Module: Pdfrb::ImageLoader::TIFF

Defined in:
lib/pdfrb/image_loader/tiff.rb

Overview

TIFF image loader. Pure-Ruby support for:

* Format detection via the II/MM byte-order magic.
* Header + IFD parsing to extract width, height, bpc, sample
count, photometric interpretation, compression, strip
offsets + byte counts.
* Single-strip, uncompressed (compression == 1) RGB or
grayscale pixel decode into raw image bytes suitable for
embedding as a /FlateDecode image XObject.

For multi-strip, tiled, YCbCr, or compressed TIFFs (LZW, PackBits, CCITT, JPEG-in-TIFF), the loader emits a metadata-only stub XObject. Callers should pre-convert those to PNG/JPEG.

Constant Summary collapse

COMPRESSION_NONE = 
1
COMPRESSION_CCITT_RLE = 
2
COMPRESSION_CCITT_FAX3 = 
3
COMPRESSION_CCITT_FAX4 = 
4
COMPRESSION_LZW = 
5
COMPRESSION_OJPEG = 
6
COMPRESSION_JPEG = 
7
COMPRESSION_DEFLATE = 
8
COMPRESSION_PACKBITS = 
32773

Class Method Summary collapse

Class Method Details

.build_compressed_image(document, info, pixels) ⇒ Object 



136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 136

def build_compressed_image(document, info, pixels)
  compressed = Zlib.deflate(pixels)
  image = document.add(
    {
      Type: :XObject, Subtype: :Image,
      Width: info[:width], Height: info[:height],
      BitsPerComponent: 8,
      ColorSpace: info[:color_space],
      Filter: :FlateDecode,
      Length: compressed.bytesize
    },
    type: Pdfrb::Model::Type::XObjectImage
  )
  image.stream = compressed
  image
end

.build_stub_image(document, info) ⇒ Object 



153
154
155
156
157
158
159
160
161
162
163
164
165
166
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 153

def build_stub_image(document, info)
  image = document.add(
    {
      Type: :XObject, Subtype: :Image,
      Width: info[:width], Height: info[:height],
      BitsPerComponent: info[:bits_per_sample] || 8,
      ColorSpace: info[:color_space] || :DeviceRGB,
      Length: 0
    },
    type: Pdfrb::Model::Type::XObjectImage
  )
  image.stream = +""
  image
end

.call(document, data, **_opts) ⇒ Object 



33
34
35
36
37
38
39
40
41
42
43
44
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 33

def call(document, data, **_opts)
  data = data.read if data.is_a?(IO) || data.is_a?(StringIO)
  info = parse_header(data)
  return nil if info.empty?

  pixels = decode_pixels(data, info)
  if pixels
    build_compressed_image(document, info, pixels)
  else
    build_stub_image(document, info)
  end
end

.color_space_for_photometric(photometric) ⇒ Object 



246
247
248
249
250
251
252
253
254
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 246

def color_space_for_photometric(photometric)
  case photometric
  when 0, 1 then :DeviceGray
  when 5 then :DeviceCMYK
  # Default covers RGB (2), unknown photometric interpretations,
  # and nil photometric (which we treat as RGB by convention).
  else :DeviceRGB
  end
end

.decode_pixels(data, info) ⇒ Object

Decode pixel data when the TIFF is single-strip uncompressed RGB or grayscale. Returns binary pixel bytes or nil if the compression/photometric combo isn't supported here.



90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 90

def decode_pixels(data, info)
  return nil unless info[:compression] == COMPRESSION_NONE
  return nil unless info[:strip_offsets] && info[:strip_byte_counts]

  # Only handle 8-bit single-sample (gray) or 3-sample (RGB)
  # in photometric 0/1/2.
  return nil unless info[:bits_per_sample] == 8
  return nil unless [1, 3].include?(info[:samples_per_pixel])
  return nil unless [0, 1, 2].include?(info[:photometric])

  bytes = read_strips(data, info)
  return nil unless bytes

  # TIFF rows are padded to word boundaries; PDF image XObjects
  # don't need this padding. Strip per row.
  channels = info[:samples_per_pixel]
  row_bytes = info[:width] * channels
  return nil if row_bytes.zero?

  rows = info[:height]
  out = +"".b
  rows.times do |r|
    offset = r * row_bytes
    out << bytes.byteslice(offset, row_bytes)
  end
  out
end

.interpret_ifd(entries) ⇒ Object 



220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 220

def interpret_ifd(entries)
  width = entries[256]&.[](:value)
  height = entries[257]&.[](:value)
  bps = entries[258]&.[](:value) || 8
  bps = bps.first if bps.is_a?(::Array)
  spp = entries[277]&.[](:value) || 1
  spp = spp.first if spp.is_a?(::Array)
  photometric = entries[262]&.[](:value) || 1
  compression = entries[259]&.[](:value) || COMPRESSION_NONE
  strip_offsets = entries[273]&.[](:value)
  strip_byte_counts = entries[279]&.[](:value)
  rows_per_strip = entries[278]&.[](:value) || height
  {
    width: width,
    height: height,
    bits_per_sample: bps,
    samples_per_pixel: spp,
    photometric: photometric,
    compression: compression,
    strip_offsets: strip_offsets,
    strip_byte_counts: strip_byte_counts,
    rows_per_strip: rows_per_strip,
    color_space: color_space_for_photometric(photometric),
  }
end

.parse_header(data) ⇒ Object

Parse the TIFF header (8 bytes) + first IFD. Returns a Hash with width, height, bits_per_sample, samples_per_pixel, photometric, compression, strip_offsets, strip_byte_counts, rows_per_strip, color_space, or empty Hash if not TIFF.



50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 50

def parse_header(data)
  return {} unless data.is_a?(::String) && data.bytesize >= 8

  byte_order = data.byteslice(0, 2)
  case byte_order
  when "II" then little_endian = true
  when "MM" then little_endian = false
  else return {}
  end

  magic = read_u16(data, 2, little_endian)
  return {} unless [42, 43].include?(magic)
  return {} unless magic == 42

  ifd_offset = read_u32(data, 4, little_endian)
  parse_ifd(data, ifd_offset, little_endian)
end

.parse_ifd(data, offset, little_endian) ⇒ Object 



68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 68

def parse_ifd(data, offset, little_endian)
  return {} if offset.nil? || offset + 2 > data.bytesize

  count = read_u16(data, offset, little_endian)
  entries = {}
  count.times do |i|
    entry_off = offset + 2 + (i * 12)
    break if entry_off + 12 > data.bytesize

    tag = read_u16(data, entry_off, little_endian)
    type_id = read_u16(data, entry_off + 2, little_endian)
    count_val = read_u32(data, entry_off + 4, little_endian)
    value = read_ifd_value(data, entry_off + 8, type_id, count_val, little_endian)
    entries[tag] = { type: type_id, count: count_val, value: value }
  end

  interpret_ifd(entries)
end

.read_ifd_value(data, offset, type_id, count, little_endian) ⇒ Object 



168
169
170
171
172
173
174
175
176
177
178
179
180
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 168

def read_ifd_value(data, offset, type_id, count, little_endian)
  # For multi-value entries (count > 1 for SHORT/LONG), the
  # value field is a pointer to the actual data. For count <= 1
  # the value lives inline in the 4-byte value field.
  case type_id
  when 3 # SHORT
    count <= 1 ? read_u16(data, offset, little_endian) : read_short_array(data, offset, count, little_endian)
  when 4 # LONG
    count <= 1 ? read_u32(data, offset, little_endian) : read_long_array(data, offset, count, little_endian)
  else
    count <= 1 ? read_u16(data, offset, little_endian) : read_u32(data, offset, little_endian)
  end
end

.read_long_array(data, value_field_offset, count, little_endian) ⇒ Object 



193
194
195
196
197
198
199
200
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 193

def read_long_array(data, value_field_offset, count, little_endian)
  if count * 4 <= 4
    (0...count).map { |i| read_u32(data, value_field_offset + (i * 4), little_endian) }
  else
    ptr = read_u32(data, value_field_offset, little_endian)
    (0...count).map { |i| read_u32(data, ptr + (i * 4), little_endian) }
  end
end

.read_short_array(data, value_field_offset, count, little_endian) ⇒ Object 



182
183
184
185
186
187
188
189
190
191
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 182

def read_short_array(data, value_field_offset, count, little_endian)
  # If count * 2 <= 4, values live in the value field; otherwise
  # the field holds a pointer.
  if count * 2 <= 4
    (0...count).map { |i| read_u16(data, value_field_offset + (i * 2), little_endian) }
  else
    ptr = read_u32(data, value_field_offset, little_endian)
    (0...count).map { |i| read_u16(data, ptr + (i * 2), little_endian) }
  end
end

.read_strips(data, info) ⇒ Object

Read all strips concatenated. Single-strip case is most common for small images; multi-strip is supported when present.



120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 120

def read_strips(data, info)
  offsets = Array(info[:strip_offsets])
  counts = Array(info[:strip_byte_counts])
  return nil if offsets.length != counts.length

  bytes = +"".b
  offsets.each_with_index do |off, i|
    n = counts[i]
    piece = data.byteslice(off, n)
    return nil unless piece && piece.bytesize == n

    bytes << piece
  end
  bytes
end

.read_u16(data, offset, little_endian) ⇒ Object 



202
203
204
205
206
207
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 202

def read_u16(data, offset, little_endian)
  bytes = data.bytes[offset, 2]
  return 0 unless bytes && bytes.length == 2

  little_endian ? ((bytes[1] << 8) | bytes[0]) : ((bytes[0] << 8) | bytes[1])
end

.read_u32(data, offset, little_endian) ⇒ Object 



209
210
211
212
213
214
215
216
217
218
 
# File 'lib/pdfrb/image_loader/tiff.rb', line 209

def read_u32(data, offset, little_endian)
  bytes = data.bytes[offset, 4]
  return 0 unless bytes && bytes.length == 4

  if little_endian
    (bytes[3] << 24) | (bytes[2] << 16) | (bytes[1] << 8) | bytes[0]
  else
    (bytes[0] << 24) | (bytes[1] << 16) | (bytes[2] << 8) | bytes[3]
  end
end