Script to validate entities in configuration files

validate_entities.py

import sys
import re
import requests

# USAGE: python3 validate_entities.py < flows.json

# Configuration
# Replace with your Home Assistant instance's address
BASE_URL = "http://homeassistant.local:8123"

# Replace with your API token
TOKEN = ""

# Add here entity ids you'd like to except
INVALID_ENTITY_IDS = []

HEADERS = {
    "Authorization": f"Bearer {TOKEN}",
    "Content-Type": "application/json",
}

def get_entities(base_url, headers):
    # Endpoint for fetching states (which includes entities)
    url = f"{base_url}/api/states"
    
    response = requests.get(url, headers=headers)
    
    if response.status_code != 200:
        print("Failed to retrieve entities. Ensure your TOKEN and BASE_URL are correct.")
        return
    
    states = response.json()
    
    # Extract entity ids from states
    entities = [state['entity_id'] for state in states]
    
    return entities

def get_services(base_url, headers):
    url = f"{base_url}/api/services"
    response = requests.get(url, headers=headers)
    
    if response.status_code != 200:
        print("Failed to retrieve services. Ensure your TOKEN and BASE_URL are correct.")
        return []
    
    services = []
    services_data = response.json()
    for service_data in services_data:
        domain = service_data["domain"]
        for service in service_data["services"]:
            services.append(f"{domain}.{service}")

    # Extracting and returning list of services in "domain.service" format
    return services

def get_event_types(base_url, headers):
    url = f"{base_url}/api/events"
    response = requests.get(url, headers=headers)
    
    if response.status_code != 200:
        print("Failed to retrieve event types. Ensure your TOKEN and BASE_URL are correct.")
        return []
    
    events_data = response.json()
    
    # Extracting and returning list of event types
    return [event['event'] for event in events_data]


def extract_domains_from_entities(entities):
    return set(entity.split('.')[0] for entity in entities)

def extract_patterns(s):
    # The regex pattern looks for sequences of characters that are separated by a dot.
    # \w+ captures one or more word characters (equivalent to [a-zA-Z0-9_])
    pattern = r'\b\w+\.\w+\b'
    return re.findall(pattern, s)


if __name__ == "__main__":
    flows_entities = []
    services = get_services(BASE_URL, HEADERS)
    entities = get_entities(BASE_URL, HEADERS)
    events = get_event_types(BASE_URL, HEADERS)
    valid_domains = extract_domains_from_entities(entities)
    
    # Read flows_data from stdin
    flows_data = sys.stdin.read()
    flows_entities = [eid for eid in extract_patterns(flows_data) if eid.split('.')[0] in valid_domains]
    
    flows_entities = list(set(flows_entities))
    flows_entities = [eid for eid in flows_entities if eid not in INVALID_ENTITY_IDS]
    flows_entities = [eid for eid in flows_entities if eid not in services]
    flows_entities = [eid for eid in flows_entities if eid not in events]
    
    for flow_entity in flows_entities:
        if flow_entity not in entities:
            print(f"Entity {flow_entity} not found in Home Assistant.")