82 | }
|
---|
83 |
|
---|
84 | if (startByte != 0x1c) {
|
---|
85 | directory.addError("Invalid start to IPTC tag");
|
---|
86 | break;
|
---|
87 | }
|
---|
88 |
|
---|
89 | // we need at least five bytes left to read a tag
|
---|
90 | if (offset + 5 >= reader.getLength()) {
|
---|
91 | directory.addError("Too few bytes remain for a valid IPTC tag");
|
---|
92 | break;
|
---|
93 | }
|
---|
94 |
|
---|
95 | offset++;
|
---|
96 |
|
---|
97 | int directoryType;
|
---|
98 | int tagType;
|
---|
99 | int tagByteCount;
|
---|
100 | try {
|
---|
101 | directoryType = reader.getUInt8(offset++);
|
---|
102 | tagType = reader.getUInt8(offset++);
|
---|
103 | tagByteCount = reader.getUInt16(offset);
|
---|
104 | offset += 2;
|
---|
105 | } catch (BufferBoundsException e) {
|
---|
106 | directory.addError("IPTC data segment ended mid-way through tag descriptor");
|
---|
107 | return;
|
---|
108 | }
|
---|
109 |
|
---|
110 | if (offset + tagByteCount > reader.getLength()) {
|
---|
111 | directory.addError("Data for tag extends beyond end of IPTC segment");
|
---|
112 | break;
|
---|
113 | }
|
---|
114 |
|
---|
115 | try {
|
---|
116 | processTag(reader, directory, directoryType, tagType, offset, tagByteCount);
|
---|
117 | } catch (BufferBoundsException e) {
|
---|
118 | directory.addError("Error processing IPTC tag");
|
---|
119 | break;
|
---|
120 | }
|
---|
121 |
|
---|
122 | offset += tagByteCount;
|
---|
123 | }
|
---|
124 | }
|
---|
125 |
|
---|
126 | private void processTag(@NotNull BufferReader reader, @NotNull Directory directory, int directoryType, int tagType, int offset, int tagByteCount) throws BufferBoundsException
|
---|
127 | {
|
---|
128 | int tagIdentifier = tagType | (directoryType << 8);
|
---|
129 |
|
---|
130 | switch (tagIdentifier) {
|
---|
131 | case IptcDirectory.TAG_APPLICATION_RECORD_VERSION:
|
---|
132 | // short
|
---|
133 | int shortValue = reader.getUInt16(offset);
|
---|
134 | directory.setInt(tagIdentifier, shortValue);
|
---|
135 | return;
|
---|
136 | case IptcDirectory.TAG_URGENCY:
|
---|
137 | // byte
|
---|
138 | directory.setInt(tagIdentifier, reader.getUInt8(offset));
|
---|
139 | return;
|
---|
140 | case IptcDirectory.TAG_RELEASE_DATE:
|
---|
141 | case IptcDirectory.TAG_DATE_CREATED:
|
---|
142 | // Date object
|
---|
143 | if (tagByteCount >= 8) {
|
---|
144 | String dateStr = reader.getString(offset, tagByteCount);
|
---|
145 | try {
|
---|
146 | int year = Integer.parseInt(dateStr.substring(0, 4));
|
---|
147 | int month = Integer.parseInt(dateStr.substring(4, 6)) - 1;
|
---|
148 | int day = Integer.parseInt(dateStr.substring(6, 8));
|
---|
149 | Date date = new java.util.GregorianCalendar(year, month, day).getTime();
|
---|
150 | directory.setDate(tagIdentifier, date);
|
---|
151 | return;
|
---|
152 | } catch (NumberFormatException e) {
|
---|
153 | // fall through and we'll store whatever was there as a String
|
---|
154 | }
|
---|
155 | }
|
---|
156 | case IptcDirectory.TAG_RELEASE_TIME:
|
---|
157 | case IptcDirectory.TAG_TIME_CREATED:
|
---|
158 | // time...
|
---|
159 | default:
|
---|
160 | // fall through
|
---|
161 | }
|
---|
162 |
|
---|
163 | // If we haven't returned yet, treat it as a string
|
---|
164 | String str;
|
---|
165 | if (tagByteCount < 1) {
|
---|
166 | str = "";
|
---|
167 | } else {
|
---|
168 | str = reader.getString(offset, tagByteCount, System.getProperty("file.encoding")); // "ISO-8859-1"
|
---|
169 | }
|
---|
170 |
|
---|
171 | if (directory.containsTag(tagIdentifier)) {
|
---|
172 | // this fancy string[] business avoids using an ArrayList for performance reasons
|
---|
173 | String[] oldStrings = directory.getStringArray(tagIdentifier);
|
---|
174 | String[] newStrings;
|
---|
175 | if (oldStrings == null) {
|
---|
176 | newStrings = new String[1];
|
---|
177 | } else {
|
---|
178 | newStrings = new String[oldStrings.length + 1];
|
---|
179 | System.arraycopy(oldStrings, 0, newStrings, 0, oldStrings.length);
|
---|
180 | }
|
---|
181 | newStrings[newStrings.length - 1] = str;
|
---|
182 | directory.setStringArray(tagIdentifier, newStrings);
|
---|
183 | } else {
|
---|
184 | directory.setString(tagIdentifier, str);
|
---|
185 | }
|
---|
186 | }
|
---|
187 | }
|
---|