README.md

Getting started with the project

Questions??: Feel free to contact Atharva Kashyap ([email protected])

How to get the project onto local machine and run the app?

1. First, clone the repository onto your preferred folder inside your local machine.
2. Then, cd into the repository folder on your local machine.
3. If you wish to create a branch and work inside that branch, make sure you perform git checkout -b "<your-branch-name>". This is to create a new branch. You can then perform git status to check if your branch was created successfully.
4. Now, perform yarn install. This is to make sure all the yarn packages are installed into your local project.
5. Then, change directory into server, which you can do by running cd server.
6. Now, perform yarn install again inside the server folder to make sure to install necessary modules inside this folder.
7. This should set-up your project on the local machine. You can now change directory by going back one level. You can do this by performing cd ... Then, you should be able to run yarn start and a reactapp should get started on your web browser. (If you get an error see note below)

NOTE: If you get an error which mentions Attempting to bind to HOST environment variable: x86_64-conda-linux-gnu, it means that the HOST env variable needs to be set. For testing run HOST=localhost in the current shell before yarn start to run the server locally, otherwise HOST should be set to aurmr-control (assuming you're on the aurmr-control workstation) so that other things can find the server.

How to start graphql playground?

For Your Information: graphql playground is something that runs on localhost:4000 that provides a space for you to write and test out queries and mutations before making it official and adding it to your app. It also provides you a nice UI that prompts you suggestions when writing your queries and mutations. To get this working, please follow the steps below.

1. First, perform cd server so that you are in your server folder.
2. Then, perform npx prisma generate
3. Finally, perform node src/index.js

Important things to keep in mind:

1. You should stop, run npx prisma generate, and then re-run node src/index.js after any edit within your server folder. Then refresh your playground to see changes.
2. If you make any edits to schema.prisma, it is important to run prisma migrate dev --name <give-your-migration-a-name>. This is because you made some edits to your database and you want those changes to be reflected.
3. The column names of models that you define in your schema must be exactly same as the models you define in your graphql.schema. In other words, the column names that you have for your prisma table needs to be exactly the same as the column names that you give to your graphql schema.
4. Extra (high-level) information about how the server side of things are working: graphql serves to provide a "front-end" experience, which means this is what you use in the reactapp to query and get information from. graphql then contacts prisma through a mode called resolvers.js, which then contacts the database to obtain the data.

Schema Outline

Here is the link to the Schema file: Schema

AmazonProduct

Each row of this table keeps track of one Amazon Product. It has columns that keeps track of asin, size, weight, and list of Attributes.

model AmazonProduct {
  id           Int                      @id @default(autoincrement())
  createdAt    DateTime                 @default(now())
  asin         String                   @unique
  name         String
  size_length  Float?
  size_width   Float?
  size_height  Float?
  size_units   String?
  weight       Float?
  weight_units String?
  attributes   AmazonProductAttribute[]
  bins         ProductBin[]
}

Attribute

Each row of this table keeps track of a particular attribute. The value of the attribute is unique. In other words, having two attributes with the name value is not possible.

model Attribute {
  id             Int                      @id @default(autoincrement())
  value          String                   @unique
  AmazonProducts AmazonProductAttribute[]
}

AmazonProductAttribute

Connects products to attributes. Each product can have multiple Attributes.

model AmazonProductAttribute {
  AttributeId     Int
  AmazonProductId Int
  attribute       Attribute?     @relation(fields: [AttributeId], references: [id])
  amazonProduct   AmazonProduct? @relation(fields: [AmazonProductId], references: [id])

  @@id([AttributeId, AmazonProductId])
}

Overview of Evaluation

Evaluation: An "Experiment" that has various products in various bins and various picks. It can be thought of as a setup.

Bin: Each row of this table keeps track of a BinId (as specified by the QR code stickers on the bins). It also keeps track of a list of ProductBin combination of all the products inside the bin.

ProductBin: This keeps track of a Product, the bin its placed in, the evaluation for which it is part of, and a unique id. There can be multiple of the same product in the same bin (but each of them will have a unique id).

Pick: A pick that specifies a particular ProductBin's unique id to simulate a pick.

Pick

model Pick {
  id              Int            @id @default(autoincrement())
  ProductBinId Int
  ProductFromBin ProductBin @relation(fields: [ProductBinId], references: [id])
  Outcome         Boolean?
  TimeTakenSec    Float?
}

Bin

model Bin {
  id             Int            @id @default(autoincrement())
  BinId          String         @unique
  AmazonProducts ProductBin[]
}

ProductBin

model ProductBin {
  id Int @id @default(autoincrement())
  AmazonProductId Int
  binId        String
  evalId          Int
  amazonProduct   AmazonProduct? @relation(fields: [AmazonProductId], references: [id])
  bin             Bin?           @relation(fields: [binId], references: [BinId])
  evaluation      Evaluation?    @relation(fields: [evalId], references: [id])
  picks Pick[]
}

Evaluation

model Evaluation {
  id    Int            @id @default(autoincrement())
  name  String         @unique
  Setup ProductBin[]
}

How to query and mutate the database from Python?

You can find some example scripts that was used for interacting with the database here.

• AddBin -> Provide a csv file with bins and its sizes use AddBin script to add them to the database.
• AddProd -> Provide a csv file with products and its details and use the AddProd scrpt to add them to the database.
• getProductDetailsScrape_Updated -> Provide a csv with amazon product links and scrape the web to get the product details. Make sure the time for each scrape query is spaced out by at least 2 minutes.

File System Outline

Backend

In the server folder, inside the src folder, there is schema.graphql, which has all the queries, mutations, and models (as defined #Schema Outline). So, if you want to add a query or a mutation (to edit the database), then you will need to define that in either type Query or type Mutation first.

Then, in the server/src folder, there is resolvers.js. After defining the queries and mutations in the schema.graphql, you will need to say what those queries and mutations should do in-terms of interacting with the database. For this, you will need to define the functionalities in the resolvers.

Finally, in the server/prisma/schema.prisma, you can define the tables for the database side. If you were to add more tables or more columns to existing tables, then you will need to edit it here first and then add the same to schema.graphql.

Frontend

In the src, inside the components folder, the important ones are the following:

• ManualEval.tsx -> Frontend for handling the stowing process.
• PickHandler.tsx -> Frontend for handling the picking process.
• Evaluator.tsx -> Frontend for random generation of product configurations in bins. Not currently in use but could be built on for future use.

In the src, inside the GraphQLQueriesMuts folder, the important ones are the following:

• Query.tsx -> All the queries that are used for Stowing and Picking
• Mutation.tsx -> All the mutations that are used for Stowing and Picking