Mahboob-A/Algocode

Algocode, coding context and practice platform, is a Django based microservices architecture backend for online judge platforms just like Leetcode.

Algocode - The Leetcode for Hackers

Algocode is a DSA practice platform just like Leetcode!

Read the blog »

Algocode Auth Service . Code Manager Service . RCE Engine Service

General Information

Algocode is an online data structure and algorithm practice backend built in microservices architecture.

Algocode currently has three services: Algocode Auth Service . Code Manager Service . and RCE Engine Service

• Algocode Auth Service Auth Service handles user management.

• Code Manager Service Code Manager Service handles code submission, code exec event creation to Message Queue, and saving and caching code executing result to Database.

• RCE Engine Service RCE Engine Service contains the C++ code execution Judge. The Judge is completely isolated and it is only accessible using events.
  RCE Engine service for Java Judge and Python Judge are under development.
  
  

The C++ Judge of RCE Engine can execute C++ codes and run test cases against the code output. It can currently handle the below events:

a. AC (Accepted) 
b. WA (Wrong Answer)
c. Compilation Error 
d. Time Limit Exceed 
e. Memory Limit Exceed 
f. Segmentation Fault

NOTE

A. The C++ Judge in RCE Engine is a pure docker implementation and no other 3rd party API or service has been used.

Please refer to the respective services to learn more about it. You will find rich and detailed documentation in the respective service's github repository, I promise!

B. Rate Limit Alert

The API endpoint to submit solution to the Algocode platfrom (API in Code Manager Service ) is rate limited to 1 request per 20 seconds.

    POST https://codemanager.algocode.site/api/v1/code/submit/

You can learn more on this in the API Guideline - Code Submission in Algocode section.

C. Documentation

All the services have dedicated documentation page. Please visit the respective service repository to learn more.

Architecture and Workflow

Architecture of Algocode

Algocode Services

The Algocode platform is comprised of 3 services.
Algocode Auth Service For user management.
Code Manager Service For interacting with client, data processing and event creation for RCE Engine.
RCE Engine Service For executing the code submitted by the user and produce result for Code Manager Service to consume.

The mentioned RCE Engine is responsible to execute C++ code. The RCE Engine for Java and Python is under development.

Workflow of Algocode - The Workflow below is Sequential How Algocode Works (See the Diagram)

A. Algocode Auth Service

• The access to the Algocode begins from the Auth Service.

• Algocode Auth Service is responsible for user management.

• The user creates an account or logs in into the Algocode platform to complete the authentication/authorization process.

• PostgreSQL database is used for user management. The database is self-hosted and capable of backup and restore data periodically.

B. Code Manager Service

• The Code Manager service is the true manager which is responsible to interact with users to submit the code submission request and return them the answer of their solution.

• The user sends a POST request to the

    POST https://codemanager.algocode.site/api/v1/code/submit/

API to submit their code solution.

• Code Manager verifies the authenticity of the user, processes the submitted data, publishes the data in a RabbbitMQ instance and the queue is based on the programming language the solution is written, and returns a submission_id to the client.

• Code Manager also listens to event produced by RCE Engine. The RCE Engine executes the solution and publishes the result to an unified result queue.

• Once message is consumed by Code Manager from result queue, Code Manager caches the result in Redis and stores the result in MongoDB database.

• The client can send a GET request to the result check

    GET  https://codemanager.algocode.site/api/v1/result/check/<submission_id>/

API to get the code execution result. The API is a short polling request for 5 seconds.

• The GET request checks the cache first for the result, then checks in database and if the result is still not found, then the API process waits for 5 seconds for availability of the data.

• If the data is available within the5 seconds wait time, it returns the result, otherwise, the below response will be returned.

Response Key	Type	Value/Description
status	string	pending
data	string	null

C. RCE Engine Service

• The RCE Engine comprises either C++ Judge, or Java Judge, or Python Judge.

• The RCE Engine consumes messages from respective queue. Each RCE Engine consumes data based on the programming language queue type i.e. RCE Engine for C++ Judge consumes data from C++ queue.

• The RCE Engine processes the consumed data, prepares the data for execution and passes the data to the Judge.

• The Judge creates a secure and unprivileged docker container to run the user submitted code.

• The Judge stores the output of the code in a file after the code execution is completed.

• Once the code execution by the Judge is completed, the RCE Engine compares the output with testcases.

• After the output comparison with testcases, the RCE Engine formats the result data and publish the result data to an unified result queue in a RabbitMQ instance which is consumed by Code Manager Service.

• In case, you want to learn more about the Judge Engine, please check Online Judge - a light-weight solution to test the Judge locally.

Watch In Action

A. Long Video (Describes all the features and architecture)

• Watch from 16:30 for code execution begin and 18:30 for code submission result.

B. Short Video (Only core features)

• Watch from 09:30 for code execution begin and 11:30 for code submission result.

Deployment and Tech Stack

Deployment Information

Code Manager Service is deployed on AWS EC2 in Ubuntu 22.04 server.

Algocode Auth Service and RCE Engine Service are deployed on Azure VM in Ubuntu 22.04 server.

The DNS for the platform www.algocode.site is hosted on Cloudflare.

Tech Stack

1. Proxy and Monitoring:
a. Nginx as webserver.
b. Nginx Proxy Manager to manage Nginx.
c. Portainer to manage and monitor docker container in Auth Service.

2.  Backend and Server:
b. Gunicorn as application server.
e. Celery to process tasks asynchronously.
f. Flower to monitor celery worker.
g. Resend for email service.
h. Django as backend.
i. Django Rest Framework for API.

3. Databases and Cache:
j. PostgreSQL for user management database.
k. MondoDB for code execution result backend.
l. Redis for celery backend and cache.

4. Containerization and Code Execution:
m. Docker for containerization. 
n. Sibling Docker for secure foreign code execution.
o. RabbitMQ for asynchronous processing. 
p. Rate Limit (Token Bucket Algorighm, Django Middleware) for rate limit code submission API.

5. Deployment and Testing:
q. AWS EC2 instance for deployment
r. Cloudflare for DNS hosting. 
s. Postman for API testing.

Problem Lists

Small Note

As of today I have built the backend in microservices, there's no frontend for the project. I am fully focusing on the advanced backend engineering, hence, if you want to contribute or want to build a frontend for the project, please do not hesitate to email me here:

Problem Lists in Algocode

I have prepared a Notion Page for all the problems currently available in the Algocode project. Please visit this link Algocode Problem Lists to get all the available problems.

You will need to visit this page in order to submit your solution. Just read the problem statements, understand the problem, and copy the problem id. That's all you need to submit a solution!

Please read API Guideline - Registration in Algocode for account details and API Guideline - Code Submission in Algocode for detailed guide on how to submit a solution in the Algocode platform.

• Although you can fetch all the problems and any specific problems using the below APIs.

• To get all the Problems

    GET https://codemanager.algocode.site/api/v1/problem/all/

• The API responses with paginated result.

• To get a single Problem

    GET https://codemanager.algocode.site/api/v1/problem/<question_uuid>/

To learn more on the Code Manager APIs and documentation, please visit the documentation of Code Manager Service

API Guideline - Registration in Algocode

Prerequisites

You will need postman or insomnia to submit a solution to the project. I know you do have postman or insomnia!

Please open postman or insomnia, and continue reading ...

Registration in the Algocode Platform

The Algocode Auth Service is handling the user management.

To Know more about the Algocode Auth Service please visit the repository.

You have two ways to submit a solution to the Algocode platform. You can create your own account, or you can use the already available verified account.

A. By Using Ready to Use Account Details

Please copy the account details and the login endpoint.

  POST https://auth.algocode.site/api/v1/auth/login/

Parameter	Type	Value
email	string	connect.mahboobalam@gmail.com
password	string	12345678@1

Send a POST request to the endpoint, and you'll receive access token and refresh token.

Copy the access token and head on to the API Guideline - Code Submission in Algocode section below.

B. By Registering in the Algocode Platform

If you want to use your own account to submit solutions, please create an account.

Creating account involves two steps:

• submit the from

• verify your email address

a. Submit The Form

Copy the below submit endpoint along with the JSON data format, and send a POST request.

Please provide your authentic email address as you'll receive a token in your email from the Algocode Auth domain auth.algocode.site.

    POST https://auth.algocode.site/api/v1/auth/registration/

Parameter	Type	Description
username	string	Required Your username for the account.
email	string	Required Your valid email address.
password1	string	Required Your password.
password2	string	Required Confirm your password.
first_name	string	Required Your first name.
last_name	string	Required Your last name.

b. Check Email for Token

At this time please check your email, you will have received an email from auth.algocode.site domain.

The email will have an email validation URL to validate your account activation. Do not click the URL

As Algocode does not have any frontend, you need to verify the account manually.

Example Token URL:

To confirm this is correct, go to https://auth.algocode.site/api/v1/auth/registration/account-confirm-email/Mw:1sFEOr:xsqvnifUV5ppDFqbQBrmp2sIXGoRY63BmnFddsYTau4/

Copy everything after the account validation URL endpoint i.e. https://auth.algocode.site/api/v1/auth/registration/account-confirm-email/

For the above example, copy Mw:1sFEOr:xsqvnifUV5ppDFqbQBrmp2sIXGoRY63BmnFddsYTau4.

Send a POST request to the below endpoint to validate your account creation.

    POST  https://auth.algocode.site/api/v1/auth/registration/verify-email/

Parameter	Type	Value
key	string	Required. Your copied token from your email

Response

Response Key	Type	Value	Description
detail	string	OK	Registration is successful!
detail	string	Any value other than OK	Registration is unsuccessful. Check the token again.

If you have received the response as ok, congratulation on creating your account on Algocode!

You can now login to your account with the login endpoint mentioned above to get access token and

head on to the API Guideline - Code Submission in Algocode section below.

But in case you were not able to create account, please raise an issue and use the ready to use account instead.

API Guideline - Code Submission in Algocode

Prerequisites

• I hope you have noted the readily available user account or you have created an account in Algocode.

If you have not done this, please check API Guideline - Registration in Algocode section for guideline.

• Get the acces token using the user credentials as per API Guideline - Registration in Algocode section.

• To submit a solution, you should visit the notion page Algocode Problem Lists for the available problems in the Algocode.

Go through the problems, and copy the problem_id of the problem you want to submit a solution.

Why notion? Did you ask?

Please check the Problem Lists section for information.

Code Submission in Algocode

The Code Manager Service is handling all the activities mentioned in this section.

To Know more about the Algocode Code Manager Service please visit the repository.

Please follow the below steps to submit a solution in Algocode.

Please Note - Rate Limit Alert

• The /api/v1/code/submit/ API endpoint to submit code solutions is Rate Limited.

The API endpoint is rate limited with one request per 20 seconds i.e. 3 requests per minute.

• Did you ask why?

Well, I am using the free AWS and Azure servers with only 1GB of RAM! Hence I have no choice but to Rate Limit the API.

• I have implemented the Rate Limit using Token Bucket Algorithm and I have applied the Rate Limiter class Project Wide for the quoted API using Middleware in Django.

• Click Here to see the implementation of the Rate Limiter Class.

Step 01

Write your solution for the problem you have chosen in C++, as currently C++ is supported. Now, you have to convert it into JSON format. As the Algocode does not have client till now, to share data to the backend services, you need to convert it into JSON. Here's how you can convert the code in JSON in python.

import json

code = """

# paste your code here.

"""
code = json.dumps(code)

print(code)

Copy the printed code from terminal and and continue reading to learn about the request body.

Step 02

As you have copied the access token as per the API Guideline - Registration in Algocode section, please paste the access token in Bearer Token option in Auth section in Postman or Insomnia.

Now, you are ready to submit the solution to the Algocode platform. Please send a POST request to the below API following the API reference.

    POST https://codemanager.algocode.site/api/v1/code/submit/

Parameter	Type	Value/Description
problem_id	string	Required. The problem_id of the problem you are submitting the solution.
lang	string	Required. cpp. Currently cpp is supported. java RCE Engine is under development.
code	string	Required. Your solution for the problem in JSON format.

Example Payload

Here's an example payload for one of a problem in Algocode Sqrt(X). The problem is same as this Leetcode Problem.

Example Payload

{
    "problem_id": "f17f511a-8c53-41d3-b750-7673d25835af", 
    "lang": "cpp", 
    "code": "#include <iostream>\n\nint mySqrt(int x) {\n    if (x == 0) return 0;\n    int left = 1, right = x, result = 0;\n    while (left <= right) {\n        
     int mid = left + (right - left) / 2;\n        if (mid <= x / mid) {\n            result = mid;\n            left = mid + 1;\n        } else {\n            right = mid - 1;\n        }\n    
    }\n    return result;\n}\n\nint main() {\n    int t;\n    std::cin >> t;\n    while (t--) {\n        int x;\n        std::cin >> x;\n        std::cout << mySqrt(x) << 
    std::endl;\n    }\n    return 0;\n}\n"
}

Response

Once you have submitted your solution to the Algocode Code Manager Service, you will receive a response with a submission_id. You would be able to see the result of your solution using the submission_id.

Response Key	Type	Value/Description
result	dict[dict]]	A dictionary containing the request's response.
detail	string	Your response has been submitted
submission_id	string	The submission_id for your solution submission.

Example Response

{
    "result": {
        "detail": "Your response has been submitted.",
        "submission_id": "5993d00f-62fa-437c-8724-ee588265175b"
    }
}

Please check the API Guideline - Result Check in Algocode section to learn how to check the code submission result.

API Guideline - Code Result Check in Algocode

Code Submission Result Check in Algocode

Please make a GET request to the below API to get the result of your solution.

    GET  https://codemanager.algocode.site/api/v1/result/check/<submission_id>/

Parameter	Type	Description
submission_id	string	Required. The submission_id of your solution as per the API Guideline - Code Submission in Algocode section.

You would see the result of your solution.

Code Submission Result Examples

Some Code Submission Result Snapshots

A. AC Solution

B. WA Solution

C. Compilation Error

D. Segmentation Fault

E. Memory Limit Exceed

F. Time Limit Exceed

Contributing and Run Locally

Contribution and Development

If you want to contribute or you want to run locally, then you can fork the development branch on each service mentioned in the Algocode Platform.

Please follow the .envs-examples to know the env-variables you would need to run the project locally.

All the services are dockerized project. You just need to cd src, create virtual environment, activate it, and run make docker-up and That's it!

This will run the project for you.

Please follow the service that you want to contribute or run locally to get detailed guideline on local development.

Lessons Learnt and Challenges

The Backstage

The project itself was a challenge for me!

Once one of my mentors told me

Do the hard things while you are learning, so that the implementation becomes easier for you.

I completely agree with this statement. I enjoy dealing with complex stuff, and bugs give me the kick I enjoy!

Well, enough praise of myself.

But I was not lying. When I thought to build the project 2 months ago (it took 1.5 months to wrap up the project, two weeks was idle). I had zero knowledge whether I would be able to do it, but I had confidence that somehow I would do it. My thought process was simple - "I do not know exactly how, but I would do it somehow for sure!"

And I am writing this Readme today that I have completed the project, and somehow I have made it! That's my motivation.

I know something is not simple as it sounds, but I know, somehow I would manage it!

Challenges

• The initial challenge was the design. Designing a complex project like Online Judge in microservices to build from scratch was not easy as it sounds.

• The communication between microservices were fun discovery. I was searching for optimal solution and I learnt RabbitMQ for this cause, and I ended up writing a nice blog on RabbitMQ 101. Read RabbitMQ 101 Here. The Algocode platform is using an RabbitMQ instance from CloudAMQP platform.

• Building a Online Judge from scratch was the hardest part to accomplish. At times, I though to abandon the idea of building the online Judge from scratch and use 3rd party APIs, but my instinct did not allow it!

• However, after countless hours of debugging and reading internals of docker networking, docker volume, docker containers, docker pyhton sdk, docker in docker and sibling docker, security in docker, linux internals - I have explored countless number of internals and fixed hundred of bugs that taught me a lot. This resulted an Online Judge built from scratch using docker. I have learnt so many new concepts while building the project just proper by googling!.

• The next struggle was the production build for the project. I had to shift from development build to production build and eventually to deploy the project in AWS and Azure servers.

• AWS closed my two new account to deploy the project, but after countless of emailing the aws support, they activated my account and finally I could deploy my project.

• To wrap up, the entire project was a challenge for me: The design stage to research stage, to development stage to production stage to deployment stage - I had to repeat all these steps with all the 3 (currently) services behind the Algocode platform.

Learnings

• I have gained practical experience with RabbitMQ building this project.

• I have gained deep knowledge on docker, docker volumes, docker networking, Lunux internals and many more.

• As the project is heavily dealing with files, I have gained valuable experience with file handling with python.

• As I have built the project from research, design, dev, production to deployment, I have gained invaluable knowledge on design, development, production and deploy the project in cloud services like AWS or Azure.

• As the project is in microservises architecture, I have gained practical knowledge on communication, networking, between all other services; experience with cloud providers such a AWS, Azure and onverall dev to production of a SDLC.

Please visit the service repositories mentioned in the introduciton section for detailed guild on the services of Algocode platform.

🔗 Links