autokattis

Updated Kattis API wrapper as of May 2023 after the major UI/UX change.

Setup

Simply install it as a Python package.

$ pip install autokattis

Use Cases

For now, this package supports OpenKattis and NUSKattis.

Login

Construct an OpenKattis object that takes in the username and the password.

from autokattis import OpenKattis
kt = OpenKattis('username', 'password')
kt = OpenKattis('username') # which will then prompt you for the password

where 'username' is your Kattis username/email and 'password' is your Kattis account password. Both should be provided as Python strings.

Similarly if you want to login to NUS Kattis.

from autokattis import NUSKattis
kt = NUSKattis('username', 'password')
kt = NUSKattis('username')

OpenKattis

Due to backwards compatibility, you can still use Kattis as a shorthand form of OpenKattis.

Problem-specific

kt.problems()                               # problems you have solved so far
kt.problems(show_partial=False)             # exclude partial submissions
kt.problems(low_detail_mode=False)          # include more data for each problem
kt.problems(*[True]*4)                      # show literally all problems on Open Kattis

kt.plot_problems()                          # plot the points distribution
kt.plot_problems(filepath='plot.png')       # save to a filepath
kt.plot_problems(show_partial=False)        # plot fully solved submissions

kt.problem('2048')                          # fetch info about a problem
kt.problem(['2048', 'abinitio', 'dasort'])  # fetch multiple in one
kt.problem({'2048', 'abinitio', 'dasort'})  # tuples or sets also allowed
kt.problem('2048', download_files=True)     # download files too

kt.stats()                                  # your best submission for each problem
kt.stats('Java')                            # all your Java submissions
kt.stats(('Python3', 'Cpp'))                # multiple languages

kt.suggest()                                # what's the next problem for me?
kt.achievements()                           # do I have any?
kt.problem_authors()                        # list down all problem authors
kt.problem_sources()                        # list down all problem sources

Ranklist

kt.ranklist()                                                           # people around you

kt.user_ranklist()                                                      # top 100 users in general ladder
kt.challenge_ranklist()                                                 # top 100 users in challenge ladder

kt.country_ranklist()                                                   # top 100 countries
kt.country_ranklist('Singapore')                                        # specific country
kt.country_ranklist('SGP')                                              # use country code instead

kt.university_ranklist()                                                # top 100 universities
kt.university_ranklist(university='National University of Singapore')   # specific university
kt.university_ranklist(university='nus.edu.sg')                         # use university domain instead

NUSKattis

Problem-specific

kt.problems()                               # problems you have solved so far, only supports low detail mode
kt.problems(show_solved=False)              # show literally all problems on NUS Kattis

kt.problem('2048')                          # fetch info about a problem
kt.problem(['2048', 'abinitio', 'dasort'])  # fetch multiple in one
kt.problem({'2048', 'abinitio', 'dasort'})  # tuples or sets also allowed
kt.problem('2048', download_files=True)     # download files too

kt.stats()                                  # your best submission for each problem
kt.stats('Java')                            # all your Java submissions
kt.stats(('Python3', 'Cpp'))                # multiple languages

Course-specific

kt.courses()                                    # current and recently ended courses
kt.offerings('CS3233')                          # course offerings
kt.assignments('CS3233_S2_AY2223')              # offering assignments but course ID not provided
kt.assignments('CS3233_S2_AY2223', 'CS3233')    # offering assignments

Convert to DataFrame

As simple as adding .to_df()!

kt.problems().to_df()
kt.ranklist().to_df()

Other Scenarios

Some scenarios you can perform when using autokattis:

1. Mapping problem ID with its difficulty

   okt = OpenKattis(...)
   df = okt.problems().to_df()
   diff_map = dict(zip(df.id, df.difficulty))

2. Find the number of questions on every assignment on an NUS course offering

   nkt = NUSKattis(...)
   df = nkt.assignments('CS3233_S2_AY2324').to_df()
   df['n_problems'] = df['problems'].apply(lambda x: len(x.split(',')))
   df[['name', 'n_problems']]

3. Find the average difficulty of every assignment on an NUS course offering

   diff_map = ... # see scenario 1
   
   nkt = NUSKattis(...)
   avg_2dp = lambda x: round(sum(y:=[v for v in x if v != None])/max(len(y), 1), 2)
   df = nkt.assignments('CS3233_S2_AY2324').to_df()
   df['avg_diff'] = df['problems'].apply(lambda x: avg_2dp(map(diff_map.get, x.split(','))))
   df[['name', 'avg_diff']]

4. Group top 100 users by country

   okt = OpenKattis(...)
   okt.user_ranklist().to_df().groupby('country').size()

More Information

The docstrings might be a great help if you want to know more about the JSON return values!

from autokattis import OpenKattis
help(OpenKattis)

Testing

The test directory is provided within this repository. You are free to test autokattis with these anytime.

>>> python test/openkattis.py
...
>>> python test/nuskattis.py
...

Useful References

• Old UI Kattis API wrapper: https://github.com/terror/kattis-api

  Most of the work in autokattis is heavily inspired and motivated by this repository.

• Kattis official CLI tool: https://github.com/Kattis/kattis-cli

  Since Kattis has provided an official tool to automate submissions, there won't be such feature in autokattis.

Contributing

Feel free to suggest anything or add on some implementation by simply creating a pull request!