Place Suggestion API

This project aims to develop a RESTful API that uses Gemini AI to suggest places based on user prompts. The API will return detailed place information, including names, locations, and multimedia content, and will be deployed in a Docker environment on AWS.

Features

• AI-Powered Recommendations: Suggest places based on user input.
• Docker Deployment: Containerized for easy deployment.
• Detailed Place Information: Includes scraped names, prices, starts, place features, locations, and multimedia etc.

Planned Features

• Google Maps API Implementation: Google Maps API will be optional for fast and more accurate results.
• Cloud Deployment: Planned deployment on services such as AWS. (The project was deployed on AWS ECS and Lambda but this feature has been postponed to a later date due to financial reasons.)

Usage

Pull and run from Docker Hub:

    docker pull elymsyr/place-suggestion-api:latest
    docker run -d -p 8000:8000 elymsyr/place-suggestion-api

Test container:

http://0.0.0.0:8000/scrap/?query=I%20need%20some%20quiet%20places%20to%20spend%20time%20in%20Amsterdam&gemini_api_key={YOUR_API_KEY}&language=en&max_worker={MAX_WORKER_NUMBER}

API Reference

Get suggestion data

See, API.

  GET /scrape

Parameter	Type	Description
query	string	Required Query seach
gemini_api_key	string	Required GEMINI API key Get your key
maps_api_key	string	Enter to use Google Maps API. (Planned Feature)
language	string	Defalut = str: 'en'
max_worker	string	Defalut = int: 1

Usage/Examples

Results with the Query I need some quiet places in Amsterdam to spend time :

Hortus Botanicus: 
    {
        "url": "https://www.google.com/maps/place/Hortus+Botanicus/@52.3669189,4.907623,17z/dat...",
        "place_name": "Hortus Botanicus",
        "place_type": "Botanical garden",
        "price": "$$",
        "coordinate": [52.3669189, 4.907623],
        ...
    }
Vondelpark: 
    {
        "url": "https://www.google.com/maps/place/Vondelpark/@52.3579946,4.8686484,17z/dat..",
        "place_name": "Vondelpark",
        "place_type": "Park",
        "price": "$$",
        "coordinate": [52.3579946, 4.8686484],
        ...
    }
    ...

Optimizations

Fast response with asynchronous processing : API scraps data as the Gemini produces response. Also, data scraping is proccessed asynchronously.

WebDriver Pool Tested : Creating a pool for web drivers in order to avoid creating a new one for each dscraping process was tested but the results has shown that creating a new one is faster and cost efficient.

Running Tests

See test. Run pytest -s Test/test_main.py for testing the api.

Contributing

Contributions are welcome! Please check back for development and contributing guidelines and adhere to this project's code of conduct.

License

GNU GENERAL PUBLIC LICENSE. See the LICENSE file for details.

Authors

• @elymsyr