{"id":2764,"date":"2025-08-31T15:10:51","date_gmt":"2025-08-31T15:10:51","guid":{"rendered":"https:\/\/www.markus-schall.de\/?p=2764"},"modified":"2025-11-24T09:17:41","modified_gmt":"2025-11-24T09:17:41","slug":"kendi-verileri-icin-evrensel-arama-motoru-olarak-ollama-ve-qdrant-ile-rag","status":"publish","type":"post","link":"https:\/\/www.markus-schall.de\/tr\/2025\/08\/kendi-verileri-icin-evrensel-arama-motoru-olarak-ollama-ve-qdrant-ile-rag\/","title":{"rendered":"Kendi verileri i\u00e7in evrensel bir arama motoru olarak Ollama ve Qdrant ile RAG"},"content":{"rendered":"<p>Giderek karma\u015f\u0131kla\u015fan bilgi d\u00fcnyas\u0131nda, kendi veritabanlar\u0131n\u0131z\u0131 hedefe y\u00f6nelik bir \u015fekilde aranabilir hale getirmek giderek daha \u00f6nemli hale geliyor - klasik tam metin aramalar\u0131 yoluyla de\u011fil, anlamsal olarak alakal\u0131 yan\u0131tlar yoluyla. \u0130\u015fte tam da bu noktada RAG veritaban\u0131 prensibi devreye giriyor - iki temel bile\u015fenden olu\u015fan yapay zeka destekli bir arama \u00e7\u00f6z\u00fcm\u00fc:<!--more--><\/p>\n<hr \/>\n\n\t\t\t<div class=\"display-post-types\">\n\n\t\t\t\t\t\t\t<style type=\"text\/css\">\n\t\t\t#dpt-wrapper-694 { --dpt-text-align: left;--dpt-image-crop: center;--dpt-border-radius: 5px;--dpt-small-grid-column: 33.33%;--dpt-large-grid-column: 33.3333333333%;--dpt-h-gutter: 10px;--dpt-v-gutter: 10px; }\t\t\t<\/style>\n\t\t\t<style type=\"text\/css\">#dpt-wrapper-694 { --dpt-title-font-style:normal;--dpt-title-font-weight:600;--dpt-title-line-height:1.5;--dpt-title-text-decoration:none;--dpt-title-text-transform:none;--dpt-excerpt-font-style:normal;--dpt-excerpt-font-weight:400;--dpt-excerpt-line-height:1.5;--dpt-excerpt-text-decoration:none;--dpt-excerpt-text-transform:none;--dpt-meta1-font-style:normal;--dpt-meta1-font-weight:400;--dpt-meta1-line-height:1.9;--dpt-meta1-text-decoration:none;--dpt-meta1-text-transform:none;--dpt-meta2-font-style:normal;--dpt-meta2-font-weight:400;--dpt-meta2-line-height:1.9;--dpt-meta2-text-decoration:none;--dpt-meta2-text-transform:none; }<\/style><div class=\"dpt-main-header\">\n\t\t\t\t\t\t<div class=\"dpt-main-title\">\n\t\t\t\t\t\t\t<span class=\"dpt-main-title-text\">Yapay zeka \u00fczerine g\u00fcncel konular<\/span>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t<\/div>\t\t\t\n\t\t\t\t<div id=\"dpt-wrapper-694\" class=\"dpt-wrapper dpt-grid1 multi-col dpt-mason-wrap\" >\n\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"dpt-entry has-thumbnail\" data-title=\"k\u00fcnstliche intelligenz: welche jobs in gefahr sind, und wie wir uns jetzt wappnen k\u00f6nnen\" data-id=\"2940\"  data-category=\"allgemein b\u00fccher gesellschaft ki-systeme\" data-post_tag=\"buch k\u00fcnstliche intelligenz llama llm mistral mlx ollama ratgeber sprachmodell\">\n\t\t\t\t\t\t\t<div class=\"dpt-entry-wrapper\"><div class=\"dpt-featured-content\"><div class=\"dpt-permalink\"><a href=\"https:\/\/www.markus-schall.de\/tr\/2025\/09\/yapay-zeka-hangi%cc%87-i%cc%87sler-tehli%cc%87kede-ve-si%cc%87mdi%cc%87-kendi%cc%87mi%cc%87zi%cc%87-nasil-koruyabi%cc%87li%cc%87ri%cc%87z\/\" class=\"dpt-permalink\"><span class=\"screen-reader-text\">Yapay zeka: hangi i\u015fler risk alt\u0131nda ve \u015fimdi kendimizi nas\u0131l silahland\u0131rabiliriz<\/span><\/a><\/div><div class=\"dpt-thumbnail\"><div class=\"dpt-thumbnail-inner\"><img width=\"1536\" height=\"1024\" class=\"attachment-full size-full\" alt=\"Gelecekte hangi i\u015fler yapay zeka taraf\u0131ndan ortadan kald\u0131r\u0131lacak?\" context=\"dpt\" data-dpt-src=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/welche-jobs-fallen-durch-ki-weg.jpg\" data-dpt-sizes=\"(max-width: 1536px) 100vw, 1536px\" data-dpt-srcset=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/welche-jobs-fallen-durch-ki-weg.jpg 1536w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/welche-jobs-fallen-durch-ki-weg-300x200.jpg 300w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/welche-jobs-fallen-durch-ki-weg-1024x683.jpg 1024w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/welche-jobs-fallen-durch-ki-weg-768x512.jpg 768w\" \/><\/div><span class=\"dpt-thumbnail-aspect-ratio\" style=\"padding-top: 67%\"><\/span><\/div><\/div><div class=\"sub-entry\"><h3 class=\"dpt-title\"><a class=\"dpt-title-link\" href=\"https:\/\/www.markus-schall.de\/tr\/2025\/09\/yapay-zeka-hangi%cc%87-i%cc%87sler-tehli%cc%87kede-ve-si%cc%87mdi%cc%87-kendi%cc%87mi%cc%87zi%cc%87-nasil-koruyabi%cc%87li%cc%87ri%cc%87z\/\" rel=\"bookmark\">Yapay zeka: hangi i\u015fler risk alt\u0131nda ve \u015fimdi kendimizi nas\u0131l silahland\u0131rabiliriz<\/a><\/h3><\/div><\/div>\n\t\t\t\t\t\t<\/div><!-- .dpt-entry -->\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"dpt-entry has-thumbnail\" data-title=\"mit ki dialogisch denken lernen: warum gute fragen wichtiger sind als gute modelle\" data-id=\"4700\"  data-category=\"allgemein ki-systeme tipps &amp; anleitungen\" data-post_tag=\"denkmodelle k\u00fcnstliche intelligenz lernen llm prozesse ratgeber sprachmodell\">\n\t\t\t\t\t\t\t<div class=\"dpt-entry-wrapper\"><div class=\"dpt-featured-content\"><div class=\"dpt-permalink\"><a href=\"https:\/\/www.markus-schall.de\/tr\/2026\/02\/ki-ile-diyalog-icinde-duesuenmeyi-oegrenmek-iyi-sorularin-neden-iyi-modellerden-daha-oenemli-oldugu\/\" class=\"dpt-permalink\"><span class=\"screen-reader-text\">Yapay zeka ile diyalog i\u00e7inde d\u00fc\u015f\u00fcnmeyi \u00f6\u011frenmek: \u0130yi sorular neden iyi modellerden daha \u00f6nemlidir?<\/span><\/a><\/div><div class=\"dpt-thumbnail\"><div class=\"dpt-thumbnail-inner\"><img width=\"1536\" height=\"1024\" class=\"attachment-full size-full\" alt=\"Yapay zeka ile diyalog i\u00e7inde d\u00fc\u015f\u00fcnmeyi \u00f6\u011frenmek\" context=\"dpt\" data-dpt-src=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/Denken-lernen-mit-KI.jpg\" data-dpt-sizes=\"(max-width: 1536px) 100vw, 1536px\" data-dpt-srcset=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/Denken-lernen-mit-KI.jpg 1536w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Denken-lernen-mit-KI-300x200.jpg 300w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Denken-lernen-mit-KI-1024x683.jpg 1024w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Denken-lernen-mit-KI-768x512.jpg 768w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Denken-lernen-mit-KI-18x12.jpg 18w\" \/><\/div><span class=\"dpt-thumbnail-aspect-ratio\" style=\"padding-top: 67%\"><\/span><\/div><\/div><div class=\"sub-entry\"><h3 class=\"dpt-title\"><a class=\"dpt-title-link\" href=\"https:\/\/www.markus-schall.de\/tr\/2026\/02\/ki-ile-diyalog-icinde-duesuenmeyi-oegrenmek-iyi-sorularin-neden-iyi-modellerden-daha-oenemli-oldugu\/\" rel=\"bookmark\">Yapay zeka ile diyalog i\u00e7inde d\u00fc\u015f\u00fcnmeyi \u00f6\u011frenmek: \u0130yi sorular neden iyi modellerden daha \u00f6nemlidir?<\/a><\/h3><\/div><\/div>\n\t\t\t\t\t\t<\/div><!-- .dpt-entry -->\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"dpt-entry has-thumbnail\" data-title=\"cloud-ki als oberlehrer: warum die zukunft des arbeitens bei lokaler ki liegt\" data-id=\"3887\"  data-category=\"apple macos hardware ki-systeme\" data-post_tag=\"datenschutz digitales eigentum k\u00fcnstliche intelligenz llama llm meinungsfreiheit mistral mlx neo4j ollama sprachmodell\">\n\t\t\t\t\t\t\t<div class=\"dpt-entry-wrapper\"><div class=\"dpt-featured-content\"><div class=\"dpt-permalink\"><a href=\"https:\/\/www.markus-schall.de\/tr\/2025\/12\/bulut-ki-en-i%cc%87yi%cc%87-oegretmen-olarak-i%cc%87si%cc%87n-gelecegi%cc%87-neden-yerel-kide-yatiyor\/\" class=\"dpt-permalink\"><span class=\"screen-reader-text\">Ba\u015f\u00f6\u011fretmen olarak bulut yapay zeka: \u0130\u015fin gelece\u011fi neden yerel yapay zekada yat\u0131yor?<\/span><\/a><\/div><div class=\"dpt-thumbnail\"><div class=\"dpt-thumbnail-inner\"><img width=\"1536\" height=\"1024\" class=\"attachment-full size-full\" alt=\"Bulut yapay zekas\u0131 ba\u015f\u00f6\u011fretmen oluyor\" context=\"dpt\" data-dpt-src=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/cloud-ki-oberlehrer.jpg\" data-dpt-sizes=\"(max-width: 1536px) 100vw, 1536px\" data-dpt-srcset=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/cloud-ki-oberlehrer.jpg 1536w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/cloud-ki-oberlehrer-300x200.jpg 300w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/cloud-ki-oberlehrer-1024x683.jpg 1024w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/cloud-ki-oberlehrer-768x512.jpg 768w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/cloud-ki-oberlehrer-18x12.jpg 18w\" \/><\/div><span class=\"dpt-thumbnail-aspect-ratio\" style=\"padding-top: 67%\"><\/span><\/div><\/div><div class=\"sub-entry\"><h3 class=\"dpt-title\"><a class=\"dpt-title-link\" href=\"https:\/\/www.markus-schall.de\/tr\/2025\/12\/bulut-ki-en-i%cc%87yi%cc%87-oegretmen-olarak-i%cc%87si%cc%87n-gelecegi%cc%87-neden-yerel-kide-yatiyor\/\" rel=\"bookmark\">Ba\u015f\u00f6\u011fretmen olarak bulut yapay zeka: \u0130\u015fin gelece\u011fi neden yerel yapay zekada yat\u0131yor?<\/a><\/h3><\/div><\/div>\n\t\t\t\t\t\t<\/div><!-- .dpt-entry -->\n\t\t\t\t\t\t\n\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\n<hr \/>\n<ol>\n<li>herhangi bir i\u00e7eri\u011fin say\u0131sal vekt\u00f6rler olarak depoland\u0131\u011f\u0131 bir vekt\u00f6r veritaban\u0131 (Qdrant gibi),<\/li>\n<li>ve ilgili talepleri uygun i\u00e7erikle ak\u0131ll\u0131 bir \u015fekilde birle\u015ftiren bir dil modeli (\u00f6rn. Ollama arac\u0131l\u0131\u011f\u0131yla).<\/li>\n<\/ol>\n<p>Modelin \"tahmin etmesine\" izin vermek yerine, bu mimari kendi bilgi kaynaklar\u0131n\u0131 kullan\u0131r - \u00f6rne\u011fin:<\/p>\n<ul>\n<li>kendi yazd\u0131\u011f\u0131 belgeseller,<\/li>\n<li>Web sitelerinin i\u00e7eri\u011fi,<\/li>\n<li>teknik k\u0131lavuzlar,<\/li>\n<li>Veritabanlar\u0131n\u0131 destekleyin,<\/li>\n<li>SSS listeleri,<\/li>\n<li>veya ar\u015fivlenmi\u015f metin kaynaklar\u0131 (\u00f6rne\u011fin eski veritabanlar\u0131ndan).<\/li>\n<\/ul>\n<p>Belirleyici fakt\u00f6r: T\u00fcm bu kaynaklar \u00f6nceden haz\u0131rlanabilir ve daha sonra bir kullan\u0131c\u0131 sorusu i\u00e7in en alakal\u0131 metin al\u0131nt\u0131lar\u0131n\u0131 sa\u011flayabilmek i\u00e7in \"par\u00e7alara ayr\u0131labilir\" (yani k\u00fc\u00e7\u00fck metin birimlerine ayr\u0131labilir).<\/p>\n<p>\u0130ster kendi bilgi veritaban\u0131n\u0131z\u0131, ister \u015firket i\u00e7i dok\u00fcmantasyonunuzu veya t\u00fcm bir \u00fcr\u00fcn ar\u015fivini analiz edilebilir hale getirmek isteyin - <a href=\"https:\/\/ollama.com\" target=\"_blank\" rel=\"noopener\"><strong>Ollama<\/strong><\/a> + <a href=\"https:\/\/qdrant.tech\" target=\"_blank\" rel=\"noopener\"><strong>Qdrant<\/strong><\/a> Bunu kendi Mac'inizde, herhangi bir bulut k\u0131s\u0131tlamas\u0131 olmadan ve veriler \u00fczerinde tam kontrol sahibi olarak yapabilirsiniz.<\/p>\n<h2>RAG veritaban\u0131 nedir - ve neden \"y\u0131\u011f\u0131nlama\"?<\/h2>\n<p>RAG, Retrieval-Augmented Generation'\u0131n k\u0131saltmas\u0131d\u0131r - ba\u015fka bir deyi\u015fle: destekli bilgi al\u0131m\u0131 ile metin \u00fcreten yapay zeka. GPT, Mistral veya LLaMA gibi bir dil modelini yaln\u0131zca \"bildikleri\" \u00fczerinde e\u011fitmek yerine, ba\u011fl\u0131 bir bilgi veritaban\u0131 (genellikle vekt\u00f6r veritaban\u0131 olarak adland\u0131r\u0131l\u0131r) arac\u0131l\u0131\u011f\u0131yla kendi ek bilgilerine eri\u015febilir.<\/p>\n<p><strong>\u00d6rnek:<\/strong><\/p>\n<blockquote><p>Bir dil modeline \"2023 vergi beyannamemde ne var?\" diye sorarsan\u0131z, orijinal verilere eri\u015fimi olmadan tahmin etmek zorunda kalacakt\u0131r. Ancak, bu belgenin yerel olarak depolanm\u0131\u015f, vekt\u00f6r tabanl\u0131 bir temsiline eri\u015fimi varsa, ilgili bilgileri alabilir ve yan\u0131t\u0131na dahil edebilir.<\/p><\/blockquote>\n<h3>\u0130\u00e7erik neden \"y\u0131\u011f\u0131nlan\u0131r\"?<\/h3>\n<p>Belgeler, web siteleri veya kitaplar genellikle tek seferde i\u015flenemeyecek veya aranamayacak kadar uzundur. Modern dil modellerinin de belirte\u00e7 s\u0131n\u0131rlar\u0131 vard\u0131r - yani bir kerede anlayabilecekleri s\u0131n\u0131rl\u0131 bir metin uzunlu\u011fu (genellikle yakla\u015f\u0131k 4.000-8.000 belirte\u00e7, hatta daha yeni modellerde 32.000 veya daha fazla).<\/p>\n<p>Bu y\u00fczden RAG a\u015fa\u011f\u0131daki hileyi kullan\u0131r:<\/p>\n<ol>\n<li>Orijinal metin k\u00fc\u00e7\u00fck b\u00f6l\u00fcmlere (par\u00e7alara) ayr\u0131lm\u0131\u015ft\u0131r.<\/li>\n<li>Her bir y\u0131\u011f\u0131n bir dil modeli (g\u00f6mme) taraf\u0131ndan bir vekt\u00f6re d\u00f6n\u00fc\u015ft\u00fcr\u00fcl\u00fcr.<\/li>\n<li>Bu vekt\u00f6rler Qdrant gibi bir veritaban\u0131nda saklan\u0131r.<\/li>\n<li>Kullan\u0131c\u0131 bir istekte bulundu\u011funda, istem de bir vekt\u00f6re \u00e7evrilir ve en benzer par\u00e7alar al\u0131n\u0131r.<\/li>\n<li>Bu i\u00e7erik daha sonra dil modeline eklenir - \u00f6rne\u011fin bir sistem istemi veya ba\u011flam enjeksiyonu yoluyla.<\/li>\n<\/ol>\n<p>Bu, bellek gibi davranan bir sistem yarat\u0131r - ancak klasik anahtar kelimeler veya tam metin aramas\u0131 olmadan, tamamen anlam tabanl\u0131 (semantik).<\/p>\n<h2>Gereksinimler ve hedef<\/h2>\n<p>A\u015fa\u011f\u0131dakilerden olu\u015fan yerel bir RAG sistemi kuruyoruz:<\/p>\n<ul>\n<li>Ollama arac\u0131l\u0131\u011f\u0131yla yerel bir LLM<\/li>\n<li>Qdrant adl\u0131 bir vekt\u00f6r veritaban\u0131<\/li>\n<li>metinleri par\u00e7alara ay\u0131ran, vekt\u00f6rle\u015ftiren ve veritaban\u0131na ekleyen bir Python beti\u011fi<\/li>\n<li>\u0130ste\u011fe ba\u011fl\u0131: sorgulama i\u00e7in basit bir aray\u00fcz veya API<\/li>\n<\/ul>\n<p><strong>Hedef platform: macOS (Intel veya Apple Silicon)<\/strong><\/p>\n<p>Bu bir \u00f6n ko\u015fuldur:<\/p>\n<ul>\n<li>macOS 12 veya daha yeni (Monterey veya \u00fcst\u00fc)<\/li>\n<li>Temel terminal bilgisi<\/li>\n<li>Python 3.10 veya daha yeni s\u00fcr\u00fcm<\/li>\n<li>\u0130ste\u011fe ba\u011fl\u0131: Homebrew installiert<\/li>\n<\/ul>\n<h3>Ad\u0131m 1: Ollama 1TP12Hayvan<\/h3>\n<p>Ollama, Mistral, LLaMA, Gemma veya Codellama gibi yerel dil modellerini internet olmadan da kendi bilgisayar\u0131n\u0131zda \u00e7al\u0131\u015ft\u0131rman\u0131z\u0131 sa\u011flayan yal\u0131n bir ara\u00e7t\u0131r.<\/p>\n<p>Mac \u00fczerinde kurulum:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">curl -fsSL https:\/\/ollama.com\/install.sh | sh<\/pre>\n<p>Alternatif olarak Ollama, Homebrew install \u00fczerinden de atanabilir:<br \/>\nbrew install ollama<\/p>\n<p>Kurulumdan sonra:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">ollama run mistral<\/pre>\n<p>Bu, Mistral 7B modelini indirir ve yerel olarak ba\u015flat\u0131r. Ollama, daha sonra vekt\u00f6rle\u015ftirme i\u00e7in kullanaca\u011f\u0131m\u0131z bir REST API ile birlikte gelir. Elbette Gemma3 (12B), Mistral Small (24B) veya di\u011fer LLM'ler gibi di\u011fer modelleri de kullanabilirsiniz.<\/p>\n<h3>Ad\u0131m 2: Qdrant 1TP12 hayvanlar\u0131 (yerel vekt\u00f6r veritaban\u0131)<\/h3>\n<p>Qdrant, Rust tabanl\u0131 y\u0131ld\u0131r\u0131m h\u0131z\u0131nda bir vekt\u00f6r veritaban\u0131d\u0131r. \u00dccretsiz, a\u00e7\u0131k kaynak kodlu ve Mac'te ba\u015flatmas\u0131 kolay - tercihen Docker arac\u0131l\u0131\u011f\u0131yla. Docker'\u0131 hen\u00fcz Mac 1TP12'nize y\u00fcklemediyseniz, \u015fu adresten indirebilirsiniz <a href=\"https:\/\/www.docker.com\" target=\"_blank\" rel=\"noopener\"><strong>Docker web sitesi<\/strong><\/a> \u00fccretsiz olarak y\u00fckleyebilir ve Mac'inizde normal bir masa\u00fcst\u00fc uygulamas\u0131 installieren olarak \u00e7al\u0131\u015ft\u0131rabilirsiniz. Alternatif olarak, zaten Homebrew kullan\u0131yorsan\u0131z Docker'\u0131 Homebrew install \u00fczerinden de y\u00fckleyebilirsiniz:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">brew install --cask docker<\/pre>\n<p>Ard\u0131ndan Docker arac\u0131l\u0131\u011f\u0131yla Qdrant'\u0131 ba\u015flat\u0131n:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">docker run -p 6333:6333 -v qdrant_storage:\/qdrant\/storage qdrant\/qdrant<\/pre>\n<p>Qdrant'a daha sonra \u015fu adresten ula\u015f\u0131labilir<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">http:\/\/localhost:6333<\/pre>\n<p>Test i\u00e7in:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">curl http:\/\/localhost:6333\/collections<\/pre>\n<h3>Ad\u0131m 3: Python ortam\u0131n\u0131 haz\u0131rlay\u0131n<\/h3>\n<p>Y\u0131\u011f\u0131nlama, g\u00f6mme ve Qdrant ile ileti\u015fim i\u00e7in Python'a ihtiyac\u0131m\u0131z var.<\/p>\n<p>Haz\u0131rl\u0131k:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">python3 -m venv rag-env\r\nsource rag-env\/bin\/activate\r\npip install qdrant-client sentence-transformers ollama numpy<\/pre>\n<p>E\u011fer ollama bir Python paketi olarak tan\u0131nm\u0131yorsa, REST API'yi do\u011frudan istekler arac\u0131l\u0131\u011f\u0131yla kullan\u0131n:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">pip install requests<\/pre>\n<h3>Ad\u0131m 4: Y\u0131\u011f\u0131nlama ve g\u00f6mme<\/h3>\n<p>A\u015fa\u011f\u0131da, bir metin belgesini par\u00e7alara ay\u0131ran, Ollama arac\u0131l\u0131\u011f\u0131yla kat\u0131\u015ft\u0131rmalar olu\u015fturan ve bunlar\u0131 Qdrant'a ekleyen \u00f6rnek bir kod bulacaks\u0131n\u0131z:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">import requests\r\nfrom qdrant_client import QdrantClient\r\nfrom qdrant_client.models import PointStruct\r\nimport uuid\r\n\r\n# Konfiguration\r\nCHUNK_SIZE = 500 # Zeichen\r\nCOLLECTION_NAME = \"mein_rag_wissen\"\r\n\r\n# Text vorbereiten\r\nwith open(\"mein_text.txt\", \"r\") as f:\r\ntext = f.read()\r\n\r\nchunks = [text[i:i+CHUNK_SIZE] for i in range(0, len(text), CHUNK_SIZE)]\r\n\r\n# Qdrant vorbereiten\r\nclient = QdrantClient(\"localhost\", port=6333)\r\n\r\n# Neue Collection anlegen (falls noch nicht vorhanden)\r\nclient.recreate_collection(\r\ncollection_name=COLLECTION_NAME,\r\nvectors_config={\"size\": 4096, \"distance\": \"Cosine\"}\r\n)\r\n\r\ndef get_embedding_ollama(text):\r\nresponse = requests.post(\r\n\"http:\/\/localhost:11434\/api\/embeddings\",\r\njson={\"model\": \"mistral\", \"prompt\": text}\r\n)\r\nreturn response.json()[\"embedding\"]\r\n\r\n# Embeddings erzeugen und in Qdrant speichern\r\npoints = []\r\nfor i, chunk in enumerate(chunks):\r\nvector = get_embedding_ollama(chunk)\r\npoints.append(PointStruct(\r\nid=str(uuid.uuid4()),\r\nvector=vector,\r\npayload={\"text\": chunk}\r\n))\r\n\r\nclient.upsert(collection_name=COLLECTION_NAME, points=points)\r\nprint(f\"{len(points)} Chunks erfolgreich eingef\u00fcgt.\")<\/pre>\n<h3>Ad\u0131m 5: Anlamsal arama yoluyla sorgular<\/h3>\n<p>Art\u0131k sorgular\u0131 Qdrant'a bir vekt\u00f6r olarak g\u00f6nderebilir ve en alakal\u0131 metin b\u00f6l\u00fcmlerini bulmas\u0131n\u0131 sa\u011flayabilirsiniz:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">query = \"Wie funktioniert ein RAG-System?\"\r\nquery_vector = get_embedding_ollama(query)\r\n\r\nresults = client.search(\r\ncollection_name=COLLECTION_NAME,\r\nquery_vector=query_vector,\r\nlimit=3\r\n)\r\n\r\nfor r in results:\r\nprint(r.payload[\"text\"])<\/pre>\n<p>Daha sonra bu par\u00e7alar\u0131 \u00f6rne\u011fin bir sistem istemi arac\u0131l\u0131\u011f\u0131yla Ollama'ye iletebilir ve bunlar\u0131n ba\u011flamla ilgili bir yan\u0131t olarak form\u00fcle edilmesini sa\u011flayabilirsiniz.<\/p>\n<h2>FileMaker ve di\u011fer veritabanlar\u0131na y\u0131\u011f\u0131nlama + JSON aktar\u0131m\u0131<\/h2>\n<p>Bir\u00e7ok durumda, y\u0131\u011f\u0131nlama mevcut bir veritaban\u0131 \u00e7\u00f6z\u00fcm\u00fcnde - \u00f6rne\u011fin FileMaker'de - zaten ger\u00e7ekle\u015ftirilebilir. Kendi \u00e7al\u0131\u015fma ortam\u0131mda tam olarak bu \u015fekilde \u00e7al\u0131\u015f\u0131yor: kaynak veriler - web sitesi i\u00e7eri\u011fi, destek giri\u015fleri veya teknik makaleler gibi - FileMaker'de yap\u0131land\u0131r\u0131lm\u0131\u015f bi\u00e7imde zaten mevcut.<\/p>\n<p>S\u00fcre\u00e7 bu \u015fekilde i\u015flemektedir:<\/p>\n<ol>\n<li>Metinler, FileMaker'nin kendi \u00f6bekleme mant\u0131\u011f\u0131 kullan\u0131larak \u00f6rne\u011fin 300-500 karakterlik b\u00f6l\u00fcmlere ayr\u0131l\u0131r.<\/li>\n<li>Her y\u0131\u011f\u0131na kendi kimli\u011fi ve varsa meta verileri (ba\u015fl\u0131k, kategori, kaynak, dil, vb.) verilir.<\/li>\n<li>T\u00fcm par\u00e7alar otomatik olarak JSON dosyalar\u0131 olarak d\u0131\u015fa aktar\u0131l\u0131r - \u00f6rne\u011fin bir a\u011f s\u00fcr\u00fcc\u00fcs\u00fcndeki belirli bir dizine veya do\u011frudan AI sunucusunun sabit s\u00fcr\u00fcc\u00fcs\u00fcne.<\/li>\n<li>Sunucudaki bir Python beti\u011fi bu JSON dosyalar\u0131n\u0131 okur ve Qdrant veritaban\u0131na kaydeder.<\/li>\n<\/ol>\n<h3>D\u0131\u015fa aktar\u0131lan bir y\u0131\u011f\u0131n dosyas\u0131 \u00f6rne\u011fi (chunk_00017.json)<\/h3>\n<pre class=\"notranslate\" data-no-translation=\"\">{\r\n\"id\": \"00017\",\r\n\"text\": \"Dies ist ein einzelner Textabschnitt mit ca. 400 Zeichen, der aus einer gr\u00f6\u00dferen Quelle stammt. Er wurde in FileMaker vorbereitet und enth\u00e4lt alle relevanten Inhalte, die f\u00fcr eine semantische Suche ben\u00f6tigt werden.\",\r\n\"metadata\": {\r\n\"source\": \"support_center\",\r\n\"category\": \"Fehlermeldung\",\r\n\"language\": \"de\",\r\n\"title\": \"Drucker wird nicht erkannt\"\r\n}\r\n}<\/pre>\n<p>Daha sonra i\u00e7e aktarma komut dosyas\u0131 otomatik olarak veya terminal \u00fczerinden d\u00fczenli olarak \u00e7al\u0131\u015ft\u0131r\u0131labilir - \u00f6rne\u011fin bir cron i\u015fi veya manuel \u00e7a\u011fr\u0131 yoluyla:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">python3 import_json_chunks.py \/Users\/markus\/Desktop\/chunks\/<\/pre>\n<p>Kod her bir JSON y\u0131\u011f\u0131n\u0131n\u0131 okur, ilgili vekt\u00f6r\u00fc olu\u015fturur (\u00f6rne\u011fin Ollama veya SentenceTransformers arac\u0131l\u0131\u011f\u0131yla) ve giri\u015fi Qdrant veritaban\u0131na aktar\u0131r.<\/p>\n<p>Bu y\u00f6ntem sadece \u015feffaf olmakla kalmaz, ayn\u0131 zamanda mevcut BT yap\u0131lar\u0131yla da \u00e7ok iyi bir \u015fekilde birle\u015ftirilebilir - \u00f6zellikle halihaz\u0131rda FileMaker kullanan veya s\u00fcre\u00e7 netli\u011fi nedeniyle her \u015feyi merkezi ve g\u00f6rsel olarak kontrol etmek isteyen \u015firketlerde.<\/p>\n<h2>T\u00fcm veritabanlar\u0131n\u0131 yerel yapay zekan\u0131za ba\u011flay\u0131n<\/h2>\n<p>Ollama ve Qdrant ile Mac \u00fczerinde k\u0131sa s\u00fcrede eksiksiz, y\u00fcksek performansl\u0131 bir RAG sistemi kurulabilir:<\/p>\n<ul>\n<li>Yerel, bulut veya abonelik olmadan<\/li>\n<li>Kendi i\u00e7eri\u011finizle geni\u015fletilebilir<\/li>\n<li>Bilgisayardan hi\u00e7bir \u015fey \u00e7\u0131kmad\u0131\u011f\u0131 i\u00e7in veriler g\u00fcvende<\/li>\n<li>Qdrant b\u00fcy\u00fck miktarda veride bile h\u0131zl\u0131 kald\u0131\u011f\u0131 i\u00e7in verimli<\/li>\n<\/ul>\n<p>Yapay zekan\u0131z\u0131 sadece sohbet etmek i\u00e7in de\u011fil, ger\u00e7ek bir bilgi ve haf\u0131za sistemi olarak kullanmak istiyorsan\u0131z, bu kombinasyon bir zorunluluktur. Ve kendi verileriniz \u00fczerinde tam kontrol ile \u00e7ok az \u00e7abayla \u00e7al\u0131\u015f\u0131r.<\/p>\n<hr \/>\n<h3>Yerel yapay zeka sistemlerinin kullan\u0131m\u0131na ili\u015fkin g\u00fcncel anket<\/h3>\n<div class='bootstrap-yop yop-poll-mc'>\n\t\t\t\t\t\t\t<div class=\"basic-yop-poll-container\" style=\"background-color:#ffffff; border:0px; border-style:solid; border-color:#000000; border-radius:5px; padding:0px 5px;\" data-id=\"9\" data-temp=\"basic-pretty\" data-skin=\"square\" data-cscheme=\"blue\" data-cap=\"0\" data-access=\"guest\" data-tid=\"\" data-uid=\"e34d2e8929d0f2a0884ab4abecd1d0a1\" data-pid=\"3887\" data-resdet=\"votes-number,percentages\" data-show-results-to=\"guest\" data-show-results-moment=\"after-vote\" data-show-results-only=\"false\" data-show-message=\"true\" data-show-results-as=\"bar\" data-sort-results-by=\"as-defined\" data-sort-results-rule=\"asc\"data-is-ended=\"0\" data-percentages-decimals=\"2\" data-gdpr=\"no\" data-gdpr-sol=\"consent\" data-css=\".basic-yop-poll-container[data-uid] .basic-vote {\t\t\t\t\t\t\t\t\ttext-align: center;\t\t\t\t\t\t\t\t}\" data-counter=\"0\" data-load-with=\"1\" data-notification-section=\"top\"><div class=\"row\"><div class=\"col-md-12\"><div class=\"basic-inner\"><div class=\"basic-message hide\" style=\"border-left: 10px solid #008000; padding: 0px 10px;\" data-error=\"#ff0000\" data-success=\"#008000\"><p class=\"basic-message-text\" style=\"color:#000000; font-size:14px; font-weight:normal;\"><\/p><\/div><div class=\"basic-overlay hide\"><div class=\"basic-vote-options\"><\/div><div class=\"basic-preloader\"><div class=\"basic-windows8\"><div class=\"basic-wBall basic-wBall_1\"><div class=\"basic-wInnerBall\"><\/div><\/div><div class=\"basic-wBall basic-wBall_2\"><div class=\"basic-wInnerBall\"><\/div><\/div><div class=\"basic-wBall basic-wBall_3\"><div class=\"basic-wInnerBall\"><\/div><\/div><div class=\"basic-wBall basic-wBall_4\"><div class=\"basic-wInnerBall\"><\/div><\/div><div class=\"basic-wBall basic-wBall_5\"><div class=\"basic-wInnerBall\"><\/div><\/div><\/div><\/div><\/div><form class=\"basic-form\" action=\"\"><input type=\"hidden\" name=\"_token\" value=\"dd098a18a0\" autocomplete=\"off\"><div class=\"basic-elements\"><div class=\"basic-element basic-question basic-question-text-vertical\" data-id=\"9\" data-uid=\"6c659558c6c7ca151051a13509ff0845\" data-type=\"question\" data-question-type=\"text\" data-required=\"yes\" data-allow-multiple=\"no\" data-min=\"1\" data-max=\"7\" data-display=\"vertical\" data-colnum=\"\" data-display-others=\"no\" data-others-color=\"\" data-others=\"\" data-others-max-chars=\"0\"><div class=\"basic-question-title\"><h5 style=\"color:#000000; font-size:16px; font-weight:normal; text-align:left;\">MLX veya Ollama gibi yerel olarak \u00e7al\u0131\u015fan yapay zeka yaz\u0131l\u0131mlar\u0131 hakk\u0131nda ne d\u00fc\u015f\u00fcn\u00fcyorsunuz?<\/h5><\/div><ul class=\"basic-answers\"><li class=\"basic-answer\" style=\"padding:0px 0px;\" data-id=\"45\" data-type=\"text\" data-vn=\"125\" data-color=\"#000000\" data-make-link=\"no\" data-link=\"\"><div class=\"basic-answer-content basic-text-vertical\"><label for=\"answer[45]\" class=\"basic-answer-label\"><input type=\"radio\" id=\"answer[45]\" name=\"answer[9]\" value=\"45\"><span class=\"basic-text\" style=\"color: #000000; font-size: 14px; font-weight: normal;\">Ingenious - nihayet buluttan ba\u011f\u0131ms\u0131z<\/span><\/label><\/div><\/li><li class=\"basic-answer\" style=\"padding:0px 0px;\" data-id=\"46\" data-type=\"text\" data-vn=\"24\" data-color=\"#000000\" data-make-link=\"no\" data-link=\"\"><div class=\"basic-answer-content basic-text-vertical\"><label for=\"answer[46]\" class=\"basic-answer-label\"><input type=\"radio\" id=\"answer[46]\" name=\"answer[9]\" value=\"46\"><span class=\"basic-text\" style=\"color: #000000; font-size: 14px; font-weight: normal;\">\u0130lgin\u00e7, ancak (hala) \u00e7ok karma\u015f\u0131k<\/span><\/label><\/div><\/li><li class=\"basic-answer\" style=\"padding:0px 0px;\" data-id=\"47\" data-type=\"text\" data-vn=\"25\" data-color=\"#000000\" data-make-link=\"no\" data-link=\"\"><div class=\"basic-answer-content basic-text-vertical\"><label for=\"answer[47]\" class=\"basic-answer-label\"><input type=\"radio\" id=\"answer[47]\" name=\"answer[9]\" value=\"47\"><span class=\"basic-text\" style=\"color: #000000; font-size: 14px; font-weight: normal;\">Yak\u0131nda deneyece\u011fim.<\/span><\/label><\/div><\/li><li class=\"basic-answer\" style=\"padding:0px 0px;\" data-id=\"48\" data-type=\"text\" data-vn=\"5\" data-color=\"#000000\" data-make-link=\"no\" data-link=\"\"><div class=\"basic-answer-content basic-text-vertical\"><label for=\"answer[48]\" class=\"basic-answer-label\"><input type=\"radio\" id=\"answer[48]\" name=\"answer[9]\" value=\"48\"><span class=\"basic-text\" style=\"color: #000000; font-size: 14px; font-weight: normal;\">\u0130htiyac\u0131m yok - bulut benim i\u00e7in yeterli<\/span><\/label><\/div><\/li><li class=\"basic-answer\" style=\"padding:0px 0px;\" data-id=\"49\" data-type=\"text\" data-vn=\"3\" data-color=\"#000000\" data-make-link=\"no\" data-link=\"\"><div class=\"basic-answer-content basic-text-vertical\"><label for=\"answer[49]\" class=\"basic-answer-label\"><input type=\"radio\" id=\"answer[49]\" name=\"answer[9]\" value=\"49\"><span class=\"basic-text\" style=\"color: #000000; font-size: 14px; font-weight: normal;\">Bunun tam olarak ne hakk\u0131nda oldu\u011funu bilmiyorum.<\/span><\/label><\/div><\/li><\/ul><\/div><div class=\"clearfix\"><\/div><\/div><div class=\"basic-vote\"><a href=\"#\" class=\"button basic-vote-button\" role=\"button\" style=\"background:#027bb8; border:0px; border-style: solid; border-color:#000000; border-radius:5px; padding:10px 10px; color:#ffffff; font-size:14px; font-weight:normal;\">Oylama<\/a><\/div><input type=\"hidden\" name=\"trp-form-language\" value=\"tr\"\/><\/form><\/div><\/div><\/div><\/div>\n\t\t\t\t\t\t<\/div>\n<hr \/>\n<h2>G\u00f6r\u00fcn\u00fcm: RAG, Ollama ve Qdrant ile neler m\u00fcmk\u00fcn<\/h2>\n<p>Bu makalede anlat\u0131lan d\u00fczenek, bilgiyle ba\u015fa \u00e7\u0131kman\u0131n yeni bir yolunun teknik temelini olu\u015fturuyor - yerel, kontroll\u00fc ve esnek bir \u015fekilde geni\u015fletilebilir. Ancak yolculuk hi\u00e7bir \u015fekilde burada bitmiyor. Y\u0131\u011f\u0131nlama, g\u00f6mme, anlamsal arama ve dil modellerinin etkile\u015fimini anlad\u0131ktan sonra, bu mimarinin pratikte ne kadar \u00e7ok y\u00f6nl\u00fc oldu\u011funu \u00e7abucak fark edeceksiniz.<\/p>\n<h3>1. kendi veritabanlar\u0131na ba\u011flant\u0131<\/h3>\n<p>\u0130ster FileMaker, MySQL, PostgreSQL veya MongoDB olsun - herhangi bir i\u00e7erik d\u00fczenli olarak ay\u0131klanabilir, par\u00e7alara ayr\u0131labilir ve hedeflenen sorgular kullan\u0131larak vekt\u00f6r veritaban\u0131na otomatik olarak eklenebilir. Bu, klasik bir veritaban\u0131n\u0131 semantik olarak aranabilir bir bilgi kayna\u011f\u0131na d\u00f6n\u00fc\u015ft\u00fcr\u00fcr. \u00d6zellikle destek sistemlerinde, \u00fcr\u00fcn ar\u015fivlerinde veya dijital k\u00fct\u00fcphanelerde bu, \u00e7al\u0131\u015fanlar veya m\u00fc\u015fteriler i\u00e7in tamamen yeni eri\u015fim se\u00e7eneklerinin \u00f6n\u00fcn\u00fc a\u00e7ar.<\/p>\n<h3>2. Web sayfalar\u0131n\u0131n, PDF'lerin veya belgelerin otomatik olarak i\u00e7e aktar\u0131lmas\u0131<\/h3>\n<p>\u0130\u00e7eri\u011fin manuel olarak aktar\u0131lmas\u0131 gerekmez. BeautifulSoup, readability, pdfplumber veya docx2txt gibi ara\u00e7larla t\u00fcm web siteleri, PDF k\u0131lavuzlar\u0131 veya Word belgeleri otomatik olarak i\u00e7e aktar\u0131labilir, metin formuna d\u00f6n\u00fc\u015ft\u00fcr\u00fclebilir ve y\u0131\u011f\u0131nlama i\u00e7in haz\u0131rlanabilir. \u00d6rne\u011fin, teknik wiki'ler, m\u00fc\u015fteri portallar\u0131 veya \u00e7evrimi\u00e7i belgeler d\u00fczenli olarak g\u00fcncellenebilir ve RAG veritaban\u0131na beslenebilir.<\/p>\n<h3>3. Yap\u0131land\u0131rma yoluyla uzun vadeli bilgi birikimi olu\u015fturma<\/h3>\n<p>Her soruyla s\u0131f\u0131rdan ba\u015flayan klasik bir yapay zeka uygulamas\u0131n\u0131n aksine, bir RAG kurulumu, temel bilginin ad\u0131m ad\u0131m geni\u015fletilmesine ve iyile\u015ftirilmesine olanak tan\u0131r. Par\u00e7alar\u0131n hedefli se\u00e7imi ve haz\u0131rlanmas\u0131, her giri\u015fte daha de\u011ferli hale gelen kendi semantik haf\u0131zas\u0131n\u0131 yarat\u0131r.<\/p>\n<h3>4. Bilgi grafikleri ile ba\u011flant\u0131 (Neo4j)<\/h3>\n<p>Bir ad\u0131m daha ileri gitmek isterseniz, bilgileri yaln\u0131zca semantik olarak depolamakla kalmaz, ayn\u0131 zamanda mant\u0131ksal olarak da ba\u011flayabilirsiniz. Bir grafik veritaban\u0131 olan Neo4j ile terimler, ki\u015filer, konular veya kategoriler aras\u0131ndaki ili\u015fkiler g\u00f6rselle\u015ftirilebilir ve \u00f6zel olarak sorgulanabilir. Bu, bir metin koleksiyonunu hem insanlar hem de yapay zeka taraf\u0131ndan kullan\u0131labilen yap\u0131land\u0131r\u0131lm\u0131\u015f bir bilgi grafi\u011fine d\u00f6n\u00fc\u015ft\u00fcr\u00fcr - \u00f6rne\u011fin nedensel zincirleri, zamansal dizileri veya tematik k\u00fcmeleri g\u00f6rselle\u015ftirmek i\u00e7in.<\/p>\n<h3>5. Kendi ara\u00e7lar\u0131n\u0131zda, uygulamalar\u0131n\u0131zda veya sohbet botlar\u0131n\u0131zda kullan\u0131n<\/h3>\n<p>Bir kez kurulduktan sonra, RAG mant\u0131\u011f\u0131 hemen hemen her uygulamaya entegre edilebilir: dahili bir web uygulamas\u0131nda semantik bir arama i\u015flevi olarak, bir CRM sisteminde ak\u0131ll\u0131 bir giri\u015f yard\u0131m\u0131 olarak veya \u015firket web sitesinde kendi uzmanl\u0131\u011f\u0131na sahip bir sohbet robotu olarak. Yerel API'ler (\u00f6rne\u011fin Ollama REST ve Qdrant gRPC) kullan\u0131larak, t\u00fcm bile\u015fenler esnek ve geni\u015fletilebilir kal\u0131r - geleneksel \u015firket s\u0131n\u0131rlar\u0131n\u0131n \u00f6tesinde bile.<\/p>\n<p>Bu ara\u00e7lara a\u015fina olma cesaretini g\u00f6sterenler, kontrol, egemenlik ve teknik netlik ruhu i\u00e7inde, ger\u00e7ek kullan\u0131m de\u011ferine sahip ba\u011f\u0131ms\u0131z, yerel yapay zeka sistemlerinin temelini olu\u015ftururlar.<\/p>\n<hr \/>\n\n\t\t\t<div class=\"display-post-types\">\n\n\t\t\t\t\t\t\t<style type=\"text\/css\">\n\t\t\t#dpt-wrapper-695 { --dpt-text-align: left;--dpt-image-crop: center;--dpt-border-radius: 5px;--dpt-small-grid-column: 33.33%;--dpt-large-grid-column: 33.3333333333%;--dpt-h-gutter: 10px;--dpt-v-gutter: 10px; }\t\t\t<\/style>\n\t\t\t<style type=\"text\/css\">#dpt-wrapper-695 { --dpt-title-font-style:normal;--dpt-title-font-weight:600;--dpt-title-line-height:1.5;--dpt-title-text-decoration:none;--dpt-title-text-transform:none;--dpt-excerpt-font-style:normal;--dpt-excerpt-font-weight:400;--dpt-excerpt-line-height:1.5;--dpt-excerpt-text-decoration:none;--dpt-excerpt-text-transform:none;--dpt-meta1-font-style:normal;--dpt-meta1-font-weight:400;--dpt-meta1-line-height:1.9;--dpt-meta1-text-decoration:none;--dpt-meta1-text-transform:none;--dpt-meta2-font-style:normal;--dpt-meta2-font-weight:400;--dpt-meta2-line-height:1.9;--dpt-meta2-text-decoration:none;--dpt-meta2-text-transform:none; }<\/style><div class=\"dpt-main-header\">\n\t\t\t\t\t\t<div class=\"dpt-main-title\">\n\t\t\t\t\t\t\t<span class=\"dpt-main-title-text\">ERP yaz\u0131l\u0131m\u0131 ile ilgili g\u00fcncel konular<\/span>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t<\/div>\t\t\t\n\t\t\t\t<div id=\"dpt-wrapper-695\" class=\"dpt-wrapper dpt-grid1 multi-col dpt-mason-wrap\" >\n\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"dpt-entry has-thumbnail\" data-title=\"sehen wir uns auf der filemaker-konferenz fmk 2025 in hamburg?\" data-id=\"2670\"  data-category=\"filemaker &amp; erp\" data-post_tag=\"erp-software filemaker gfm-business\">\n\t\t\t\t\t\t\t<div class=\"dpt-entry-wrapper\"><div class=\"dpt-featured-content\"><div class=\"dpt-permalink\"><a href=\"https:\/\/www.markus-schall.de\/tr\/2025\/08\/hamburgdaki%cc%87-fmk-2025-fi%cc%87lemaker-konferansinda-goeruesmek-uezere\/\" class=\"dpt-permalink\"><span class=\"screen-reader-text\">Hamburg'daki FileMaker konferans\u0131 FMK 2025'te g\u00f6r\u00fc\u015fmek \u00fczere?<\/span><\/a><\/div><div class=\"dpt-thumbnail\"><div class=\"dpt-thumbnail-inner\"><img width=\"2560\" height=\"491\" class=\"attachment-full size-full\" alt=\"FileMaker Konferans FMK2025\" context=\"dpt\" data-dpt-src=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-scaled.jpg\" data-dpt-sizes=\"(max-width: 2560px) 100vw, 2560px\" data-dpt-srcset=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-scaled.jpg 2560w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-300x58.jpg 300w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-1024x196.jpg 1024w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-768x147.jpg 768w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-1536x295.jpg 1536w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-2048x393.jpg 2048w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/filemaker-konferenz-fmk2025-500x96.jpg 500w\" \/><\/div><span class=\"dpt-thumbnail-aspect-ratio\" style=\"padding-top: 19%\"><\/span><\/div><\/div><div class=\"sub-entry\"><h3 class=\"dpt-title\"><a class=\"dpt-title-link\" href=\"https:\/\/www.markus-schall.de\/tr\/2025\/08\/hamburgdaki%cc%87-fmk-2025-fi%cc%87lemaker-konferansinda-goeruesmek-uezere\/\" rel=\"bookmark\">Hamburg'daki FileMaker konferans\u0131 FMK 2025'te g\u00f6r\u00fc\u015fmek \u00fczere?<\/a><\/h3><\/div><\/div>\n\t\t\t\t\t\t<\/div><!-- .dpt-entry -->\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"dpt-entry has-thumbnail\" data-title=\"digitales eigentum erkl\u00e4rt \u2013 so entstehen nachhaltige online-verm\u00f6genswerte\" data-id=\"4766\"  data-category=\"allgemein b\u00fccher featured gesellschaft ki-systeme\" data-post_tag=\"buch datenbanken denkmodelle digitales eigentum erp-software k\u00fcnstliche intelligenz prozesse publishing\">\n\t\t\t\t\t\t\t<div class=\"dpt-entry-wrapper\"><div class=\"dpt-featured-content\"><div class=\"dpt-permalink\"><a href=\"https:\/\/www.markus-schall.de\/tr\/2026\/02\/di%cc%87ji%cc%87tal-muelki%cc%87yet-suerdueruelebi%cc%87li%cc%87r-cevri%cc%87mi%cc%87ci%cc%87-varliklarin-nasil-olusturuldugunu-acikladi\/\" class=\"dpt-permalink\"><span class=\"screen-reader-text\">Dijital sahiplik a\u00e7\u0131kland\u0131 - S\u00fcrd\u00fcr\u00fclebilir \u00e7evrimi\u00e7i varl\u0131klar nas\u0131l olu\u015fturulur?<\/span><\/a><\/div><div class=\"dpt-thumbnail\"><div class=\"dpt-thumbnail-inner\"><img width=\"1536\" height=\"1024\" class=\"attachment-full size-full\" alt=\"Dijital m\u00fclk nedir\" context=\"dpt\" data-dpt-src=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/Digitales-Eigentum.jpg\" data-dpt-sizes=\"(max-width: 1536px) 100vw, 1536px\" data-dpt-srcset=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/Digitales-Eigentum.jpg 1536w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Digitales-Eigentum-300x200.jpg 300w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Digitales-Eigentum-1024x683.jpg 1024w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Digitales-Eigentum-768x512.jpg 768w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/Digitales-Eigentum-18x12.jpg 18w\" \/><\/div><span class=\"dpt-thumbnail-aspect-ratio\" style=\"padding-top: 67%\"><\/span><\/div><\/div><div class=\"sub-entry\"><h3 class=\"dpt-title\"><a class=\"dpt-title-link\" href=\"https:\/\/www.markus-schall.de\/tr\/2026\/02\/di%cc%87ji%cc%87tal-muelki%cc%87yet-suerdueruelebi%cc%87li%cc%87r-cevri%cc%87mi%cc%87ci%cc%87-varliklarin-nasil-olusturuldugunu-acikladi\/\" rel=\"bookmark\">Dijital sahiplik a\u00e7\u0131kland\u0131 - S\u00fcrd\u00fcr\u00fclebilir \u00e7evrimi\u00e7i varl\u0131klar nas\u0131l olu\u015fturulur?<\/a><\/h3><\/div><\/div>\n\t\t\t\t\t\t<\/div><!-- .dpt-entry -->\n\t\t\t\t\t\t\t\t\t\t\t\t<div class=\"dpt-entry has-thumbnail\" data-title=\"forum zur gfm-business erp-software und claris filemaker\" data-id=\"2393\"  data-category=\"filemaker &amp; erp\" data-post_tag=\"erp-software filemaker gfm-business\">\n\t\t\t\t\t\t\t<div class=\"dpt-entry-wrapper\"><div class=\"dpt-featured-content\"><div class=\"dpt-permalink\"><a href=\"https:\/\/www.markus-schall.de\/tr\/2023\/10\/gfm-is-erp-yazilimi-ve-claris-filemaker-uezerine-forum\/\" class=\"dpt-permalink\"><span class=\"screen-reader-text\">gFM-Business ERP yaz\u0131l\u0131m\u0131 ve Claris FileMaker i\u00e7in forum<\/span><\/a><\/div><div class=\"dpt-thumbnail\"><div class=\"dpt-thumbnail-inner\"><img width=\"1920\" height=\"1515\" class=\"attachment-full size-full\" alt=\"ERP yaz\u0131l\u0131m\u0131 ve FileMaker Forumu\" context=\"dpt\" data-dpt-src=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum.jpg\" data-dpt-sizes=\"(max-width: 1920px) 100vw, 1920px\" data-dpt-srcset=\"https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum.jpg 1920w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum-300x237.jpg 300w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum-1024x808.jpg 1024w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum-768x606.jpg 768w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum-1536x1212.jpg 1536w, https:\/\/www.markus-schall.de\/wp-content\/uploads\/erp-software-filemaker-forum-380x300.jpg 380w\" \/><\/div><span class=\"dpt-thumbnail-aspect-ratio\" style=\"padding-top: 79%\"><\/span><\/div><\/div><div class=\"sub-entry\"><h3 class=\"dpt-title\"><a class=\"dpt-title-link\" href=\"https:\/\/www.markus-schall.de\/tr\/2023\/10\/gfm-is-erp-yazilimi-ve-claris-filemaker-uezerine-forum\/\" rel=\"bookmark\">gFM-Business ERP yaz\u0131l\u0131m\u0131 ve Claris FileMaker i\u00e7in forum<\/a><\/h3><\/div><\/div>\n\t\t\t\t\t\t<\/div><!-- .dpt-entry -->\n\t\t\t\t\t\t\n\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\n<hr \/>\n<h2>Ollama + Qdrant ile RAG hakk\u0131nda s\u0131k\u00e7a sorulan sorular<\/h2>\n<h3>1. RAG veritaban\u0131 nedir ve ne i\u015fe yarar?<\/h3>\n<p>Bir RAG veritaban\u0131 (Retrieval Augmented Generation) bir vekt\u00f6r veritaban\u0131n\u0131 bir dil modeliyle birle\u015ftirir. Yapay zeka modellerinin \u00f6zellikle kendi veritaban\u0131n\u0131z\u0131n ilgili b\u00f6l\u00fcmlerine eri\u015febilmesi i\u00e7in kendi i\u00e7eri\u011finizi (\u00f6rne\u011fin belgeler veya web siteleri) semantik olarak aranabilir hale getirmenizi m\u00fcmk\u00fcn k\u0131lar.<\/p>\n<h3>2 Bu ba\u011flamda \"y\u0131\u011f\u0131nlama\" ne anlama gelmektedir?<\/h3>\n<p>Par\u00e7alara ay\u0131rma, uzun metinleri daha k\u00fc\u00e7\u00fck, anlaml\u0131 bir \u015fekilde tutarl\u0131 b\u00f6l\u00fcmlere (par\u00e7alara) ay\u0131rmak anlam\u0131na gelir - genellikle 200 ila 500 karakter aras\u0131nda. Bu, tek tek metin b\u00f6l\u00fcmlerinin vekt\u00f6r veritaban\u0131na verimli bir \u015fekilde kaydedilmesini ve sorular ortaya \u00e7\u0131kt\u0131\u011f\u0131nda daha sonra geri getirilmesini sa\u011flar.<\/p>\n<h3>3. Neden t\u00fcm metinleri Qdrant'a kaydedemiyorsunuz?<\/h3>\n<p>\u00c7\u00fcnk\u00fc yapay zeka modelleri ve vekt\u00f6r aramalar\u0131 s\u0131n\u0131rl\u0131 metin uzunluklar\u0131yla \u00e7al\u0131\u015f\u0131r. B\u00fcy\u00fck belgeler \u00f6nemli i\u00e7eri\u011fi \"gizler\" ya da hatal\u0131 hale getirir. Y\u0131\u011f\u0131nlama do\u011frulu\u011fu art\u0131r\u0131r \u00e7\u00fcnk\u00fc tam metinler yerine belirli b\u00f6l\u00fcmler kar\u015f\u0131la\u015ft\u0131r\u0131l\u0131r.<\/p>\n<h3>4. Herhangi bir kaynaktan i\u00e7erik kullanabilir miyim?<\/h3>\n<p>Evet, metinleri d\u00fczenlenebilir bir bi\u00e7imde (\u00f6rne\u011fin d\u00fcz metin, HTML, Markdown, PDF, FileMaker girdileri vb. olarak) ald\u0131\u011f\u0131n\u0131z s\u00fcrece, bunlar\u0131 haz\u0131rlayabilir, par\u00e7alara ay\u0131rabilir ve Qdrant'a entegre edebilirsiniz. Kar\u0131\u015f\u0131k kaynaklar da m\u00fcmk\u00fcnd\u00fcr.<\/p>\n<h3>5. B\u00f6yle bir sistem kurmak i\u00e7in programlama yapabilmem gerekir mi?<\/h3>\n<p>Temel Terminal ve Python bilgisi yard\u0131mc\u0131 olur, ancak gerekli de\u011fildir. Bir\u00e7ok ad\u0131m (\u00f6rne\u011fin FileMaker'de y\u0131\u011f\u0131nlama, JSON d\u0131\u015fa aktarma) g\u00f6rsel ve otomatik olarak uygulanabilir. Qdrant i\u00e7e aktarma beti\u011fi kolayca \u00f6zelle\u015ftirilebilir.<\/p>\n<h3>6. Birka\u00e7 belge veya kategoriyi de y\u00f6netebilir miyim?<\/h3>\n<p>Evet, her y\u0131\u011f\u0131n meta veri i\u00e7erebilir - \u00f6rne\u011fin ba\u015fl\u0131k, kaynak, dil veya kategori. Bunlar, sonu\u00e7lar\u0131 daha spesifik olarak filtrelemek i\u00e7in arama s\u0131ras\u0131nda dikkate al\u0131nabilir.<\/p>\n<h3>7 Yerle\u015ftirme \u00fcretimi i\u00e7in hangi modeller uygundur?<\/h3>\n<p>Ollama arac\u0131l\u0131\u011f\u0131yla yerel bir model (\u00f6rne\u011fin mistral, llama2, gemma) ya da c\u00fcmle d\u00f6n\u00fc\u015ft\u00fcr\u00fcc\u00fclerden all-MiniLM gibi ayr\u0131 bir g\u00f6mme modeli kullanabilirsiniz. Modelin g\u00f6mme \u00e7\u0131kt\u0131lar\u0131n\u0131 vekt\u00f6r olarak \u00fcretmesi \u00f6nemlidir.<\/p>\n<h3>8. Mac'te Qdrant'\u0131 nas\u0131l ba\u015flatabilirim?<\/h3>\n<p>En kolay yol Docker komutudur:<\/p>\n<pre class=\"notranslate\" data-no-translation=\"\">docker run -p 6333:6333 -v qdrant_storage:\/qdrant\/storage qdrant\/qdrant<\/pre>\n<p>Qdrant daha sonra http:\/\/localhost:6333 alt\u0131nda \u00e7al\u0131\u015f\u0131r<\/p>\n<h3>9. veri\u0307 hacmi\u0307m ne kadar b\u00fcy\u00fck olabi\u0307li\u0307r?<\/h3>\n<p>Qdrant \u00e7ok performansl\u0131d\u0131r ve on binlerce veya y\u00fcz binlerce par\u00e7ay\u0131 kolayca y\u00f6netebilir. Ana s\u0131n\u0131rlama RAM ve depolama alan\u0131d\u0131r, say\u0131 de\u011fil.<\/p>\n<h3>10. Bu FileMaker ile de \u00e7al\u0131\u015f\u0131yor mu?<\/h3>\n<p>Evet, t\u00fcm y\u0131\u011f\u0131nlama ve JSON d\u0131\u015fa aktarma i\u015flemlerini do\u011frudan FileMaker'de yapabilirsiniz. Par\u00e7alar ayr\u0131 JSON dosyalar\u0131 olarak d\u0131\u015fa aktar\u0131l\u0131r ve bunlar daha sonra bir Python beti\u011fi arac\u0131l\u0131\u011f\u0131yla Qdrant'a aktar\u0131l\u0131r - orijinal sistemden tamamen ba\u011f\u0131ms\u0131z olarak.<\/p>\n<h3>11. Bunu Mac yerine ba\u015fka bir sunucuda da \u00e7al\u0131\u015ft\u0131rabilir miyim?<\/h3>\n<p>Kesinlikle. Bu kurulum Linux sunucular\u0131nda, Raspberry Pi'de veya bulutta (istenirse) da \u00e7al\u0131\u015f\u0131r. Docker bunu platformdan ba\u011f\u0131ms\u0131z hale getirir. Verimli kullan\u0131m i\u00e7in genellikle daha fazla RAM ve GPU deste\u011fine sahip bir sunucu \u00f6nerilir.<\/p>\n<h3>12. Vekt\u00f6r aramas\u0131n\u0131 Ollama ile nas\u0131l birle\u015ftirebilirim?<\/h3>\n<p>\u00d6nce Ollama (Embedding API) arac\u0131l\u0131\u011f\u0131yla bir kullan\u0131c\u0131 sorusu i\u00e7in bir vekt\u00f6r olu\u015fturursunuz, bunu Qdrant'ta en alakal\u0131 par\u00e7alar\u0131 aramak i\u00e7in kullan\u0131rs\u0131n\u0131z ve bunlar\u0131 ba\u011flam olarak dil modeline verirsiniz. Ollama daha sonra soruyu + ba\u011flamla ilgili bilgileri i\u015fler ve sa\u011flam temelli bir yan\u0131t olu\u015fturur.<\/p>\n<p>Resim (c) geralt @ pixabay<\/p>","protected":false},"excerpt":{"rendered":"<p>Giderek karma\u015f\u0131kla\u015fan bilgi d\u00fcnyas\u0131nda, kendi veritabanlar\u0131n\u0131z\u0131 hedefe y\u00f6nelik bir \u015fekilde aranabilir hale getirmek giderek daha \u00f6nemli hale geliyor - klasik tam metin aramalar\u0131 yoluyla de\u011fil, anlamsal olarak alakal\u0131 yan\u0131tlar yoluyla. \u0130\u015fte tam da bu noktada RAG veritaban\u0131 prensibi devreye giriyor - iki temel bile\u015fenden olu\u015fan yapay zeka destekli bir arama \u00e7\u00f6z\u00fcm\u00fc:<\/p>","protected":false},"author":1,"featured_media":2768,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"iawp_total_views":692,"footnotes":""},"categories":[431,3],"tags":[440,452,471,435,433,437,432,450,451],"class_list":["post-2764","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ki-systeme","category-filemaker","tag-datenbanken","tag-docker","tag-kuenstliche-intelligenz","tag-llama","tag-llm","tag-mistral","tag-ollama","tag-qdrant","tag-vektordatenbank"],"_links":{"self":[{"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/posts\/2764","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/comments?post=2764"}],"version-history":[{"count":14,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/posts\/2764\/revisions"}],"predecessor-version":[{"id":3401,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/posts\/2764\/revisions\/3401"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/media\/2768"}],"wp:attachment":[{"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/media?parent=2764"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/categories?post=2764"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.markus-schall.de\/tr\/wp-json\/wp\/v2\/tags?post=2764"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}