HTTPS SSH

What is ineffable?

Simple command-line grader for local grading of solutions for problems of competitive programming contests. It is written in Python and it uses timeout Perl library to limit resources (time and memory limits) of program being tested.

It works on Ubuntu, and I guess it should work on any Linux/Unix system as long as Python and Perl are installed.

Screenshots


(Original quote is "Success consists of going from failure to failure without loss of enthusiasm")

Installation

Run the following commands on your terminal:

$ mkdir ~/.ineffable
$ cd ~/.ineffable
$ wget https://bitbucket.org/silap/ineffable/downloads/ineffable-1.0.zip
$ unzip master.zip
$ rm master.zip
$ mv silap-ineffable-*/* .
$ rm -r silap-ineffable-*/
$ sudo ln -s ~/.ineffable/ineffable /usr/local/bin/ineffable

Or alternatively you may run this command if you have curl installed:

$ bash < <(curl -sL https://bitbucket.org/silap/ineffable/raw/f208b207160eacd8ba4c30c078d1255e309c6b40/install.sh)

Configuration

Whenever you run ineffable it looks for ineff.json file. You can also tell ineffable which configuration file to use using this syntax: inefabble another.json.

Configuring ineffable is easy and quite intuitive. After you install ineffable you can view the example configuration file that is located at ~/.ineffable/ineff.json.example.

Available configuration options are as follows:

key example value Description
pkdesc "Balkan OI 2014 Day 2 Problem Ephesus" Package description. (Optional)
pbcode "steeple" Problem code. Ineffable searches for string {PROBLEM} in keys warmup, execute, inp, out and tstdir, and replaces them with pbcode value. (Optional)
tl 1000 Time limit (in milliseconds). (Required)
ml 65536 Memory limit (in kilobytes). (Required)
warmup "g++ -O2 -o solution solution.cpp" Command to be executed before the grading starts. For example, it can be used to compile C++ code. Grading stops if the command exits with non-zero code (such as, when compilation fails). (Optional)
execute "./solution" or "python solution.py" Command that runs the solution. <span style="color:red">Solution needs to read from stdin and write to stdout.</span> (Required)
tstdir "steeple_tests" Directory where the test files are located. If ommited, it is assumed that tests are located in the current directory. (Optional)

There are 2 different ways to show ineffable which files are input files and which files are output files:

  1. Smart stars:

    "inp": "input*.txt",
    "out": "output*.txt"
    
  2. Listing files.

    # Syntax 1
    
    "iopairs": {
      "input1.txt": "output1.txt",
      "input2.txt": "output2.txt",
      "input3.txt": "output3.txt"
    }
    
    # Syntax 2
    
    "iopairs": {
      "inp": {"input1.txt", "input2.txt", "input3.txt"},
      "out": {"output1.txt", "output2.txt", "output3.txt"}
    }
    

See it at work (sample session)

Let's solve problem Cow Steeplechase of USACO 2011 November Contest.

$ # Create folder `steeple` in your desktop and another folder `steeple_tests` inside it.
$ mkdir Desktop/steeple
$ cd Desktop/steeple
$ mkdir steeple_tests
$ cd steeple_tests
$
$ # Download and extract test data into `steeple/steeple_tests/`.
$ wget -q http://usaco.org/current/data/steeple.zip
$ unzip -q steeple.zip
$ rm steeple.zip
$ cd ..
$
$ # Create file `ineff.json` and configure.
$ touch ineff.json
$ echo '{
  "pkdesc": "Cow Steeplechase (USACO Gold November 2011 Contest)",
  "pbcode": "steeple",
  "tl": 3000,
  "ml": 65536,
  "warmup": "g++ -std=c++11 -O2 -o {PROBLEM} {PROBLEM}.cpp",
  "execute": "./{PROBLEM}",
  "tstdir": "{PROBLEM}_tests/",
  "inp": "I.*",
  "out": "O.*"
}' >ineff.json
$
$ # Create `steeple.cpp` and open it with your favorite editor to code your solution.
$ touch steeple.cpp
$ vim steeple.cpp
$ # ...
$
$ # Test your solution
$ ineffable

As I haven't spent much effort in making ineffable, there may be some bugs. I won't be able to fix the bugs at least till the IOI 2015 ends.

Ineffable is a simple project and anyone who knows some Python can alter it for their own purpose or even add some features. Project is available on Bitbucket (sorry no Github, it is blocked in my country :/).