vector-db-ops

Overview

Vector database operations for AI applications. Covers embedding generation, index creation, metadata filtering, hybrid search, and production deployment across Pinecone, Weaviate, Qdrant, and ChromaDB.

Capabilities

• Generate and store vector embeddings from text, images, and code
• Create and manage collections with metadata schemas
• Perform semantic similarity search with filters
• Implement hybrid search (dense + sparse vectors)
• Optimize index parameters for speed and recall
• Manage vector database lifecycle (backup, scaling, monitoring)

When to Use

Trigger phrases:

• "vector db ops"

• "Vector database operations — Pinecone, Weaviate, Qdrant, ChromaDB"

• Building RAG (Retrieval-Augmented Generation) systems

• Implementing semantic search for documents or products

• Creating recommendation engines based on similarity

• Building memory systems for AI agents

• Implementing image/code similarity search

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Pseudo Code

# Example workflow for this skill
def execute(input_data):
    # Step 1: Validate input
    if not input_data:
        raise ValueError("Input data is required")

    # Step 2: Process core logic
    result = process(input_data)

    # Step 3: Validate output
    validate_output(result)

    return result

Pinecone

import pinecone
from openai import OpenAI

# Initialize
pc = pinecone.Pinecone(api_key="YOUR_API_KEY")
index = pc.Index("my-index")

# Upsert vectors
openai = OpenAI()
response = openai.embeddings.create(input=["text"], model="text-embedding-3-small")
embeddings = response.data[0].embedding

index.upsert(vectors=[{
    "id": "doc-1",
    "values": embeddings,
    "metadata": {"source": "pdf", "page": 42}
}])

# Query
results = index.query(vector=query_embedding, top_k=10, include_metadata=True,
                       filter={"source": {"$eq": "pdf"}})