codint

codint is a collection of programming problems with a standardized test script.

Use this tool to practice for your coding interview. You can also use it to pick up new programming languages quickly!

Requirements

• python3 and pytest

How to Use

Getting Started

Clone this repo to a local directory.

Inside this directory, initialize test scripts:

bash setup.sh clear && bash setup.sh distribute

• The above command sends the testtemplate.py in the repo root to each problem's subdirectory as a symlink, to make sure they can be recognized by pytest.

Now, run pytest -slv to run a diagnosis of all existing coding problems and test exising solutions for each problem.

Read a Problem

cd into any of the top-level directories. Inside you will find subdirectories each named d[x]_[pName], where [x] is the difficulty rating of the problem ([0-2], the lower, the easier), and [pName] is the name of the problem. Each of these contains material for a programming problem.

• For example, d1_traverseBST is a valid subdirectory name in this format.

cd into one of these. Read PS.md for the problem statement.

Write Solutions

Write a solution to the problem as described in the PS.md writeup. You can write in any of the supported languages. Your solution must be a single file. The naming convention is [pName].[suffix], where [pName] is the problem name as inferred from this directory name (without the difficulty rating), and [suffix] is that of your programming language.

• If the problem directory is d1_traverseBST and you decide to write C code, your solution should be called traverseBST.c.

• If a solution already exists, just delete it. It's in repo history, anyways.

• If you write multiple solutions, starting with the second solution you may append a number after pName. e.g. traverseBST2.c, traverseBST3.c. These will all be recognized by the test script.

  • The numbers are unlimited but they cannot be skipped: Do not write the 3rd solution without the 2nd solution.

  • Numbers do not overlap across languages. If you write two solutions in C and Python each, they should be called traverseBST.c, traverseBST2.c, traverseBST.py, traverseBST2.py.

Your solution should read all input from stdin. By default any single-line input will be postfixed with a newline character.

Your solution should write all output to stdout. Make sure the end of the output contains a newline character.

Test Solutions

In the problem directory, run pytest -slv to test all solutions for this problem following the correct naming convention.

Test cases are located in d[x]_[pName]/tests/. The naming convention is t[n]-in and t[n]-out, where [n] is the test case number, which must be named in order, starting from 1.

• For example, t1-in and t1-out together form a complete test case.

• When a solution takes in the contents of t[n]-in via stdin, it is expected to print content equal to t[n]-out to stdout.

• Unless, if the PS.md file of this problem includes a -anyOrder flag on the second line, that will make the test script ignore line order when comparing solutions. In that case a test case will pass as long as each line of t[n]-out is printed, no matter the order.

The test script will test all solutions it can find in all supported languages (according to the solution naming convention) against all test cases it can find (according to the test case naming convention).

If a test case fails, pytest will output these variables:

• sol which is the solution being run
• thisIn and thisOut which is the failing test case
• thisSolOut which is the solution's output
• thisRefOut which is the solution's output

For compiled languages (such as C), the test script will manage the compilation and clearance of the compiled executable after testing finishes.

Current Development

Supported Languages

The following language-specific requirements are only required if you plan to write and test code in that language.

• Python3
• C
  • gcc must invoke a C compiler.
  • On macOS, gcc invokes clang by default. This also works.
• C#
  • mcs must invoke a C# compiler, and mono must invoke a C# runtime.
  • These can be installed with the mono package.

Supported Flags

• -anyOrder: The test script will ignore line order when testing solutions. A test case will pass as long as each line of the correct output is printed, no matter the order.

Contributing

Adding new test cases for an existing problem

Just create new t[n]-in and t[n]-out files in the tests directory. Remember that test case numbers cannot be skipped. Do not write the 9th pair of test case if there's no 8th yet.

It is advisable to put easier test cases first (in smaller numbers) as the test script tests them in increasing order.

Notice that input strings must end with "\n". For a test case input to be an empty string, do not simply create an empty file. Make sure that there is a newline character. In vim this can be done by inserting a new line, and deleting the second line with dd. This step is not necessary for non-empty strings.

Adding new programming problems

In an appropriate category directory, create a sub-directory to host the new problem. Name it d[x]_[pName], as outlined.

• [x] is the difficulty rating. Roughly 0 means easy, 1 means medium, 2 means hard.

• For [pName] follow the camelCase naming convention.

In the new problem directory, make sure there is:

• PS.md: Reference the format from a PS.md from another file.

  • If it's needed to supply any flags, specify them on the second line with a space between any two adjacent flag strings, but no space at the start of line. Here's an example of a PS.md on the second line:

    -anyOrder -otherFlag -anotherFlag
    

  • If no flag is needed, leave the second line blank.

• tests/: A directory to host the test cases. This is required for the setup script to recognize this directory as a problem directory.

  • For things to keep in mind while designing test cases, see Adding new test cases for an existing problem.

When the directory is prepared, go to the repo's root, and run bash setup.sh distribute to redistribute the test script symlink:

Now run pytest -slv and confirm that the new programming problem is included in the list of all problems.

Adding support for new programming languages

Study testtemplate.py in the repo root directory. Add a new function test_languageName(capfd, dirInfo) in the script. Most components can be copied over from existing test functions for other languages, other than the ext extension name and the specific compile and cleanup commands.

Existing language-specific test functions in testtemplate.py can be renamed without any consequences.

When this new language's test function is added to testtemplate.py, no change should be needed to update the individual symlinks in problem directories.

Now run pytest -slv. If there is no programming files in that language extension yet, it should display relevant tests as 'SKIPPED'.

Adding new flags

Flags are parsed in the readFlags(workDir) function in testtemplate.py, and are passed to individual test_languageName functions to influence specific test routines. Add customized compile/test commands there, or elsewhere as seen fit. Be sure to expand every language-specific test function to account for the new flag.

Other functionalities to contribute

• List language-specific and problem-specific stats (How many problems are solved in this many languages?) Look into setup.sh.

• Package boilerplate testing function for each language, shortening each testing function

• Allow different solutions and test cases to carry textual names

• Remove symlinks from repo, add to gitignore, and update use instructions