Design Patterns MCP Server

{ "id": "embeddings-pattern", "name": "Embeddings Pattern", "category": "AI/ML", "description": "Converts text/data into high-dimensional vectors for similarity and search", "when_to_use": "Semantic search\nSimilarity matching\nClustering", "benefits": "Semantic understanding\nEfficient search\nSimilarity computation\nClustering capability", "drawbacks": "Storage requirements\nDimensionality choice\nQuality dependence", "use_cases": "Search engines\nRecommendation systems\nDocument clustering", "complexity": "Medium", "tags": [ "embeddings", "vectors", "similarity" ], "examples": { "python": { "language": "python", "code": "# Embeddings: convert text to semantic vectors\n\nfrom sentence_transformers import SentenceTransformer\nimport numpy as np\n\nclass EmbeddingsSystem:\n def __init__(self, model_name='all-MiniLM-L6-v2'):\n self.model = SentenceTransformer(model_name)\n \n def embed(self, texts: list[str]):\n # Convert texts to vectors\n return self.model.encode(texts)\n \n def similarity(self, text1: str, text2: str):\n emb1, emb2 = self.embed([text1, text2])\n # Cosine similarity\n return np.dot(emb1, emb2) / (\n np.linalg.norm(emb1) * np.linalg.norm(emb2)\n )\n \n def find_similar(self, query: str, documents: list[str], top_k=3):\n query_emb = self.embed([query])[0]\n doc_embs = self.embed(documents)\n \n # Calculate similarities\n similarities = [\n (i, np.dot(query_emb, doc_emb) / (\n np.linalg.norm(query_emb) * np.linalg.norm(doc_emb)\n ))\n for i, doc_emb in enumerate(doc_embs)\n ]\n \n # Return top-k\n similarities.sort(key=lambda x: x[1], reverse=True)\n return [(documents[i], score) for i, score in similarities[:top_k]]\n\n# Usage: semantic search and similarity\nembeddings = EmbeddingsSystem()\nsimilar_docs = embeddings.find_similar(\n \"machine learning\",\n [\"AI and ML\", \"cooking recipes\", \"neural networks\"]\n)\nprint(similar_docs) # [(\"AI and ML\", 0.85), (\"neural networks\", 0.72)]" } } }