| 1 | #if 0 |
|---|
| 2 | # ----------------------------------------------------------------------- |
|---|
| 3 | # $Id: jpginfo.py,v 1.21 2005/04/16 17:08:40 dischi Exp $ |
|---|
| 4 | # ----------------------------------------------------------------------- |
|---|
| 5 | # $Log: jpginfo.py,v $ |
|---|
| 6 | # Revision 1.21 2005/04/16 17:08:40 dischi |
|---|
| 7 | # read metadata from epeg |
|---|
| 8 | # |
|---|
| 9 | # Revision 1.20 2005/04/16 15:01:15 dischi |
|---|
| 10 | # convert exif tags to str |
|---|
| 11 | # |
|---|
| 12 | # Revision 1.19 2004/05/20 15:56:31 dischi |
|---|
| 13 | # use Python Imaging for more info and gif/bmp support |
|---|
| 14 | # |
|---|
| 15 | # Revision 1.18 2003/06/30 13:17:19 the_krow |
|---|
| 16 | # o Refactored mediainfo into factory, synchronizedobject |
|---|
| 17 | # o Parsers now register directly at mmpython not at mmpython.mediainfo |
|---|
| 18 | # o use mmpython.Factory() instead of mmpython.mediainfo.get_singleton() |
|---|
| 19 | # o Bugfix in PNG parser |
|---|
| 20 | # o Renamed disc.AudioInfo into disc.AudioDiscInfo |
|---|
| 21 | # o Renamed disc.DataInfo into disc.DataDiscInfo |
|---|
| 22 | # |
|---|
| 23 | # Revision 1.17 2003/06/20 19:17:22 dischi |
|---|
| 24 | # remove filename again and use file.name |
|---|
| 25 | # |
|---|
| 26 | # Revision 1.16 2003/06/10 13:02:32 the_krow |
|---|
| 27 | # Softened JPEG parser a little so it accepts jpegs that do not end in FFD9. |
|---|
| 28 | # |
|---|
| 29 | # Revision 1.15 2003/06/09 14:31:57 the_krow |
|---|
| 30 | # fixes on the mpeg parser |
|---|
| 31 | # resolutions, fps and bitrate should be reported correctly now |
|---|
| 32 | # |
|---|
| 33 | # Revision 1.14 2003/06/09 12:50:08 the_krow |
|---|
| 34 | # mp3 now fills tables |
|---|
| 35 | # |
|---|
| 36 | # Revision 1.13 2003/06/08 19:55:22 dischi |
|---|
| 37 | # added bins metadata support |
|---|
| 38 | # |
|---|
| 39 | # Revision 1.12 2003/06/08 16:48:08 dischi |
|---|
| 40 | # Changed self.exif to exif_info and the same for ipc. We should extract |
|---|
| 41 | # everything we need to self.xxx and don't remember where the info came |
|---|
| 42 | # from (it's bad for the cache, we cache everything twice). |
|---|
| 43 | # Also added thumbnail to the list of things we want |
|---|
| 44 | # |
|---|
| 45 | # Revision 1.11 2003/06/08 13:44:57 dischi |
|---|
| 46 | # Changed all imports to use the complete mmpython path for mediainfo |
|---|
| 47 | # |
|---|
| 48 | # Revision 1.10 2003/06/08 13:11:51 dischi |
|---|
| 49 | # removed print at the end and moved it into register |
|---|
| 50 | # |
|---|
| 51 | # Revision 1.9 2003/06/07 21:48:47 the_krow |
|---|
| 52 | # Added Copying info |
|---|
| 53 | # started changing riffinfo to new AV stuff |
|---|
| 54 | # |
|---|
| 55 | # Revision 1.8 2003/05/13 18:28:17 the_krow |
|---|
| 56 | # JPEG Resolution |
|---|
| 57 | # |
|---|
| 58 | # Revision 1.7 2003/05/13 17:49:41 the_krow |
|---|
| 59 | # IPTC restructured\nEXIF Height read correctly\nJPEG Endmarker read |
|---|
| 60 | # |
|---|
| 61 | # Revision 1.6 2003/05/13 15:52:42 the_krow |
|---|
| 62 | # Caption added |
|---|
| 63 | # |
|---|
| 64 | # Revision 1.5 2003/05/13 15:23:59 the_krow |
|---|
| 65 | # IPTC |
|---|
| 66 | # |
|---|
| 67 | # Revision 1.4 2003/05/13 15:00:23 the_krow |
|---|
| 68 | # Tiff parsing |
|---|
| 69 | # |
|---|
| 70 | # Revision 1.3 2003/05/13 12:31:11 the_krow |
|---|
| 71 | # + GNU Copyright Notice |
|---|
| 72 | # + PNG Parsing |
|---|
| 73 | # |
|---|
| 74 | # ----------------------------------------------------------------------- |
|---|
| 75 | # MMPython - Media Metadata for Python |
|---|
| 76 | # Copyright (C) 2003 Thomas Schueppel |
|---|
| 77 | # |
|---|
| 78 | # This program is free software; you can redistribute it and/or modify |
|---|
| 79 | # it under the terms of the GNU General Public License as published by |
|---|
| 80 | # the Free Software Foundation; either version 2 of the License, or |
|---|
| 81 | # (at your option) any later version. |
|---|
| 82 | # |
|---|
| 83 | # This program is distributed in the hope that it will be useful, but |
|---|
| 84 | # WITHOUT ANY WARRANTY; without even the implied warranty of MER- |
|---|
| 85 | # CHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General |
|---|
| 86 | # Public License for more details. |
|---|
| 87 | # |
|---|
| 88 | # You should have received a copy of the GNU General Public License along |
|---|
| 89 | # with this program; if not, write to the Free Software Foundation, Inc., |
|---|
| 90 | # 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
|---|
| 91 | # |
|---|
| 92 | # ----------------------------------------------------------------------- |
|---|
| 93 | #endif |
|---|
| 94 | |
|---|
| 95 | |
|---|
| 96 | from mmpython import mediainfo |
|---|
| 97 | import mmpython |
|---|
| 98 | import IPTC |
|---|
| 99 | import EXIF |
|---|
| 100 | import struct |
|---|
| 101 | |
|---|
| 102 | import ImageInfo |
|---|
| 103 | |
|---|
| 104 | # interesting file format info: |
|---|
| 105 | # http://www.dcs.ed.ac.uk/home/mxr/gfx/2d-hi.html |
|---|
| 106 | # http://www.funducode.com/freec/Fileformats/format3/format3b.htm |
|---|
| 107 | |
|---|
| 108 | SOF = { 0xC0 : "Baseline", |
|---|
| 109 | 0xC1 : "Extended sequential", |
|---|
| 110 | 0xC2 : "Progressive", |
|---|
| 111 | 0xC3 : "Lossless", |
|---|
| 112 | 0xC5 : "Differential sequential", |
|---|
| 113 | 0xC6 : "Differential progressive", |
|---|
| 114 | 0xC7 : "Differential lossless", |
|---|
| 115 | 0xC9 : "Extended sequential, arithmetic coding", |
|---|
| 116 | 0xCA : "Progressive, arithmetic coding", |
|---|
| 117 | 0xCB : "Lossless, arithmetic coding", |
|---|
| 118 | 0xCD : "Differential sequential, arithmetic coding", |
|---|
| 119 | 0xCE : "Differential progressive, arithmetic coding", |
|---|
| 120 | 0xCF : "Differential lossless, arithmetic coding", |
|---|
| 121 | } |
|---|
| 122 | |
|---|
| 123 | _debug = mediainfo._debug |
|---|
| 124 | |
|---|
| 125 | class JPGInfo(mediainfo.ImageInfo): |
|---|
| 126 | |
|---|
| 127 | def __init__(self,file): |
|---|
| 128 | mediainfo.ImageInfo.__init__(self) |
|---|
| 129 | iptc_info = None |
|---|
| 130 | self.mime = 'image/jpeg' |
|---|
| 131 | self.type = 'jpeg image' |
|---|
| 132 | self.valid = 1 |
|---|
| 133 | if file.read(2) != '\xff\xd8': |
|---|
| 134 | self.valid = 0 |
|---|
| 135 | return |
|---|
| 136 | file.seek(-2,2) |
|---|
| 137 | if file.read(2) != '\xff\xd9': |
|---|
| 138 | # Normally an JPEG should end in ffd9. This does not however |
|---|
| 139 | # we assume it's an jpeg for now |
|---|
| 140 | mediainfo._debug("Wrong encode found for jpeg") |
|---|
| 141 | file.seek(2) |
|---|
| 142 | app = file.read(4) |
|---|
| 143 | self.meta = {} |
|---|
| 144 | while (len(app) == 4): |
|---|
| 145 | (ff,segtype,seglen) = struct.unpack(">BBH", app) |
|---|
| 146 | if ff != 0xff: break |
|---|
| 147 | _debug("SEGMENT: 0x%x%x, len=%d" % (ff,segtype,seglen)) |
|---|
| 148 | if segtype == 0xd9: |
|---|
| 149 | break |
|---|
| 150 | elif SOF.has_key(segtype): |
|---|
| 151 | data = file.read(seglen-2) |
|---|
| 152 | (precision,self.height,self.width,num_comp) = struct.unpack('>BHHB', data[:6]) |
|---|
| 153 | #_debug("H/W: %i / %i" % (self.height, self.width)) |
|---|
| 154 | elif segtype == 0xed: |
|---|
| 155 | app = file.read(seglen-2) |
|---|
| 156 | iptc_info = IPTC.flatten(IPTC.parseiptc(app)) |
|---|
| 157 | break |
|---|
| 158 | elif segtype == 0xe7: |
|---|
| 159 | # information created by libs like epeg |
|---|
| 160 | data = file.read(seglen-2) |
|---|
| 161 | if data.count('\n') == 1: |
|---|
| 162 | key, value = data.split('\n') |
|---|
| 163 | self.meta[key] = value |
|---|
| 164 | else: |
|---|
| 165 | file.seek(seglen-2,1) |
|---|
| 166 | app = file.read(4) |
|---|
| 167 | file.seek(0) |
|---|
| 168 | exif_info = EXIF.process_file(file) |
|---|
| 169 | if exif_info: |
|---|
| 170 | self.setitem( 'date', exif_info, 'Image DateTime', True ) |
|---|
| 171 | self.setitem( 'artist', exif_info, 'Image Artist', True ) |
|---|
| 172 | self.setitem( 'hardware', exif_info, 'Image Model', True ) |
|---|
| 173 | self.setitem( 'software', exif_info, 'Image Software', True ) |
|---|
| 174 | self.setitem( 'thumbnail', exif_info, 'JPEGThumbnail', True ) |
|---|
| 175 | self.appendtable( 'EXIF', exif_info ) |
|---|
| 176 | if iptc_info: |
|---|
| 177 | self.setitem( 'title', iptc_info, 517, True ) |
|---|
| 178 | self.setitem( 'date' , iptc_info, 567, True ) |
|---|
| 179 | self.setitem( 'comment', iptc_info, 617, True ) |
|---|
| 180 | self.setitem( 'keywords', iptc_info, 537, True ) |
|---|
| 181 | self.setitem( 'artist', iptc_info, 592, True ) |
|---|
| 182 | self.setitem( 'country', iptc_info, 612, True ) |
|---|
| 183 | self.setitem( 'caption', iptc_info, 632, True ) |
|---|
| 184 | self.appendtable( 'IPTC', iptc_info ) |
|---|
| 185 | |
|---|
| 186 | if len(self.meta.keys()): |
|---|
| 187 | self.appendtable( 'JPGMETA', self.meta ) |
|---|
| 188 | for key, value in self.meta.items(): |
|---|
| 189 | if key.startswith('Thumb:') or key == 'Software': |
|---|
| 190 | setattr(self, key, value) |
|---|
| 191 | if not key in self.keys: |
|---|
| 192 | self.keys.append(key) |
|---|
| 193 | |
|---|
| 194 | ImageInfo.add(file.name, self) |
|---|
| 195 | return |
|---|
| 196 | |
|---|
| 197 | |
|---|
| 198 | mmpython.registertype( 'image/jpeg', ('jpg','jpeg'), mediainfo.TYPE_IMAGE, JPGInfo ) |
|---|
