{"id":30572,"date":"2025-09-29T11:31:34","date_gmt":"2025-09-29T08:31:34","guid":{"rendered":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/"},"modified":"2025-09-29T11:31:34","modified_gmt":"2025-09-29T08:31:34","slug":"ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli","status":"publish","type":"post","link":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/","title":{"rendered":"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#8217;ler) Nedir? Neden \u00d6nemli?"},"content":{"rendered":"<h1>Laravel ile Ollama Paketi Geli\u015ftirme: Kapsaml\u0131 Rehber<\/h1>\n<p>Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM&#8217;ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM&#8217;leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP&#8217;nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.<\/p>\n<p>Son y\u0131llarda yapay zeka d\u00fcnyas\u0131nda \u00e7\u0131\u011f\u0131r a\u00e7an B\u00fcy\u00fck Dil Modelleri (Large Language Models &#8211; LLM&#8217;ler), do\u011fal dil i\u015fleme yetenekleriyle metin olu\u015fturma, \u00f6zetleme, \u00e7eviri ve karma\u015f\u0131k sorular\u0131 yan\u0131tlama gibi bir\u00e7ok alanda etkileyici sonu\u00e7lar sunmaktad\u0131r. OpenAI&#8217;nin GPT serisi veya Google&#8217;\u0131n Gemini gibi modelleri, bu alandaki en bilinen \u00f6rneklerdendir. Ancak, bu modellerin \u00e7o\u011fu bulut tabanl\u0131 hizmetler arac\u0131l\u0131\u011f\u0131yla sunulur. Bu durum, veri gizlili\u011fi, maliyet ve eri\u015fim h\u0131z\u0131 gibi konularda baz\u0131 k\u0131s\u0131tlamalar\u0131 beraberinde getirir. \u0130\u015fte burada Ollama devreye giriyor.<\/p>\n<p>Ollama, a\u00e7\u0131k kaynakl\u0131 bir platform olup, \u00e7e\u015fitli LLM&#8217;leri kendi bilgisayar\u0131n\u0131zda veya sunucunuzda, yani yerel olarak \u00e7al\u0131\u015ft\u0131rman\u0131za olanak tan\u0131r. Docker gibi bir sanalla\u015ft\u0131rma deneyimi sunarak, pop\u00fcler modelleri (\u00f6rne\u011fin Llama 2, Mistral, Gemma) kolayca indirip kurman\u0131z\u0131 ve bir API arac\u0131l\u0131\u011f\u0131yla bunlarla etkile\u015fim kurman\u0131z\u0131 sa\u011flar. Bu yakla\u015f\u0131m\u0131n en b\u00fcy\u00fck avantajlar\u0131ndan biri, verilerinizin \u015firketinizin veya ki\u015fisel kontrol\u00fcn\u00fcz alt\u0131nda kalmas\u0131d\u0131r. Hassas bilgileri \u00fc\u00e7\u00fcnc\u00fc taraf bulut sa\u011flay\u0131c\u0131lar\u0131na g\u00f6ndermek zorunda kalmazs\u0131n\u0131z, bu da \u00f6zellikle finans, sa\u011fl\u0131k veya yasal sekt\u00f6rlerde \u00e7al\u0131\u015fan kurumlar i\u00e7in kritik bir \u00f6neme sahiptir. Ayr\u0131ca, internet ba\u011flant\u0131s\u0131 kesildi\u011finde bile LLM&#8217;leri kullanmaya devam edebilir, bulut maliyetlerinden tasarruf edebilirsiniz. \u00d6rne\u011fin, bir \u00fcretim band\u0131nda anl\u0131k hata tespiti yapan bir kamera sistemi, g\u00f6r\u00fcnt\u00fc i\u015fleme sonu\u00e7lar\u0131n\u0131 yerel bir LLM&#8217;e g\u00f6ndererek do\u011fal dilde a\u00e7\u0131klama alabilir. Bu senaryoda internet ba\u011f\u0131ml\u0131l\u0131\u011f\u0131 ciddi bir kesinti riski olu\u015ftururken, Ollama ile tamamen ba\u011f\u0131ms\u0131z bir \u00e7\u00f6z\u00fcm sunulabilir. Yerel \u00e7al\u0131\u015fma yetene\u011fi ayn\u0131 zamanda gecikmeyi (latency) \u00f6nemli \u00f6l\u00e7\u00fcde azaltarak daha h\u0131zl\u0131 yan\u0131t s\u00fcreleri sa\u011flar, bu da ger\u00e7ek zamanl\u0131 uygulamalar i\u00e7in b\u00fcy\u00fck bir avantajd\u0131r. \u00d6rne\u011fin, canl\u0131 bir m\u00fc\u015fteri destek sohbet robotunda her saniye \u00f6nemlidir ve Ollama, bulut servislerinin getirece\u011fi ekstra gecikmeyi ortadan kald\u0131r\u0131r. T\u00fcm bu nedenlerle, geli\u015ftiricilerin Ollama gibi yerel LLM \u00e7al\u0131\u015ft\u0131rma \u00e7\u00f6z\u00fcmlerine olan ilgisi giderek artmaktad\u0131r ve bu g\u00fcc\u00fc Laravel gibi modern bir web \u00e7at\u0131s\u0131na entegre etmek, bir\u00e7ok yenilik\u00e7i uygulaman\u0131n kap\u0131s\u0131n\u0131 aralayabilir.<\/p>\n<h2>Laravel Paketi Geli\u015ftirmeye Nas\u0131l Ba\u015flan\u0131r? Temeller<\/h2>\n<p>Laravel, PHP ekosistemindeki en g\u00fc\u00e7l\u00fc ve esnek \u00e7er\u00e7evelerden biridir. Kapsaml\u0131 \u00f6zellik setine ek olarak, paket sistemi sayesinde uygulama geli\u015ftiricileri, ortak i\u015flevleri mod\u00fcler ve yeniden kullan\u0131labilir bile\u015fenler halinde ay\u0131rabilirler. Ollama entegrasyonu gibi belirli bir g\u00f6revi yerine getirecek bir paket olu\u015fturmak, hem kod tekrar\u0131n\u0131 \u00f6nler hem de gelecekteki projelerde ayn\u0131 i\u015flevselli\u011fi kolayca kullanman\u0131z\u0131 sa\u011flar. \u00d6ncelikle, yeni bir paket olu\u015fturma s\u00fcrecini ele alal\u0131m.<\/p>\n<p>Bir Laravel paketi asl\u0131nda bir Composer paketidir. Dolay\u0131s\u0131yla, i\u015fe bir Composer projesi ba\u015flatarak ba\u015flayaca\u011f\u0131z. Paketinizi ana Laravel uygulaman\u0131zdan ayr\u0131 bir dizinde, \u00f6rne\u011fin <code>packages\/vendor-name\/package-name<\/code> \u015feklinde konumland\u0131rman\u0131z en iyi uygulamad\u0131r. \u0130lk ad\u0131m olarak, paketin k\u00f6k dizininde <code>composer init<\/code> komutunu \u00e7al\u0131\u015ft\u0131rarak bir <code>composer.json<\/code> dosyas\u0131 olu\u015fturmal\u0131y\u0131z:<\/p>\n<pre><code class=\"language-bash\">\ncd packages\/yazaradi\/ollama-paketi\ncomposer init\n<\/pre>\n<p><\/code><\/p>\n<p>Bu komut, size paket ad\u0131 (\u00f6rne\u011fin <code>yazaradi\/ollama-paketi<\/code>), a\u00e7\u0131klama, yazar gibi bilgileri soracakt\u0131r. En \u00f6nemli k\u0131s\u0131m ise \"PSR-4 autoloading\" ayar\u0131d\u0131r. Burada paketin ana dizinini ve kar\u015f\u0131l\u0131k gelen PHP namespace'ini belirtmeniz gerekir. Genellikle <code>src\/<\/code> dizinini ve <code>YazarAdi\\OllamaPaketi\\<\/code> namespace'ini kullan\u0131r\u0131z.<\/p>\n<pre><code class=\"language-json\">\n{\n    \"name\": \"yazaradi\/ollama-paketi\",\n    \"description\": \"Laravel uygulamalar\u0131 i\u00e7in Ollama entegrasyonu.\",\n    \"type\": \"library\",\n    \"license\": \"MIT\",\n    \"autoload\": {\n        \"psr-4\": {\n            \"YazarAdi\\\\OllamaPaketi\\\\\": \"src\/\"\n        }\n    },\n    \"authors\": [\n        {\n            \"name\": \"Ad\u0131n\u0131z Soyad\u0131n\u0131z\",\n            \"email\": \"email@example.com\"\n        }\n    ],\n    \"require\": {\n        \"php\": \"^8.2\",\n        \"illuminate\/support\": \"^10.0 || ^11.0\",\n        \"guzzlehttp\/guzzle\": \"^7.0\"\n    },\n    \"require-dev\": {\n        \"orchestra\/testbench\": \"^8.0 || ^9.0\",\n        \"phpunit\/phpunit\": \"^10.0 || ^11.0\"\n    },\n    \"extra\": {\n        \"laravel\": {\n            \"providers\": [\n                \"YazarAdi\\\\OllamaPaketi\\\\OllamaPaketiServiceProvider\"\n            ]\n        }\n    },\n    \"minimum-stability\": \"dev\",\n    \"prefer-stable\": true\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Yukar\u0131daki <code>composer.json<\/code> dosyas\u0131, paketin temel ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131 (<code>illuminate\/support<\/code> ve <code>guzzlehttp\/guzzle<\/code> gibi) ve geli\u015ftirme ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131 (<code>orchestra\/testbench<\/code> gibi test ara\u00e7lar\u0131) tan\u0131mlar. Ayr\u0131ca, Laravel'in paketinizi otomatik olarak ke\u015ffetmesini sa\u011flayan <code>extra.laravel.providers<\/code> b\u00f6l\u00fcm\u00fcne dikkat edin. Burada paketin ana servis sa\u011flay\u0131c\u0131s\u0131n\u0131 tan\u0131ml\u0131yoruz. Bir servis sa\u011flay\u0131c\u0131 (Service Provider), paketinizi Laravel \u00e7er\u00e7evesiyle entegre eden temel bile\u015fendir. Ba\u011f\u0131ml\u0131l\u0131klar\u0131 kaydetme, yap\u0131land\u0131rma dosyalar\u0131n\u0131 yay\u0131nlama, rotalar\u0131 tan\u0131mlama ve hatta paket komutlar\u0131n\u0131 kaydetme gibi g\u00f6revleri \u00fcstlenir. <code>src<\/code> dizini alt\u0131nda <code>OllamaPaketiServiceProvider.php<\/code> ad\u0131nda bir dosya olu\u015fturarak ba\u015flayabiliriz:<\/p>\n<pre><code class=\"language-php\">\nnamespace YazarAdi\\OllamaPaketi;\n\nuse Illuminate\\Support\\ServiceProvider;\n\nclass OllamaPaketiServiceProvider extends ServiceProvider\n{\n    \/**\n     * Paket servislerini kaydet.\n     *\n     * @return void\n     *\/\n    public function register(): void\n    {\n        $this->mergeConfigFrom(__DIR__.'\/..\/config\/ollama.php', 'ollama');\n\n        $this->app->singleton('ollama', function ($app) {\n            $config = $app['config']->get('ollama');\n            return new OllamaClient($config['base_url'], $config['api_key'] ?? null);\n        });\n    }\n\n    \/**\n     * Uygulama servislerini \u00f6ny\u00fckle.\n     *\n     * @return void\n     *\/\n    public function boot(): void\n    {\n        $this->publishes([\n            __DIR__.'\/..\/config\/ollama.php' => config_path('ollama.php'),\n        ], 'ollama-config');\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p><code>register()<\/code> metodu i\u00e7inde, paketin yap\u0131land\u0131rma dosyas\u0131n\u0131 birle\u015ftiriyor ve <code>OllamaClient<\/code> s\u0131n\u0131f\u0131m\u0131z\u0131n bir singleton \u00f6rne\u011fini Laravel'in hizmet kapsay\u0131c\u0131s\u0131na ba\u011fl\u0131yoruz. Bu sayede, uygulaman\u0131z\u0131n herhangi bir yerinden <code>app('ollama')<\/code> veya ba\u011f\u0131ml\u0131l\u0131k enjeksiyonu yoluyla bu istemciye eri\u015febileceksiniz. <code>boot()<\/code> metodu ise, paketin yap\u0131land\u0131rma dosyas\u0131n\u0131n (<code>config\/ollama.php<\/code>) kullan\u0131c\u0131 uygulamas\u0131na yay\u0131nlanmas\u0131n\u0131 sa\u011flar. Bu sayede kullan\u0131c\u0131lar, paketin davran\u0131\u015f\u0131n\u0131 kendi ihtiya\u00e7lar\u0131na g\u00f6re \u00f6zelle\u015ftirebilirler. <code>config<\/code> dizini alt\u0131nda <code>ollama.php<\/code> ad\u0131nda basit bir yap\u0131land\u0131rma dosyas\u0131 olu\u015fturmay\u0131 da unutmay\u0131n:<\/p>\n<pre><code class=\"language-php\">\n<?php\n\nreturn [\n    'base_url' => env('OLLAMA_BASE_URL', 'http:\/\/localhost:11434\/api'),\n    'api_key' => env('OLLAMA_API_KEY'), \/\/ E\u011fer Ollama i\u00e7in bir anahtar gerekirse\n    'timeout' => env('OLLAMA_TIMEOUT', 60),\n    'verify_ssl' => env('OLLAMA_VERIFY_SSL', false),\n];\n<\/pre>\n<p><\/code><\/p>\n<p>Bu temel yap\u0131, Laravel uygulaman\u0131zda Ollama ile ileti\u015fim kuracak bir paket i\u00e7in sa\u011flam bir temel olu\u015fturur. Art\u0131k Composer'\u0131n otoloading mekanizmas\u0131n\u0131 g\u00fcncellemek i\u00e7in paketin k\u00f6k dizininde (<code>packages\/yazaradi\/ollama-paketi<\/code>) <code>composer dump-autoload<\/code> komutunu \u00e7al\u0131\u015ft\u0131rabiliriz. Ana Laravel projenizde paketi kullanmak i\u00e7inse, <code>composer.json<\/code> dosyan\u0131z\u0131n <code>repositories<\/code> b\u00f6l\u00fcm\u00fcne paketinizi ekleyip <code>require<\/code> etmeniz yeterlidir:<\/p>\n<pre><code class=\"language-json\">\n\/\/ Ana Laravel projesi composer.json\n{\n    \"repositories\": [\n        {\n            \"type\": \"path\",\n            \"url\": \"packages\/yazaradi\/ollama-paketi\",\n            \"options\": {\n                \"symlink\": true\n            }\n        }\n    ],\n    \"require\": {\n        \"php\": \"^8.2\",\n        \/\/ ... di\u011fer ba\u011f\u0131ml\u0131l\u0131klar\n        \"yazaradi\/ollama-paketi\": \"@dev\"\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Daha sonra ana Laravel projenizin dizininde <code>composer update<\/code> komutunu \u00e7al\u0131\u015ft\u0131rarak paketi y\u00fckleyebilirsiniz. Bu sayede, Laravel'in otomatik paket ke\u015fif \u00f6zelli\u011fi sayesinde servis sa\u011flay\u0131c\u0131n\u0131z otomatik olarak kaydedilecek ve paketinizi kullanmaya ba\u015flayabileceksiniz.<\/p>\n<aside class=\"uzman-ipucu\">\n  <strong>Uzman \u0130pucu:<\/strong> Geli\u015ftirme s\u00fcrecinde paketinizi ana Laravel projenize symlink ile ba\u011flamak (<code>\"symlink\": true<\/code>), paket \u00fczerinde yapt\u0131\u011f\u0131n\u0131z de\u011fi\u015fikliklerin an\u0131nda ana projeye yans\u0131mas\u0131n\u0131 sa\u011flar. Bu, her de\u011fi\u015fiklikte Composer g\u00fcncellemesi \u00e7al\u0131\u015ft\u0131rmak zorunda kalmadan h\u0131zl\u0131 iterasyon yapman\u0131za olanak tan\u0131r ve geli\u015ftirme h\u0131z\u0131n\u0131z\u0131 ciddi oranda art\u0131r\u0131r.<br \/>\n<\/aside>\n<h2>Ollama API \u0130stemcisi Nas\u0131l Olu\u015fturulur? HTTP \u0130stemci ve \u0130stekler<\/h2>\n<p>Bir Laravel paketi i\u00e7inde Ollama ile etkile\u015fime ge\u00e7menin kalbi, Ollama'n\u0131n RESTful API'sine istek g\u00f6nderecek bir HTTP istemcisidir. Laravel, g\u00fc\u00e7l\u00fc ve kullan\u0131m\u0131 kolay bir HTTP istemcisi olan Guzzle \u00fczerine kurulu kendi <code>Illuminate\\Http\\Client\\Factory<\/code> s\u0131n\u0131f\u0131n\u0131 sunar. Bu, harici API'lerle entegrasyonu olduk\u00e7a basit hale getirir. Ollama, belirli bir portta (varsay\u0131lan olarak 11434) \u00e7al\u0131\u015f\u0131r ve JSON tabanl\u0131 istekler bekler. Temel olarak, metin olu\u015fturma (generate), sohbet (chat) ve vekt\u00f6r g\u00f6mme (embeddings) gibi i\u015flevler i\u00e7in farkl\u0131 u\u00e7 noktalara sahiptir.<\/p>\n<p>\u00d6ncelikle, paketimiz i\u00e7inde Ollama API'si ile ileti\u015fim kuracak ana s\u0131n\u0131f\u0131, <code>src\/OllamaClient.php<\/code> dosyas\u0131n\u0131 olu\u015ftural\u0131m. Bu s\u0131n\u0131f, Laravel'in HTTP istemcisini kullanarak Ollama API'sine istekleri sarmalayacakt\u0131r. Servis sa\u011flay\u0131c\u0131da bu s\u0131n\u0131f\u0131 zaten singleton olarak ba\u011flad\u0131\u011f\u0131m\u0131z i\u00e7in, bu istemciye kolayca eri\u015febilece\u011fiz.<\/p>\n<pre><code class=\"language-php\">\nnamespace YazarAdi\\OllamaPaketi;\n\nuse Illuminate\\Http\\Client\\PendingRequest;\nuse Illuminate\\Support\\Facades\\Http;\nuse YazarAdi\\OllamaPaketi\\Exceptions\\OllamaClientException;\n\nclass OllamaClient\n{\n    protected PendingRequest $httpClient;\n    protected ?string $apiKey; \/\/ Ollama'n\u0131n gelecekte API anahtar\u0131 kullanmas\u0131 ihtimaline kar\u015f\u0131l\u0131k\n\n    public function __construct(string $baseUrl, ?string $apiKey = null, int $timeout = 60, bool $verifySsl = false)\n    {\n        $this->apiKey = $apiKey;\n\n        $this->httpClient = Http::baseUrl($baseUrl)\n                                ->timeout($timeout)\n                                ->withOptions(['verify' => $verifySsl]);\n\n        if ($this->apiKey) {\n            $this->httpClient->withToken($this->apiKey); \/\/ E\u011fer Ollama API anahtar\u0131 kullan\u0131rsa\n        }\n    }\n\n    \/**\n     * Belirtilen bir model ile metin olu\u015fturur.\n     *\n     * @param string $model\n     * @param string $prompt\n     * @param array $options\n     * @return array\n     * @throws OllamaClientException\n     *\/\n    public function generate(string $model, string $prompt, array $options = []): array\n    {\n        $payload = [\n            'model' => $model,\n            'prompt' => $prompt,\n            'stream' => false, \/\/ \u0130lk etapta stream'i devre d\u0131\u015f\u0131 b\u0131rakal\u0131m\n            'options' => $options,\n        ];\n\n        try {\n            $response = $this->httpClient->post('generate', $payload);\n\n            if ($response->successful()) {\n                return $response->json();\n            }\n\n            throw new OllamaClientException(\n                'Ollama generate iste\u011fi ba\u015far\u0131s\u0131z oldu: ' . $response->status() . ' - ' . $response->body(),\n                $response->status()\n            );\n\n        } catch (\\Throwable $e) {\n            throw new OllamaClientException('Ollama generate iste\u011fi s\u0131ras\u0131nda bir hata olu\u015ftu: ' . $e->getMessage(), 0, $e);\n        }\n    }\n\n    \/**\n     * Sohbet tamamlama (chat completion) iste\u011fi g\u00f6nderir.\n     *\n     * @param string $model\n     * @param array $messages\n     * @param array $options\n     * @return array\n     * @throws OllamaClientException\n     *\/\n    public function chat(string $model, array $messages, array $options = []): array\n    {\n        $payload = [\n            'model' => $model,\n            'messages' => $messages,\n            'stream' => false,\n            'options' => $options,\n        ];\n\n        try {\n            $response = $this->httpClient->post('chat', $payload);\n\n            if ($response->successful()) {\n                return $response->json();\n            }\n\n            throw new OllamaClientException(\n                'Ollama chat iste\u011fi ba\u015far\u0131s\u0131z oldu: ' . $response->status() . ' - ' . $response->body(),\n                $response->status()\n            );\n\n        } catch (\\Throwable $e) {\n            throw new OllamaClientException('Ollama chat iste\u011fi s\u0131ras\u0131nda bir hata olu\u015ftu: ' . $e->getMessage(), 0, $e);\n        }\n    }\n\n    \/**\n     * Belirtilen metin i\u00e7in embedding (vekt\u00f6r g\u00f6mme) olu\u015fturur.\n     *\n     * @param string $model\n     * @param string $prompt\n     * @param array $options\n     * @return array\n     * @throws OllamaClientException\n     *\/\n    public function embeddings(string $model, string $prompt, array $options = []): array\n    {\n        $payload = [\n            'model' => $model,\n            'prompt' => $prompt,\n            'options' => $options,\n        ];\n\n        try {\n            $response = $this->httpClient->post('embeddings', $payload);\n\n            if ($response->successful()) {\n                return $response->json();\n            }\n\n            throw new OllamaClientException(\n                'Ollama embeddings iste\u011fi ba\u015far\u0131s\u0131z oldu: ' . $response->status() . ' - ' . $response->body(),\n                $response->status()\n            );\n\n        } catch (\\Throwable $e) {\n            throw new OllamaClientException('Ollama embeddings iste\u011fi s\u0131ras\u0131nda bir hata olu\u015ftu: ' . $e->getMessage(), 0, $e);\n        }\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu <code>OllamaClient<\/code> s\u0131n\u0131f\u0131, Ollama API'sinin temel fonksiyonlar\u0131n\u0131 sarmalar: metin olu\u015fturma (<code>\/api\/generate<\/code>), sohbet tamamlama (<code>\/api\/chat<\/code>) ve metin g\u00f6mme (<code>\/api\/embeddings<\/code>). Constructor'da, Ollama sunucusunun temel URL'sini (genellikle <code>http:\/\/localhost:11434\/api<\/code>), zaman a\u015f\u0131m\u0131 s\u00fcresini ve SSL do\u011frulama ayar\u0131n\u0131 alarak Laravel HTTP istemcisini yap\u0131land\u0131r\u0131yoruz. Her metot (<code>generate<\/code>, <code>chat<\/code>, <code>embeddings<\/code>), ilgili Ollama API u\u00e7 noktas\u0131na bir POST iste\u011fi g\u00f6nderir. Gerekli parametreleri (model ad\u0131, prompt\/mesajlar, se\u00e7enekler) bir payload olarak iletir. \u00d6zellikle, Ollama API'sinin yan\u0131t\u0131 JSON format\u0131nda oldu\u011fu i\u00e7in <code>$response->json()<\/code> metodunu kullanarak do\u011frudan PHP dizisine d\u00f6n\u00fc\u015ft\u00fcr\u00fcyoruz. Ba\u015far\u0131l\u0131 olmayan (4xx veya 5xx) yan\u0131tlar veya a\u011f hatalar\u0131 durumunda ise \u00f6zel bir <code>OllamaClientException<\/code> f\u0131rlatarak hata y\u00f6netimini merkezi hale getiriyoruz. Bu, uygulaman\u0131z\u0131n Ollama ile ilgili sorunlar\u0131 daha temiz bir \u015fekilde ele almas\u0131n\u0131 sa\u011flar. <code>src\/Exceptions\/OllamaClientException.php<\/code> dosyas\u0131n\u0131 da \u015fu \u015fekilde olu\u015fturabiliriz:<\/p>\n<pre><code class=\"language-php\">\nnamespace YazarAdi\\OllamaPaketi\\Exceptions;\n\nuse Exception;\n\nclass OllamaClientException extends Exception\n{\n    \/\/ \u00d6zel hata mesajlar\u0131 veya durum kodlar\u0131 eklenebilir.\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu yap\u0131 sayesinde, paketimiz art\u0131k Ollama ile temel d\u00fczeyde ileti\u015fim kurabilir ve ana Laravel uygulamam\u0131zda bu i\u015flevleri kullanabiliriz. \u0130stemci s\u0131n\u0131f\u0131m\u0131z\u0131n esnekli\u011fi sayesinde, Ollama'n\u0131n sundu\u011fu farkl\u0131 modelleri ve onlar\u0131n \u00f6zel parametrelerini kolayca destekleyebiliriz. \u00d6rne\u011fin, bir <code>chat<\/code> iste\u011fi g\u00f6nderirken, LLM'in ge\u00e7mi\u015f konu\u015fmalar\u0131 hat\u0131rlamas\u0131n\u0131 sa\u011flayan bir mesaj dizisi (<code>messages<\/code> array) g\u00f6nderebiliriz. Bu, daha ba\u011flamsal ve tutarl\u0131 yan\u0131tlar elde etmemizi sa\u011flar. Bir sonraki b\u00f6l\u00fcmde, bu istemciyi bir servis katman\u0131na nas\u0131l entegre edece\u011fimizi ve uygulaman\u0131z\u0131n i\u00e7inden nas\u0131l daha kolay eri\u015filebilir hale getirece\u011fimizi inceleyece\u011fiz.<\/p>\n<aside class=\"uzman-ipucu\">\n  <strong>Uzman \u0130pucu:<\/strong> Geli\u015ftirme ortam\u0131nda, Ollama sunucusunun \u00e7al\u0131\u015ft\u0131\u011f\u0131ndan emin olun. <code>ollama run llama2<\/code> gibi bir komutla bir model ba\u015flatarak test edebilir ve <code>http:\/\/localhost:11434\/api\/tags<\/code> adresine taray\u0131c\u0131n\u0131zdan giderek y\u00fckl\u00fc modellerinizi g\u00f6rebilirsiniz. E\u011fer bir proxy arkas\u0131nda \u00e7al\u0131\u015f\u0131yorsan\u0131z, <code>http_proxy<\/code> ve <code>https_proxy<\/code> ortam de\u011fi\u015fkenlerini yap\u0131land\u0131rman\u0131z gerekebilir. Ayr\u0131ca, <code>guzzlehttp\/guzzle<\/code> paketi, HTTP isteklerini taklit etmek i\u00e7in harika test ara\u00e7lar\u0131 sunar; bu da API entegrasyonlar\u0131n\u0131 test ederken ger\u00e7ek bir sunucuya ba\u011flanma ihtiyac\u0131n\u0131 ortadan kald\u0131r\u0131r.<br \/>\n<\/aside>\n<h2>Paket \u0130\u00e7inde Ollama Fonksiyonelli\u011fi Nas\u0131l Kaps\u00fcllenir?<\/h2>\n<p><code>OllamaClient<\/code> s\u0131n\u0131f\u0131m\u0131z, Ollama API'si ile do\u011frudan ileti\u015fim kurmaktan sorumludur, ancak Laravel uygulaman\u0131z\u0131n i\u015f mant\u0131\u011f\u0131 katman\u0131nda do\u011frudan bu istemciyi kullanmak her zaman en temiz yakla\u015f\u0131m olmayabilir. Uygulama geli\u015ftiricilerin daha y\u00fcksek seviyeli, model odakl\u0131 veya g\u00f6rev odakl\u0131 aray\u00fczler arac\u0131l\u0131\u011f\u0131yla Ollama yeteneklerine eri\u015fmelerini sa\u011flamak, paketin kullan\u0131m\u0131n\u0131 daha sezgisel hale getirir ve gelecekteki de\u011fi\u015fikliklere kar\u015f\u0131 daha diren\u00e7li olmas\u0131n\u0131 sa\u011flar. Bu nedenle, OllamaClient'\u0131 sarmalayacak ve i\u015f mant\u0131\u011f\u0131n\u0131 kaps\u00fclleyecek bir ana \"y\u00f6netici\" veya \"servis\" s\u0131n\u0131f\u0131 olu\u015fturmak iyi bir fikirdir.<\/p>\n<p>Paketimiz i\u00e7inde, \u00f6rne\u011fin <code>src\/OllamaManager.php<\/code> ad\u0131nda bir s\u0131n\u0131f olu\u015fturabiliriz. Bu s\u0131n\u0131f, <code>OllamaClient<\/code> \u00f6rne\u011fini constructor'\u0131nda ba\u011f\u0131ml\u0131l\u0131k olarak alacak ve uygulaman\u0131n belirli i\u015f ak\u0131\u015flar\u0131na y\u00f6nelik daha anlaml\u0131 metotlar sunacakt\u0131r. Bu yakla\u015f\u0131m, Facade deseniyle birlikte kullan\u0131ld\u0131\u011f\u0131nda, paketinize eri\u015fimi son derece kolayla\u015ft\u0131racakt\u0131r.<\/p>\n<pre><code class=\"language-php\">\nnamespace YazarAdi\\OllamaPaketi;\n\nuse YazarAdi\\OllamaPaketi\\Exceptions\\OllamaClientException;\n\nclass OllamaManager\n{\n    protected OllamaClient $client;\n    protected string $defaultModel;\n\n    public function __construct(OllamaClient $client, string $defaultModel = 'llama2')\n    {\n        $this->client = $client;\n        $this->defaultModel = $defaultModel;\n    }\n\n    \/**\n     * Metin olu\u015fturma i\u015flemi yapar.\n     *\n     * @param string $prompt\n     * @param string|null $model\n     * @param array $options\n     * @return string\n     * @throws OllamaClientException\n     *\/\n    public function generateText(string $prompt, ?string $model = null, array $options = []): string\n    {\n        $usedModel = $model ?? $this->defaultModel;\n        $response = $this->client->generate($usedModel, $prompt, $options);\n\n        \/\/ Ollama'dan gelen yan\u0131t\u0131n yap\u0131s\u0131na g\u00f6re uygun alan\u0131 d\u00f6nd\u00fcr\u00fcn.\n        \/\/ Genellikle 'response' veya 'content' anahtar\u0131 alt\u0131nda bulunur.\n        return $response['response'] ?? $response['content'] ?? '';\n    }\n\n    \/**\n     * Sohbet tamamlama i\u015flemi yapar.\n     *\n     * @param array $messages [{ role: 'user', content: '...' }]\n     * @param string|null $model\n     * @param array $options\n     * @return array\n     * @throws OllamaClientException\n     *\/\n    public function chatCompletion(array $messages, ?string $model = null, array $options = []): array\n    {\n        $usedModel = $model ?? $this->defaultModel;\n        $response = $this->client->chat($usedModel, $messages, $options);\n\n        \/\/ Ollama'dan gelen chat yan\u0131t\u0131n\u0131n yap\u0131s\u0131na g\u00f6re uygun alan\u0131 d\u00f6nd\u00fcr\u00fcn.\n        \/\/ Genellikle son mesaj\u0131n 'content' alan\u0131d\u0131r.\n        if (isset($response['message']['content'])) {\n            return $response['message'];\n        }\n\n        return $response; \/\/ Tam yan\u0131t\u0131 d\u00f6nd\u00fcrmek daha esnek olabilir\n    }\n\n    \/**\n     * Metin i\u00e7in embedding (vekt\u00f6r g\u00f6mme) olu\u015fturur.\n     *\n     * @param string $text\n     * @param string|null $model\n     * @param array $options\n     * @return array\n     * @throws OllamaClientException\n     *\/\n    public function embedText(string $text, ?string $model = null, array $options = []): array\n    {\n        $usedModel = $model ?? $this->defaultModel;\n        $response = $this->client->embeddings($usedModel, $text, $options);\n\n        return $response['embedding'] ?? [];\n    }\n\n    \/**\n     * Varsay\u0131lan modeli de\u011fi\u015ftirir.\n     *\n     * @param string $model\n     * @return $this\n     *\/\n    public function setDefaultModel(string $model): self\n    {\n        $this->defaultModel = $model;\n        return $this;\n    }\n\n    \/**\n     * Temeldeki OllamaClient \u00f6rne\u011fine do\u011frudan eri\u015fim sa\u011flar.\n     *\n     * @return OllamaClient\n     *\/\n    public function getClient(): OllamaClient\n    {\n        return $this->client;\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p><code>OllamaManager<\/code> s\u0131n\u0131f\u0131, <code>OllamaClient<\/code>'\u0131 constructor'\u0131nda alarak bir ba\u011f\u0131ml\u0131l\u0131k enjeksiyonu \u00f6rne\u011fi sunar. Ayr\u0131ca, varsay\u0131lan bir LLM modelini de kabul ederek, her istekte modeli belirtme zorunlulu\u011funu ortadan kald\u0131r\u0131r. <code>generateText<\/code>, <code>chatCompletion<\/code> ve <code>embedText<\/code> gibi metotlar, <code>OllamaClient<\/code>'\u0131n metotlar\u0131n\u0131 \u00e7a\u011f\u0131r\u0131r ancak gelen yan\u0131tlar\u0131 daha kullan\u0131c\u0131 dostu bir formata d\u00f6n\u00fc\u015ft\u00fcr\u00fcr. \u00d6rne\u011fin, <code>generateText<\/code> metodu sadece olu\u015fturulan metin i\u00e7eri\u011fini d\u00f6nd\u00fcr\u00fcr, b\u00f6ylece uygulaman\u0131n di\u011fer k\u0131s\u0131mlar\u0131 ham API yan\u0131t\u0131n\u0131n detaylar\u0131yla u\u011fra\u015fmak zorunda kalmaz. Bu kaps\u00fclleme, paketin kullan\u0131m\u0131n\u0131 basitle\u015ftirir ve uygulaman\u0131z\u0131n i\u015f mant\u0131\u011f\u0131n\u0131 Ollama API'sinin spesifik detaylar\u0131ndan soyutlar.<\/p>\n<p>Bu <code>OllamaManager<\/code>'\u0131 Laravel'in hizmet kapsay\u0131c\u0131s\u0131na ba\u011flamak i\u00e7in <code>OllamaPaketiServiceProvider<\/code> s\u0131n\u0131f\u0131m\u0131z\u0131 g\u00fcncelleyelim:<\/p>\n<pre><code class=\"language-php\">\nnamespace YazarAdi\\OllamaPaketi;\n\nuse Illuminate\\Support\\ServiceProvider;\n\nclass OllamaPaketiServiceProvider extends ServiceProvider\n{\n    public function register(): void\n    {\n        $this->mergeConfigFrom(__DIR__.'\/..\/config\/ollama.php', 'ollama');\n\n        \/\/ OllamaClient'\u0131 hizmet kapsay\u0131c\u0131s\u0131na ba\u011flama\n        $this->app->singleton(OllamaClient::class, function ($app) {\n            $config = $app['config']->get('ollama');\n            return new OllamaClient(\n                $config['base_url'],\n                $config['api_key'] ?? null,\n                $config['timeout'] ?? 60,\n                $config['verify_ssl'] ?? false\n            );\n        });\n\n        \/\/ OllamaManager'\u0131 hizmet kapsay\u0131c\u0131s\u0131na ba\u011flama, OllamaClient'\u0131 ba\u011f\u0131ml\u0131l\u0131k olarak enjekte etme\n        $this->app->singleton('ollama', function ($app) {\n            $config = $app['config']->get('ollama');\n            return new OllamaManager(\n                $app->make(OllamaClient::class),\n                $config['default_model'] ?? 'llama2'\n            );\n        });\n    }\n\n    public function boot(): void\n    {\n        $this->publishes([\n            __DIR__.'\/..\/config\/ollama.php' => config_path('ollama.php'),\n        ], 'ollama-config');\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Yap\u0131land\u0131rma dosyam\u0131za (<code>config\/ollama.php<\/code>) varsay\u0131lan model i\u00e7in bir giri\u015f eklemeyi unutmay\u0131n:<\/p>\n<pre><code class=\"language-php\">\n\/\/ config\/ollama.php\nreturn [\n    'base_url' => env('OLLAMA_BASE_URL', 'http:\/\/localhost:11434\/api'),\n    'api_key' => env('OLLAMA_API_KEY'),\n    'timeout' => env('OLLAMA_TIMEOUT', 60),\n    'verify_ssl' => env('OLLAMA_VERIFY_SSL', false),\n    'default_model' => env('OLLAMA_DEFAULT_MODEL', 'llama2'),\n];\n<\/pre>\n<p><\/code><\/p>\n<p>Son olarak, paketinize daha kolay ve statik eri\u015fim sa\u011flamak i\u00e7in bir Laravel Facade olu\u015fturabiliriz. <code>src\/Facades\/Ollama.php<\/code> ad\u0131nda bir dosya olu\u015fturarak bunu yapabiliriz:<\/p>\n<pre><code class=\"language-php\">\nnamespace YazarAdi\\OllamaPaketi\\Facades;\n\nuse Illuminate\\Support\\Facades\\Facade;\n\n\/**\n * @method static string generateText(string $prompt, ?string $model = null, array $options = [])\n * @method static array chatCompletion(array $messages, ?string $model = null, array $options = [])\n * @method static array embedText(string $text, ?string $model = null, array $options = [])\n * @method static \\YazarAdi\\OllamaPaketi\\OllamaManager setDefaultModel(string $model)\n * @method static \\YazarAdi\\OllamaPaketi\\OllamaClient getClient()\n *\n * @see \\YazarAdi\\OllamaPaketi\\OllamaManager\n *\/\nclass Ollama extends Facade\n{\n    protected static function getFacadeAccessor(): string\n    {\n        return 'ollama';\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu Facade, <code>OllamaManager<\/code> s\u0131n\u0131f\u0131na statik metotlar arac\u0131l\u0131\u011f\u0131yla eri\u015fim sa\u011flar. Laravel uygulaman\u0131z\u0131n <code>config\/app.php<\/code> dosyas\u0131ndaki <code>aliases<\/code> dizisine paketinizi ekledi\u011finizde (veya Laravel 5.5+ ile otomatik ke\u015fif sayesinde), do\u011frudan <code>Ollama::generateText('...')<\/code> \u015feklinde kullanabileceksiniz. Bu katmanlama, paketinizi hem g\u00fc\u00e7l\u00fc hem de kullan\u0131m\u0131 kolay hale getirirken, i\u015f mant\u0131\u011f\u0131n\u0131z\u0131 API detaylar\u0131ndan ay\u0131r\u0131r. Bu esneklik, paketinizi farkl\u0131 Ollama API de\u011fi\u015fikliklerine veya yeni LLM \u00f6zelliklerine daha kolay uyarlaman\u0131za olanak tan\u0131r. Art\u0131k ana Laravel uygulaman\u0131zda bu paketi kullanmaya haz\u0131rs\u0131n\u0131z!<\/p>\n<h2>Laravel Uygulaman\u0131zda Ollama Paketini Nas\u0131l Kullan\u0131rs\u0131n\u0131z?<\/h2>\n<p>Paketimiz haz\u0131rland\u0131\u011f\u0131na g\u00f6re, \u015fimdi s\u0131ra onu bir Laravel uygulamas\u0131nda nas\u0131l kullanaca\u011f\u0131m\u0131z\u0131 \u00f6\u011frenmeye geldi. Bu b\u00f6l\u00fcm, paketin kurulumu, yap\u0131land\u0131r\u0131lmas\u0131 ve \u00e7e\u015fitli ger\u00e7ek d\u00fcnya senaryolar\u0131nda nas\u0131l entegre edilece\u011fine dair ad\u0131m ad\u0131m rehberlik edecektir.<\/p>\n<h3>Kurulum ve Temel Yap\u0131land\u0131rma<\/h3>\n<p>Daha \u00f6nce belirtildi\u011fi gibi, paketinizi Composer \u00fczerinden projenize dahil ettikten sonra (Composer'\u0131n <code>repositories<\/code> k\u0131sm\u0131n\u0131 kullanarak yerel paketler i\u00e7in veya gelecekte Packagist'e yay\u0131nlad\u0131\u011f\u0131n\u0131zda do\u011frudan), ilk ad\u0131m yap\u0131land\u0131rma dosyas\u0131n\u0131 yay\u0131nlamakt\u0131r. Bu, Laravel uygulaman\u0131z\u0131n <code>config<\/code> dizinine paketin <code>ollama.php<\/code> yap\u0131land\u0131rma dosyas\u0131n\u0131 kopyalar ve b\u00f6ylece ayarlar\u0131 projenize \u00f6zel olarak \u00f6zelle\u015ftirebilirsiniz.<\/p>\n<pre><code class=\"language-bash\">\nphp artisan vendor:publish --tag=ollama-config\n<\/pre>\n<p><\/code><\/p>\n<p>Bu komut, <code>config\/ollama.php<\/code> dosyas\u0131n\u0131 olu\u015fturacakt\u0131r. Bu dosyada, Ollama sunucunuzun temel URL'sini ve varsay\u0131lan modeli ayarlayabilirsiniz. Genellikle bu ayarlar\u0131 <code>.env<\/code> dosyan\u0131zda tutmak ve yap\u0131land\u0131rma dosyas\u0131nda <code>env()<\/code> helper'\u0131 ile \u00e7ekmek en iyi uygulamad\u0131r:<\/p>\n<pre><code class=\"language-dotenv\">\nOLLAMA_BASE_URL=\"http:\/\/localhost:11434\/api\"\nOLLAMA_DEFAULT_MODEL=\"llama2\"\nOLLAMA_TIMEOUT=120\nOLLAMA_VERIFY_SSL=false\n# OLLAMA_API_KEY=\"your_ollama_api_key_if_needed\"\n<\/pre>\n<p><\/code><\/p>\n<p>Unutmay\u0131n, bu ayarlar\u0131 de\u011fi\u015ftirdikten sonra Laravel'in yap\u0131land\u0131rma \u00f6nbelle\u011fini temizlemeniz gerekebilir:<\/p>\n<pre><code class=\"language-bash\">\nphp artisan config:clear\n<\/pre>\n<p><\/code><\/p>\n<h3>Basit Bir Kullan\u0131m \u00d6rne\u011fi: Metin Olu\u015fturma<\/h3>\n<p>Paketiniz haz\u0131r ve yap\u0131land\u0131r\u0131ld\u0131\u011f\u0131nda, <code>Ollama<\/code> Facade'\u0131n\u0131 kullanarak \u00e7e\u015fitli Ollama i\u015flevlerine kolayca eri\u015febilirsiniz. \u00d6rne\u011fin, bir metin olu\u015fturma i\u015flemi i\u00e7in Controller veya Artisan komutunuzda a\u015fa\u011f\u0131daki gibi kullanabilirsiniz:<\/p>\n<pre><code class=\"language-php\">\nnamespace App\\Http\\Controllers;\n\nuse Illuminate\\Http\\Request;\nuse YazarAdi\\OllamaPaketi\\Facades\\Ollama;\nuse YazarAdi\\OllamaPaketi\\Exceptions\\OllamaClientException;\n\nclass AiController extends Controller\n{\n    public function generateStory(Request $request)\n    {\n        $prompt = $request->input('prompt', 'Uzayda kaybolan bir kedinin hikayesini yaz.');\n        $model = $request->input('model', 'mistral'); \/\/ Kullan\u0131c\u0131 farkl\u0131 bir model se\u00e7ebilir\n\n        try {\n            $story = Ollama::generateText($prompt, $model, [\n                'temperature' => 0.7, \/\/ Yarat\u0131c\u0131l\u0131k derecesi\n                'num_predict' => 500  \/\/ Maksimum token say\u0131s\u0131\n            ]);\n\n            return response()->json(['success' => true, 'story' => $story]);\n\n        } catch (OllamaClientException $e) {\n            \/\/ Ollama API's\u0131ndan d\u00f6nen hatay\u0131 yakala ve uygun \u015fekilde ele al\n            return response()->json(['success' => false, 'message' => $e->getMessage()], 500);\n        } catch (\\Exception $e) {\n            \/\/ Di\u011fer beklenmedik hatalar\u0131 yakala\n            return response()->json(['success' => false, 'message' => 'Bir hata olu\u015ftu: ' . $e->getMessage()], 500);\n        }\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu \u00f6rnekte, bir HTTP iste\u011fi ile gelen prompt ve model bilgilerini al\u0131yor, <code>Ollama::generateText()<\/code> metodunu \u00e7a\u011f\u0131rarak Ollama'dan bir hikaye olu\u015fturmas\u0131n\u0131 istiyoruz. Ayr\u0131ca, <code>temperature<\/code> ve <code>num_predict<\/code> gibi model parametrelerini de \u00fc\u00e7\u00fcnc\u00fc bir ba\u011f\u0131ms\u0131z de\u011fi\u015fken olarak iletiyoruz. Bu, LLM'in davran\u0131\u015f\u0131n\u0131 daha detayl\u0131 kontrol etmenizi sa\u011flar. Ollama istemcisinden gelebilecek olas\u0131 hatalar\u0131 (\u00f6rne\u011fin, sunucunun kapal\u0131 olmas\u0131 veya modelin bulunamamas\u0131 gibi) yakalamak i\u00e7in <code>try-catch<\/code> bloklar\u0131n\u0131 kullanmak kritik \u00f6neme sahiptir.<\/p>\n<h3>Ger\u00e7ek D\u00fcnya Senaryosu: Geli\u015fmi\u015f Bir Chatbot Entegrasyonu<\/h3>\n<p>Bir e-ticaret sitesinde m\u00fc\u015fteri hizmetleri i\u00e7in geli\u015fmi\u015f bir chatbot entegrasyonu d\u00fc\u015f\u00fcnebiliriz. Bu chatbot, kullan\u0131c\u0131n\u0131n sorular\u0131n\u0131 do\u011fal dilde anlay\u0131p yan\u0131tlayabilir. Ollama'n\u0131n <code>chat<\/code> API'si bu senaryo i\u00e7in idealdir.<\/p>\n<pre><code class=\"language-php\">\nnamespace App\\Http\\Controllers;\n\nuse Illuminate\\Http\\Request;\nuse Illuminate\\Support\\Facades\\Session;\nuse YazarAdi\\OllamaPaketi\\Facades\\Ollama;\nuse YazarAdi\\OllamaPaketi\\Exceptions\\OllamaClientException;\n\nclass ChatbotController extends Controller\n{\n    private const CHAT_HISTORY_KEY = 'ollama_chat_history';\n\n    public function sendMessage(Request $request)\n    {\n        $userMessage = $request->input('message');\n        $model = Ollama::setDefaultModel('llama2:7b'); \/\/ Veya kullan\u0131c\u0131n\u0131n se\u00e7ti\u011fi model\n\n        \/\/ Sohbet ge\u00e7mi\u015fini al veya bo\u015f ba\u015flat\n        $history = Session::get(self::CHAT_HISTORY_KEY, []);\n\n        \/\/ Kullan\u0131c\u0131n\u0131n mesaj\u0131n\u0131 ge\u00e7mi\u015fe ekle\n        $history[] = ['role' => 'user', 'content' => $userMessage];\n\n        try {\n            \/\/ Ollama'ya sohbet ge\u00e7mi\u015fi ile birlikte mesajlar\u0131 g\u00f6nder\n            $response = Ollama::chatCompletion($history, null, [\n                'temperature' => 0.6,\n                'top_p' => 0.9,\n            ]);\n\n            $ollamaResponseContent = $response['content'] ?? '\u00dczg\u00fcn\u00fcm, sorunuzu anlayamad\u0131m.';\n\n            \/\/ Ollama'n\u0131n yan\u0131t\u0131n\u0131 ge\u00e7mi\u015fe ekle\n            $history[] = ['role' => 'assistant', 'content' => $ollamaResponseContent];\n\n            \/\/ G\u00fcncellenmi\u015f ge\u00e7mi\u015fi oturuma kaydet\n            Session::put(self::CHAT_HISTORY_KEY, $history);\n\n            return response()->json(['success' => true, 'reply' => $ollamaResponseContent, 'history' => $history]);\n\n        } catch (OllamaClientException $e) {\n            return response()->json(['success' => false, 'message' => 'AI ile ileti\u015fim hatas\u0131: ' . $e->getMessage()], 500);\n        } catch (\\Exception $e) {\n            return response()->json(['success' => false, 'message' => 'Beklenmedik bir hata olu\u015ftu: ' . $e->getMessage()], 500);\n        }\n    }\n\n    public function getChatHistory()\n    {\n        return response()->json(['history' => Session::get(self::CHAT_HISTORY_KEY, [])]);\n    }\n\n    public function clearChatHistory()\n    {\n        Session::forget(self::CHAT_HISTORY_KEY);\n        return response()->json(['message' => 'Sohbet ge\u00e7mi\u015fi temizlendi.']);\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu chatbot \u00f6rne\u011fi, Laravel'in oturum (Session) mekanizmas\u0131n\u0131 kullanarak sohbet ge\u00e7mi\u015fini tutar. Her yeni mesaj geldi\u011finde, t\u00fcm ge\u00e7mi\u015fi Ollama'ya g\u00f6nderir, b\u00f6ylece LLM konu\u015fman\u0131n ba\u011flam\u0131n\u0131 koruyabilir. Bu, daha tutarl\u0131 ve ak\u0131c\u0131 bir sohbet deneyimi sa\u011flar. <code>setDefaultModel<\/code> ile anl\u0131k olarak kullan\u0131lacak modeli de\u011fi\u015ftirebiliriz. M\u00fc\u015fteri destek sistemleri, e\u011fitim platformlar\u0131 veya i\u00e7erik olu\u015fturma ara\u00e7lar\u0131 gibi bir\u00e7ok alanda bu t\u00fcr bir entegrasyon b\u00fcy\u00fck faydalar sa\u011flayabilir. Bu \u00f6rnek, paketin temel fonksiyonelli\u011fini ger\u00e7ek bir uygulaman\u0131n ihtiya\u00e7lar\u0131na g\u00f6re nas\u0131l uyarlayabilece\u011finizi g\u00f6stermektedir. Paketin esnekli\u011fi sayesinde, Ollama'n\u0131n sundu\u011fu t\u00fcm yetenekleri Laravel uygulaman\u0131z\u0131n kalbine yerle\u015ftirebilirsiniz.<\/p>\n<h2>Ollama Entegrasyonunda \u0130leri D\u00fczey Konular ve Optimizasyonlar Nelerdir?<\/h2>\n<p>Ollama paketinizin temel i\u015flevselli\u011fini olu\u015fturdunuz ve Laravel uygulaman\u0131za entegre ettiniz. Ancak, daha b\u00fcy\u00fck \u00f6l\u00e7ekli veya performans odakl\u0131 projelerde, paketinizi daha g\u00fc\u00e7l\u00fc ve dayan\u0131kl\u0131 hale getirmek i\u00e7in baz\u0131 ileri d\u00fczey konular\u0131 ve optimizasyon stratejilerini g\u00f6z \u00f6n\u00fcnde bulundurman\u0131z gerekir. Bu konular aras\u0131nda veri ak\u0131\u015f\u0131 (streaming), \u00f6nbellekleme (caching), g\u00fcvenlik, test s\u00fcre\u00e7leri ve hatta mobil uyumlu \u00e7\u0131kt\u0131lar\u0131n y\u00f6netimi yer almaktad\u0131r.<\/p>\n<h3>Veri Ak\u0131\u015f\u0131 (Streaming) ve E\u015fzamans\u0131z \u0130\u015flemler<\/h3>\n<p>B\u00fcy\u00fck dil modellerinden yan\u0131tlar bazen birka\u00e7 saniye s\u00fcrebilir. Kullan\u0131c\u0131 deneyimini iyile\u015ftirmek i\u00e7in, yan\u0131t\u0131n tamam\u0131n\u0131 beklemek yerine metnin par\u00e7a par\u00e7a ak\u0131\u015f\u0131n\u0131 sa\u011flamak (streaming) \u00f6nemlidir. Ollama API'si <code>stream: true<\/code> parametresiyle bu \u00f6zelli\u011fi destekler. Laravel'in HTTP istemcisi, Guzzle'\u0131n e\u015fzamans\u0131z yeteneklerini kullanarak ak\u0131\u015fl\u0131 yan\u0131tlar\u0131 i\u015fleyebilir. <code>OllamaClient<\/code> s\u0131n\u0131f\u0131m\u0131zda <code>generate<\/code> veya <code>chat<\/code> metotlar\u0131n\u0131 ak\u0131\u015fl\u0131 yan\u0131tlara destek verecek \u015fekilde g\u00fcncelleyebiliriz:<\/p>\n<pre><code class=\"language-php\">\n\/\/ src\/OllamaClient.php i\u00e7indeki generate metodu \u00f6rne\u011fi (stream i\u00e7in g\u00fcncellenmi\u015f)\n\npublic function generateStream(string $model, string $prompt, array $options = []): \\Generator\n{\n    $payload = [\n        'model' => $model,\n        'prompt' => $prompt,\n        'stream' => true, \/\/ Streaming'i etkinle\u015ftir\n        'options' => $options,\n    ];\n\n    try {\n        $response = $this->httpClient->withOptions([\n            'stream' => true \/\/ Guzzle'a ak\u0131\u015fl\u0131 yan\u0131t beklenildi\u011fini bildir\n        ])->post('generate', $payload);\n\n        \/\/ Yan\u0131t\u0131n ak\u0131\u015f\u0131n\u0131 sat\u0131r sat\u0131r oku\n        foreach ($response->toPsrResponse()->getBody() as $chunk) {\n            $data = json_decode($chunk, true);\n            if (json_last_error() === JSON_ERROR_NONE) {\n                yield $data; \/\/ Her JSON objesini d\u00f6nd\u00fcr\n            }\n        }\n\n    } catch (\\Throwable $e) {\n        throw new OllamaClientException('Ollama stream generate iste\u011fi s\u0131ras\u0131nda hata: ' . $e->getMessage(), 0, $e);\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu <code>generateStream<\/code> metodu, PHP'nin <code>Generator<\/code> \u00f6zelli\u011fini kullanarak her bir yan\u0131t par\u00e7as\u0131n\u0131 (chunk) d\u00f6nd\u00fcr\u00fcr. Laravel kontrolc\u00fcn\u00fczde bu generator'\u0131 kullanarak anl\u0131k yan\u0131tlar\u0131 istemciye (\u00f6rne\u011fin, bir SSE - Server-Sent Events ba\u011flant\u0131s\u0131 \u00fczerinden) g\u00f6nderebilirsiniz. Bu, kullan\u0131c\u0131 aray\u00fcz\u00fcn\u00fcz\u00fcn LLM yan\u0131t\u0131 geldik\u00e7e g\u00fcncellenmesini sa\u011flar, t\u0131pk\u0131 ChatGPT gibi sohbet uygulamalar\u0131nda oldu\u011fu gibi. E\u015fzamans\u0131z i\u015flemleri desteklemek, \u00f6zellikle kullan\u0131c\u0131 aray\u00fcz\u00fc ile etkile\u015fimli \u00e7al\u0131\u015fan uygulamalarda performans\u0131 ve kullan\u0131c\u0131 memnuniyetini art\u0131r\u0131r. Laravel'in queue (kuyruk) sistemi, arka planda uzun s\u00fcreli LLM i\u015flemlerini y\u00fcr\u00fctmek i\u00e7in de kullan\u0131labilir, b\u00f6ylece HTTP iste\u011fi kilitlenmez.<\/p>\n<h3>\u00d6nbellekleme (Caching) Stratejileri<\/h3>\n<p>Belirli Ollama istekleri (\u00f6zellikle s\u0131k\u00e7a sorulan sorulara verilen yan\u0131tlar veya sabit metin g\u00f6mmeleri) ayn\u0131 sonucu tekrar tekrar d\u00f6nd\u00fcrebilir. Bu t\u00fcr istekler i\u00e7in \u00f6nbellekleme kullanmak, hem LLM'e yap\u0131lan gereksiz \u00e7a\u011fr\u0131lar\u0131 azalt\u0131r hem de yan\u0131t s\u00fcrelerini h\u0131zland\u0131r\u0131r. Laravel'in g\u00fc\u00e7l\u00fc \u00f6nbellekleme s\u00fcr\u00fcc\u00fcleri (Redis, Memcached, File vb.) bu ama\u00e7la kullan\u0131labilir.<\/p>\n<pre><code class=\"language-php\">\n\/\/ OllamaManager s\u0131n\u0131f\u0131nda \u00f6rnek bir \u00f6nbellekleme entegrasyonu\nuse Illuminate\\Support\\Facades\\Cache;\n\npublic function generateTextWithCache(string $prompt, ?string $model = null, array $options = []): string\n{\n    $cacheKey = 'ollama_generate_' . md5($prompt . json_encode($options) . ($model ?? $this->defaultModel));\n\n    return Cache::remember($cacheKey, now()->addHours(24), function () use ($prompt, $model, $options) {\n        return $this->generateText($prompt, $model, $options);\n    });\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu \u00f6rnek, ayn\u0131 prompt ve se\u00e7eneklerle yap\u0131lan istekler i\u00e7in 24 saatlik bir \u00f6nbellek s\u00fcresi belirler. E\u011fer \u00f6nbellekte bir yan\u0131t bulunursa, Ollama API'sine hi\u00e7 istek yap\u0131lmaz. Bu, \u00f6zellikle maliyetli veya zaman al\u0131c\u0131 i\u015flemler i\u00e7in b\u00fcy\u00fck bir optimizasyon sa\u011flar. Hangi isteklerin \u00f6nbelle\u011fe al\u0131naca\u011f\u0131na, verinin g\u00fcncelli\u011fi gereksinimine g\u00f6re karar vermeniz \u00f6nemlidir.<\/p>\n<h3>G\u00fcvenlik: API Anahtarlar\u0131 ve Yetkilendirme<\/h3>\n<p>Ollama yerel olarak \u00e7al\u0131\u015fsa da, API'sine d\u0131\u015far\u0131dan eri\u015fim sa\u011flanabilir. E\u011fer Ollama sunucunuz internete a\u00e7\u0131k bir ortamdaysa, yetkisiz eri\u015fimi engellemek i\u00e7in g\u00fcvenlik \u00f6nlemleri alman\u0131z \u015fartt\u0131r. \u015eu anda Ollama'n\u0131n kendisi g\u00fc\u00e7l\u00fc bir API anahtar\u0131 veya kimlik do\u011frulama mekanizmas\u0131 sunmamaktad\u0131r; ancak, API eri\u015fimini sunucu seviyesinde bir proxy (\u00f6rne\u011fin Nginx) ile k\u0131s\u0131tlayabilir veya Laravel paketinizi kullanarak kendi yetkilendirme katman\u0131n\u0131z\u0131 olu\u015fturabilirsiniz. \u00d6rne\u011fin, Laravel'in Middleware'leri arac\u0131l\u0131\u011f\u0131yla gelen isteklerin belirli bir API anahtar\u0131 i\u00e7erdi\u011fini do\u011frulayabilirsiniz. Bu anahtar, Laravel uygulaman\u0131zdan Ollama sunucusuna yap\u0131lan her istekte bir HTTP ba\u015fl\u0131\u011f\u0131 olarak eklenebilir. Ayr\u0131ca, Laravel paketi, Ollama sunucusuyla ileti\u015fim kurarken SSL\/TLS kullanacak \u015fekilde yap\u0131land\u0131r\u0131lmal\u0131d\u0131r (<code>verify_ssl<\/code> ayar\u0131). Bu, ileti\u015fimin \u015fifrelenmesini ve dinlenmesini \u00f6nler.<\/p>\n<h3>Test S\u00fcre\u00e7leri: Unit ve Feature Testleri<\/h3>\n<p>Paketinizin beklenen \u015fekilde \u00e7al\u0131\u015ft\u0131\u011f\u0131ndan emin olmak i\u00e7in kapsaml\u0131 testler yazmak hayati \u00f6neme sahiptir. Laravel paketleri genellikle Orchestra Testbench ile test edilir. Bu, paketinizi minimum bir Laravel uygulamas\u0131 i\u00e7inde test etmenizi sa\u011flar. OllamaClient gibi d\u0131\u015f API'lerle etkile\u015fime giren s\u0131n\u0131flar i\u00e7in, ger\u00e7ek API \u00e7a\u011fr\u0131lar\u0131n\u0131 yapmadan test etmek amac\u0131yla HTTP istemcisini taklit etmeniz (mocking) gerekir. Laravel'in <code>Http::fake()<\/code> \u00f6zelli\u011fi, bu i\u015flemi olduk\u00e7a basitle\u015ftirir.<\/p>\n<pre><code class=\"language-php\">\n\/\/ tests\/Feature\/OllamaClientTest.php \u00f6rne\u011fi\nuse YazarAdi\\OllamaPaketi\\OllamaClient;\nuse Illuminate\\Support\\Facades\\Http;\nuse Orchestra\\Testbench\\TestCase;\n\nclass OllamaClientTest extends TestCase\n{\n    \/** @test *\/\n    public function it_can_generate_text_successfully()\n    {\n        Http::fake([\n            'localhost:11434\/api\/generate' => Http::response([\n                'model' => 'llama2',\n                'created_at' => '2023-01-01T00:00:00Z',\n                'response' => 'Merhaba, ben bir test yan\u0131t\u0131y\u0131m.',\n                'done' => true,\n            ], 200),\n        ]);\n\n        $client = new OllamaClient('http:\/\/localhost:11434\/api');\n        $response = $client->generate('llama2', 'Merhaba d\u00fcnya');\n\n        $this->assertEquals('Merhaba, ben bir test yan\u0131t\u0131y\u0131m.', $response['response']);\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>Bu test \u00f6rne\u011fi, <code>Http::fake()<\/code> kullanarak Ollama API'sine yap\u0131lan bir <code>generate<\/code> iste\u011fini taklit eder. Ger\u00e7ek bir HTTP \u00e7a\u011fr\u0131s\u0131 yap\u0131lmadan, istemcinin yan\u0131t\u0131 do\u011fru \u015fekilde i\u015fleyip i\u015flemedi\u011fini kontrol edebiliriz. Bu yakla\u015f\u0131m, testlerin h\u0131zl\u0131 ve tutarl\u0131 \u00e7al\u0131\u015fmas\u0131n\u0131 sa\u011flar.<\/p>\n<h3>Mobil Uyumlu HTML \u00c7\u0131kt\u0131s\u0131 D\u00fc\u015f\u00fcnceleri<\/h3>\n<p>E\u011fer Ollama paketiniz, do\u011frudan veya dolayl\u0131 olarak kullan\u0131c\u0131 aray\u00fcz\u00fcne y\u00f6nelik HTML i\u00e7eri\u011fi \u00fcretiyorsa (\u00f6rne\u011fin, bir y\u00f6netim paneli widget'\u0131 veya bir chatbot aray\u00fcz\u00fc), bu \u00e7\u0131kt\u0131lar\u0131n mobil cihazlarda da iyi g\u00f6r\u00fcnmesini sa\u011flamak \u00f6nemlidir. Modern web geli\u015ftirme pratikleri, duyarl\u0131 tasar\u0131m (responsive design) prensiplerini benimser. Bu, CSS Media Queries kullanarak ekran boyutuna g\u00f6re farkl\u0131 stiller uygulayarak ba\u015far\u0131l\u0131r. Paketin kendisi genellikle do\u011frudan HTML \u00fcretmese de, Laravel uygulaman\u0131z\u0131n genel g\u00f6r\u00fcn\u00fcm ve hissini etkileyen blade bile\u015fenleri veya View Composer'lar sa\u011flayabilir. \u0130\u015fte basit bir duyarl\u0131 CSS \u00f6rne\u011fi:<\/p>\n<pre><code class=\"language-css\">\n\/* resources\/assets\/css\/ollama-widget.css (\u00f6rne\u011fin) *\/\n\n.ollama-card {\n    border: 1px solid #e2e8f0;\n    border-radius: 0.5rem;\n    padding: 1rem;\n    margin-bottom: 1rem;\n    background-color: #fff;\n    box-shadow: 0 1px 3px 0 rgba(0, 0, 0, 0.1), 0 1px 2px 0 rgba(0, 0, 0, 0.06);\n}\n\n.ollama-card-title {\n    font-size: 1.25rem;\n    font-weight: bold;\n    color: #2d3748;\n    margin-bottom: 0.75rem;\n}\n\n\/* Mobil cihazlar i\u00e7in stil ayarlar\u0131 *\/\n@media (max-width: 768px) {\n    .ollama-card {\n        padding: 0.75rem;\n        margin-bottom: 0.75rem;\n        border-radius: 0.25rem;\n    }\n\n    .ollama-card-title {\n        font-size: 1rem;\n        margin-bottom: 0.5rem;\n    }\n}\n\n\/* K\u00fc\u00e7\u00fck cihazlar i\u00e7in ek ayarlar *\/\n@media (max-width: 480px) {\n    .ollama-card {\n        padding: 0.5rem;\n    }\n\n    .ollama-card-title {\n        font-size: 0.9rem;\n    }\n}\n<\/pre>\n<p><\/code><\/p>\n<p>E\u011fer paketiniz bir Blade bile\u015feni sunuyorsa, bu CSS dosyalar\u0131n\u0131 kendi <code>ServiceProvider<\/code>'\u0131n\u0131zdan yay\u0131nlayabilir ve uygulama taraf\u0131ndan kullan\u0131lmas\u0131n\u0131 sa\u011flayabilirsiniz. \u00d6rne\u011fin, <code>boot()<\/code> metodunda <code>$this->publishes([__DIR__.'\/..\/resources\/assets' => public_path('vendor\/yazaradi\/ollama')], 'ollama-assets');<\/code> komutuyla CSS dosyalar\u0131n\u0131 yay\u0131nlayabilirsiniz. B\u00f6ylece, paketinize ait t\u00fcm kullan\u0131c\u0131 aray\u00fcz\u00fc bile\u015fenleri otomatik olarak duyarl\u0131 hale gelecektir. Duyarl\u0131 tasar\u0131m, kullan\u0131c\u0131lar\u0131n\u0131z\u0131n herhangi bir cihazdan uygulaman\u0131za eri\u015fti\u011finde sorunsuz bir deneyim ya\u015famas\u0131n\u0131 garantiler.<\/p>\n<p>Bu ileri d\u00fczey konular\u0131 dikkate alarak, Ollama Laravel paketinizi sadece i\u015flevsel de\u011fil, ayn\u0131 zamanda performansl\u0131, g\u00fcvenli, test edilebilir ve kullan\u0131c\u0131 dostu hale getirebilirsiniz. Bu, paketinizi geni\u015f bir kullan\u0131m alan\u0131na yayg\u0131nla\u015ft\u0131rmak i\u00e7in anahtar niteli\u011findedir.<\/p>\n<h2>Sonu\u00e7: Yerel Yapay Zeka G\u00fcc\u00fcn\u00fc Laravel'e Ta\u015f\u0131y\u0131n<\/h2>\n<p>Bu kapsaml\u0131 rehber boyunca, a\u00e7\u0131k kaynakl\u0131 Ollama platformunun g\u00fcc\u00fcn\u00fc Laravel uygulamalar\u0131n\u0131zla nas\u0131l birle\u015ftirebilece\u011finizi ad\u0131m ad\u0131m \u00f6\u011frendiniz. Yerel B\u00fcy\u00fck Dil Modellerini \u00e7al\u0131\u015ft\u0131rma yetene\u011fi sunan Ollama, geli\u015ftiricilere bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin getirdi\u011fi maliyet, veri gizlili\u011fi ve gecikme endi\u015felerini ortadan kald\u0131ran cazip bir alternatif sunmaktad\u0131r. Kendi Laravel paketinizi olu\u015fturarak, Ollama'n\u0131n sundu\u011fu metin olu\u015fturma, sohbet tamamlama ve metin g\u00f6mme gibi geli\u015fmi\u015f yapay zeka \u00f6zelliklerini Laravel ekosistemine entegre etmenin temel prensiplerini ke\u015ffettik. Composer ile paket yap\u0131s\u0131 olu\u015fturmaktan, Laravel'in HTTP istemcisini kullanarak Ollama API'si ile ileti\u015fim kurmaya, servis sa\u011flay\u0131c\u0131lar\u0131 ve Facade'lar arac\u0131l\u0131\u011f\u0131yla paketinizi uygulaman\u0131zda kolayca eri\u015filebilir k\u0131lmaya kadar t\u00fcm s\u00fcre\u00e7leri ele ald\u0131k.<\/p>\n<p>Paket geli\u015ftirme s\u00fcrecinde, hem i\u015f mant\u0131\u011f\u0131n\u0131 API detaylar\u0131ndan soyutlamak hem de kodu daha mod\u00fcler hale getirmek i\u00e7in <code>OllamaClient<\/code> ve <code>OllamaManager<\/code> s\u0131n\u0131flar\u0131n\u0131 nas\u0131l katmanland\u0131raca\u011f\u0131m\u0131z\u0131 g\u00f6rd\u00fck. Ger\u00e7ek d\u00fcnya senaryolar\u0131yla, \u00f6rne\u011fin bir chatbot entegrasyonuyla, bu paketin potansiyel kullan\u0131m alanlar\u0131n\u0131 somutla\u015ft\u0131rd\u0131k. Ayr\u0131ca, streaming ile ger\u00e7ek zamanl\u0131 yan\u0131tlar alma, \u00f6nbellekleme ile performans\u0131 art\u0131rma, g\u00fcvenlik \u00f6nlemleri alma ve test s\u00fcre\u00e7lerini uygulama gibi ileri d\u00fczey konulara de\u011finerek paketinizi daha g\u00fc\u00e7l\u00fc ve dayan\u0131kl\u0131 hale getirme yollar\u0131n\u0131 inceledik. Mobil uyumlu HTML \u00e7\u0131kt\u0131lar\u0131 \u00fcretme konusundaki d\u00fc\u015f\u00fcnceler, paketinizi kullan\u0131c\u0131 aray\u00fcz\u00fcne y\u00f6nelik bile\u015fenler sunmas\u0131 durumunda nas\u0131l duyarl\u0131 hale getirebilece\u011finizi g\u00f6sterdi.<\/p>\n<p>Ollama'n\u0131n yerel LLM \u00e7al\u0131\u015ft\u0131rma yetene\u011fi, Laravel'in geli\u015ftirici dostu yap\u0131s\u0131yla birle\u015fti\u011finde, sonsuz yarat\u0131c\u0131 ve pratik uygulama f\u0131rsatlar\u0131 sunar. \u0130\u00e7erik otomasyonundan ak\u0131ll\u0131 m\u00fc\u015fteri destek sistemlerine, veri analizinden ki\u015fiselle\u015ftirilmi\u015f \u00f6\u011frenme deneyimlerine kadar bir\u00e7ok alanda \u00e7\u0131\u011f\u0131r a\u00e7\u0131c\u0131 projeler geli\u015ftirebilirsiniz. Kendi kontrol\u00fcn\u00fczdeki yapay zeka g\u00fcc\u00fcyle, hem esnek hem de maliyet etkin \u00e7\u00f6z\u00fcmler \u00fcretebilir, veri gizlili\u011fi endi\u015felerini minimize edebilirsiniz. Bu rehberin, Laravel ve Ollama ile yapay zeka yolculu\u011funuza ilham vermesini ve sizi kendi yenilik\u00e7i \u00e7\u00f6z\u00fcmlerinizi geli\u015ftirmeye te\u015fvik etmesini umuyoruz. Gelecek, yerel yapay zeka ile daha da ki\u015fiselle\u015ftirilebilir ve verimli uygulamalar in\u015fa etme f\u0131rsatlar\u0131 sunuyor.<\/p>\n<h3>S\u0131k\u00e7a Sorulan Sorular (SSS)<\/h3>\n<p><strong>S1: Ollama paketini kullanmak i\u00e7in hangi gereksinimler var?<\/strong><\/p>\n<p>C1: Ollama paketini kullanmak i\u00e7in \u00f6ncelikle yerel makinenizde veya sunucunuzda Ollama uygulamas\u0131n\u0131n kurulu ve \u00e7al\u0131\u015f\u0131yor olmas\u0131 gerekir. Ayr\u0131ca, Laravel uygulaman\u0131z\u0131n PHP 8.2 veya \u00fczeri bir s\u00fcr\u00fcme sahip olmas\u0131 ve Composer ile paket ba\u011f\u0131ml\u0131l\u0131klar\u0131n\u0131n y\u00f6netilmesi gerekmektedir. Y\u00fckl\u00fc bir LLM modeli (\u00f6rne\u011fin <code>llama2<\/code> veya <code>mistral<\/code>) Ollama \u00fczerinde \u00e7al\u0131\u015f\u0131r durumda olmal\u0131d\u0131r.<\/p>\n<p><strong>S2: Farkl\u0131 LLM'leri paket i\u00e7inde nas\u0131l y\u00f6netebilirim?<\/strong><\/p>\n<p>C2: Paket, Ollama'n\u0131n destekledi\u011fi herhangi bir modeli kullanman\u0131za olanak tan\u0131r. <code>config\/ollama.php<\/code> dosyas\u0131ndaki <code>default_model<\/code> ayar\u0131n\u0131 de\u011fi\u015ftirerek varsay\u0131lan modeli belirleyebilirsiniz. Ayr\u0131ca, <code>Ollama::generateText('prompt', 'yeni-model')<\/code> veya <code>Ollama::chatCompletion(messages, 'yeni-model')<\/code> gibi metotlarda ikinci parametre olarak do\u011frudan model ad\u0131n\u0131 belirterek anl\u0131k olarak farkl\u0131 bir model kullanabilirsiniz.<\/p>\n<p><strong>S3: Paket performans\u0131 nas\u0131l optimize edilir?<\/strong><\/p>\n<p>C3: Performans\u0131 optimize etmek i\u00e7in birka\u00e7 yol vard\u0131r:<\/p>\n<ol>\n<li><strong>\u00d6nbellekleme:<\/strong> S\u0131k\u00e7a tekrar eden veya statik yan\u0131tlar i\u00e7in Laravel'in \u00f6nbellekleme sistemini kullan\u0131n.<\/li>\n<li><strong>Streaming:<\/strong> Uzun yan\u0131tlar i\u00e7in veri ak\u0131\u015f\u0131 (streaming) \u00f6zelli\u011fini kullanarak kullan\u0131c\u0131 deneyimini iyile\u015ftirin.<\/li>\n<li><strong>E\u015fzamans\u0131z \u0130\u015flemler\/Kuyruklar:<\/strong> Uzun s\u00fcreli LLM i\u015flemlerini Laravel kuyruk sistemine (queues) atarak ana HTTP iste\u011fini serbest b\u0131rak\u0131n.<\/li>\n<li><strong>Model Se\u00e7imi:<\/strong> Daha k\u00fc\u00e7\u00fck ve daha h\u0131zl\u0131 modelleri tercih edin veya belirli g\u00f6revler i\u00e7in optimize edilmi\u015f modelleri kullan\u0131n.<\/li>\n<\/ol>\n<p><strong>S4: Streaming \u00f6zelli\u011fi ne i\u015fe yarar ve nas\u0131l kullan\u0131l\u0131r?<\/strong><\/p>\n<p>C4: Streaming \u00f6zelli\u011fi, B\u00fcy\u00fck Dil Modellerinden gelen yan\u0131tlar\u0131n tamam\u0131n\u0131 beklemek yerine, metnin par\u00e7a par\u00e7a (chunk by chunk) akmas\u0131n\u0131 sa\u011flar. Bu, \u00f6zellikle sohbet robotlar\u0131 gibi etkile\u015fimli uygulamalarda kullan\u0131c\u0131ya daha h\u0131zl\u0131 geri bildirim sunarak bekleme s\u00fcresi alg\u0131s\u0131n\u0131 azalt\u0131r. Paketinizde, <code>OllamaClient<\/code> i\u00e7indeki <code>generateStream<\/code> gibi bir metot olu\u015fturarak ve PHP'nin Generator'lar\u0131n\u0131 kullanarak bu \u00f6zelli\u011fi etkinle\u015ftirebilirsiniz. \u0130stemci taraf\u0131nda ise genellikle Server-Sent Events (SSE) veya WebSockets ile bu ak\u0131\u015f\u0131 dinleyerek kullan\u0131c\u0131 aray\u00fcz\u00fcn\u00fc g\u00fcncelleyebilirsiniz.<\/p>\n<p><strong>S5: Kendi \u00f6zel modellerimi paketle entegre edebilir miyim?<\/strong><\/p>\n<p>C5: Evet, Ollama'n\u0131n en b\u00fcy\u00fck avantajlar\u0131ndan biri, kendi \u00f6zelle\u015ftirilmi\u015f veya ince ayarl\u0131 modellerinizi (fine-tuned models) kolayca i\u00e7e aktarman\u0131za ve \u00e7al\u0131\u015ft\u0131rman\u0131za izin vermesidir. Kendi modelinizi Ollama'ya ekledikten sonra, t\u0131pk\u0131 standart modeller gibi, bu paketi kullanarak modelinizin ad\u0131n\u0131 belirterek do\u011frudan onunla etkile\u015fim kurabilirsiniz. Paket, Ollama API'sinin genel yap\u0131s\u0131na dayand\u0131\u011f\u0131 i\u00e7in, hangi modelin kullan\u0131ld\u0131\u011f\u0131 konusunda esneklik sunar.<\/p>\n<p><\/body><\/p>\n","protected":false},"excerpt":{"rendered":"Laravel ile Ollama Paketi Geli\u015ftirme: Kapsaml\u0131 Rehber Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na&hellip;","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"csco_page_header_type":"","csco_page_load_nextpost":"","csco_page_subscribe_form":"","csco_page_contact_form":"","footnotes":""},"categories":[1],"tags":[],"class_list":{"0":"post-30572","1":"post","2":"type-post","3":"status-publish","4":"format-standard","6":"category-genel","7":"cs-entry","8":"cs-video-wrap"},"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v20.5 (Yoast SEO v25.3.1) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#039;ler) Nedir? Neden \u00d6nemli?<\/title>\n<meta name=\"description\" content=\"Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM&#039;ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM&#039;leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP&#039;nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\" \/>\n<meta property=\"og:locale\" content=\"tr_TR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#039;ler) Nedir? Neden \u00d6nemli?\" \/>\n<meta property=\"og:description\" content=\"Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM&#039;ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM&#039;leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP&#039;nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\" \/>\n<meta property=\"og:site_name\" content=\"Kodlar\u0131n Gizemli D\u00fcnyas\u0131\" \/>\n<meta property=\"article:published_time\" content=\"2025-09-29T08:31:34+00:00\" \/>\n<meta name=\"author\" content=\"Fatih Soysal\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Yazan:\" \/>\n\t<meta name=\"twitter:data1\" content=\"Fatih Soysal\" \/>\n\t<meta name=\"twitter:label2\" content=\"Tahmini okuma s\u00fcresi\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 dakika\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\"},\"author\":{\"name\":\"Fatih Soysal\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"headline\":\"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#8217;ler) Nedir? Neden \u00d6nemli?\",\"datePublished\":\"2025-09-29T08:31:34+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\"},\"wordCount\":4373,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#respond\"]}],\"copyrightYear\":\"2025\",\"copyrightHolder\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\",\"name\":\"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM'ler) Nedir? Neden \u00d6nemli?\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\"},\"datePublished\":\"2025-09-29T08:31:34+00:00\",\"description\":\"Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM'ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM'leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP'nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.\",\"breadcrumb\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#breadcrumb\"},\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Anasayfa\",\"item\":\"https:\/\/fatihsoysal.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#8217;ler) Nedir? Neden \u00d6nemli?\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/\",\"name\":\"Fatihsoysal.com\",\"description\":\"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim\",\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"tr\"},{\"@type\":[\"Person\",\"Organization\"],\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\",\"name\":\"Fatih Soysal\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"tr\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"contentUrl\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"width\":512,\"height\":512,\"caption\":\"Fatih Soysal\"},\"logo\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\"},\"description\":\"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM'ler) Nedir? Neden \u00d6nemli?","description":"Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM'ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM'leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP'nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/","og_locale":"tr_TR","og_type":"article","og_title":"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM'ler) Nedir? Neden \u00d6nemli?","og_description":"Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM'ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM'leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP'nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.","og_url":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/","og_site_name":"Kodlar\u0131n Gizemli D\u00fcnyas\u0131","article_published_time":"2025-09-29T08:31:34+00:00","author":"Fatih Soysal","twitter_card":"summary_large_image","twitter_misc":{"Yazan:":"Fatih Soysal","Tahmini okuma s\u00fcresi":"6 dakika"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#article","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/"},"author":{"name":"Fatih Soysal","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"headline":"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#8217;ler) Nedir? Neden \u00d6nemli?","datePublished":"2025-09-29T08:31:34+00:00","mainEntityOfPage":{"@id":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/"},"wordCount":4373,"commentCount":0,"publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"inLanguage":"tr","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#respond"]}],"copyrightYear":"2025","copyrightHolder":{"@id":"https:\/\/fatihsoysal.com\/blog\/#organization"}},{"@type":"WebPage","@id":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/","url":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/","name":"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM'ler) Nedir? Neden \u00d6nemli?","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/#website"},"datePublished":"2025-09-29T08:31:34+00:00","description":"Yapay zeka teknolojileri h\u0131zla geli\u015firken, geli\u015ftiriciler de bu g\u00fcc\u00fc kendi uygulamalar\u0131na entegre etmenin yollar\u0131n\u0131 ar\u0131yor. B\u00fcy\u00fck Dil Modelleri (LLM'ler) alan\u0131nda ya\u015fanan devrim niteli\u011findeki geli\u015fmeler, pek \u00e7ok kap\u0131y\u0131 aralasa da, bulut tabanl\u0131 \u00e7\u00f6z\u00fcmlerin maliyeti ve veri gizlili\u011fi endi\u015feleri alternatif aray\u0131\u015flar\u0131n\u0131 beraberinde getiriyor. \u0130\u015fte tam bu noktada, yerel olarak LLM'leri \u00e7al\u0131\u015ft\u0131rabilen a\u00e7\u0131k kaynakl\u0131 bir platform olan Ollama devreye giriyor. Bu makalede, PHP'nin pop\u00fcler \u00e7at\u0131s\u0131 Laravel ile Ollama entegrasyonunu nas\u0131l sa\u011flayaca\u011f\u0131n\u0131z\u0131, s\u0131f\u0131rdan bir Laravel paketi geli\u015ftirerek ad\u0131m ad\u0131m \u00f6\u011freneceksiniz. Kendi uygulamalar\u0131n\u0131zda g\u00fc\u00e7l\u00fc yapay zeka yeteneklerini yerel ve g\u00fcvenli bir \u015fekilde kullanmak i\u00e7in gerekli t\u00fcm ad\u0131mlar\u0131 ve ipu\u00e7lar\u0131n\u0131 bu kapsaml\u0131 rehberde bulacaks\u0131n\u0131z.","breadcrumb":{"@id":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#breadcrumb"},"inLanguage":"tr","potentialAction":[{"@type":"ReadAction","target":["https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/fatihsoysal.com\/blog\/ollama-ve-buyuk-dil-modelleri-llmler-nedir-neden-onemli\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Anasayfa","item":"https:\/\/fatihsoysal.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Ollama ve B\u00fcy\u00fck Dil Modelleri (LLM&#8217;ler) Nedir? Neden \u00d6nemli?"}]},{"@type":"WebSite","@id":"https:\/\/fatihsoysal.com\/blog\/#website","url":"https:\/\/fatihsoysal.com\/blog\/","name":"Fatihsoysal.com","description":"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim","publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"tr"},{"@type":["Person","Organization"],"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1","name":"Fatih Soysal","image":{"@type":"ImageObject","inLanguage":"tr","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","contentUrl":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","width":512,"height":512,"caption":"Fatih Soysal"},"logo":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/"},"description":"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.","url":"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/"}]}},"yoast_meta":{"yoast_wpseo_title":"","yoast_wpseo_metadesc":"","yoast_wpseo_canonical":""},"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/30572","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/comments?post=30572"}],"version-history":[{"count":0,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/30572\/revisions"}],"wp:attachment":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/media?parent=30572"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/categories?post=30572"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/tags?post=30572"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}