[](https://hamel.dev/blog/posts/field-guide/#cb2-1)def generate_search_query(scenario, persona, listing_db):
[](https://hamel.dev/blog/posts/field-guide/#cb2-2)    """Generate a realistic user query about listings"""
[](https://hamel.dev/blog/posts/field-guide/#cb2-3)    # Pull real listing data to ground the generation
[](https://hamel.dev/blog/posts/field-guide/#cb2-4)    sample_listings = listing_db.get_sample_listings(
[](https://hamel.dev/blog/posts/field-guide/#cb2-5)        price_range=persona.price_range,
[](https://hamel.dev/blog/posts/field-guide/#cb2-6)        location=persona.preferred_areas
[](https://hamel.dev/blog/posts/field-guide/#cb2-7)    )
[](https://hamel.dev/blog/posts/field-guide/#cb2-8)    
[](https://hamel.dev/blog/posts/field-guide/#cb2-9)    # Verify we have listings that will trigger our scenario
[](https://hamel.dev/blog/posts/field-guide/#cb2-10)    if scenario == "multiple_matches" and len(sample_listings) < 2:
[](https://hamel.dev/blog/posts/field-guide/#cb2-11)        raise ValueError("Need multiple listings for this scenario")
[](https://hamel.dev/blog/posts/field-guide/#cb2-12)    if scenario == "no_matches" and len(sample_listings) > 0:
[](https://hamel.dev/blog/posts/field-guide/#cb2-13)        raise ValueError("Found matches when testing no-match scenario")
[](https://hamel.dev/blog/posts/field-guide/#cb2-14)    
[](https://hamel.dev/blog/posts/field-guide/#cb2-15)    prompt = f"""
[](https://hamel.dev/blog/posts/field-guide/#cb2-16)    You are an expert real estate agent who is searching for listings. You are given a customer type and a scenario.
[](https://hamel.dev/blog/posts/field-guide/#cb2-17)    
[](https://hamel.dev/blog/posts/field-guide/#cb2-18)    Your job is to generate a natural language query you would use to search these listings.
[](https://hamel.dev/blog/posts/field-guide/#cb2-19)    
[](https://hamel.dev/blog/posts/field-guide/#cb2-20)    Context:
[](https://hamel.dev/blog/posts/field-guide/#cb2-21)    - Customer type: {persona.description}
[](https://hamel.dev/blog/posts/field-guide/#cb2-22)    - Scenario: {scenario}
[](https://hamel.dev/blog/posts/field-guide/#cb2-23)    
[](https://hamel.dev/blog/posts/field-guide/#cb2-24)    Use these actual listings as reference:
[](https://hamel.dev/blog/posts/field-guide/#cb2-25)    {format_listings(sample_listings)}
[](https://hamel.dev/blog/posts/field-guide/#cb2-26)    
[](https://hamel.dev/blog/posts/field-guide/#cb2-27)    The query should reflect the customer type and the scenario.
[](https://hamel.dev/blog/posts/field-guide/#cb2-28)
[](https://hamel.dev/blog/posts/field-guide/#cb2-29)    Example query: Find homes in the 75019 zip code, 3 bedrooms, 2 bathrooms, price range $750k - $1M for an investor.
[](https://hamel.dev/blog/posts/field-guide/#cb2-30)    """
[](https://hamel.dev/blog/posts/field-guide/#cb2-31)    return generate_with_llm(prompt)