Protokoll #29680

{
    "event": "PreToolUse",
    "tool_name": "mcp__mcp-docs__docs_get",
    "tool_input": {
        "id": 21
    }
}

{
    "tool_response": [
        {
            "type": "text",
            "text": "{\n  \"success\": true,\n  \"doc\": {\n    \"id\": 21,\n    \"parent_id\": 16,\n    \"slug\": \"ki-embedding\",\n    \"path\": \"\/ki-system\/ki-embedding\",\n    \"title\": \"Embedding\",\n    \"description\": \"Vektorerzeugung für semantische Suche und RAG.\",\n    \"content\": \"<nav class=\\\"breadcrumb\\\">\\n    <a href=\\\"\/docs\\\">Dokumentation<\/a> &raquo; <a href=\\\"\/docs\/ki-system\\\">KI-System<\/a> &raquo; Embedding\\n<\/nav>\\n\\n<h1>Embedding<\/h1>\\n<p class=\\\"doc-meta\\\"><strong>Erstellt:<\/strong> 2025-12-20 | <strong>Aktualisiert:<\/strong> 2025-12-20<\/p>\\n\\n<p>Vektorerzeugung für semantische Suche und RAG.<\/p>\\n\\n<table>\\n    <tr><th>Model<\/th><td>mxbai-embed-large<\/td><\/tr>\\n    <tr><th>Dimensionen<\/th><td>1024<\/td><\/tr>\\n    <tr><th>Provider<\/th><td>Ollama (lokal)<\/td><\/tr>\\n    <tr><th>Fallback<\/th><td>OpenAI (optional)<\/td><\/tr>\\n<\/table>\\n\\n<h2>Qdrant Collections<\/h2>\\n<table>\\n    <tr><th>Collection<\/th><th>Zweck<\/th><th>Dimensionen<\/th><\/tr>\\n    <tr><td>documents<\/td><td>Dokument-Chunks<\/td><td>1024<\/td><\/tr>\\n    <tr><td>mail<\/td><td>E-Mail-Inhalte<\/td><td>1024<\/td><\/tr>\\n    <tr><td>entities<\/td><td>Entitäten-Embeddings<\/td><td>1024<\/td><\/tr>\\n<\/table>\\n\\n<h2>Qdrant-Konfiguration<\/h2>\\n<pre><code>{\\n  \\\"vectors\\\": {\\n    \\\"size\\\": 1024,\\n    \\\"distance\\\": \\\"Cosine\\\"\\n  },\\n  \\\"hnsw_config\\\": {\\n    \\\"m\\\": 16,\\n    \\\"ef_construct\\\": 100\\n  }\\n}<\/code><\/pre>\\n\\n<h2>Model installieren<\/h2>\\n<pre><code>ollama pull mxbai-embed-large\\nollama list<\/code><\/pre>\\n\\n<h2>API-Aufruf<\/h2>\\n<pre><code>curl http:\/\/localhost:11434\/api\/embeddings -d '{\\n  \\\"model\\\": \\\"mxbai-embed-large\\\",\\n  \\\"prompt\\\": \\\"Text zum Embedden\\\"\\n}'<\/code><\/pre>\\n\\n<h2>Python-Integration<\/h2>\\n<pre><code>import requests\\n\\ndef get_embedding(text: str) -> list[float]:\\n    response = requests.post(\\n        'http:\/\/localhost:11434\/api\/embeddings',\\n        json={\\n            'model': 'mxbai-embed-large',\\n            'prompt': text\\n        }\\n    )\\n    return response.json()['embedding']<\/code><\/pre>\\n\\n<h2>Qdrant-Speicherung<\/h2>\\n<pre><code>from qdrant_client import QdrantClient\\nfrom qdrant_client.models import PointStruct\\n\\nclient = QdrantClient(host=\\\"localhost\\\", port=6333)\\n\\nclient.upsert(\\n    collection_name=\\\"documents\\\",\\n    points=[\\n        PointStruct(\\n            id=uuid4().hex,\\n            vector=embedding,\\n            payload={\\n                \\\"document_id\\\": 123,\\n                \\\"chunk_id\\\": 1,\\n                \\\"content_preview\\\": text[:200]\\n            }\\n        )\\n    ]\\n)<\/code><\/pre>\",\n    \"status\": \"published\",\n    \"sort_order\": 5,\n    \"depth\": 1,\n    \"created_at\": \"2025-12-20T11:13:58\",\n    \"updated_at\": \"2025-12-27T15:58:19\"\n  }\n}"
        }
    ]
}