Constrained Query Rewriting Logic

U

@

·

Query rewriting with strict token and latency constraints

15 copies0 forks
Implement query rewriting for RAG within these strict constraints.

CONSTRAINTS:
- Maximum rewritten queries: {{max_queries}}
- Token budget per query: {{token_limit}}
- Latency overhead: <{{latency_budget_ms}}ms
- Must preserve original intent

Original Query:
{{user_query}}

Query Context:
{{context_metadata}}

OUTPUT FORMAT:
```json
{
  "original": "<query>",
  "rewrites": [
    {
      "query": "<rewritten>",
      "strategy": "expansion|decomposition|synonym|specificity",
      "token_count": <number>,
      "expected_improvement": "<reason>"
    }
  ],
  "total_tokens": <number>,
  "constraint_compliance": true
}
```

Details

Category

Coding

Use Cases

Query optimizationSearch enhancementRAG improvement

Works Best With

claude-sonnet-4-20250514gpt-4o
Created Shared

Create your own prompt vault and start sharing