|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace LeKoala\Multilingual; |
|
4
|
|
|
|
|
5
|
|
|
use Exception; |
|
6
|
|
|
|
|
7
|
|
|
/** |
|
8
|
|
|
* A simple ollama client to query towerinstruct model |
|
9
|
|
|
* @link https://ollama.com/thinkverse/towerinstruct |
|
10
|
|
|
* @link https://huggingface.co/Unbabel/TowerInstruct-7B-v0.2 |
|
11
|
|
|
*/ |
|
12
|
|
|
class OllamaTowerInstruct |
|
13
|
|
|
{ |
|
14
|
|
|
final public const BASE_URL = 'http://localhost:11434'; |
|
15
|
|
|
final public const BASE_MODEL = 'thinkverse/towerinstruct'; |
|
16
|
|
|
|
|
17
|
|
|
protected ?string $model; |
|
18
|
|
|
protected ?string $url; |
|
19
|
|
|
|
|
20
|
|
|
public function __construct(?string $model = null, ?string $url = null) |
|
21
|
|
|
{ |
|
22
|
|
|
$this->model = $model ?? self::BASE_MODEL; |
|
23
|
|
|
$this->url = $url ?? self::BASE_URL; |
|
24
|
|
|
} |
|
25
|
|
|
|
|
26
|
|
|
public function expandLang(string $lang): string |
|
27
|
|
|
{ |
|
28
|
|
|
$lc = strtolower($lang); |
|
29
|
|
|
// English, Portuguese, Spanish, French, German, Dutch, Italian, Korean, Chinese, Russian |
|
30
|
|
|
return match ($lc) { |
|
31
|
|
|
'en' => 'English', |
|
32
|
|
|
'fr' => 'French', |
|
33
|
|
|
'nl' => 'Dutch', |
|
34
|
|
|
'it' => 'Italian', |
|
35
|
|
|
'de' => 'German', |
|
36
|
|
|
'pt' => 'Portuguese', |
|
37
|
|
|
'ko' => 'Korean', |
|
38
|
|
|
'zh' => 'Chinese', |
|
39
|
|
|
'ru' => 'Russian', |
|
40
|
|
|
default => $lang |
|
41
|
|
|
}; |
|
42
|
|
|
} |
|
43
|
|
|
|
|
44
|
|
|
public function translate(?string $string, string $to, string $from) |
|
45
|
|
|
{ |
|
46
|
|
|
/* |
|
47
|
|
|
messages = [ |
|
48
|
|
|
{"role": "user", "content": "Translate the following text from Portuguese into English.\nPortuguese: Um grupo de investigadores lançou um novo modelo para tarefas relacionadas com tradução.\nEnglish:"}, |
|
49
|
|
|
] |
|
50
|
|
|
*/ |
|
51
|
|
|
|
|
52
|
|
|
$string = $string ?? ''; |
|
53
|
|
|
$from = $this->expandLang($from); |
|
54
|
|
|
$to = $this->expandLang($to); |
|
55
|
|
|
|
|
56
|
|
|
$prompt = "Translate the following text from $from into $to and keep variables between {} as is.\n$from: $string\n$to:"; |
|
57
|
|
|
|
|
58
|
|
|
$result = $this->generate($prompt); |
|
59
|
|
|
|
|
60
|
|
|
$response = $result['response'] ?? ''; |
|
61
|
|
|
|
|
62
|
|
|
// Avoid extra space |
|
63
|
|
|
$response = trim($response); |
|
64
|
|
|
|
|
65
|
|
|
// Make sure we don't get any extra ending dot |
|
66
|
|
|
$endsWithDot = str_ends_with($string, '.'); |
|
67
|
|
|
$translationEndsWithDot = str_ends_with($response, '.'); |
|
68
|
|
|
if (!$endsWithDot && $translationEndsWithDot) { |
|
69
|
|
|
$response = rtrim($response, '.'); |
|
70
|
|
|
} |
|
71
|
|
|
|
|
72
|
|
|
// No crazy {} |
|
73
|
|
|
$includesParen = str_contains($string, '{}'); |
|
74
|
|
|
$translationIncludesParen = str_contains($string, '{}'); |
|
75
|
|
|
if (!$includesParen && $translationIncludesParen) { |
|
76
|
|
|
$response = trim(str_replace('{}', '', $response)); |
|
77
|
|
|
} |
|
78
|
|
|
|
|
79
|
|
|
// No spaces in { } |
|
80
|
|
|
$response = str_replace('{ ', '{', $response); |
|
81
|
|
|
$response = str_replace(' }', '}', $response); |
|
82
|
|
|
|
|
83
|
|
|
return $response; |
|
84
|
|
|
} |
|
85
|
|
|
|
|
86
|
|
|
/** |
|
87
|
|
|
* @param null|array<int> $context |
|
88
|
|
|
* @return array{model:string,created_at:string,response:string,done:bool,done_reason:string,context:array<int>,total_duration:int,load_duration:int,prompt_eval_count:int,prompt_eval_duration:int,eval_count:int,eval_duration:int} |
|
89
|
|
|
*/ |
|
90
|
|
|
public function generate(string $prompt, ?array $context = null) |
|
91
|
|
|
{ |
|
92
|
|
|
$data = [ |
|
93
|
|
|
'model' => $this->model, |
|
94
|
|
|
'prompt' => $prompt, |
|
95
|
|
|
'stream' => false, |
|
96
|
|
|
]; |
|
97
|
|
|
if (!empty($context)) { |
|
98
|
|
|
$data['context'] = $context; |
|
99
|
|
|
} |
|
100
|
|
|
|
|
101
|
|
|
$url = $this->url . '/api/generate'; |
|
102
|
|
|
$ch = curl_init(); |
|
103
|
|
|
|
|
104
|
|
|
curl_setopt($ch, CURLOPT_URL, $url); |
|
105
|
|
|
curl_setopt($ch, CURLOPT_POST, true); |
|
106
|
|
|
curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($data)); |
|
107
|
|
|
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true); |
|
108
|
|
|
|
|
109
|
|
|
$output = curl_exec($ch); |
|
110
|
|
|
|
|
111
|
|
|
curl_close($ch); |
|
112
|
|
|
|
|
113
|
|
|
$decoded = json_decode($output, true); |
|
|
|
|
|
|
114
|
|
|
|
|
115
|
|
|
if (!$decoded) { |
|
116
|
|
|
throw new Exception("Failed to decode json: " . json_last_error_msg()); |
|
117
|
|
|
} |
|
118
|
|
|
|
|
119
|
|
|
//@phpstan-ignore-next-line |
|
120
|
|
|
return $decoded; |
|
121
|
|
|
} |
|
122
|
|
|
} |
|
123
|
|
|
|
lekoala /
silverstripe-multilingual
Loading history...