Garedami(ガレダミ)

Garademi is auto grading system that use for judging user's source code.

Garedami are from 2 words.

• Gareda(ガレダ) is Grader
• mi(ミ) is stand for Miracle

Ability

• Compile and run in any programming language.
• Flexible Configurable Problem

Requirement

• python 3.8
  • yaml
  • json

To use it you have to call function Judge in judge.py and require 4+2 arguments.

• idTask (as str) ... self-explanatory but we don't use them much.
• proLang (as str) is programming language that user what to compile
• problemDir (as str) is Directory of that problem.
• src (as str) is source code that user want to judge
• timeJudge (as int) (optional)
• memJudge (as int) (optional) is memory that use for during Judge (available for linux)

how it's working

After call by judge.py, it will do 4 main Process

1. Gathering problem info
2. Checking src lang with Problem
3. Compile src
4. Run and Judge src

Configulation

You can config anything you want in Config Folder

If it not found, it will create by itself for first time

Config folder will look like this

Config/
├── Langs/
|   ├── C.yaml
|   └── Cpp.yaml
|   ...
├── Grader.yaml
├── ProblemDefault.yaml
└── Verdict.yaml

Langs

in each lang you can config 3 things

• BIN_FILE: Path for Binary file
• BIN_PATH: Path for Binary Folder
• TIME_FACTOR: Time factor for each language (eg. C = 1, Python = 5,Java = 1.5) Because each language have different performance

Grader

• COMPILE_TIME: Compiler time (in ms)
• JUDGE_TIME: Judge time (in ms)
• MAX_DISPLAY: Max verdict to display. If it over, we will use short form.
• MAX_TEST_CASE: self explain

Default_Problem()

is config for each problem

• timeLimit: is Time limit in each testcase (in ms)
• memLimit: is Memory limit in each testcase (in md)
• judging: is cmd for judging in each testcase
• compiling: is cmd for compile in each language
• running: is cmd for run in each language

Verdict

use to convert verdict symbol to Full form.For example

• P: Pass
• -: Wrong answer
• T: Time Limit Exceed

...

• ?: use for unknown verdict symbol

Problem or Task

In this Garademi you can make a lot of thing in one problem. So I will show example Basic problem to FULL custom problem

Basic Problem (No Custom Judge)

It just use for compare between src output and solution in each test-case. Here is stucture for Basic Problem.

SomeProblem/
├── CompileSpace
|   |-- Custom_Header...
|   └── Src_Program
|   
├── 1.in
├── 1.sol
├── 2.in
├── 2.sol
├── ...
├── 10.sol
└── Config.yaml (Not require but better than noting)

Every time when you use grader.User's src will go in CompileSpace\Src_Program and run it.

It will use input from ?.in and check output with ?.sol.In each ?.in and ?.sol we call it Testcase.You can have any Testcase Number that you want.(But not over than MAX_TEST_CASE in Config\Grader)

Fully Customed Problem.

In Basic Problem you must output exactly as ?.sol.But some problem you there are more than one answer.You can use Custom Judge by config them in Config.yaml.You can look example from StandardJudge\StdJudge which just use in Basic Problem.

SomeProblem/
├── Shadow
|   └── Original Header
├── CompileSpace (will copy from Shadow)
|   |-- Custom_Header...
|   └── Src_Program
|   
├── 1.in
├── 1.sol
├── 2.in
├── 2.sol
├── ...
├── 10.sol
├── Custom_Judge
└── Config.yaml 

It's not different from Basic Problem

But you can custom in Config.yaml and Custom_Judge