Provayderdan xabardor Prompt Caching

Superdav AI Agent v1.12.0 provayderdan xabardor prompt cachingni taqdim etadi; u promptlarni turli LLM provayderlari bo‘ylab keshlash orqali API xarajatlari va kechikishni optimallashtiradi. Har bir provayderning keshlash mexanizmlari va konfiguratsiyalari turlicha.

Umumiy ko‘rinish

Prompt caching sizga quyidagilarni beradi:

• Katta, tez-tez ishlatiladigan promptlarni keshlash
• Ortiqcha qayta ishlashdan qochib, API xarajatlarini kamaytirish
• Keshlangan so‘rovlar uchun kechikishni yaxshilash
• Kesh hayot siklini aniq boshqarish

Turli provayderlar keshlashni turlicha amalga oshiradi:

• Google Gemini: cachedContents API
• Azure OpenAI: TTL bilan prompt caching
• OpenRouter: Provayderga xos keshlash
• Vertex Anthropic: Kesh boshqaruvi bilan prompt caching

Google Gemini: cachedContents API

Google Gemini cachedContents API orqali aniq kesh boshqaruvini taqdim etadi.

Konfiguratsiya

$config = [
    'provider' => 'google-gemini',
    'model' => 'gemini-2.0-flash',
    'caching' => [
        'enabled' => true,
        'ttl' => 3600, // 1 hour in seconds
        'max_tokens' => 1000000, // Max tokens to cache
    ],
];

Keshlangan prompt yaratish

use Superdav\AI\Providers\GoogleGemini;

$gemini = new GoogleGemini( $config );

$cached_content = $gemini->create_cached_content(
    [
        'system_prompt' => 'You are a helpful assistant...',
        'context' => 'Large context document...',
        'ttl' => 3600,
    ]
);

// Returns: ['cache_id' => 'abc123', 'expires_at' => timestamp]

Keshlangan promptdan foydalanish

$response = $gemini->generate(
    [
        'cache_id' => 'abc123',
        'prompt' => 'User question here',
    ]
);

Kesh hayot sikli

// List cached contents
$caches = $gemini->list_cached_contents();

// Get cache details
$cache = $gemini->get_cached_content( 'abc123' );

// Extend cache TTL
$gemini->update_cached_content(
    'abc123',
    ['ttl' => 7200] // Extend to 2 hours
);

// Delete cache
$gemini->delete_cached_content( 'abc123' );

Gemini uchun eng yaxshi amaliyotlar

• Mos TTL belgilang: Xarajatni tejash va keshning eskirishi o‘rtasida muvozanat saqlang
• Tizim promptlarini keshlang: Bir xil tizim promptini so‘rovlar bo‘ylab qayta ishlating
• Keshdan foydalanishni kuzating: Qaysi keshlar eng ko‘p ishlatilishini kuzatib boring
• Muddati o‘tgan keshlarni tozalang: Ishlatilmaydigan keshlarni vaqti-vaqti bilan o‘chiring

Azure OpenAI: Prompt Caching

Azure OpenAI avtomatik TTL boshqaruvi bilan prompt cachingni qo‘llab-quvvatlaydi.

Konfiguratsiya

$config = [
    'provider' => 'azure-openai',
    'model' => 'gpt-4-turbo',
    'api_version' => '2024-08-01-preview',
    'caching' => [
        'enabled' => true,
        'cache_control' => 'max_age=3600',
    ],
];

Keshlashni yoqish

use Superdav\AI\Providers\AzureOpenAI;

$azure = new AzureOpenAI( $config );

$response = $azure->generate(
    [
        'system_prompt' => 'You are a helpful assistant...',
        'context' => 'Large context document...',
        'prompt' => 'User question here',
        'cache_control' => 'max_age=3600',
    ]
);

// Response includes cache usage:
// [
//     'content' => '...',
//     'cache_creation_input_tokens' => 1000,
//     'cache_read_input_tokens' => 500,
// ]

Kesh headerlari

Azure OpenAI kesh boshqaruvi uchun HTTP headerlardan foydalanadi:

Cache-Control: max_age=3600

Qo‘llab-quvvatlanadigan qiymatlar:

• max_age=<seconds>: Belgilangan muddatga keshlash
• no_cache: Bu so‘rovni keshlamang
• no_store: Keshlamang va qayta ishlatmang

Keshdan foydalanishni kuzatish

$response = $azure->generate( [...] );

$cache_tokens = $response['cache_creation_input_tokens'] ?? 0;
$cache_hits = $response['cache_read_input_tokens'] ?? 0;

echo "Cache creation: $cache_tokens tokens\n";
echo "Cache hits: $cache_hits tokens\n";

Azure OpenAI uchun eng yaxshi amaliyotlar

• Izchil promptlardan foydalaning: Bir xil promptlar keshlashdan foyda ko‘radi
• Oqilona TTL belgilang: Xarajat va yangilik o‘rtasida muvozanat saqlang
• Kesh metrikalarini kuzating: Kesh yaratilishi va hitlarni kuzatib boring
• O‘xshash so‘rovlarni paketlang: Kesh hitlarini maksimal qilish uchun so‘rovlarni guruhlang

OpenRouter: Provayderga xos keshlash

OpenRouter asosiy provayderlar (OpenAI, Anthropic va boshqalar) orqali keshlashni qo‘llab-quvvatlaydi.

Konfiguratsiya

$config = [
    'provider' => 'openrouter',
    'model' => 'openai/gpt-4-turbo',
    'caching' => [
        'enabled' => true,
        'provider_cache' => 'openai', // Use OpenAI's caching
    ],
];

OpenRouter keshlashidan foydalanish

use Superdav\AI\Providers\OpenRouter;

$router = new OpenRouter( $config );

$response = $router->generate(
    [
        'system_prompt' => 'You are a helpful assistant...',
        'context' => 'Large context document...',
        'prompt' => 'User question here',
        'cache_control' => 'max_age=3600',
    ]
);

Provayderga xos variantlar

Turli provayderlarning keshlash mexanizmlari turlicha:

// OpenAI-compatible caching
$response = $router->generate(
    [
        'model' => 'openai/gpt-4-turbo',
        'cache_control' => 'max_age=3600',
    ]
);

// Anthropic-compatible caching
$response = $router->generate(
    [
        'model' => 'anthropic/claude-3-opus',
        'cache_control' => [
            'type' => 'ephemeral',
            'max_tokens' => 1000000,
        ],
    ]
);

OpenRouter uchun eng yaxshi amaliyotlar

• Provayderingizning keshlashini biling: Har bir provayderning mexanizmlari turlicha
• Keshlash xatti-harakatini sinang: Tanlangan provayderingiz bilan keshlash ishlashini tekshiring
• Xarajatlarni kuzating: Keshlashdan olingan tejashni kuzatib boring
• Izchil modellardan foydalaning: Modellarni almashtirish kesh hitlarini buzadi

Vertex Anthropic: Kesh boshqaruvi bilan Prompt Caching

Vertex Anthropic (Google Cloud) aniq kesh boshqaruvi bilan prompt cachingni qo‘llab-quvvatlaydi.

Konfiguratsiya

$config = [
    'provider' => 'vertex-anthropic',
    'model' => 'claude-3-opus',
    'project_id' => 'your-gcp-project',
    'region' => 'us-central1',
    'caching' => [
        'enabled' => true,
        'cache_control' => [
            'type' => 'ephemeral',
            'max_tokens' => 1000000,
        ],
    ],
];

Vertex Anthropic keshlashidan foydalanish

use Superdav\AI\Providers\VertexAnthropic;

$vertex = new VertexAnthropic( $config );

$response = $vertex->generate(
    [
        'system_prompt' => 'You are a helpful assistant...',
        'context' => 'Large context document...',
        'prompt' => 'User question here',
        'cache_control' => [
            'type' => 'ephemeral',
            'max_tokens' => 1000000,
        ],
    ]
);

// Response includes cache metrics:
// [
//     'content' => '...',
//     'usage' => [
//         'input_tokens' => 1000,
//         'cache_creation_input_tokens' => 500,
//         'cache_read_input_tokens' => 300,
//     ],
// ]

Kesh nazorati turlari

• ephemeral: So‘rov davomida keshda saqlash (standart)
• persistent: Bir nechta so‘rovlar bo‘ylab keshda saqlash (agar qo‘llab-quvvatlansa)

Keshdan foydalanishni monitoring qilish

$response = $vertex->generate( [...] );

$usage = $response['usage'];
$cache_created = $usage['cache_creation_input_tokens'] ?? 0;
$cache_read = $usage['cache_read_input_tokens'] ?? 0;

echo "Cache created: $cache_created tokens\n";
echo "Cache read: $cache_read tokens\n";

Vertex Anthropic uchun eng yaxshi amaliyotlar

• Ephemeral keshlashdan foydalaning: Yagona sessiyali keshlash uchun yaxshi
• max_tokens ni mos ravishda belgilang: Kesh hajmi va xarajat o‘rtasida muvozanat saqlang
• Kesh metrikalarini monitoring qiling: Kesh samaradorligini kuzatib boring
• Ish yuklamangiz bilan sinab ko‘ring: Keshlash foydalanish holatingizga foyda berishini tekshiring

Provayderlararo keshlash strategiyasi

Yagona konfiguratsiya

$config = [
    'caching' => [
        'enabled' => true,
        'default_ttl' => 3600,
        'providers' => [
            'google-gemini' => [
                'ttl' => 3600,
                'max_tokens' => 1000000,
            ],
            'azure-openai' => [
                'cache_control' => 'max_age=3600',
            ],
            'vertex-anthropic' => [
                'cache_control' => [
                    'type' => 'ephemeral',
                    'max_tokens' => 1000000,
                ],
            ],
        ],
    ],
];

Provayderni aniqlash

$provider = $config['provider'];

$cache_config = $config['caching']['providers'][ $provider ]
    ?? $config['caching'];

// Use provider-specific caching configuration

Zaxira strategiyasi

try {
    // Try caching with primary provider
    $response = $primary_provider->generate( $request );
} catch ( CacheException $e ) {
    // Fall back to non-cached request
    $response = $primary_provider->generate(
        array_merge( $request, ['cache_control' => 'no_cache'] )
    );
}

Xarajatlarni optimallashtirish

Tejashni hisoblash

$cache_created_tokens = $response['cache_creation_input_tokens'] ?? 0;
$cache_read_tokens = $response['cache_read_input_tokens'] ?? 0;
$regular_tokens = $response['input_tokens'] ?? 0;

// Typical pricing (varies by provider):
$cache_creation_cost = $cache_created_tokens * 0.00001; // 10x cheaper
$cache_read_cost = $cache_read_tokens * 0.000001; // 100x cheaper
$regular_cost = $regular_tokens * 0.00001;

$total_cost = $cache_creation_cost + $cache_read_cost + $regular_cost;
$savings = ($regular_tokens * 0.00001) - $total_cost;

echo "Estimated savings: \$$savings\n";

Optimallashtirish bo‘yicha maslahatlar

• Katta tizim promptlarini keshda saqlang: Eng katta xarajat tejalishi
• Kontekstdan qayta foydalaning: Tez-tez ishlatiladigan kontekst hujjatlarini keshda saqlang
• So‘rovlarni paketlang: Keshga tushishlarni maksimal oshirish uchun o‘xshash so‘rovlarni guruhlang
• Kesh samaradorligini monitoring qiling: Haqiqiy tejashni kuzatib boring
• TTL ni sozlang: Xarajat va yangilik o‘rtasida muvozanat saqlang

Muammolarni bartaraf etish

Kesh ishlatilmayapti

• Konfiguratsiyada keshlash yoqilganini tekshiring
• Promptlar bir xil ekanini tekshiring (keshlash aniq moslikni talab qiladi)
• Kesh muddati tugamaganini tekshiring
• Provayderga xos kesh cheklovlarini tekshiring

Kesh yaratish muvaffaqiyatsiz bo‘lmoqda

• Kesh hajmi provayder limitlari doirasida ekanini tekshiring
• Kesh nazorati sintaksisi to‘g‘ri ekanini tekshiring
• Provayder modelingiz uchun keshlashni qo‘llab-quvvatlashiga ishonch hosil qiling
• Cheklovlar uchun provayder hujjatlarini ko‘rib chiqing

Kutilmagan xarajatlar

• Kesh yaratish va keshdan o‘qish tokenlarini monitoring qiling
• Kesh haqiqatda ishlatilayotganini tekshiring
• Prompt o‘zgarishlari tufayli keshga tushmaslik holatlarini tekshiring
• TTL yoki kesh strategiyasini sozlashni ko‘rib chiqing

Provayderlarni taqqoslash

Xususiyat	Gemini	Azure OpenAI	OpenRouter	Vertex Anthropic
Kesh API	cachedContents	HTTP headers	Provayderga xos	Kesh nazorati
TTL nazorati	Aniq	Headerlar orqali	Provayderga bog‘liq	Ephemeral/persistent
Maksimal kesh hajmi	1M token	Provayderga bog‘liq	Provayderga bog‘liq	1M token
Xarajat kamayishi	90%	90%	Provayderga bog‘liq	90%
Monitoring	Batafsil	Metrikalar orqali	Provayderga bog‘liq	Foydalanish orqali

Keyingi qadamlar

1. Provayderingizni tanlang: Ehtiyojlaringiz asosida tanlang
2. Keshlashni sozlang: Provayderga xos keshlashni sozlang
3. Keshlashni sinab ko‘ring: Promptlaringiz bilan ishlashini tekshiring
4. Foydalanishni monitoring qiling: Keshga tushishlar va xarajat tejalishini kuzatib boring
5. Optimallashtiring: Natijalar asosida TTL va kesh strategiyasini sozlang