format_parser 1.6.0 → 1.7.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +6 -0
- data/README.md +1 -1
- data/lib/format_parser/version.rb +1 -1
- data/lib/parsers/arw_parser.rb +50 -0
- data/lib/parsers/tiff_parser.rb +5 -6
- data/spec/parsers/arw_parser_spec.rb +119 -0
- data/spec/parsers/tiff_parser_spec.rb +9 -15
- metadata +8 -6
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 8b3cef665ae16efd68e8da952fd4656e2d9403f3899bd58839da3d8026db91f4
|
4
|
+
data.tar.gz: 7b0ec88efc2ea62f526699a4041cb3f1b3062994d2a6e0b24c2cfdf247aaf532
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 24c6379ef4fd3b5a9f061c6fc40fd8c0498ad33213684d08dd27a8b8994ba40a98bf1fa18a6d6b3b8189aa71436ec9bb394e3b8d41a8dd3ca90a5b93d0f1718a
|
7
|
+
data.tar.gz: a2d3df2c17d2559aa99f52f04624032a9243915f2a1b28a6f3626bd3b9112eb8c325b0c9a286864d25c2b4e92a44a8939448a85d1004ac5d48c2f81f747749c1
|
data/CHANGELOG.md
CHANGED
data/README.md
CHANGED
@@ -194,7 +194,7 @@ Unless specified otherwise in this section the fixture files are MIT licensed an
|
|
194
194
|
- `IMG_9266_*.tif` and all it's variations were created by the project maintainers
|
195
195
|
|
196
196
|
### ARW
|
197
|
-
- ARW
|
197
|
+
- ARW examples are downloaded from http://www.rawsamples.ch/ and are Creative Common Licensed.
|
198
198
|
|
199
199
|
### ZIP
|
200
200
|
- The .zip fixture files have been created by the project maintainers
|
@@ -0,0 +1,50 @@
|
|
1
|
+
require_relative 'exif_parser'
|
2
|
+
|
3
|
+
class FormatParser::ARWParser
|
4
|
+
include FormatParser::IOUtils
|
5
|
+
include FormatParser::EXIFParser
|
6
|
+
|
7
|
+
# Standard TIFF headers
|
8
|
+
MAGIC_LE = [0x49, 0x49, 0x2A, 0x0].pack('C4')
|
9
|
+
MAGIC_BE = [0x4D, 0x4D, 0x0, 0x2A].pack('C4')
|
10
|
+
HEADER_BYTES = [MAGIC_LE, MAGIC_BE]
|
11
|
+
ARW_MIME_TYPE = 'image/x-sony-arw'
|
12
|
+
|
13
|
+
def likely_match?(filename)
|
14
|
+
filename =~ /\.arw$/i
|
15
|
+
end
|
16
|
+
|
17
|
+
def call(io)
|
18
|
+
io = FormatParser::IOConstraint.new(io)
|
19
|
+
|
20
|
+
return unless HEADER_BYTES.include?(safe_read(io, 4))
|
21
|
+
exif_data = exif_from_tiff_io(io)
|
22
|
+
|
23
|
+
return unless valid?(exif_data)
|
24
|
+
|
25
|
+
w = exif_data.width || exif_data.pixel_x_dimension
|
26
|
+
h = exif_data.height || exif_data.pixel_y_dimension
|
27
|
+
|
28
|
+
FormatParser::Image.new(
|
29
|
+
format: :arw,
|
30
|
+
width_px: w,
|
31
|
+
height_px: h,
|
32
|
+
display_width_px: exif_data.rotated? ? h : w,
|
33
|
+
display_height_px: exif_data.rotated? ? w : h,
|
34
|
+
orientation: exif_data.orientation_sym,
|
35
|
+
intrinsics: { exif: exif_data },
|
36
|
+
content_type: ARW_MIME_TYPE,
|
37
|
+
)
|
38
|
+
rescue EXIFR::MalformedTIFF
|
39
|
+
nil
|
40
|
+
end
|
41
|
+
|
42
|
+
def valid?(exif_data)
|
43
|
+
# taken directly from tiff_parser.rb
|
44
|
+
# Similar to how exiftool determines the image type as ARW, we are implementing a check here
|
45
|
+
# https://github.com/exiftool/exiftool/blob/e969456372fbaf4b980fea8bb094d71033ac8bf7/lib/Image/ExifTool/Exif.pm#L929
|
46
|
+
exif_data.compression == 6 && exif_data.new_subfile_type == 1 && exif_data.make&.start_with?('SONY')
|
47
|
+
end
|
48
|
+
|
49
|
+
FormatParser.register_parser new, natures: :image, formats: :arw
|
50
|
+
end
|
data/lib/parsers/tiff_parser.rb
CHANGED
@@ -6,7 +6,6 @@ class FormatParser::TIFFParser
|
|
6
6
|
MAGIC_BE = [0x4D, 0x4D, 0x0, 0x2A].pack('C4')
|
7
7
|
HEADER_BYTES = [MAGIC_LE, MAGIC_BE]
|
8
8
|
TIFF_MIME_TYPE = 'image/tiff'
|
9
|
-
ARW_MIME_TYPE = 'image/x-sony-arw'
|
10
9
|
|
11
10
|
def likely_match?(filename)
|
12
11
|
filename =~ /\.tiff?$/i
|
@@ -28,20 +27,20 @@ class FormatParser::TIFFParser
|
|
28
27
|
exif_data = exif_from_tiff_io(io)
|
29
28
|
return unless exif_data
|
30
29
|
|
30
|
+
return if arw?(exif_data)
|
31
|
+
|
31
32
|
w = exif_data.width || exif_data.pixel_x_dimension
|
32
33
|
h = exif_data.height || exif_data.pixel_y_dimension
|
33
34
|
|
34
|
-
format = arw?(exif_data) ? :arw : :tif
|
35
|
-
mime_type = arw?(exif_data) ? ARW_MIME_TYPE : TIFF_MIME_TYPE
|
36
35
|
FormatParser::Image.new(
|
37
|
-
format:
|
36
|
+
format: :tif,
|
38
37
|
width_px: w,
|
39
38
|
height_px: h,
|
40
39
|
display_width_px: exif_data.rotated? ? h : w,
|
41
40
|
display_height_px: exif_data.rotated? ? w : h,
|
42
41
|
orientation: exif_data.orientation_sym,
|
43
42
|
intrinsics: {exif: exif_data},
|
44
|
-
content_type:
|
43
|
+
content_type: TIFF_MIME_TYPE,
|
45
44
|
)
|
46
45
|
rescue EXIFR::MalformedTIFF
|
47
46
|
nil
|
@@ -55,7 +54,7 @@ class FormatParser::TIFFParser
|
|
55
54
|
# Similar to how exiftool determines the image type as ARW, we are implementing a check here
|
56
55
|
# https://github.com/exiftool/exiftool/blob/e969456372fbaf4b980fea8bb094d71033ac8bf7/lib/Image/ExifTool/Exif.pm#L929
|
57
56
|
def arw?(exif_data)
|
58
|
-
exif_data.compression == 6 && exif_data.new_subfile_type == 1 && exif_data.make
|
57
|
+
exif_data.compression == 6 && exif_data.new_subfile_type == 1 && exif_data.make&.start_with?('SONY')
|
59
58
|
end
|
60
59
|
|
61
60
|
FormatParser.register_parser new, natures: :image, formats: :tif
|
@@ -0,0 +1,119 @@
|
|
1
|
+
require 'spec_helper'
|
2
|
+
|
3
|
+
describe FormatParser::ARWParser do
|
4
|
+
shared_examples 'likely_match for file' do |filename_with_extension|
|
5
|
+
it "matches '#{filename_with_extension}'" do
|
6
|
+
expect(subject.likely_match?(filename_with_extension)).to be_truthy
|
7
|
+
end
|
8
|
+
end
|
9
|
+
|
10
|
+
shared_examples 'no likely_match for file' do |filename_with_extension|
|
11
|
+
it "does not match '#{filename_with_extension}'" do
|
12
|
+
expect(subject.likely_match?(filename_with_extension)).to be_falsey
|
13
|
+
end
|
14
|
+
end
|
15
|
+
|
16
|
+
describe 'likely_match' do
|
17
|
+
filenames = ['raw_file', 'another raw file', 'and.another', 'one-more']
|
18
|
+
valid_extensions = ['.arw', '.Arw', '.aRw', '.arW', '.ARw', '.ArW', '.aRW', '.ARW']
|
19
|
+
invalid_extensions = ['.tiff', '.cr2', '.new', '.jpeg']
|
20
|
+
filenames.each do |filename|
|
21
|
+
valid_extensions.each do |extension|
|
22
|
+
include_examples 'likely_match for file', filename + extension
|
23
|
+
end
|
24
|
+
invalid_extensions.each do |extension|
|
25
|
+
include_examples 'no likely_match for file', filename + extension
|
26
|
+
end
|
27
|
+
end
|
28
|
+
end
|
29
|
+
|
30
|
+
describe 'parses Sony ARW fixtures as arw format file' do
|
31
|
+
expected_parsed_dimensions = {
|
32
|
+
'RAW_SONY_A100.ARW' => {
|
33
|
+
width_px: 3872,
|
34
|
+
height_px: 2592,
|
35
|
+
display_width_px: 3872,
|
36
|
+
display_height_px: 2592,
|
37
|
+
orientation: :top_left
|
38
|
+
},
|
39
|
+
'RAW_SONY_A700.ARW' => {
|
40
|
+
width_px: 4288,
|
41
|
+
height_px: 2856,
|
42
|
+
display_width_px: 4288,
|
43
|
+
display_height_px: 2856,
|
44
|
+
orientation: :top_left
|
45
|
+
},
|
46
|
+
'RAW_SONY_A900.ARW' => {
|
47
|
+
width_px: 6080,
|
48
|
+
height_px: 4048,
|
49
|
+
display_width_px: 6080,
|
50
|
+
display_height_px: 4048,
|
51
|
+
orientation: :top_left
|
52
|
+
},
|
53
|
+
# rotated 90 degree image
|
54
|
+
'RAW_SONY_DSC-RX100M2.ARW' => {
|
55
|
+
width_px: 5472,
|
56
|
+
height_px: 3648,
|
57
|
+
display_width_px: 3648,
|
58
|
+
display_height_px: 5472,
|
59
|
+
orientation: :right_top,
|
60
|
+
},
|
61
|
+
'RAW_SONY_ILCE-7RM2.ARW' => {
|
62
|
+
width_px: 7952,
|
63
|
+
height_px: 5304,
|
64
|
+
display_width_px: 7952,
|
65
|
+
display_height_px: 5304,
|
66
|
+
orientation: :top_left,
|
67
|
+
},
|
68
|
+
'RAW_SONY_NEX7.ARW' => {
|
69
|
+
width_px: 6000,
|
70
|
+
height_px: 4000,
|
71
|
+
display_width_px: 6000,
|
72
|
+
display_height_px: 4000,
|
73
|
+
orientation: :top_left,
|
74
|
+
},
|
75
|
+
'RAW_SONY_SLTA55V.ARW' => {
|
76
|
+
width_px: 4928,
|
77
|
+
height_px: 3280,
|
78
|
+
display_width_px: 4928,
|
79
|
+
display_height_px: 3280,
|
80
|
+
orientation: :top_left,
|
81
|
+
},
|
82
|
+
}
|
83
|
+
|
84
|
+
Dir.glob(fixtures_dir + '/ARW/*.ARW').each do |arw_path|
|
85
|
+
it "is able to parse #{File.basename(arw_path)}" do
|
86
|
+
expected_dimension = expected_parsed_dimensions[File.basename(arw_path)]
|
87
|
+
# error if a new .arw test file is added without specifying the expected dimensions
|
88
|
+
expect(expected_dimension).not_to be_nil
|
89
|
+
|
90
|
+
parsed = subject.call(File.open(arw_path, 'rb'))
|
91
|
+
expect(parsed).not_to be_nil
|
92
|
+
expect(parsed.nature).to eq(:image)
|
93
|
+
expect(parsed.format).to eq(:arw)
|
94
|
+
expect(parsed.intrinsics[:exif]).not_to be_nil
|
95
|
+
expect(parsed.content_type).to eq('image/x-sony-arw')
|
96
|
+
|
97
|
+
expect(parsed.width_px).to eq(expected_dimension[:width_px])
|
98
|
+
expect(parsed.height_px).to eq(expected_dimension[:height_px])
|
99
|
+
expect(parsed.display_width_px).to eq(expected_dimension[:display_width_px])
|
100
|
+
expect(parsed.display_height_px).to eq(expected_dimension[:display_height_px])
|
101
|
+
expect(parsed.orientation).to eq(expected_dimension[:orientation])
|
102
|
+
end
|
103
|
+
end
|
104
|
+
|
105
|
+
shared_examples 'invalid filetype' do |filetype, fixture_path|
|
106
|
+
it "should fail to parse #{filetype}" do
|
107
|
+
file_path = fixtures_dir + fixture_path
|
108
|
+
parsed = subject.call(File.open(file_path, 'rb'))
|
109
|
+
expect(parsed).to be_nil
|
110
|
+
end
|
111
|
+
end
|
112
|
+
|
113
|
+
include_examples 'invalid filetype', 'NEF', '/NEF/RAW_NIKON_1S2.NEF'
|
114
|
+
include_examples 'invalid filetype', 'TIFF', '/TIFF/Shinbutsureijoushuincho.tiff'
|
115
|
+
include_examples 'invalid filetype', 'JPG', '/JPEG/orient_6.jpg'
|
116
|
+
include_examples 'invalid filetype', 'PNG', '/PNG/cat.png'
|
117
|
+
include_examples 'invalid filetype', 'CR2', '/CR2/RAW_CANON_1DM2.CR2'
|
118
|
+
end
|
119
|
+
end
|
@@ -47,21 +47,6 @@ describe FormatParser::TIFFParser do
|
|
47
47
|
expect(parsed.intrinsics[:exif]).not_to be_nil
|
48
48
|
end
|
49
49
|
|
50
|
-
it 'parses Sony ARW fixture as arw format file' do
|
51
|
-
arw_path = fixtures_dir + '/ARW/RAW_SONY_ILCE-7RM2.ARW'
|
52
|
-
|
53
|
-
parsed = subject.call(File.open(arw_path, 'rb'))
|
54
|
-
|
55
|
-
expect(parsed).not_to be_nil
|
56
|
-
expect(parsed.nature).to eq(:image)
|
57
|
-
expect(parsed.format).to eq(:arw)
|
58
|
-
|
59
|
-
expect(parsed.width_px).to eq(7952)
|
60
|
-
expect(parsed.height_px).to eq(5304)
|
61
|
-
expect(parsed.intrinsics[:exif]).not_to be_nil
|
62
|
-
expect(parsed.content_type).to eq('image/x-sony-arw')
|
63
|
-
end
|
64
|
-
|
65
50
|
describe 'correctly extracts dimensions from various TIFF flavors of the same file' do
|
66
51
|
Dir.glob(fixtures_dir + '/TIFF/IMG_9266*.tif').each do |tiff_path|
|
67
52
|
it "is able to parse #{File.basename(tiff_path)}" do
|
@@ -100,4 +85,13 @@ describe FormatParser::TIFFParser do
|
|
100
85
|
end
|
101
86
|
end
|
102
87
|
end
|
88
|
+
|
89
|
+
describe 'bails out on ARW files, such as' do
|
90
|
+
Dir.glob(fixtures_dir + '/ARW/*.ARW').each do |arw_path|
|
91
|
+
it "skips #{File.basename(arw_path)}" do
|
92
|
+
parsed = subject.call(File.open(arw_path, 'rb'))
|
93
|
+
expect(parsed).to be_nil
|
94
|
+
end
|
95
|
+
end
|
96
|
+
end
|
103
97
|
end
|
metadata
CHANGED
@@ -1,15 +1,15 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: format_parser
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 1.
|
4
|
+
version: 1.7.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Noah Berman
|
8
8
|
- Julik Tarkhanov
|
9
|
-
autorequire:
|
9
|
+
autorequire:
|
10
10
|
bindir: exe
|
11
11
|
cert_chain: []
|
12
|
-
date: 2022-
|
12
|
+
date: 2022-10-05 00:00:00.000000000 Z
|
13
13
|
dependencies:
|
14
14
|
- !ruby/object:Gem::Dependency
|
15
15
|
name: ks
|
@@ -232,6 +232,7 @@ files:
|
|
232
232
|
- lib/parsers/aac_parser.rb
|
233
233
|
- lib/parsers/aac_parser/adts_header_info.rb
|
234
234
|
- lib/parsers/aiff_parser.rb
|
235
|
+
- lib/parsers/arw_parser.rb
|
235
236
|
- lib/parsers/bmp_parser.rb
|
236
237
|
- lib/parsers/cr2_parser.rb
|
237
238
|
- lib/parsers/dpx_parser.rb
|
@@ -278,6 +279,7 @@ files:
|
|
278
279
|
- spec/parsers/aac_parser_spec.rb
|
279
280
|
- spec/parsers/adts_header_info_spec.rb
|
280
281
|
- spec/parsers/aiff_parser_spec.rb
|
282
|
+
- spec/parsers/arw_parser_spec.rb
|
281
283
|
- spec/parsers/bmp_parser_spec.rb
|
282
284
|
- spec/parsers/cr2_parser_spec.rb
|
283
285
|
- spec/parsers/dpx_parser_spec.rb
|
@@ -310,7 +312,7 @@ licenses:
|
|
310
312
|
- MIT (Hippocratic)
|
311
313
|
metadata:
|
312
314
|
allowed_push_host: https://rubygems.org
|
313
|
-
post_install_message:
|
315
|
+
post_install_message:
|
314
316
|
rdoc_options: []
|
315
317
|
require_paths:
|
316
318
|
- lib
|
@@ -325,8 +327,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
325
327
|
- !ruby/object:Gem::Version
|
326
328
|
version: '0'
|
327
329
|
requirements: []
|
328
|
-
rubygems_version: 3.
|
329
|
-
signing_key:
|
330
|
+
rubygems_version: 3.3.7
|
331
|
+
signing_key:
|
330
332
|
specification_version: 4
|
331
333
|
summary: A library for efficient parsing of file metadata
|
332
334
|
test_files: []
|