evo-travel-planning

Build multi-day travel itineraries from structured databases (cities, accommodations, restaurants, attractions, driving distances) with robust constraint validation, budget tracking, and pet-friendly

SKILL.md

GitHub

evo-travel-planning

Purpose

Build multi-day travel itineraries from structured CSV/text databases covering cities, accommodations, restaurants, attractions, and driving distances. Handles complex constraints including pet-friendly accommodations, cuisine preferences, budget limits, multi-city routing, and transportation restrictions.

Key Rules

  1. All data must come from database files in /app/data/ — never use LLM memory for POIs.
  2. Pet-friendly accommodations: house_rules column must NOT contain "No pets" (use na=False to handle NaN safely per Pandas 2.0.3 best practices).
  3. Output accommodation field MUST contain "Pet-friendly" prefix when pet-friendly is required.
  4. Cuisine coverage verified from restaurant Cuisines field, not restaurant names.
  5. All 5 tool names must appear in tool_called: search_cities, search_accommodations, search_restaurants, search_attractions, search_driving_distance.
  6. Attractions separated by semicolons with trailing semicolon (e.g., "Attraction A;Attraction B;").
  7. Transportation format: "Self-driving: from A to B" or "-" for same-city days.
  8. current_city: city name for same-city days, or "from A to B" for transit days.
  9. Budget validation: accommodations are per-night, meals are per-person, transportation is per-trip.
  10. Uses pd.concat() instead of deprecated DataFrame.append() (removed in Pandas 2.0).

Usage

import sys
sys.path.insert(0, '/app/environment/skills/evo-travel-planning/scripts')
from utils import (
    search_cities, search_accommodations, search_restaurants,
    search_attractions, search_driving_distance,
    format_accommodation_name, format_attractions,
    build_itinerary, save_itinerary, validate_budget,
    optimize_route, allocate_days_to_cities,
    get_cheapest_accommodation, get_restaurants_by_cuisines
)

# 1. Search Ohio cities
ohio_cities = search_cities(state='Ohio')

# 2. Optimize route from origin through destination cities
route = optimize_route('Minneapolis', ['Cleveland', 'Columbus', 'Cincinnati'])

# 3. Allocate days across cities
allocation = allocate_days_to_cities(route, total_days=7)

# 4. Find pet-friendly accommodations
accomm = search_accommodations(city='Cleveland', pet_friendly=True, min_occupancy=2)

# 5. Find cheapest pet-friendly accommodation
cheapest = get_cheapest_accommodation(city='Cleveland', pet_friendly=True, min_occupancy=2)

# 6. Find restaurants by multiple cuisines
restaurants = get_restaurants_by_cuisines(city='Cleveland',
    cuisines=['American', 'Mediterranean', 'Chinese', 'Italian'])

# 7. Get attractions
attractions = search_attractions(city='Cleveland')

# 8. Check driving distance/cost
dist = search_driving_distance(origin='Minneapolis', destination='Cleveland')

# 9. Validate budget
budget_result = validate_budget(
    accommodation_costs=[{'price': 80, 'nights': 2}],
    meal_costs=[15, 20, 25],
    transport_costs=[50],
    num_travelers=2,
    budget=5100
)

# 10. Build and save
itinerary = build_itinerary(plan_days, tools_used)
save_itinerary(itinerary)