Smoothify

The basis for the code is through React Native using the Expo simulator: Create Expo app page

Note: This app works for iOS, but remains untested for Android

Overview:

This is the basis for the Smoothify app. As such, this app takes in the user's allergies, preferences, and ingredients and generates a recipe using ChatGPT. The prompt for ChatGPT includes a template based on the flavor profile the user has chosen. These templates I use are general recipes found on pages like Allrecipes, but the generated recipes will vary widely depending on the template. Something to keep in mind, since ChatGPT is used to make the recipes, I provide a disclaimer to the user to review the recipe due to the possibility of hallucinations or an accidental inclusion of an allergy into the recipe.

Below is the Data Flow Diagram:

Frontend: (React Native: Javascript, Typescript)

• The screen layout is defined at its base through ./_layout.tsx
• The app utilizes tab navigation, which is defined in ./(tabs)/_layout.tsx
• With the current construction, any file in (tabs) would create a new tab
  • These contain the majority of the visual features associated with each of the individual pages
  • (tabs) contains index.tsx, generate.tsx, and profile.tsx in addition to the layout file
  • index.tsx is the default page
  • The subpages are handled using states in the individual pages
  • Reoccurring custom components are kept in the components folder
  • Within generate.tsx and profile.tsx
    • Google Cloud Platform APIs for GET, POST, and DELETE operations
    • Open AI API call (generate only)
      • Open AI API should be called in the backend, but this iteration of the app has it in the frontend
    • Apple Authentication call (profile only)

Backend: (Google Cloud Platform: Golang)

Note: I have found the best way to create a Cloud Run function is through terminal operations

• Upload a zip into a bucket (can be done with the Cloud Bucket page)

  • gsutil cp ./archive.zip gs://[BUCKET_NAME]/

• Deploy to Cloud Run from zip (must be done with the command)

  • gcloud run deploy [SERVICE_NAME]
    --source gs://[BUCKET_NAME]/archive.zip
    --region [YOUR REGION]
    --allow-unauthenticated

• Using a single Cloud Run function for each database operation

• config

  • Establishes the database connection from .env login information

• controllers

  • The database operations themselves
  • Each operation checks that the user exists before the operation
    • Get (get all of the user's recipes)
    • Health (check the database connectivity)
    • Create_Recipe (posts a new recipe in the database)
    • Create_User (posts a new user in the database)
    • Delete (deletes a recipe from the database)

• main.go

  • This file is what is executed

• modules

  • This outlines the objects containing the columns for the information taken from the database

• router

  • This sets the routes for each database operation

Personal Experience:

Overall, this project was a fantastic experience for me. It let me explore some tools I had minimal time working with, like React Native, or languages that I have never worked with before, like Golang. The most time-consuming part for me was learning how the Google Cloud Platform operates, as well as learning the Go language for database functions. React Native has a plethora of components and tools available for you that allow for a lot of freedom to build Smoothify in the way I envisioned.