ecom-cart

ecom-cart implements a promotion engine which can compute promotion rules such as:

if you buy 2 or more trousers, you get 15% off belts and shoes
if you buy two shirts, each additional shirt costs only 45$
if you purchase 3 or more shirts, all ties are half price
buy 2 shirts and get 1 shirt free

Notice that fourth rule (buy 2 shirts and get 1 shirt free) has to be applied in such way if the customer has purchased 7 shirts then two of them should be free.

Installation and Usage

This project follows a slightly different structure from the recommended GOLang guidelines. (This can be corrected later) The packages are contained within the project folder under src folder The project depends upon:

"github.com/gorilla/mux"
"github.com/satori/go.uuid" To install the dependencies run the following commands in project folder:

go get -u github.com/satori/go.uuid
go get -u github.com/gorilla/mux"

To run the project:

go run main.go

Usage:

Create a cart Ex: cart with - two trousers, three shirts, one belt, one shoe

curl localhost:8090/carts -X POST -d '{ "items": [{"name": "t1", "category": "trousers", "price": 100}, {"name": "t2", "category": "trousers", "price": 110}, {"name": "belt-1", "category": "belts", "price": 100}, {"name": "shoe-1", "category": "shoes", "price": 100}, {"name": "s1", "category": "shirts", "price": 120}, {"name": "s2", "category": "shirts", "price": 90}, {"name": "s3", "category": "shirts", "price": 65} ] }'

Response:
{"message":"Created","id":"06f9af8a-d312-41a7-9576-f097958cc61f","_links":[{"instance":"/carts/06f9af8a-d312-41a7-9576-f097958cc61f"}]}

The cart is now created and assigned a unique id

Get a cart

curl localhost:8090/carts/06f9af8a-d312-41a7-9576-f097958cc61f

Response:
{"id":"06f9af8a-d312-41a7-9576-f097958cc61f",
"items":[
{"id":"","name":"t1","category":"trousers","price":100},
{"id":"","name":"t2","category":"trousers","price":110},
{"id":"","name":"belt-1","category":"belts","price":100},
{"id":"","name":"shoe-1","category":"shoes","price":100},
{"id":"","name":"s1","category":"shirts","price":120},
{"id":"","name":"s2","category":"shirts","price":90},
{"id":"","name":"s3","category":"shirts","price":65}
]
}

Get available promos:

curl localhost:8090/promos

Response:

[
  {
    "id": "promo2",
    "description": "if you buy two shirts, each additional shirt costs only 45$",
    "buys": [ { "category": "shirts", "count": 2 } ],
    "gets": [ { "category": "shirts", "all": true, "count": 0, "off": { "discount": null, "fixed": { "price": 45 } } ]
  },
  {
    "id": "promo3",
    "description": "if you purchase 3 or more shirts, all ties are half price",
  ...

Apply promos Note: Applies the following promos:
promo1 = if you buy 2 or more trousers, you get 15% off belts and shoes
promo2 = if you buy two shirts, each additional shirt costs only 45$

curl localhost:8090/carts/06f9af8a-d312-41a7-9576-f097958cc61f/promofied -X POST -d '{ "promos": ["promo1","promo2"] }'

Reponse:
{
"totalPrice":685,
"totalOffPrice":635,

"items":[
{"Item":{"id":"","name":"t1","category":"trousers","price":100},"sourceForPromos":{"promo1":true},"targetForPromos":{}},
{"Item":{"id":"","name":"t2","category":"trousers","price":110},"sourceForPromos":{"promo1":true},"targetForPromos":{}},
{"Item":{"id":"","name":"belt-1","category":"belts","price":100},"sourceForPromos":{},"targetForPromos":{"promo1":85}},
{"Item":{"id":"","name":"shoe-1","category":"shoes","price":100},"sourceForPromos":{},"targetForPromos":{"promo1":85}},
{"Item":{"id":"","name":"s1","category":"shirts","price":120},"sourceForPromos":{"promo2":true},"targetForPromos":{}},
{"Item":{"id":"","name":"s2","category":"shirts","price":90},"sourceForPromos":{"promo2":true},"targetForPromos":{}},
{"Item":{"id":"","name":"s3","category":"shirts","price":65},"sourceForPromos":{},"targetForPromos":{"promo2":45}}
]
}

The property sourceForPromos indicactes the list of promos for which the item is a source or a trigger. he property targetForPromos indicactes the list of promos that have been applied for the item.

If more than one promo applies to an item, the off-price (or discounted price) is taken to be the minimum of all promo applications So if for an item with sale price of 100, two promos apply, one with price = 95$ and other with prie = 85$. Then 85$ is considered as the off price.

Delete a cart

curl -X DELETE localhost:8090/carts/06f9af8a-d312-41a7-9576-f097958cc61f
Returns: 200 OK (probably ideally should 204 - No content)

Design Details

Given time, the project would follow the below design approach:

Cart Service:

Mantains carts in a distributed cache (such as Redis)
Gets promotions from promotions service and caches them
Applies the promotions on carts and caches the computation; Invalidating the cache if a cart is updated or promos are updated
In order to be aware of changes to promos, it listens to events from promotions service for:
1. New promotions
2. Deletion of promotions
3. Update of promotions
Promo calculation is CPU intensive. Further tests could be done to determine if it makes sense to movie Promo calcuation as service and provide a cluster of CPU heavy nodes

Endpoints and current design:

. Controllers (CartController, PromoController)
. CartController provides the following endpoints:
    - GET: /carts/{id} 
    - POST: /carts/  > Creates a new cart
    - DEL: /carts/{id}
    - POST /carts/{cartId}/promofied > Applies promos to a cart. The ids of promos that need to be applied has to be passed as an array in POST body:
    {
       promos: ["promo1","promo2"]
    }
. PromoController
    - GET: /promos : Gets all promos (as JSON objects) 
. PromoCalc : Computes and applies the promotions to a cart
. PromoCache: Provides the promos that can be applied. Currently the four promos have been hardcoded in the code. The PromoController can be easily extended to support addition of new promos
. Repository: Stores all carts. Currently an in-memory version of the repository has been immplemented. This needs to be changed to use Redis.

Representation of a promo:

A promo is represented as follows:

{
  id: "promo-id",
  description "Buy two shirt and a trouser and get a shirt free"
  buys: [
     { 
       category: "shirts",
       count: 2
     },
     {
       category: "trousers",
       count: 1
     }
  ],
 gets: [
   {
     category: "shirts"
     count: 1,
     off : {
       discount: { percentage: 100 } 
     }
   }
 ]

The above promo represents Buy two shirt and a trouser and get a shirt free. As it can be observed, the schema to represent a promo is quite flexible and can be used to represent different promo rules.

Examples promos as GO code:

// if you buy 2 or more trousers, you get 15% off belts and shoes
	promo1 := models.Promo{Id: "promo1",
		Description: "if you buy 2 or more trousers, you get 15 percent off belts and shoes",
		Buys:        []models.Buy{models.Buy{Category: "trousers", Count: 2}},
		Gets: []models.Get{models.Get{Category: "belts", All: true, Off: models.Off{Discount: &models.Discount{Percentage: 15}}},
			models.Get{Category: "shoes", All: true, Off: models.Off{Discount: &models.Discount{Percentage: 15}}}}}

	//if you buy two shirts, each additional shirt costs only 45$
	promo2 := models.Promo{Id: "promo2",
		Description: "if you buy two shirts, each additional shirt costs only 45$",
		Buys:        []models.Buy{models.Buy{Category: "shirts", Count: 2}},
		Gets:        []models.Get{models.Get{Category: "shirts", All: true, Off: models.Off{Fixed: &models.Fixed{Price: 45}}}}}

Algorithm for apply promos:

The algorithm to apply promos, groups items by category so that it can access items efficiently. Then for a given promo it executes the following logic

for true {
        // check if any of the items (or combination of items) can act as triggers (or sources) for current promo
	var appliedBuys = applyBuys(groupedItems, promo)
	// if none of the items were triggers (or sources) for the current promo, break and return
	if !appliedBuys.applied {
		break
	}
	groupedItems = appliedBuys.groupedItems
	// pass grouped items to apply promos 
	// applyGets identifies the target items for a given promo and applies them
	groupedItems = applyGets(groupedItems, promo)
	}

Note that the logic has to be repeatdely applied until it none of the target (or source) items match. This is required to especially implement rules such as "Buy 2 Get 1 Free"

Currently the project has been implemented with the following limitations: The project has the following limitations. These limitations can be addressed in future

Promotion rules are specified within the code. All thouugh this limitation can be easily done away with, as the rules themselves can be easily represented using JSON
Ideally promotion rules should be managed by promotion service 3.1 The changes in promotions can be communicated as events to cart service
Once promotions have been applied on cart, the resultant cart (also called as promofied cart) can be cached (unless an applied promotion changes or a new item is added, or deleted)
Currently all carts are held in memory. But this funcationaly is encapsulated behind a "repository" interface. A redis based repository needs to be implemented
No logging is implemented
Limited unit tests (only a few unit tests have been implemented; need to add more unit tests) (Ideally would use BDD testing framework)

Units tests

Currently only promocalc package has unit tests

go test promocalc -v