Caching ny Prompt Mahafantatra Provider

Superdav AI Agent v1.12.0 dia mampiditra caching ny prompt mahafantatra provider, izay manatsara ny vidin’ny API sy ny latency amin’ny alalan’ny caching ny prompts manerana ireo provider LLM samihafa. Samy manana mekanisma sy fikirakirana caching samihafa ny provider tsirairay.

Topimaso

Ny caching ny prompt dia mamela anao:

• Hanaovana cache ireo prompts lehibe sy ampiasaina matetika
• Hampihenana ny vidin’ny API amin’ny fisorohana fanodinana miverimberina
• Hanatsarana ny latency ho an’ireo fangatahana voatahiry anaty cache
• Hitantana mazava ny tsingerim-piainan’ny cache

Samy hafa ny fomba anatanterahan’ny provider samihafa ny caching:

• Google Gemini: cachedContents API
• Azure OpenAI: Caching ny prompt miaraka amin’ny TTL
• OpenRouter: Caching manokana arakaraka ny provider
• Vertex Anthropic: Caching ny prompt miaraka amin’ny fanaraha-maso cache

Google Gemini: cachedContents API

Google Gemini dia manome fitantanana cache mazava amin’ny alalan’ny cachedContents API.

Fikirakirana

$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
    ],
];

Famoronana Prompt Voatahiry anaty Cache

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]

Fampiasana Prompt Voatahiry anaty Cache

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

Tsingerim-piainan’ny Cache

// 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' );

Fomba Fanao Tsara Indrindra ho an’ny Gemini

• Mametraha TTL mety: Ampifandanjao ny fitsitsiana vola sy ny fahalasanan’ny cache
• Ataovy cache ny system prompts: Ampiasao indray ilay system prompt mitovy manerana ny fangatahana
• Araho maso ny fampiasana cache: Araho hoe iza amin’ireo cache no ampiasaina indrindra
• Diovy ireo cache lany daty: Fafao tsindraindray ireo cache tsy ampiasaina

Azure OpenAI: Caching ny Prompt

Azure OpenAI dia manohana caching ny prompt miaraka amin’ny fitantanana TTL mandeha ho azy.

Fikirakirana

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

Fampandehanana ny Caching

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,
// ]

Lohateny Cache

Azure OpenAI dia mampiasa lohateny HTTP ho an’ny fanaraha-maso cache:

Cache-Control: max_age=3600

Sanda tohanana:

• max_age=<seconds>: Ataovy cache mandritra ny faharetana voafaritra
• no_cache: Aza atao cache ity fangatahana ity
• no_store: Aza atao cache ary aza ampiasaina indray

Fanaraha-maso ny Fampiasana Cache

$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";

Fomba Fanao Tsara Indrindra ho an’ny Azure OpenAI

• Mampiasà prompts tsy miovaova: Mahazo tombony amin’ny caching ny prompts mitovy tanteraka
• Mametraha TTL antonony: Ampifandanjao ny vidiny sy ny maha-vaovao
• Araho maso ny metrika cache: Araho ny famoronana cache raha oharina amin’ny hits
• Ataovy andiany ireo fangatahana mitovy: Vondroy ny fangatahana mba hampitomboana ny cache hits

OpenRouter: Caching Manokana arakaraka ny Provider

OpenRouter dia manohana caching amin’ny alalan’ireo provider fototra (OpenAI, Anthropic, sns.).

Fikirakirana

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

Fampiasana Caching OpenRouter

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',
    ]
);

Safidy Manokana arakaraka ny Provider

Samy manana mekanisma caching samihafa ny provider samihafa:

// 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,
        ],
    ]
);

Fomba Fanao Tsara Indrindra ho an’ny OpenRouter

• Fantaro ny caching an’ny provider-nao: Samy manana mekanisma samihafa ny provider tsirairay
• Andramo ny fitondran-tenan’ny caching: Hamarino fa mandeha amin’ilay provider nofidinao ny caching
• Araho maso ny vola lany: Araho ny fitsitsiana azo avy amin’ny caching
• Mampiasà models tsy miovaova: Manapaka ny cache hits ny fifindrana models

Vertex Anthropic: Caching ny Prompt miaraka amin’ny Fanaraha-maso Cache

Vertex Anthropic (Google Cloud) dia manohana caching ny prompt miaraka amin’ny fanaraha-maso cache mazava.

Fikirakirana

$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,
        ],
    ],
];

Fampiasana Vertex Anthropic Caching

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,
//     ],
// ]

Karazana Cache Control

• ephemeral: Cache mandritra ny fangatahana (default)
• persistent: Cache manerana fangatahana maro (raha tohanana)

Fanaraha-maso ny Fampiasana Cache

$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";

Fomba Fanao Tsara Indrindra ho an'ny Vertex Anthropic

• Mampiasà caching ephemeral: Tsara ho an'ny caching amin'ny session tokana
• Apetraka araka ny tokony ho izy ny max_tokens: Ampifandanjana ny haben'ny cache sy ny vidiny
• Araho maso ny metrika cache: Araho ny fahombiazan'ny cache
• Andramo amin'ny enta-miasanao: Hamarino fa mahasoa ny tranga fampiasanao ny caching

Paikady Caching Miampita Mpanome Tolotra

Fanamboarana Mitambatra

$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,
                ],
            ],
        ],
    ],
];

Famantarana Mpanome Tolotra

$provider = $config['provider'];

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

// Use provider-specific caching configuration

Paikady Fallback

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'] )
    );
}

Fanatsarana Vidiny

Kajio ny Fitsitsiana

$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";

Torohevitra Fanatsarana

• Cache prompts rafitra lehibe: Fitsitsiana vidiny lehibe indrindra
• Ampiasao indray ny context: Cache antontan-taratasy context ampiasaina matetika
• Ataovy batch ny fangatahana: Vondrony ny fangatahana mitovy mba hampitomboana ny cache hits
• Araho maso ny fahombiazan'ny cache: Araho ny fitsitsiana tena izy
• Ahitsio ny TTL: Ampifandanjana ny vidiny sy ny havaozina

Famahana Olana

Tsy ampiasaina ny cache

• Hamarino fa alefa ao amin'ny fanamboarana ny caching
• Jereo fa mitovy tanteraka ny prompts (mitaky fitoviana tanteraka ny caching)
• Hamarino fa tsy lany daty ny cache
• Jereo ny fetran'ny cache manokana ho an'ny mpanome tolotra

Tsy mahomby ny famoronana cache

• Hamarino fa ao anatin'ny fetran'ny mpanome tolotra ny haben'ny cache
• Jereo fa marina ny syntax cache control
• Ataovy azo antoka fa manohana caching ho an'ny model-nao ny mpanome tolotra
• Avereno jerena ny documentation an'ny mpanome tolotra momba ny fetra

Vidiny tsy nampoizina

• Araho maso ny famoronana cache raha oharina amin'ny tokens vakina avy amin'ny cache
• Hamarino fa tena ampiasaina ny cache
• Jereo raha misy cache misses noho ny fiovaovan'ny prompt
• Diniho ny fanitsiana ny TTL na ny paikady cache

Fampitahana Mpanome Tolotra

Endri-javatra	Gemini	Azure OpenAI	OpenRouter	Vertex Anthropic
Cache API	cachedContents	HTTP headers	Manokana ho an'ny mpanome tolotra	Cache control
Fanaraha-maso TTL	Mazava	Amin'ny alalan'ny headers	Miankina amin'ny mpanome tolotra	Ephemeral/persistent
Haben'ny cache ambony indrindra	1M tokens	Miankina amin'ny mpanome tolotra	Miankina amin'ny mpanome tolotra	1M tokens
Fampihenana vidiny	90%	90%	Miankina amin'ny mpanome tolotra	90%
Fanaraha-maso	Amin'ny antsipiriany	Amin'ny alalan'ny metrics	Miankina amin'ny mpanome tolotra	Amin'ny alalan'ny usage

Dingana Manaraka

1. Safidio ny mpanome tolotrao: Safidio araka ny filanao
2. Amboary ny caching: Apetraho ny caching manokana ho an'ny mpanome tolotra
3. Andramo ny caching: Hamarino fa mandeha amin'ny prompts-nao izany
4. Araho maso ny fampiasana: Araho ny cache hits sy ny fitsitsiana vidiny
5. Hatsarao: Ahitsio ny TTL sy ny paikady cache araka ny vokatra