gh-itmediatech-dt-cli/.claude/commands/rag-index.md

description

description
Index or re-index the codebase for RAG search

RAG Index Command

Index or re-index the codebase to update the RAG system's knowledge.

Usage

The user wants to index the codebase. This will:

• Scan all code files in the repository
• Generate embeddings for code chunks
• Store them in the local vector database

Instructions

1. Trigger the indexing process via the MCP server
2. Show progress to the user
3. Display completion status

Implementation

import httpx
import json

print("[#] Starting codebase indexing...")
print("This may take a few minutes depending on codebase size.\n")

try:
    response = httpx.post(
        "http://127.0.0.1:8765/execute",
        json={
            "category": "rag",
            "tool_name": "rag_index",
            "parameters": {"root_path": "."}
        },
        timeout=300.0  # 5 minutes timeout
    )

    if response.status_code == 200:
        result = response.json()

        if result.get("success"):
            status = result.get("status", {})

            print("[OK] Indexing complete!\n")
            print(f"[=] Status:")
            print(f"   Indexed chunks: {status.get('indexed_chunks', 0)}")
            print(f"   Embedding model: {status.get('embedding_model', 'unknown')}")
            print(f"   Status: {status.get('status', 'unknown')}")
        else:
            print(f"Error: {result.get('error', 'Unknown error')}")
    else:
        print(f"MCP Server error: {response.status_code}")

except Exception as e:
    print(f"Error indexing codebase: {e}")
    print("Make sure the MCP server is running.")

Execute this Python code to index the codebase.