|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
|
|
4
|
|
|
namespace GDCInfo\Flows; |
|
5
|
|
|
|
|
6
|
|
|
use Carbon\Carbon; |
|
7
|
|
|
use DOMDocument; |
|
8
|
|
|
use DOMXPath; |
|
9
|
|
|
use GDCInfo\Exceptions\FindGDCInfoException; |
|
10
|
|
|
use GDCInfo\Exceptions\NotFoundGDCInfoException; |
|
11
|
|
|
use GDCInfo\GDCInfo; |
|
12
|
|
|
use Illuminate\Support\Facades\Http; |
|
13
|
|
|
use Illuminate\Support\Str; |
|
14
|
|
|
|
|
15
|
|
|
class GDCInfoFromHtmlFlow |
|
16
|
|
|
{ |
|
17
|
|
|
|
|
18
|
6 |
|
public static function make(...$arguments) |
|
19
|
|
|
{ |
|
20
|
6 |
|
return new static(...$arguments); |
|
|
|
|
|
|
21
|
|
|
} |
|
22
|
|
|
|
|
23
|
6 |
|
public function get(string $gdcNumber): GDCInfo |
|
24
|
|
|
{ |
|
25
|
6 |
|
$response = Http::timeout(20) |
|
26
|
6 |
|
->get('https://olr.gdc-uk.org/SearchRegister/SearchResult', [ |
|
27
|
6 |
|
'RegistrationNumber' => $gdcNumber, |
|
28
|
6 |
|
]); |
|
29
|
|
|
|
|
30
|
6 |
|
if (!$response->successful()) { |
|
31
|
|
|
throw new FindGDCInfoException("Response error with status [{$response->status()}]. We can't parse html"); |
|
32
|
|
|
} |
|
33
|
|
|
|
|
34
|
6 |
|
$domDocument = new DOMDocument(); |
|
35
|
6 |
|
@$domDocument->loadHTML($response->body()); |
|
|
|
|
|
|
36
|
6 |
|
$xpath = new DOMXpath($domDocument); |
|
37
|
6 |
|
$heads = $xpath->query('/html/head/title'); |
|
38
|
6 |
|
if (!$heads || !($head = $heads->item(0)) || !$head->nodeValue) { |
|
|
|
|
|
|
39
|
|
|
throw new FindGDCInfoException("'Head' elements not exists. Error on parsing data"); |
|
40
|
|
|
} |
|
41
|
|
|
|
|
42
|
6 |
|
if ($head->nodeValue === 'No Results') { |
|
43
|
3 |
|
throw new NotFoundGDCInfoException("Invalid gdc number [{$gdcNumber}]."); |
|
44
|
|
|
} |
|
45
|
|
|
|
|
46
|
3 |
|
$data = [ |
|
47
|
3 |
|
...$this->parseName($xpath), |
|
48
|
3 |
|
...$this->parseInfo($xpath), |
|
49
|
3 |
|
]; |
|
50
|
|
|
|
|
51
|
3 |
|
return GDCInfo::fromArray($data); |
|
52
|
|
|
} |
|
53
|
|
|
|
|
54
|
3 |
|
protected function parseName(DOMXPath $xpath): array |
|
55
|
|
|
{ |
|
56
|
3 |
|
$names = $xpath->query('/html/body/div/main/div/div/div/div/div/div/div/h2'); |
|
57
|
3 |
|
if (!$names || !($name = $names->item(0)) || !$name->nodeValue) { |
|
|
|
|
|
|
58
|
|
|
throw new FindGDCInfoException("'Name item not found' elements not exists. Error on parsing data"); |
|
59
|
|
|
} |
|
60
|
|
|
|
|
61
|
3 |
|
$data = []; |
|
62
|
3 |
|
$firstName = $name->childNodes->item(1)?->nodeValue; |
|
63
|
3 |
|
if ($firstName) { |
|
64
|
3 |
|
$data['first_name'] = trim($firstName); |
|
65
|
|
|
} |
|
66
|
3 |
|
$lastName = $name->childNodes->item(2)?->nodeValue; |
|
67
|
3 |
|
if ($lastName) { |
|
68
|
3 |
|
$data['last_name'] = trim($lastName); |
|
69
|
|
|
} |
|
70
|
|
|
|
|
71
|
3 |
|
return $data; |
|
72
|
|
|
} |
|
73
|
|
|
|
|
74
|
3 |
|
protected function parseInfo(DOMXPath $xpath): array |
|
75
|
|
|
{ |
|
76
|
|
|
/** @var \DOMNodeList $items */ |
|
77
|
3 |
|
$items = $xpath->query('/html/body/div/main/div/div/div/div/div/div/div/div'); |
|
78
|
3 |
|
if (!$items) { |
|
|
|
|
|
|
79
|
|
|
throw new FindGDCInfoException("'Info items not found' elements not exists. Error on parsing data"); |
|
80
|
|
|
} |
|
81
|
|
|
|
|
82
|
3 |
|
$data = []; |
|
83
|
|
|
|
|
84
|
3 |
|
for ($i = 1; $i < $items->count(); $i++) { |
|
85
|
3 |
|
$item = $items->item($i); |
|
86
|
3 |
|
if (!$item) { |
|
87
|
|
|
continue; |
|
88
|
|
|
} |
|
89
|
3 |
|
$key = rtrim(trim((string)$item->childNodes->item(1)?->nodeValue), ':'); |
|
90
|
3 |
|
$value = rtrim(trim((string)$item->childNodes->item(3)?->nodeValue), ':'); |
|
91
|
3 |
|
if (!$key || !$value) { |
|
92
|
3 |
|
continue; |
|
93
|
|
|
} |
|
94
|
|
|
|
|
95
|
3 |
|
$data[Str::snake($key)] = $value; |
|
96
|
|
|
} |
|
97
|
|
|
|
|
98
|
3 |
|
if (isset($data['first_registered_on'])) { |
|
99
|
3 |
|
$date = $data['first_registered_on']; |
|
100
|
3 |
|
unset($data['first_registered_on']); |
|
101
|
|
|
|
|
102
|
|
|
try { |
|
103
|
3 |
|
$date = Carbon::createFromFormat('d M Y', $date); |
|
104
|
3 |
|
if ($date) { |
|
105
|
3 |
|
$data['first_registered_on'] = $date; |
|
106
|
|
|
} |
|
107
|
|
|
} catch (\Exception $e) { |
|
108
|
|
|
// nothing |
|
109
|
|
|
} |
|
110
|
|
|
} |
|
111
|
|
|
|
|
112
|
3 |
|
if (isset($data['current_period_of_registration_from'])) { |
|
113
|
3 |
|
$dateData = $data['current_period_of_registration_from']; |
|
114
|
3 |
|
unset($data['current_period_of_registration_from']); |
|
115
|
3 |
|
if (Str::contains($dateData, 'until:')) { |
|
116
|
3 |
|
$dateData = array_map('trim', explode('until:', $dateData)); |
|
117
|
3 |
|
if (count($dateData) == 2) { |
|
118
|
3 |
|
$date = $dateData[0]; |
|
119
|
|
|
|
|
120
|
|
|
try { |
|
121
|
3 |
|
$date = Carbon::createFromFormat('d M Y', $date); |
|
122
|
3 |
|
if ($date) { |
|
123
|
3 |
|
$data['current_period_from'] = $date; |
|
124
|
|
|
} |
|
125
|
|
|
} catch (\Exception $e) { |
|
126
|
|
|
// nothing |
|
127
|
|
|
} |
|
128
|
|
|
|
|
129
|
3 |
|
$date = $dateData[1]; |
|
130
|
|
|
|
|
131
|
|
|
try { |
|
132
|
3 |
|
$date = Carbon::createFromFormat('d M Y', $date); |
|
133
|
3 |
|
if ($date) { |
|
134
|
3 |
|
$data['current_period_until'] = $date; |
|
135
|
|
|
} |
|
136
|
|
|
} catch (\Exception $e) { |
|
137
|
|
|
// nothing |
|
138
|
|
|
} |
|
139
|
|
|
} |
|
140
|
|
|
} |
|
141
|
|
|
} |
|
142
|
|
|
|
|
143
|
3 |
|
return $data; |
|
144
|
|
|
} |
|
145
|
|
|
} |
|
146
|
|
|
|
dev-think-one /
laravel-gdc-info
Loading history...
This check compares calls to functions or methods with their respective definitions. If the call has more arguments than are defined, it raises an issue.
If a function is defined several times with a different number of parameters, the check may pick up the wrong definition and report false positives. One codebase where this has been known to happen is Wordpress. Please note the @ignore annotation hint above.