Class: Cul::Image::Properties::Exif::EXIF_header
- Inherits:
-
Object
- Object
- Cul::Image::Properties::Exif::EXIF_header
- Defined in:
- lib/cul_image_props/image/properties/exif/types.rb
Overview
class that handles an EXIF header
Constant Summary collapse
- X00 =
[0x00].pack('C')
Instance Attribute Summary collapse
-
#endian ⇒ Object
readonly
Returns the value of attribute endian.
-
#offset ⇒ Object
readonly
Returns the value of attribute offset.
-
#tags ⇒ Object
Returns the value of attribute tags.
Class Method Summary collapse
-
.s2n_intel(src) ⇒ Object
extract multibyte integer in Intel format (little endian).
-
.s2n_motorola(src) ⇒ Object
extract multibyte integer in Motorola format (big/network endian).
Instance Method Summary collapse
-
#canon_decode_tag(value, dict) ⇒ Object
decode Canon MakerNote tag based on offset within tag see http =>//www.burren.cx/david/canon.html by David Burren.
-
#decode_maker_note ⇒ Object
Note is the data that comprises this MakerNote.
-
#dump_IFD(ifd, ifd_name, opts) ⇒ Object
def dump_IFD(ifd, ifd_name, dict=EXIF_TAGS, relative=false, stop_tag=‘UNDEF’).
-
#extract_TIFF_thumbnail(thumb_ifd) ⇒ Object
extract uncompressed TIFF thumbnail (like pulling teeth) we take advantage of the pre-existing layout in the thumbnail IFD as much as possible.
-
#first_IFD ⇒ Object
return first IFD.
-
#initialize(file, endian, offset, fake_exif, strict, detail = true) ⇒ EXIF_header
constructor
A new instance of EXIF_header.
-
#list_IFDs ⇒ Object
return list of IFDs in header.
-
#n2s(offset, length) ⇒ Object
convert offset to string.
-
#next_IFD(ifd) ⇒ Object
return pointer to next IFD.
-
#olympus_decode_tag(value, dict) ⇒ Object
XXX TODO decode Olympus MakerNote tag based on offset within tag.
-
#s2n(offset, length, signed = false) ⇒ Object
convert slice to integer, based on sign and endian flags usually this offset is assumed to be relative to the beginning of the start of the EXIF information.
- #unpack_number(src, signed = false) ⇒ Object
Constructor Details
#initialize(file, endian, offset, fake_exif, strict, detail = true) ⇒ EXIF_header
Returns a new instance of EXIF_header.
153 154 155 156 157 158 159 160 161 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 153 def initialize(file, endian, offset, fake_exif, strict, detail=true) @file = file @endian = endian @offset = offset @fake_exif = fake_exif @strict = strict @detail = detail @tags = {} end |
Instance Attribute Details
#endian ⇒ Object (readonly)
Returns the value of attribute endian.
152 153 154 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 152 def endian @endian end |
#offset ⇒ Object (readonly)
Returns the value of attribute offset.
152 153 154 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 152 def offset @offset end |
#tags ⇒ Object
Returns the value of attribute tags.
151 152 153 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 151 def @tags end |
Class Method Details
.s2n_intel(src) ⇒ Object
extract multibyte integer in Intel format (little endian)
178 179 180 181 182 183 184 185 186 187 188 189 190 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 178 def self.s2n_intel(src) x = 0 l = src.bytesize if l == 1 return src.unpack('C')[0] elsif l == 2 return src.unpack('v')[0] elsif l == 4 return src.unpack('V')[0] else raise "Unexpected packed Fixnum length: " + l.to_s end end |
.s2n_motorola(src) ⇒ Object
extract multibyte integer in Motorola format (big/network endian)
164 165 166 167 168 169 170 171 172 173 174 175 176 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 164 def self.s2n_motorola(src) x = 0 l = src.bytesize if l == 1 return src.unpack('C')[0] elsif l == 2 return src.unpack('n')[0] elsif l == 4 return src.unpack('N')[0] else raise "Unexpected packed Fixnum length: " + l.to_s end end |
Instance Method Details
#canon_decode_tag(value, dict) ⇒ Object
decode Canon MakerNote tag based on offset within tag see http =>//www.burren.cx/david/canon.html by David Burren
524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 524 def canon_decode_tag(value, dict) (1 ... len(value)).each { |i| x=dict.get(i, ['Unknown']) name=x[0] if len(x) > 1 val=x[1].get(value[i], 'Unknown') else val=value[i] end # it's not a real IFD Tag but we fake one to make everybody # happy. this will have a "proprietary" type self.['MakerNote '+name]=IFD_Tag(str(val), None, 0, None, None, None) } end |
#decode_maker_note ⇒ Object
Note is the data that comprises this MakerNote. The MakerNote will likely have pointers in it that point to other parts of the file. We’ll use self.offset as the starting point for most of those pointers, since they are relative to the beginning of the file.
If the MakerNote is in a newer format, it may use relative addressing within the MakerNote. In that case we’ll use relative addresses for the pointers.
As an aside => it’s not just to be annoying that the manufacturers use relative offsets. It’s so that if the makernote has to be moved by the picture software all of the offsets don’t have to be adjusted. Overall, this is probably the right strategy for makernotes, though the spec is ambiguous. (The spec does not appear to imagine that makernotes would follow EXIF format internally. Once they did, it’s ambiguous whether the offsets should be from the header at the start of all the EXIF info, or from the header at the start of the makernote.)
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 443 def decode_maker_note() note = self.['EXIF MakerNote'] # Some apps use MakerNote tags but do not use a format for which we # have a description, so just do a raw dump for these. make = self.['Image Make'].printable # Nikon # The maker note usually starts with the word Nikon, followed by the # type of the makernote (1 or 2, as a short). If the word Nikon is # not at the start of the makernote, it's probably type 2, since some # cameras work that way. if make.include? 'NIKON' if note.values[0,7] == [78, 105, 107, 111, 110, 0, 1] self.dump_IFD(note.field_offset+8, 'MakerNote', :dict=>MAKERNOTE_NIKON_OLDER_TAGS) elsif note.values[0, 7] == [78, 105, 107, 111, 110, 0, 2] if note.values[12,2] != [0, 42] and note.values[12,2] != [42, 0] raise "Missing marker tag '42' in MakerNote." end # skip the Makernote label and the TIFF header self.dump_IFD(note.field_offset+10+8, 'MakerNote', :dict=>MAKERNOTE_NIKON_NEWER_TAGS, :relative=>true) else # E99x or D1 self.dump_IFD(note.field_offset, 'MakerNote', :dict=>MAKERNOTE_NIKON_NEWER_TAGS) end return end # Olympus if make.index('OLYMPUS') == 0 self.dump_IFD(note.field_offset+8, 'MakerNote', :dict=>MAKERNOTE_OLYMPUS_TAGS) return end # Casio if make.include? 'CASIO' or make.include? 'Casio' self.dump_IFD(note.field_offset, 'MakerNote', :dict=>MAKERNOTE_CASIO_TAGS) return end # Fujifilm if make == 'FUJIFILM' # bug => everything else is "Motorola" endian, but the MakerNote # is "Intel" endian endian = self.endian self.endian = 'I' # bug => IFD offsets are from beginning of MakerNote, not # beginning of file header offset = self.offset self.offset += note.field_offset # process note with bogus values (note is actually at offset 12) self.dump_IFD(12, 'MakerNote', :dict=>MAKERNOTE_FUJIFILM_TAGS) # reset to correct values self.endian = endian self.offset = offset return end # Canon if make == 'Canon' self.dump_IFD(note.field_offset, 'MakerNote', :dict=>MAKERNOTE_CANON_TAGS) [['MakerNote Tag 0x0001', MAKERNOTE_CANON_TAG_0x001], ['MakerNote Tag 0x0004', MAKERNOTE_CANON_TAG_0x004]].each { |i| begin self.canon_decode_tag(self.[i[0]].values, i[1]) # gd added rescue end } return end end |
#dump_IFD(ifd, ifd_name, opts) ⇒ Object
def dump_IFD(ifd, ifd_name, dict=EXIF_TAGS, relative=false, stop_tag=‘UNDEF’)
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 264 def dump_IFD(ifd, ifd_name, opts) opts = {:dict => EXIF_TAGS, :relative => false, :stop_tag => 'UNDEF'}.merge(opts) dict = opts[:dict] relative = opts[:relative] stop_tag = opts[:stop_tag] @file.seek(@offset + ifd) entries = unpack_number(@file.read(2)) entries_ptr = ifd + 2 puts ifd_name + " had zero entries!" if entries == 0 (0 ... entries).each { |i| # entry is index of start of this IFD in the file entry = ifd + 2 + (12 * i) @file.seek(@offset + entry) tag_id = unpack_number(@file.read(2)) # get tag name early to avoid errors, help debug tag_entry = dict[tag_id] || TagEntry.new('Tag 0x%04X' % tag_id) # ignore certain tags for faster processing if not (not @detail and IGNORE_TAGS.include? tag_id) # The 12 byte Tag format is ID (short) TYPE (short) COUNT (long) VALUE (long) # if actual values would exceed 4 bytes (long), VALUE # is instead a pointer to the actual values. field_type = unpack_number(@file.read(2)) # unknown field type if 0 > field_type or field_type >= FIELD_TYPES.length if not @strict next else raise format("unknown type %d in tag 0x%04X", field_type, tag) end end typelen = FIELD_TYPES[field_type].length signed = FIELD_TYPES[field_type].signed count = unpack_number(@file.read(4)) # If the value exceeds 4 bytes, it is a pointer to values. if (count * typelen) > 4 # Note that 'relative' is a fix for the Nikon type 3 makernote. field_offset = unpack_number(@file.read(4)) if relative field_offset = field_offset + ifd - 8 if @fake_exif field_offset = field_offset + 18 end end else field_offset = entry + 8 end if field_type == 2 # special case => null-terminated ASCII string # XXX investigate # sometimes gets too big to fit in int value if count != 0 and count < (2**31) @file.seek(@offset + field_offset) values = @file.read(count) # ascii string of count bytes #print values # Drop any garbage after a null. values = values.split(X00, 1)[0] # a null-terminated string else values = '' end else values = [] # @todo investigate # some entries get too big to handle could be malformed file if count < 1000 or tag_entry.name == 'MakerNote' @file.seek(@offset + field_offset) count.times { if field_type == 5 or field_type == 10 # a ratio value = Ratio.new(unpack_number(@file.read(4), signed), unpack_number(@file.read(4), signed) ) else value = unpack_number(@file.read(typelen), signed) end values << value } end end if (tag_id == 0x8769) puts "tag_id #{tag_id} field_type #{field_type} count #{count} values #{values.inspect}" end self.[ifd_name + ' ' + tag_entry.name] = IFD_Tag.new(tag_id, field_type, tag_entry, values, field_offset, count) end if tag_entry.name == stop_tag break end } end |
#extract_TIFF_thumbnail(thumb_ifd) ⇒ Object
extract uncompressed TIFF thumbnail (like pulling teeth) we take advantage of the pre-existing layout in the thumbnail IFD as much as possible
366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 366 def extract_TIFF_thumbnail(thumb_ifd) entries = self.s2n(thumb_ifd, 2) # this is header plus offset to IFD ... if @endian == 'M' tiff = "MM#{X00}*" + [0x00,0x00,0x00,0x00,0x08].pack('C*') else tiff = 'II*' + [0x00,0x08,0x00,0x00,0x00].pack('C*') end # ... plus thumbnail IFD data plus a null "next IFD" pointer self.file.seek(self.offset+thumb_ifd) tiff += self.file.read(entries*12+2)+[0x00,0x00,0x00,0x00].pack('C*') # fix up large value offset pointers into data area (0...entries).each { |i| entry = thumb_ifd + 2 + 12 * i tag = self.s2n(entry, 2) field_type = self.s2n(entry+2, 2) typelen = FIELD_TYPES[field_type].length count = self.s2n(entry+4, 4) oldoff = self.s2n(entry+8, 4) # start of the 4-byte pointer area in entry ptr = i * 12 + 18 # remember strip offsets location if tag == 0x0111 strip_off = ptr strip_len = count * typelen end # is it in the data area? if count * typelen > 4 # update offset pointer (nasty "strings are immutable" crap) # should be able to say "tiff[ptr..ptr+4]=newoff" newoff = len(tiff) tiff = tiff[ 0..ptr] + self.n2s(newoff, 4) + tiff[ptr+4...tiff.length] # remember strip offsets location if tag == 0x0111 strip_off = newoff strip_len = 4 end # get original data and store it self.file.seek(self.offset + oldoff) tiff += self.file.read(count * typelen) end } # add pixel strips and update strip offset info old_offsets = self.['Thumbnail StripOffsets'].values old_counts = self.['Thumbnail StripByteCounts'].values (0...len(old_offsets)).each { |i| # update offset pointer (more nasty "strings are immutable" crap) offset = self.n2s(len(tiff), strip_len) tiff = tiff[ 0..strip_off] + offset + tiff[strip_off + strip_len ... tiff.length] strip_off += strip_len # add pixel strip to end self.file.seek(self.offset + old_offsets[i]) tiff += self.file.read(old_counts[i]) } self.['TIFFThumbnail'] = tiff end |
#first_IFD ⇒ Object
return first IFD
238 239 240 241 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 238 def first_IFD() @file.seek(@offset + 4) return unpack_number(@file.read(4)) end |
#list_IFDs ⇒ Object
return list of IFDs in header
252 253 254 255 256 257 258 259 260 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 252 def list_IFDs() i=self.first_IFD() a=[] while i != 0 a << i i=self.next_IFD(i) end return a end |
#n2s(offset, length) ⇒ Object
convert offset to string
224 225 226 227 228 229 230 231 232 233 234 235 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 224 def n2s(offset, length) s = '' length.times { if @endian == 'I' s = s + chr(offset & 0xFF) else s = chr(offset & 0xFF) + s end offset = offset >> 8 } return s end |
#next_IFD(ifd) ⇒ Object
return pointer to next IFD
244 245 246 247 248 249 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 244 def next_IFD(ifd) @file.seek(@offset + ifd) entries = unpack_number(@file.read(2)) @file.seek((12*entries), IO::SEEK_CUR) return unpack_number(@file.read(4)) end |
#olympus_decode_tag(value, dict) ⇒ Object
XXX TODO decode Olympus MakerNote tag based on offset within tag
519 520 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 519 def olympus_decode_tag(value, dict) end |
#s2n(offset, length, signed = false) ⇒ Object
convert slice to integer, based on sign and endian flags usually this offset is assumed to be relative to the beginning of the start of the EXIF information. For some cameras that use relative tags, this offset may be relative to some other starting point.
212 213 214 215 216 217 218 219 220 221 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 212 def s2n(offset, length, signed=false) @file.seek(@offset+offset) if @file.eof? and length != 0 # raise "Read past EOF" puts "Read past EOF" return 0 end slice=@file.read(length) return unpack_number(slice,signed) end |
#unpack_number(src, signed = false) ⇒ Object
192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 |
# File 'lib/cul_image_props/image/properties/exif/types.rb', line 192 def unpack_number(src, signed=false) if @endian == 'I' val = EXIF_header.s2n_intel(src) else val = EXIF_header.s2n_motorola(src) end # Sign extension ? if signed msb= 1 << (8*src.length-1) if val & msb val=val-(msb << 1) end end return val end |