crypto-robot-validator

Crypto Data Validator

1. About the Project
   1. Input
   2. Output
   3. Persistence
      1. Client DB
         1. Schema
         2. Operation
         3. Query
      2. Operation DB
         1. Schema
         2. Operation
         3. Query
      3. Lock DB
         1. Schema
         2. Operation
         3. Query
   4. Rules
   5. Built With
      1. Dependencies
      2. Compiler Dependencies
      3. Test Dependencies
   6. Roadmap
2. Getting Started
   1. Prerequisites
   2. Installation
   3. Requirements
      1. Deploying Local Infrastructure
   4. Usage
      1. Manual Input
      2. Docker Input
   5. Testing
3. About Me

About the Project

The objective of this project is to receive an operation event for a single client, validate it and approve operation execution.

Input

The input should be received as an SNS message sent through an SQS subscription. This message will trigger the lambda handler to perform the service. Operation events should have information needed to track client who will execute the operation, the operation type, crypto being traded, operation event time sent and available amount (if operation type is BUY amount should be in cash, otherwise if operation type is SELL, amount should be used like cryptocurrency amount).

Also, this step should validate user stop losses configuration and may lock the client until top loss is available again.

Example of how the data received should look like:

{
  "client_id": "aa324edf-99fa-4a95-b9c4-a588d1ccb441e",
  "operation": "BUY",
  "symbol": "BTC",
  "start_time": "2022-09-17T12:05:07.45066-03:00"
}

Output

Since this is an async application there is no output to be returned, but operation events are generated from the data received. Operation events should have information needed to get the operation from its db, the operation id.

Example of how the line should look like:

{
  "operation_id": "aa324edf-99fa-4a95-b9c4-a588d1ccb441e"
}

Persistence

Client DB

Client DB is the database that contains the client information and configuration needed to trigger the operations.

{
  "id": "aa324edf-99fa-4a95-b9c4-a588d1ccb441e",
  "active": true,
  "locked_until": "2022-09-17T12:05:07.45066-03:00",
  "locked": false,
  "cash_amount": 100,
  "cash_reserved": 0.00,
  "crypto_amount": 0.0000312,
  "crypto_reserved": 0.0,
  "symbols": [
    "BTC",
    "SOL"
  ],
  "buy_on": "STRONG_BUY",
  "sell_on": "SELL",
  "ops_timeout_seconds": 60,
  "operation_stop_loss": 50.00,
  "day_stop_loss": 500.00,
  "month_stop_loss": 500.00,
  "summary": [
    {
      "type": "MONTH",
      "day": 1,
      "month": 8,
      "year": 2022,
      "amount_sold": 23000.42,
      "amount_bought": 37123.42,
      "profit": 1032.32,
      "crypto": [
        {
          "symbol": "BTC",
          "average_buy_value": 230020.42,
          "average_sell_value": 235020.42,
          "amount_sold": 0.00231,
          "amount_bought": 0.00431,
          "profit": -53.00
        }
      ]
    },
    {
      "type": "DAY",
      "day": 14,
      "month": 8,
      "year": 2022,
      "amount_sold": 23000.42,
      "amount_bought": 37123.42,
      "profit": -53.00,
      "crypto": [
        {
          "symbol": "BTC",
          "average_buy_value": 230020.42,
          "average_sell_value": 235020.42,
          "amount_sold": 0.00231,
          "amount_bought": 0.00431,
          "profit": -53.00
        }
      ]
    }
  ]
}

This application supports the following operations to the Client DB:

• Read ops:

  • Used to find clients using client_id

• Write ops:

  • Used to lock clients using client_id

This is the query used to get clients from DB:

    expr, _ := expression.NewBuilder().WithFilter(
    expression.And(
    expression.Name("client_id").Equal(expression.Value(client_id))),
    ),
    ).Build()

This is the query used to update (lock/unlock) clients:

    expr, _ := expression.NewBuilder().WithFilter(
    expression.And(
    expression.Name("client_id").Equal(expression.Value(client_id))),
    ),
    ).Build()

Operation DB

Operation DB is the database that contains the created operation's information.

Operation statuses:

• CREATED
• PENDING
• COMPLETED
• ERROR

{
  "id": "aa324edf-99fa-4a95-b9c4-a588d1ccb441e",
  "status": "COMPLETED",
  "created_at": "2022-09-17T12:05:07.45066-03:00",
  "expires_at": "2022-09-17T12:05:07.45066-03:00",
  "completed_at": "2022-09-17T12:05:07.45066-03:00",
  "locked": false,
  "quote": "BTC",
  "base": "BRL",
  "type": "BUY",
  "amount": 100.00,
  "stop_loss": 50.00,
  "profit": 1.0,
  "transactions": [
    {
      "type": "BUY",
      "unitary_value": 100000.00,
      "quote_amount": 0.001,
      "base_amount": 100.00,
      "created_at": "2022-09-17T12:05:07.45066-03:00",
      "expires_at": "2022-09-17T12:05:07.45066-03:00",
      "confirmed_at": "2022-09-17T12:05:07.45066-03:00"
    },
    {
      "type": "SELL",
      "unitary_value": 101000.00,
      "quote_amount": 0.001,
      "base_amount": 101.00,
      "created_at": "2022-09-17T12:05:07.45066-03:00",
      "expires_at": "2022-09-17T12:05:07.45066-03:00",
      "confirmed_at": "2022-09-17T12:05:07.45066-03:00"
    }
  ]
}

This application supports the following operations to the Client DB:

• Write ops:
  • Used to create new operations

This is the query used to create operations in DB:

    expr, _ := expression.NewBuilder().WithFilter(
    expression.And(
    expression.Name("client_id").Equal(expression.Value(client_id))),
    ),
    ).Build()

Lock DB

Lock DB is the database that contains the client_id's locked during execution.

OBS: Redis is used for this DB.

This DB uses key value to store validator client_id's locked

{
  "crypto_robot.validator.lock.{client_id}": "{client_id}"
}

This application supports the following operations to the Lock DB:

• Read ops:

  • Used to find locked client_ids

• Write ops:

  • Used to lock client_ids

This is the query used to get locked client_id's from DB:

    expr, _ := expression.NewBuilder().WithFilter(
    expression.And(
    expression.Name("client_id").Equal(expression.Value(client_id))),
    ),
    ).Build()

This is the query used to update (lock/unlock) client_id's from DB:

    expr, _ := expression.NewBuilder().WithFilter(
    expression.And(
    expression.Name("client_id").Equal(expression.Value(client_id))),
    ),
    ).Build()

Rules

Here are some rules that need to be implemented in this application.

Not Implemented:

Client validations:

• Client must be active
• Client must not be locked
• Current date must be greater than locked_until value
• Client must have enough cash to buy minimum allowed amount of crypto
• Client must have enough crypto to sell minimum allowed amount
• Client must have the coin symbol selected inside config.symbols variable to operate it
• Buy operations should be triggered when the summary received is equal or less restricting than the config.buy_on value.
  • For example if the config value is equal to BUY and a STRONG_BUY analysis was received, the operation should be allowed, and the opposite should be denied.
• Sell operations should be triggered when the summary received is equal or less restricting than the config.sell_on value.
  • For example if the config value is equal to SELL and a STRONG_SELL analysis was received, the operation should be allowed, and the opposite should be denied.
• Operations should not be triggered if daily_summary.proffit has a negative value of more than or equal to the config.day_stop_loss value.
  • daily_summary.day value should be checked to see if current day has changed, in this case, the values should be updated to start a new day.
• Operations should not be triggered if monthly_summary.proffit has a negative value of more than or equal to the config.month_stop_loss value.
  • monthly_summary.month value should be checked to see if current month has changed, in this case, the values should be updated to start a new month.

Lock:

• Ids received should be locked on Redis for execution and unlocked after, even if error occurred.
• Clients should be locked on DynamoDB for execution and unlocked after, if error occurred after DynamoDB lock, clients should be unlocked.
• If client fails validation locked_until value could be set on DynamoDB to lock for an extended amount of time (stop loss block for example)

Operations:

• Operation should be created with status CREATED and it's id should be sent to the SNS topic for later execution.
• Operation amount should be created using client configuration and Biscoint current unitary value.

Biscoint:

• Client balance should be validated from Biscoint and updated in DynamoDB clients DB.

Built With

This application is build with Golang, code is build using a Dockerfile every deployment into the main branch in GitHub using GitHub actions. Local environment is created using localstack for testing purposes using crypto-robot-localstack.

Dependencies

• aws/aws-lambda-go: Used in Lambda Handler integration
• aws/aws-sdk-go-v2: Used in SNS and DynamoDB integration
• google/uuid: Used to generate uuids
• joho/godotenv: Used to map .env variables

Compiler Dependencies

• golangci/golangci-lint: Used to enforce coding practices

Test Dependencies

• cucumber/godog: Used to run integration tests
• stretchr/testify: Used to perform test assertions

Roadmap

• Implement Behaviour tests (BDD)
• Implement Unit tests
• Implement application logic
• Create Dockerfile
• Create Docker compose for local infrastructure
• Add biscoint mock docker image to docker compose file
• Document everything in Readme

(back to top)

Getting Started

Prerequisites

• Install Golang

  • Windows/MacOS/Linux
    • Manual
  • macOS
    • Homebrew

      brew install go
      

  • Linux
    • Via terminal

      sudo add-apt-repository ppa:longsleep/golang-backports
      sudo apt update
      sudo apt install golang-go
      

• Install Docker

  • Windows/macOS/Linux/WSL

Installation

• Run the following to install project dependencies:

  • Windows/MacOS/Linux/WSL

    go mod download
    

• Run the following to compile the project and generate executable:

  • Windows/MacOS/Linux/WSL

    go build -o bin/validator cmd/validator/main.go
    

Note: the binary generated will be available at ./bin folder.

Requirements

To run the application locally, first a local infrastructure needs to be deployed

Deploying Local Infrastructure

This requires docker to be installed. Localstack will deploy aws local integration and create the topic used by this application to send the events.

Obs: Make sure Docker is running before.

• Start the required infrastructure via localstack using docker compose command:

  • Windows/macOS/Linux/WSL

    • Run the application as a local lambda with docker

      docker-compose -f ./build/local/docker-compose.yml up
      

    • Run the infrastructure with docker but run the application locally (on the IDE or using the binary)

      docker-compose -f ./build/local/docker-compose-local.yml up
      

• To stop localstack:

  • Windows/macOS/Linux/WSL (Use docker-compose-local.yml if needed)

    docker-compose -f ./build/local/docker-compose.yml down
    

Usage

Manual Input

• Start the compiled application locally:
  • Windows/macOS/Linux/WSL

      VALIDATOR_ENV=development go run cmd/local/main_local.go 
    

• To run the application like this the local infrastructure needs to be deployed using docker-compose-local.yml
• To stop the application just press Ctrl+C

Docker Input

• In case you want to use a Docker container to run the application first you need to deploy the local infrastructure from docker-compose.tml file. This will build the docker image using the Dockerfile located in the root folder and make it available as a lambda. After that you just need to send an event to the SNS topic subscribed using the following command:

  • Windows/macOS/Linux/WSL

      ./scripts/send_message_command.sh 
    

• To run the application like this the local infrastructure needs to be deployed using docker-compose.yml

Testing

• To run the unit tests:

  • Windows/macOS/Linux/WSL

    go test ./test/unit/...
    

• To run the integration tests:

  • Windows/macOS/Linux/WSL

    go test ./test/integrated/...
    

• To run the all tests:

  • Windows/macOS/Linux/WSL

    go test ./...
    

(back to top)

About me

Hello! :)

My name is Luis Brienze, and I'm a Software Engineer.

I focus primarily on software development, but I'm also good at system architecture, mentoring other developers, etc... I've been in the IT industry for 4+ years, during this time I worked for companies like Itaú, Dock, Imagine Learning and EPAM.

I graduated from UNESP studying Automation and Control Engineering in 2022, and I also took multiple courses on Udemy and Alura.

My main stack is Java, but I'm also pretty good working with Kotlin, Typescript and Go (backend only). I have quite a good knowledge of AWS Cloud, and I'm also very conformable working with Docker.

Also, I have experience working with relational (PostgreSQL, Microsoft SQL Server, MySQL, ...) and non-relational ( DynamoDB, Redis, Cassandra, ...) databases.

During my career, while working with QA's, I've also gained a lot of valuable experience with testing applications in general from unit/integrated testing using TDD and BDD, to performance testing apps with JMeter for example.

If you want to talk to me, please fell free to reach me anytime at LinkedIn or e-mail.

(back to top)