server is crashing, please help how can i start prooject locally, i followed the insterection from read_me.md but nothing is working
and databse_url should be this DATABASE_URL="postgresql://postgres:supersecurepassword@localhost:5432/algo-arena"

Can we make the cloudflare captcha only work at production?

• Make the cloudflare captcha only work at production

Bug Description:

In the problem submission table, the number of test cases passed for each submission is always displayed as 0, regardless of the actual number of test cases that were passed. This issue persists across all problem submissions and misleads users regarding the success of their code.

Steps to Reproduce:

1. Navigate to the problem submission page.
2. Submit a solution to any coding problem.
3. Observe the "Test Cases Passed" column in the submission table

Expected Behavior:

The "Test Cases Passed" column should reflect the actual number of test cases passed for each submission.

Actual Behavior:

The "Test Cases Passed" column displays 0 test cases passed for every submission.

Screenshots:

Currently, only int can be taken as an input or output for a problem

add a way to take larger numbers as well up to 1e18

Currently the dependencies are not uptodate. Turbo version complaining on the key tasks. Which is migrated from pipeline key in the turbo@1^ to turbo@2^.

Migrate to pnpm. Update the scripta and fix all the dependencies and push the latest lock file.

If I submit twice, it doesnt go in the second time

currently in our project we don't have husky if we add husky it is very useful to check(commit message) or lint the project before the actually commit has made.

rn , the footer seems old and default, and no links are working.

please add a discord server for this project, that would be agreat help

Currently, the answers are checked by matching each token of the output with the correct and

But for questions that have more than one correct answer, this might be wrong

For eg if the question is to print an odd number between a and b there can be multiple odd numbers that can be printed

I suggest allowing the problem setters to add custom checkers which will take the output of the users and then return the verdicts

For the example of odd numbers the script will check if the output is odd and is between a and b

docker setup, with running apps/web and apps/submission-webhook together with db running locally

Anytime there is a problem submission that is made (similar to https://github.com/code100x/algorithmic-arena/pull/54/files#diff-61558a2c20b5fca331bfb30009dcd693af0267be01f204f658f960b9f50befba ), we should push this problem (hidden) in the DB via a CI/CD pipeline

Datasource "db": PostgreSQL database "algo-arena", schema "public" at "localhost:5432"
│
│ Error: P1001: Can't reach database server at localhost:5432
│
│ Please make sure your database server is running at localhost:5432.
│  ELIFECYCLE  Command failed with exit code 1.
│ command finished with error: command (/Users/veerpratap/Desktop/Cohort-3/algorithmic-arena/packages/db) /opt/homebr
│ ew/bin/pnpm run db:migrate exited (1)
└────>
ERROR run failed: command exited (1)

A single user should be able to send at max 2 requests / 10 seconds

This should be done in a distributed way, use redis to store the rate at which the user is sending requests

Right now, our test cases have to go to judge0 via an HTTP call
If the test cases are very big, then you send a very big payload in an http call

If the test cases were direclty mounted on the judge0 server, then they can be directly picked up from there

Currently, to add problems, problem setters needs to clone the repo (which will probably be private at a later stage in production, so granting permission might be an overhead in that case) and then create a PR.

How about creating an interactive UI to add problems?

Rough sketch of what I am planning

Also, I have some ideas regarding the automation of other steps regarding this, so that the effort required by admin is minimum.

Please let me know whether you guys want this feature.

Currently having both npm and yarn. Migrate fully to yarn and fix some linting issues when using yarn as the package manager.

proper auth system should be there, currently there is not proper auth

@hkirat
Is it correct way of solving this problem since Redis in single threaded so it gives us feature of mutex lock which will help to resolve this .
Since we want three things

1. Maximum time
2. Maximum memory
3. Submission Status (AC or Reject)

These all three things will be handled by submission-callback endpoint
And Nextjs API used for polling will eventually show data on basis of this Redis

I could see three services in this repository:

1. A PostgreSQL DB
2. Redis (Caching and Queue)
3. Express as a store of APIs

We should move towards writing tests for it! Unit Tests are fast but only test the business logic (they can't test whether the SQL query written is actually correct, a code passing all the unit tests might fail the code in production!) and Integration tests are slow (can become a headache for developers).
Therefore we can move to E2E Tests. These tests are faster than Integration tests and can be extended to test the action of these services!

So the route is simple:

1. We need to setup testing instances of all the services, as tests should not contact the real instances. docker compose can be a good idea for this.

2. We need to replace the the actual services with the testing instances when tests are running, that means Dependency Injection.

But there are some problems:

a) Dependency Injection Design Pattern: As far I know, there is no good library available for this in Nodejs. tsyringe is there, but it also uses a global instance of container, hence second point can't be achieved!

b) Docker might make the CI slow!

Solutions:

a) Traditional way of Dependency Injection (in JAVA) is by using Constructors and decorators on controllers but JS is a functional programming, so either we have to migrate to a new OOPs based framework like Nestjs or we might use a hack by just changing the protocol to GraphQL and using the context variable for Dependency Injection.

b) Caching in DockerFile and multi stage builds can fix this.

See this repo: https://github.com/Manik2708/Hi_Server which is implemented in Nestjs with E2E Tests and docker compose in CI.

WE need to protect a few routes to ensure user isnt able to hit submit endpoints/see problems before a contest starts.
After the contest we need to make sure user cant send an activeContest variable, and if they do, their request is rejected

read about it here - https://blog.cloudflare.com/turnstile-private-captcha-alternative

After Writing Code its not Submiting

Make our db schema such that judge0 can directly talk to it

Currently, the problem setter has to run the test cases manually and then write them to their respective files which is annoying if there are many test cases

Add a shell script which runs the solution file on the input and writes the output to their respective files

2 dimensional arrays can be added to Structure.md
Should be parsed for every language

Need to add boilerplate generators

I'll work on this