chupa-text 1.1.3 → 1.1.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/doc/text/news.md +12 -0
- data/lib/chupa-text/command/chupa-text.rb +7 -1
- data/lib/chupa-text/decomposer.rb +8 -0
- data/lib/chupa-text/decomposers/office-open-xml-document.rb +51 -0
- data/lib/chupa-text/decomposers/office-open-xml-presentation.rb +67 -0
- data/lib/chupa-text/decomposers/office-open-xml-workbook.rb +114 -0
- data/lib/chupa-text/decomposers/office-open-xml.rb +196 -0
- data/lib/chupa-text/decomposers/opendocument-presentation.rb +105 -0
- data/lib/chupa-text/decomposers/opendocument-spreadsheet.rb +134 -0
- data/lib/chupa-text/decomposers/opendocument-text.rb +89 -0
- data/lib/chupa-text/decomposers/opendocument.rb +139 -0
- data/lib/chupa-text/extractor.rb +8 -2
- data/lib/chupa-text/formatters/mime.rb +3 -2
- data/lib/chupa-text/version.rb +1 -1
- data/test/decomposers/test-office-open-xml-document.rb +144 -0
- data/test/decomposers/test-office-open-xml-presentation.rb +133 -0
- data/test/decomposers/test-office-open-xml-workbook.rb +138 -0
- data/test/decomposers/test-open-document-presentation.rb +136 -0
- data/test/decomposers/test-open-document-spreadsheet.rb +152 -0
- data/test/decomposers/test-open-document-text.rb +144 -0
- data/test/fixture/docx/attributes.docx +0 -0
- data/test/fixture/docx/multi-pages.docx +0 -0
- data/test/fixture/docx/one-page.docx +0 -0
- data/test/fixture/docx/special-characters.docx +0 -0
- data/test/fixture/odp/attributes.odp +0 -0
- data/test/fixture/odp/multi-slides.odp +0 -0
- data/test/fixture/odp/one-slide.odp +0 -0
- data/test/fixture/ods/attributes.ods +0 -0
- data/test/fixture/ods/multi-sheets.ods +0 -0
- data/test/fixture/ods/one-sheet.ods +0 -0
- data/test/fixture/odt/attributes.odt +0 -0
- data/test/fixture/odt/multi-pages.odt +0 -0
- data/test/fixture/odt/one-page.odt +0 -0
- data/test/fixture/odt/special-characters.odt +0 -0
- data/test/fixture/pptx/attributes.pptx +0 -0
- data/test/fixture/pptx/multi-slides.pptx +0 -0
- data/test/fixture/pptx/one-slide.pptx +0 -0
- data/test/fixture/xlsx/attributes.xlsx +0 -0
- data/test/fixture/xlsx/multi-sheets.xlsx +0 -0
- data/test/fixture/xlsx/one-sheet.xlsx +0 -0
- metadata +36 -2
@@ -0,0 +1,136 @@
|
|
1
|
+
# Copyright (C) 2019 Kouhei Sutou <kou@clear-code.com>
|
2
|
+
#
|
3
|
+
# This library is free software; you can redistribute it and/or
|
4
|
+
# modify it under the terms of the GNU Lesser General Public
|
5
|
+
# License as published by the Free Software Foundation; either
|
6
|
+
# version 2.1 of the License, or (at your option) any later version.
|
7
|
+
#
|
8
|
+
# This library is distributed in the hope that it will be useful,
|
9
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
10
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
11
|
+
# Lesser General Public License for more details.
|
12
|
+
#
|
13
|
+
# You should have received a copy of the GNU Lesser General Public
|
14
|
+
# License along with this library; if not, write to the Free Software
|
15
|
+
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
16
|
+
|
17
|
+
class TestDecomposersOpenDocumentPresentation < Test::Unit::TestCase
|
18
|
+
include Helper
|
19
|
+
|
20
|
+
def setup
|
21
|
+
@decomposer = ChupaText::Decomposers::OpenDocumentPresentation.new({})
|
22
|
+
end
|
23
|
+
|
24
|
+
def decompose(path)
|
25
|
+
data = ChupaText::InputData.new(path)
|
26
|
+
decomposed = []
|
27
|
+
@decomposer.decompose(data) do |decomposed_data|
|
28
|
+
decomposed << decomposed_data
|
29
|
+
end
|
30
|
+
decomposed
|
31
|
+
end
|
32
|
+
|
33
|
+
sub_test_case("#target_score") do
|
34
|
+
def test_extension
|
35
|
+
data = ChupaText::Data.new
|
36
|
+
data.body = ""
|
37
|
+
data.uri = "document.odp"
|
38
|
+
assert_equal(-1, @decomposer.target_score(data))
|
39
|
+
end
|
40
|
+
|
41
|
+
def test_mime_type
|
42
|
+
data = ChupaText::Data.new
|
43
|
+
data.mime_type = "application/vnd.oasis.opendocument.presentation"
|
44
|
+
assert_equal(-1, @decomposer.target_score(data))
|
45
|
+
end
|
46
|
+
end
|
47
|
+
|
48
|
+
sub_test_case("#decompose") do
|
49
|
+
sub_test_case("attributes") do
|
50
|
+
def decompose(attribute_name)
|
51
|
+
super(fixture_path("odp", "attributes.odp")).first[attribute_name]
|
52
|
+
end
|
53
|
+
|
54
|
+
def test_title
|
55
|
+
assert_equal("Title", decompose("title"))
|
56
|
+
end
|
57
|
+
|
58
|
+
def test_subject
|
59
|
+
assert_equal("Subject", decompose("subject"))
|
60
|
+
end
|
61
|
+
|
62
|
+
def test_keywords
|
63
|
+
assert_equal(["Keyword1", "Keyword2"], decompose("keywords"))
|
64
|
+
end
|
65
|
+
|
66
|
+
def test_created_time
|
67
|
+
assert_equal(Time,
|
68
|
+
decompose("created_time").class)
|
69
|
+
end
|
70
|
+
|
71
|
+
def test_modified_time
|
72
|
+
assert_equal(Time,
|
73
|
+
decompose("modified_time").class)
|
74
|
+
end
|
75
|
+
|
76
|
+
def test_generator
|
77
|
+
assert_equal("LibreOffice",
|
78
|
+
normalize_generator(decompose("generator")))
|
79
|
+
end
|
80
|
+
|
81
|
+
def normalize_generator(generator)
|
82
|
+
if generator.start_with?("LibreOffice")
|
83
|
+
"LibreOffice"
|
84
|
+
else
|
85
|
+
generator
|
86
|
+
end
|
87
|
+
end
|
88
|
+
end
|
89
|
+
|
90
|
+
sub_test_case("one slide") do
|
91
|
+
def decompose
|
92
|
+
super(fixture_path("odp", "one-slide.odp"))
|
93
|
+
end
|
94
|
+
|
95
|
+
def test_body
|
96
|
+
assert_equal([
|
97
|
+
[nil, ""],
|
98
|
+
[
|
99
|
+
0,
|
100
|
+
"Slide1 title\n" +
|
101
|
+
"Slide1 content\n",
|
102
|
+
],
|
103
|
+
],
|
104
|
+
decompose.collect {|data| [data["index"], data.body]})
|
105
|
+
end
|
106
|
+
end
|
107
|
+
|
108
|
+
sub_test_case("multi slides") do
|
109
|
+
def decompose
|
110
|
+
super(fixture_path("odp", "multi-slides.odp"))
|
111
|
+
end
|
112
|
+
|
113
|
+
def test_body
|
114
|
+
assert_equal([
|
115
|
+
[nil, ""],
|
116
|
+
[
|
117
|
+
0,
|
118
|
+
"Slide1 title\n" +
|
119
|
+
"Slide1 content\n",
|
120
|
+
],
|
121
|
+
[
|
122
|
+
1,
|
123
|
+
"Slide2 title\n" +
|
124
|
+
"Slide2 content\n",
|
125
|
+
],
|
126
|
+
[
|
127
|
+
2,
|
128
|
+
"Slide3 title\n" +
|
129
|
+
"Slide3 content\n",
|
130
|
+
],
|
131
|
+
],
|
132
|
+
decompose.collect {|data| [data["index"], data.body]})
|
133
|
+
end
|
134
|
+
end
|
135
|
+
end
|
136
|
+
end
|
@@ -0,0 +1,152 @@
|
|
1
|
+
# Copyright (C) 2019 Kouhei Sutou <kou@clear-code.com>
|
2
|
+
#
|
3
|
+
# This library is free software; you can redistribute it and/or
|
4
|
+
# modify it under the terms of the GNU Lesser General Public
|
5
|
+
# License as published by the Free Software Foundation; either
|
6
|
+
# version 2.1 of the License, or (at your option) any later version.
|
7
|
+
#
|
8
|
+
# This library is distributed in the hope that it will be useful,
|
9
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
10
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
11
|
+
# Lesser General Public License for more details.
|
12
|
+
#
|
13
|
+
# You should have received a copy of the GNU Lesser General Public
|
14
|
+
# License along with this library; if not, write to the Free Software
|
15
|
+
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
16
|
+
|
17
|
+
class TestDecomposersOpenDocumentSpreadsheet < Test::Unit::TestCase
|
18
|
+
include Helper
|
19
|
+
|
20
|
+
def setup
|
21
|
+
@decomposer = ChupaText::Decomposers::OpenDocumentSpreadsheet.new({})
|
22
|
+
end
|
23
|
+
|
24
|
+
def decompose(path)
|
25
|
+
data = ChupaText::InputData.new(path)
|
26
|
+
decomposed = []
|
27
|
+
@decomposer.decompose(data) do |decomposed_data|
|
28
|
+
decomposed << decomposed_data
|
29
|
+
end
|
30
|
+
decomposed
|
31
|
+
end
|
32
|
+
|
33
|
+
sub_test_case("#target_score") do
|
34
|
+
def test_extension
|
35
|
+
data = ChupaText::Data.new
|
36
|
+
data.body = ""
|
37
|
+
data.uri = "document.ods"
|
38
|
+
assert_equal(-1, @decomposer.target_score(data))
|
39
|
+
end
|
40
|
+
|
41
|
+
def test_mime_type
|
42
|
+
data = ChupaText::Data.new
|
43
|
+
data.mime_type = "application/vnd.oasis.opendocument.spreadsheet"
|
44
|
+
assert_equal(-1, @decomposer.target_score(data))
|
45
|
+
end
|
46
|
+
end
|
47
|
+
|
48
|
+
sub_test_case("#decompose") do
|
49
|
+
sub_test_case("attributes") do
|
50
|
+
def decompose(attribute_name)
|
51
|
+
super(fixture_path("ods", "attributes.ods")).first[attribute_name]
|
52
|
+
end
|
53
|
+
|
54
|
+
def test_title
|
55
|
+
assert_equal("Title", decompose("title"))
|
56
|
+
end
|
57
|
+
|
58
|
+
def test_subject
|
59
|
+
assert_equal("Subject", decompose("subject"))
|
60
|
+
end
|
61
|
+
|
62
|
+
def test_keywords
|
63
|
+
assert_equal(["Keyword1", "Keyword2"], decompose("keywords"))
|
64
|
+
end
|
65
|
+
|
66
|
+
def test_created_time
|
67
|
+
assert_equal(Time,
|
68
|
+
decompose("created_time").class)
|
69
|
+
end
|
70
|
+
|
71
|
+
def test_modified_time
|
72
|
+
assert_equal(Time,
|
73
|
+
decompose("modified_time").class)
|
74
|
+
end
|
75
|
+
|
76
|
+
def test_generator
|
77
|
+
assert_equal("LibreOffice",
|
78
|
+
normalize_generator(decompose("generator")))
|
79
|
+
end
|
80
|
+
|
81
|
+
def normalize_generator(generator)
|
82
|
+
if generator.start_with?("LibreOffice")
|
83
|
+
"LibreOffice"
|
84
|
+
else
|
85
|
+
generator
|
86
|
+
end
|
87
|
+
end
|
88
|
+
end
|
89
|
+
|
90
|
+
sub_test_case("one sheet") do
|
91
|
+
def decompose
|
92
|
+
super(fixture_path("ods", "one-sheet.ods")).collect do |data|
|
93
|
+
[
|
94
|
+
data["index"],
|
95
|
+
data["name"],
|
96
|
+
data.body,
|
97
|
+
]
|
98
|
+
end
|
99
|
+
end
|
100
|
+
|
101
|
+
def test_body
|
102
|
+
assert_equal([
|
103
|
+
[nil, nil, ""],
|
104
|
+
[
|
105
|
+
0,
|
106
|
+
"Sheet1",
|
107
|
+
"Sheet1 - A1\tSheet1 - B1\n" +
|
108
|
+
"Sheet1 - A2\tSheet1 - B2\n",
|
109
|
+
],
|
110
|
+
],
|
111
|
+
decompose)
|
112
|
+
end
|
113
|
+
end
|
114
|
+
|
115
|
+
sub_test_case("multi sheets") do
|
116
|
+
def decompose
|
117
|
+
super(fixture_path("ods", "multi-sheets.ods")).collect do |data|
|
118
|
+
[
|
119
|
+
data["index"],
|
120
|
+
data["name"],
|
121
|
+
data.body,
|
122
|
+
]
|
123
|
+
end
|
124
|
+
end
|
125
|
+
|
126
|
+
def test_body
|
127
|
+
assert_equal([
|
128
|
+
[nil, nil, ""],
|
129
|
+
[
|
130
|
+
0,
|
131
|
+
"Sheet1",
|
132
|
+
"Sheet1 - A1\tSheet1 - B1\n" +
|
133
|
+
"Sheet1 - A2\tSheet1 - B2\n",
|
134
|
+
],
|
135
|
+
[
|
136
|
+
1,
|
137
|
+
"Sheet2",
|
138
|
+
"Sheet2 - A1\tSheet2 - B1\n" +
|
139
|
+
"Sheet2 - A2\tSheet2 - B2\n",
|
140
|
+
],
|
141
|
+
[
|
142
|
+
2,
|
143
|
+
"Sheet3",
|
144
|
+
"Sheet3 - A1\tSheet3 - B1\n" +
|
145
|
+
"Sheet3 - A2\tSheet3 - B2\n",
|
146
|
+
],
|
147
|
+
],
|
148
|
+
decompose)
|
149
|
+
end
|
150
|
+
end
|
151
|
+
end
|
152
|
+
end
|
@@ -0,0 +1,144 @@
|
|
1
|
+
# Copyright (C) 2019 Kouhei Sutou <kou@clear-code.com>
|
2
|
+
#
|
3
|
+
# This library is free software; you can redistribute it and/or
|
4
|
+
# modify it under the terms of the GNU Lesser General Public
|
5
|
+
# License as published by the Free Software Foundation; either
|
6
|
+
# version 2.1 of the License, or (at your option) any later version.
|
7
|
+
#
|
8
|
+
# This library is distributed in the hope that it will be useful,
|
9
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
10
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
11
|
+
# Lesser General Public License for more details.
|
12
|
+
#
|
13
|
+
# You should have received a copy of the GNU Lesser General Public
|
14
|
+
# License along with this library; if not, write to the Free Software
|
15
|
+
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
16
|
+
|
17
|
+
class TestDecomposersOpenDocumentText < Test::Unit::TestCase
|
18
|
+
include Helper
|
19
|
+
|
20
|
+
def setup
|
21
|
+
@decomposer = ChupaText::Decomposers::OpenDocumentText.new({})
|
22
|
+
end
|
23
|
+
|
24
|
+
def decompose(path)
|
25
|
+
data = ChupaText::InputData.new(path)
|
26
|
+
decomposed = []
|
27
|
+
@decomposer.decompose(data) do |decomposed_data|
|
28
|
+
decomposed << decomposed_data
|
29
|
+
end
|
30
|
+
decomposed
|
31
|
+
end
|
32
|
+
|
33
|
+
sub_test_case("#target_score") do
|
34
|
+
def test_extension
|
35
|
+
data = ChupaText::Data.new
|
36
|
+
data.body = ""
|
37
|
+
data.uri = "document.odt"
|
38
|
+
assert_equal(-1, @decomposer.target_score(data))
|
39
|
+
end
|
40
|
+
|
41
|
+
def test_mime_type
|
42
|
+
data = ChupaText::Data.new
|
43
|
+
data.mime_type = "application/vnd.oasis.opendocument.text"
|
44
|
+
assert_equal(-1, @decomposer.target_score(data))
|
45
|
+
end
|
46
|
+
end
|
47
|
+
|
48
|
+
sub_test_case("#decompose") do
|
49
|
+
sub_test_case("attributes") do
|
50
|
+
def decompose(attribute_name)
|
51
|
+
super(fixture_path("odt", "attributes.odt")).collect do |data|
|
52
|
+
data[attribute_name]
|
53
|
+
end
|
54
|
+
end
|
55
|
+
|
56
|
+
def test_title
|
57
|
+
assert_equal(["Title"], decompose("title"))
|
58
|
+
end
|
59
|
+
|
60
|
+
def test_author
|
61
|
+
assert_equal([nil], decompose("author"))
|
62
|
+
end
|
63
|
+
|
64
|
+
def test_subject
|
65
|
+
assert_equal(["Subject"], decompose("subject"))
|
66
|
+
end
|
67
|
+
|
68
|
+
def test_keywords
|
69
|
+
assert_equal([["Keyword1", "Keyword2"]], decompose("keywords"))
|
70
|
+
end
|
71
|
+
|
72
|
+
def test_created_time
|
73
|
+
assert_equal([Time],
|
74
|
+
decompose("created_time").collect(&:class))
|
75
|
+
end
|
76
|
+
|
77
|
+
def test_modified_time
|
78
|
+
assert_equal([Time],
|
79
|
+
decompose("modified_time").collect(&:class))
|
80
|
+
end
|
81
|
+
|
82
|
+
def test_generator
|
83
|
+
assert_equal(["LibreOffice"],
|
84
|
+
normalize_generators(decompose("generator")))
|
85
|
+
end
|
86
|
+
|
87
|
+
def normalize_generators(generators)
|
88
|
+
generators.collect do |generator|
|
89
|
+
normalize_generator(generator)
|
90
|
+
end
|
91
|
+
end
|
92
|
+
|
93
|
+
def normalize_generator(generator)
|
94
|
+
if generator.start_with?("LibreOffice")
|
95
|
+
"LibreOffice"
|
96
|
+
else
|
97
|
+
generator
|
98
|
+
end
|
99
|
+
end
|
100
|
+
|
101
|
+
def test_creation_date
|
102
|
+
assert_equal([nil], decompose("creation_date"))
|
103
|
+
end
|
104
|
+
end
|
105
|
+
|
106
|
+
sub_test_case("one page") do
|
107
|
+
def decompose
|
108
|
+
super(fixture_path("odt", "one-page.odt"))
|
109
|
+
end
|
110
|
+
|
111
|
+
def test_body
|
112
|
+
assert_equal(["Page1\n"], decompose.collect(&:body))
|
113
|
+
end
|
114
|
+
end
|
115
|
+
|
116
|
+
sub_test_case("multi pages") do
|
117
|
+
def decompose
|
118
|
+
super(fixture_path("odt", "multi-pages.odt"))
|
119
|
+
end
|
120
|
+
|
121
|
+
def test_body
|
122
|
+
assert_equal([<<-BODY], decompose.collect(&:body))
|
123
|
+
Page1
|
124
|
+
Page2
|
125
|
+
BODY
|
126
|
+
end
|
127
|
+
end
|
128
|
+
|
129
|
+
sub_test_case("special characters") do
|
130
|
+
def decompose
|
131
|
+
super(fixture_path("odt", "special-characters.odt"))
|
132
|
+
end
|
133
|
+
|
134
|
+
def test_body
|
135
|
+
assert_equal([<<-BODY], decompose.collect(&:body))
|
136
|
+
Ampersand: &
|
137
|
+
Reference: &
|
138
|
+
HTML: <a href="">
|
139
|
+
Single quote: ''
|
140
|
+
BODY
|
141
|
+
end
|
142
|
+
end
|
143
|
+
end
|
144
|
+
end
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|
Binary file
|