|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* @file This file is part of the PdfParser library. |
|
5
|
|
|
* |
|
6
|
|
|
* @author Konrad Abicht <[email protected]> |
|
7
|
|
|
* |
|
8
|
|
|
* @date 2024-04-19 |
|
9
|
|
|
* |
|
10
|
|
|
* @license LGPLv3 |
|
11
|
|
|
* |
|
12
|
|
|
* @url <https://github.com/smalot/pdfparser> |
|
13
|
|
|
* |
|
14
|
|
|
* PdfParser is a pdf library written in PHP, extraction oriented. |
|
15
|
|
|
* Copyright (C) 2017 - Sébastien MALOT <[email protected]> |
|
16
|
|
|
* |
|
17
|
|
|
* This program is free software: you can redistribute it and/or modify |
|
18
|
|
|
* it under the terms of the GNU Lesser General Public License as published by |
|
19
|
|
|
* the Free Software Foundation, either version 3 of the License, or |
|
20
|
|
|
* (at your option) any later version. |
|
21
|
|
|
* |
|
22
|
|
|
* This program is distributed in the hope that it will be useful, |
|
23
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
24
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
25
|
|
|
* GNU Lesser General Public License for more details. |
|
26
|
|
|
* |
|
27
|
|
|
* You should have received a copy of the GNU Lesser General Public License |
|
28
|
|
|
* along with this program. |
|
29
|
|
|
* If not, see <http://www.pdfparser.org/sites/default/LICENSE.txt>. |
|
30
|
|
|
*/ |
|
31
|
|
|
|
|
32
|
|
|
namespace PHPUnitTests\Integration; |
|
33
|
|
|
|
|
34
|
|
|
use PHPUnitTests\TestCase; |
|
35
|
|
|
use Smalot\PdfParser\Document; |
|
36
|
|
|
use Smalot\PdfParser\Element\ElementArray; |
|
37
|
|
|
use Smalot\PdfParser\Font; |
|
38
|
|
|
use Smalot\PdfParser\Header; |
|
39
|
|
|
use Smalot\PdfParser\Page; |
|
40
|
|
|
use Smalot\PdfParser\Pages; |
|
41
|
|
|
use Smalot\PdfParser\Parser; |
|
42
|
|
|
|
|
43
|
|
|
/** |
|
44
|
|
|
* @internal only for test purposes |
|
45
|
|
|
*/ |
|
46
|
|
|
class PagesDummy extends Pages |
|
47
|
|
|
{ |
|
48
|
|
|
/** |
|
49
|
|
|
* The purpose of this function is to bypass the tedious |
|
50
|
|
|
* work to setup instances which lead to a valid $fonts variable. |
|
51
|
|
|
* |
|
52
|
|
|
* @param array<\Smalot\PdfParser\Font> $fonts |
|
53
|
|
|
* |
|
54
|
|
|
* @return void |
|
55
|
|
|
*/ |
|
56
|
|
|
public function setFonts($fonts) |
|
57
|
|
|
{ |
|
58
|
|
|
$this->fonts = $fonts; |
|
59
|
|
|
} |
|
60
|
|
|
} |
|
61
|
|
|
|
|
62
|
|
|
class PagesTest extends TestCase |
|
63
|
|
|
{ |
|
64
|
|
|
/** |
|
65
|
|
|
* If fonts are not stored in Page instances but in the Pages instance. |
|
66
|
|
|
* |
|
67
|
|
|
* Pages |
|
68
|
|
|
* | `--- fonts = Font[] <=== will be used to override fonts in Page1 ... |
|
69
|
|
|
* | |
|
70
|
|
|
* | |
|
71
|
|
|
* `--+ Page1 |
|
72
|
|
|
* | `--- fonts = null <=== Will be overwritten with the content of Pages.fonts |
|
73
|
|
|
* `--+ ... |
|
74
|
|
|
* |
|
75
|
|
|
* @see https://github.com/smalot/pdfparser/pull/698 |
|
76
|
|
|
*/ |
|
77
|
|
|
public function testPullRequest698NoFontsSet(): void |
|
78
|
|
|
{ |
|
79
|
|
|
$document = $this->createMock(Document::class); |
|
80
|
|
|
|
|
81
|
|
|
// Create a Page mock and tell PHPUnit that its setFonts has to be called once |
|
82
|
|
|
// otherwise an error is raised |
|
83
|
|
|
$page1 = $this->createMock(Page::class); |
|
84
|
|
|
$page1->expects($this->once())->method('setFonts'); |
|
85
|
|
|
|
|
86
|
|
|
// setup header |
|
87
|
|
|
$header = new Header([ |
|
88
|
|
|
'Kids' => new ElementArray([ |
|
89
|
|
|
$page1, |
|
90
|
|
|
]), |
|
91
|
|
|
], $document); |
|
92
|
|
|
|
|
93
|
|
|
$font1 = $this->createMock(Font::class); |
|
94
|
|
|
|
|
95
|
|
|
// Preset fonts variable so we don't have to prepare all the |
|
96
|
|
|
// prerequisites manually (like creating a Ressources instance |
|
97
|
|
|
// with Font instances, see Pages::setupFonts()) |
|
98
|
|
|
$pages = new PagesDummy($document, $header); |
|
99
|
|
|
$pages->setFonts([$font1]); |
|
100
|
|
|
|
|
101
|
|
|
// We expect setFonts is called on $page1, therefore no assertion here |
|
102
|
|
|
$pages->getPages(true); |
|
103
|
|
|
} |
|
104
|
|
|
|
|
105
|
|
|
/** |
|
106
|
|
|
* Dont override fonts list in a Page instance, if available. |
|
107
|
|
|
* |
|
108
|
|
|
* Pages |
|
109
|
|
|
* | `--- fonts = Font[] <=== Has to be ignored because fonts in Page1 is set |
|
110
|
|
|
* | |
|
111
|
|
|
* | |
|
112
|
|
|
* `--+ Page1 |
|
113
|
|
|
* | `--- fonts = Font[] <=== must not be overwritten |
|
114
|
|
|
* `--+ ... |
|
115
|
|
|
* |
|
116
|
|
|
* @see https://github.com/smalot/pdfparser/pull/698 |
|
117
|
|
|
*/ |
|
118
|
|
|
public function testPullRequest698DontOverride(): void |
|
119
|
|
|
{ |
|
120
|
|
|
$document = $this->createMock(Document::class); |
|
121
|
|
|
|
|
122
|
|
|
// Setup an empty Page instance and insert a Font instance. |
|
123
|
|
|
// We wanna see later on, if $font2 is overwritten by $font1. |
|
124
|
|
|
$font2 = new Font($document); |
|
125
|
|
|
$page1 = new Page($document); |
|
126
|
|
|
$page1->setFonts([$font2]); |
|
127
|
|
|
|
|
128
|
|
|
// setup header |
|
129
|
|
|
$header = new Header([ |
|
130
|
|
|
'Kids' => new ElementArray([ |
|
131
|
|
|
$page1, |
|
132
|
|
|
]), |
|
133
|
|
|
], $document); |
|
134
|
|
|
|
|
135
|
|
|
$font1 = $this->createMock(Font::class); |
|
136
|
|
|
|
|
137
|
|
|
$pages = new PagesDummy($document, $header); |
|
138
|
|
|
$pages->setFonts([$font1]); |
|
139
|
|
|
|
|
140
|
|
|
// Trigger setupFonts method in $pages |
|
141
|
|
|
$pages->getPages(true); |
|
142
|
|
|
|
|
143
|
|
|
// Note: |
|
144
|
|
|
// $font1 and $font2 are intenionally not both of the same type. |
|
145
|
|
|
// One is a mock and the other one a real instance of Font. |
|
146
|
|
|
// This way we can simply check the return value of getFonts here. |
|
147
|
|
|
// If both were one of the other, we had to use a different assertation approach. |
|
148
|
|
|
$this->assertEquals([$font2], $page1->getFonts()); |
|
149
|
|
|
} |
|
150
|
|
|
|
|
151
|
|
|
/** |
|
152
|
|
|
* In this example a Document instance is created, which has one Pages instance with a few Font instances. |
|
153
|
|
|
* With the new functionality, related Font instances are passed down to related Page instances. |
|
154
|
|
|
* |
|
155
|
|
|
* @see https://github.com/smalot/pdfparser/pull/698 |
|
156
|
|
|
*/ |
|
157
|
|
|
public function testPullRequest698WithoutUsageOfSetFonts(): void |
|
158
|
|
|
{ |
|
159
|
|
|
$path = $this->rootDir.'/samples/grouped-by-generator/RichDocument_Generated_by_Libreoffice-6.4_PDF-v1.4.pdf'; |
|
160
|
|
|
$document = (new Parser())->parseFile($path); |
|
161
|
|
|
|
|
162
|
|
|
// get Pages instance from generated Document instance |
|
163
|
|
|
$objectsOfTypePages = $document->getObjectsByType('Pages'); |
|
164
|
|
|
$this->assertEquals(1, \count($objectsOfTypePages)); |
|
165
|
|
|
|
|
166
|
|
|
$pagesInstance = array_values($objectsOfTypePages)[0]; |
|
167
|
|
|
|
|
168
|
|
|
// collect Font instance(s) of Page instances |
|
169
|
|
|
$fonts = []; |
|
170
|
|
|
foreach ($pagesInstance->getPages() as $page) { |
|
171
|
|
|
foreach ($page->getFonts() as $font) { |
|
172
|
|
|
$fonts[$font->getName()] = $font; |
|
173
|
|
|
} |
|
174
|
|
|
} |
|
175
|
|
|
|
|
176
|
|
|
// collect Font instance(s) of Pages instance itself |
|
177
|
|
|
$list = $pagesInstance->get('Resources')->get('Font')->getHeader()->getElements(); |
|
178
|
|
|
$fontsToCompareAgainst = []; |
|
179
|
|
|
foreach ($list as $font) { |
|
180
|
|
|
$fontsToCompareAgainst[$font->getName()] = $font; |
|
181
|
|
|
} |
|
182
|
|
|
|
|
183
|
|
|
// check if font names are equal |
|
184
|
|
|
$this->assertEquals( |
|
185
|
|
|
array_keys($fonts), |
|
186
|
|
|
array_keys($fontsToCompareAgainst) |
|
187
|
|
|
); |
|
188
|
|
|
} |
|
189
|
|
|
} |
|
190
|
|
|
|
smalot /
pdfparser
