Place index is a Python library to search, index and merge places from different providers.
For instance, if you want to search for specific places in an area, you can use this library to search on Google Maps, TripAdvisor and OpenStreet Map to get the best results, and aggregate / merge them into a single list of places.
I wanted an easy way to look for restaurants near my workplace, and I needed an exhaustive list of places to choose from.
Create a .env file in the place_index folder with the following content:
GMAPS_API_KEY=your_google_maps_api_key
TRIP_API_KEY=your_tripadvisor_api_key
OPEN_AI_API_KEY=your_openai_api_keyIn this code snippet, we fetch restaurants around the Eiffel Tower, using Google Maps and TripAdvisor as providers.
from place_index.fetcher.fetcher import Fetcher
from place_index.fetcher.provider import ProviderSource
# Add the providers you want to use (ex Google Maps or TripAdvisor)
fetcher = Fetcher(ProviderSource.GOOGLE_MAPS, ProviderSource.TRIPADVISOR)
generic_places = fetcher.fetch_all(coordinates=[(48.858265, 2.294494)])The data previously fetched may contain duplicates. We are using a nearest neighbor algorithm to merge the data, with the optional help of a LLM (Large Language Model).
Note: The LLM is not required, but it helps to improve the quality of the merged data, especially for places tags.
from place_index.merger.merger import Merger
merger = Merger(use_llm=False)
merger.add_restaurants(generic_places)
print(merger.places)Data can be easily exported to JSON format.
import json
from dataclasses import asdict
with open("output.json", "w") as f:
json.dump(
{restaurant.name: asdict(restaurant) for restaurant in merger.places.values()},
f,
)