|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* @file This file is part of the PdfParser library. |
|
5
|
|
|
* |
|
6
|
|
|
* @author Konrad Abicht <[email protected]> |
|
7
|
|
|
* |
|
8
|
|
|
* @date 2020-06-01 |
|
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\Parser; |
|
36
|
|
|
|
|
37
|
|
|
/** |
|
38
|
|
|
* Document related tests which focus on certain PDF generators. |
|
39
|
|
|
*/ |
|
40
|
|
|
class DocumentGeneratorFocusTest extends TestCase |
|
41
|
|
|
{ |
|
42
|
|
|
/** |
|
43
|
|
|
* Test getText result. |
|
44
|
|
|
* |
|
45
|
|
|
* PDF generated with Chromium 116 via SaveAs-dialog. |
|
46
|
|
|
*/ |
|
47
|
|
|
public function testGetTextPull634Chromium(): void |
|
48
|
|
|
{ |
|
49
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/R2RML-Spec_Generated_by_Chromium-SaveAs-PDF.pdf'); |
|
50
|
|
|
|
|
51
|
|
|
self::assertStringContainsString('R2RML: RDB to RDF Mapping Language', $document->getText()); |
|
52
|
|
|
} |
|
53
|
|
|
|
|
54
|
|
|
/** |
|
55
|
|
|
* Test getText result. |
|
56
|
|
|
* |
|
57
|
|
|
* PDF (v 1.4) generated with Inkscape 0.92. |
|
58
|
|
|
*/ |
|
59
|
|
|
public function testGetTextPull634InkscapePDF14(): void |
|
60
|
|
|
{ |
|
61
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/SimpleImage_Generated_by_Inkscape-0.92_PDF-v1.4.pdf'); |
|
62
|
|
|
|
|
63
|
|
|
self::assertEquals('TEST', $document->getText()); |
|
64
|
|
|
} |
|
65
|
|
|
|
|
66
|
|
|
/** |
|
67
|
|
|
* Test getText result. |
|
68
|
|
|
* |
|
69
|
|
|
* PDF (v 1.5) generated with Inkscape 0.92. |
|
70
|
|
|
*/ |
|
71
|
|
|
public function testGetTextPull634InkscapePDF15(): void |
|
72
|
|
|
{ |
|
73
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/SimpleImage_Generated_by_Inkscape-0.92_PDF-v1.5.pdf'); |
|
74
|
|
|
|
|
75
|
|
|
self::assertEquals('TEST', $document->getText()); |
|
76
|
|
|
} |
|
77
|
|
|
|
|
78
|
|
|
/** |
|
79
|
|
|
* Test getText result. |
|
80
|
|
|
* |
|
81
|
|
|
* PDF (1.4) generated with LibreOffice Writer (6.4). |
|
82
|
|
|
* |
|
83
|
|
|
* @see https://help.libreoffice.org/6.4/en-US/text/shared/01/ref_pdf_export.html |
|
84
|
|
|
*/ |
|
85
|
|
|
public function testGetTextPull634LibreOffice(): void |
|
86
|
|
|
{ |
|
87
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/RichDocument_Generated_by_Libreoffice-6.4_PDF-v1.4.pdf'); |
|
88
|
|
|
|
|
89
|
|
|
self::assertStringContainsString( |
|
90
|
|
|
'Some currency symbols: £, €, ¥'."\n".'German characters: ÄÖÜß', |
|
91
|
|
|
$document->getText() |
|
92
|
|
|
); |
|
93
|
|
|
} |
|
94
|
|
|
|
|
95
|
|
|
/** |
|
96
|
|
|
* Test getText result. |
|
97
|
|
|
* |
|
98
|
|
|
* PDF (v 1.7) generated with Microsoft Print-to-PDF via Firefox. |
|
99
|
|
|
*/ |
|
100
|
|
|
public function testGetTextPull634MicrosoftPDF17(): void |
|
101
|
|
|
{ |
|
102
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/Wikipedia-PDF_Generated_by_Microsoft_Print-to-PDF.pdf'); |
|
103
|
|
|
|
|
104
|
|
|
$outputText = $document->getText(); |
|
105
|
|
|
|
|
106
|
|
|
self::assertStringContainsString( |
|
107
|
|
|
'Adobe PDF icon'."\n".'Filename'."\n".'extension', |
|
108
|
|
|
$outputText |
|
109
|
|
|
); |
|
110
|
|
|
|
|
111
|
|
|
self::assertStringContainsString( |
|
112
|
|
|
'are necessary to make, use, sell, and distribute PDF-compliant', |
|
113
|
|
|
$outputText |
|
114
|
|
|
); |
|
115
|
|
|
} |
|
116
|
|
|
|
|
117
|
|
|
/** |
|
118
|
|
|
* Test Document functions. |
|
119
|
|
|
* |
|
120
|
|
|
* PDF (v 1.5) generated by Microsoft Word 2016. |
|
121
|
|
|
*/ |
|
122
|
|
|
public function testGetTextPull634MicrosoftWord2016(): void |
|
123
|
|
|
{ |
|
124
|
|
|
$path = $this->rootDir.'/samples/grouped-by-generator/Mostly_German_Text_Generated_By_Microsoft_Word_2016.pdf'; |
|
125
|
|
|
$document = (new Parser())->parseFile($path); |
|
126
|
|
|
|
|
127
|
|
|
$outputText = $document->getText(); |
|
128
|
|
|
|
|
129
|
|
|
self::assertStringContainsString('(einschließlich Marktpflegequote) von 4 Mrd € angestrebt.', $outputText); |
|
130
|
|
|
|
|
131
|
|
|
// check whitespaces and tab usage |
|
132
|
|
|
self::assertStringContainsString( |
|
133
|
|
|
// ,--- here is a tab |
|
134
|
|
|
'Fälligkeit: 19. Oktober 2028 '."\n". |
|
135
|
|
|
'Zinszahlung: 19. Oktober gzj., Zinslaufbeginn 15. Juni 2023', |
|
136
|
|
|
$outputText |
|
137
|
|
|
); |
|
138
|
|
|
|
|
139
|
|
|
// get the two images at the top of the PDF |
|
140
|
|
|
$images = $document->getObjectsByType('Image'); |
|
141
|
|
|
self::assertCount(2, $images); |
|
142
|
|
|
// I am not sure if your changes are related to images too, thats why I added this failing test |
|
143
|
|
|
// so you have an example for further investigation. If your code is unrelated, just remove it. |
|
144
|
|
|
} |
|
145
|
|
|
|
|
146
|
|
|
/** |
|
147
|
|
|
* Test getText result. |
|
148
|
|
|
* |
|
149
|
|
|
* PDF (v 1.5) generated with Power PDF Create. |
|
150
|
|
|
*/ |
|
151
|
|
|
public function testGetTextPull634PowerPDFCreate(): void |
|
152
|
|
|
{ |
|
153
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/Mostly_German_Text_Generated_by_Power_PDF_Create.pdf'); |
|
154
|
|
|
|
|
155
|
|
|
$outputText = $document->getText(); |
|
156
|
|
|
|
|
157
|
|
|
// located on page 1 |
|
158
|
|
|
self::assertStringContainsString( |
|
159
|
|
|
'Index-Verhältniszahl: 1,17812 (am Valutierungstag 7. September 2023)', |
|
160
|
|
|
$outputText |
|
161
|
|
|
); |
|
162
|
|
|
|
|
163
|
|
|
// located on page 2 |
|
164
|
|
|
self::assertStringContainsString( |
|
165
|
|
|
'Einbeziehung in den '."\n". |
|
166
|
|
|
'Börsenhandel: Dienstag, 5. September 2023 '."\n". |
|
167
|
|
|
'Valutierungstag: Donnerstag, 7. September 2023', |
|
168
|
|
|
$outputText |
|
169
|
|
|
); |
|
170
|
|
|
} |
|
171
|
|
|
|
|
172
|
|
|
/** |
|
173
|
|
|
* Test getText result. |
|
174
|
|
|
* |
|
175
|
|
|
* PDF generated from .docx with SmallPDF (https://smallpdf.com) |
|
176
|
|
|
*/ |
|
177
|
|
|
public function testGetTextPull634SmallPDF(): void |
|
178
|
|
|
{ |
|
179
|
|
|
$document = (new Parser())->parseFile($this->rootDir.'/samples/grouped-by-generator/Document_Generated_by_SmallPDF.pdf'); |
|
180
|
|
|
|
|
181
|
|
|
$outputText = $document->getText(); |
|
182
|
|
|
|
|
183
|
|
|
// Actual encoded spaces in the document are preserved |
|
184
|
|
|
self::assertStringContainsString( |
|
185
|
|
|
'SmallPDF SMALLPDF SmallPDF', |
|
186
|
|
|
$outputText |
|
187
|
|
|
); |
|
188
|
|
|
|
|
189
|
|
|
// Hebrew text |
|
190
|
|
|
self::assertStringContainsString( |
|
191
|
|
|
'Hebrew Keyboard - תדלקמ תירבעב - Type Hebrew Online', |
|
192
|
|
|
$outputText |
|
193
|
|
|
); |
|
194
|
|
|
|
|
195
|
|
|
// Russian text |
|
196
|
|
|
self::assertStringContainsString( |
|
197
|
|
|
'Russian Keyboard - русская клавиатура - Type Russian', |
|
198
|
|
|
$outputText |
|
199
|
|
|
); |
|
200
|
|
|
} |
|
201
|
|
|
|
|
202
|
|
|
/** |
|
203
|
|
|
* Test getText result. |
|
204
|
|
|
* |
|
205
|
|
|
* PDF (1.6) generated by Word için Acrobat PDFMaker 17. |
|
206
|
|
|
*/ |
|
207
|
|
|
public function testGetTextPull634WordIcinAcrobatPDFMaker17(): void |
|
208
|
|
|
{ |
|
209
|
|
|
$path = $this->rootDir.'/samples/grouped-by-generator/Mostly_Turkish_Text_Generated_by_Word_için_Acrobat_PDFMaker_17.pdf'; |
|
210
|
|
|
$document = (new Parser())->parseFile($path); |
|
211
|
|
|
|
|
212
|
|
|
$outputText = $document->getText(); |
|
213
|
|
|
|
|
214
|
|
|
self::assertStringContainsString( |
|
215
|
|
|
'İhracat ve döviz kazandırıcı hizmetler reeskont kredisi günlük', |
|
216
|
|
|
$outputText |
|
217
|
|
|
); |
|
218
|
|
|
|
|
219
|
|
|
// white spaces and tabs are preserved |
|
220
|
|
|
self::assertStringContainsString( |
|
221
|
|
|
// ,--- tab |
|
222
|
|
|
'dikkate alınmasına devam edilecektir.', |
|
223
|
|
|
$outputText |
|
224
|
|
|
); |
|
225
|
|
|
|
|
226
|
|
|
// I assume it fails because of a wrongly de/encoded character after Say000 |
|
227
|
|
|
self::assertStringContainsString('Say000�: 20 23-34', $outputText); |
|
228
|
|
|
|
|
229
|
|
|
// get the image at the top of the PDF |
|
230
|
|
|
$images = $document->getObjectsByType('Image'); |
|
231
|
|
|
self::assertCount(2, $images); |
|
232
|
|
|
// I am not sure if your changes are related to images too, thats why I added this failing test |
|
233
|
|
|
// so you have an example for further investigation. If your code is unrelated, just remove it. |
|
234
|
|
|
} |
|
235
|
|
|
} |
|
236
|
|
|
|
smalot /
pdfparser
