nest-quest's Introduction

Nest Quest

This project intends to help identify if a potential home meets your requirements. Users configure their preferences in a tree structure. These preferences include conditional statements as well as arithmetic operators and distance values. This tree can then be used to evaluate a target location using various providers such as Open Street Map.

Architecture

The architecture for this project can be decomposed into several subsystems and/or components:

Client for user preference input. This could take several forms, such as a website, CLI, or browser extension. Regardless of the implementation each client would adhere to the same general structure and interface with the same APIs.
Lexer/tokenizer for user preferences, which are passed in as a string.
Parser that converts tokens into an AST.
Weight calculator using AHP, Paired Comparisons, or Reference Comparisons (http://www2.mitre.org/work/sepo/toolkits/STEP/files/ScoringMethodsContent.pdf).
Evaluators that evaluate a given location against the AST, returning a utility score. This should be generalized to allow multiple evaluator implementations, e.g. Overpass, Google Maps, Apple Maps, etc.
Datastore that manages retrieval and storage of data from external locations.

graph TD
    Client --> Location
    Client --> Preferences
    Preferences --> Tokenizer
    Preferences <--> Autocomplete
    Tokenizer --> Parser
    Parser --> AST
    AST --> Evaluator
    Location --> Evaluator
    Evaluator <--> Datastore
    Autocomplete <--> Datastore
    Datastore <--> Overpass
    Datastore <--> GoogleMaps
    Datastore <--> Zillow
    Datastore <--> Redfin
    Datastore <--> ...