populate_from_csv.py

import csv
import weaviate
import uuid
from datetime import datetime
import os
from weaviate_schema import setup_weaviate_client
from shelter_utils import geocode_address

# Format date in RFC3339 format
def rfc3339_date():
    return datetime.utcnow().strftime("%Y-%m-%dT%H:%M:%SZ")

def read_hotels_from_csv(csv_file_path):
    """Read hotel data from CSV file."""
    hotels = []
    try:
        with open(csv_file_path, 'r', encoding='utf-8') as csv_file:
            # Skip the first 3 rows (header info)
            for _ in range(3):
                next(csv_file)
            
            csv_reader = csv.reader(csv_file)
            for row in csv_reader:
                if len(row) >= 3 and row[0] and row[1]:  # Ensure we have at least hotel name and address
                    hotel = {
                        "name": row[0].strip(),
                        "address": row[1].strip(),
                        "website": row[2].strip() if len(row) > 2 and row[2] else "",
                        "phoneNumber": row[3].strip() if len(row) > 3 and row[3] else "No phone available",
                        "notes": row[4].strip() if len(row) > 4 and row[4] else ""
                    }
                    hotels.append(hotel)
    except Exception as e:
        print(f"Error reading CSV file: {e}")
    
    return hotels

def convert_hotels_to_shelters(hotels):
    """Convert hotel data to shelter format."""
    shelters = []
    
    for hotel in hotels:
        # Get geocode for the address
        location = geocode_address(hotel["address"])
        if not location:
            print(f"Could not geocode address for {hotel['name']}, skipping...")
            continue
        
        # Create shelter object
        shelter = {
            "name": hotel["name"],
            "description": f"Hotel available for evacuees. {hotel.get('notes', '')}",
            "location": location,
            "address": hotel["address"],
            "currentCapacity": 10,  # Default values since we don't have this information
            "maximumCapacity": 50,
            "hasSpace": True,
            "services": ["Temporary Housing", "Evacuation Support"],
            "amenities": ["Beds", "Bathroom"],
            "restrictions": [],
            "acceptsPets": "pet" in hotel.get("notes", "").lower(),  # Guess based on notes
            "accessibilityFeatures": ["Standard Hotel Accessibility"],
            "phoneNumber": hotel["phoneNumber"],
            "email": "",
            "website": hotel["website"],
            "operatingHours": "24/7",
            "lastUpdated": rfc3339_date(),
            "shelterType": "Emergency",
            "disasterResponse": ["Fire", "Evacuation"],
            "languages": ["English"]
        }
        shelters.append(shelter)
    
    return shelters

def add_shelters_to_weaviate(client, shelters):
    """Add shelters to Weaviate."""
    print(f"Adding {len(shelters)} shelters to Weaviate...")
    
    success_count = 0
    error_count = 0
    
    for shelter in shelters:
        try:
            # Generate a UUID based on the shelter name for consistency
            shelter_uuid = uuid.uuid5(uuid.NAMESPACE_DNS, shelter["name"])
            
            # Add the shelter to Weaviate
            client.data_object.create(
                data_object=shelter,
                class_name="Shelter",
                uuid=str(shelter_uuid)
            )
            print(f"Added shelter: {shelter['name']}")
            success_count += 1
        except Exception as e:
            print(f"Error adding shelter {shelter['name']}: {e}")
            error_count += 1
    
    print(f"Finished adding shelters. Success: {success_count}, Errors: {error_count}")

def main():
    """Main function to read CSV and populate Weaviate."""
    csv_file_path = "shelters.csv"
    
    if not os.path.exists(csv_file_path):
        print(f"Error: File {csv_file_path} not found.")
        return
    
    # Read hotels from CSV
    print(f"Reading hotels from {csv_file_path}...")
    hotels = read_hotels_from_csv(csv_file_path)
    print(f"Found {len(hotels)} hotels in CSV file.")
    
    # Convert hotels to shelters
    print("Converting hotels to shelter format...")
    shelters = convert_hotels_to_shelters(hotels)
    print(f"Converted {len(shelters)} hotels to shelter format.")
    
    # Create a Weaviate client
    client = setup_weaviate_client()
    
    # Check if we already have data
    result = client.query.get("Shelter", ["name"]).do()
    existing_shelters = len(result["data"]["Get"]["Shelter"]) if result["data"]["Get"]["Shelter"] else 0
    print(f"Found {existing_shelters} existing shelters in Weaviate.")
    
    # Add shelters to Weaviate
    add_shelters_to_weaviate(client, shelters)

if __name__ == "__main__":
    main()