CMS AOI

CMS import system for AOI.

Installation

Clone the repo

$ git clone https://gitlab.com/aoi-dev/cms-aoi-import.git
$ cd cms-aoi-import

Install requirements:

ninja-build (varies on platform)
A recent pandoc install (for !mdcompile)

(Optional) Install to virtual env (to not pollute user env)

$ python3 -m venv venv
$ source venv/bin/activate
# Must repeat the source each time using this tool!

Install the tool
```
$ pip3 install -e .
```
(Optional) Include a local CMS - needed for uploading tasks to cms
```
# in cms repo
$ pip3 install -e .
```

Commands:

build - build testcases and other files
```
$ cmsAOI build dijkstra
```
Built testcases will be put in .aoi-temp/result local folder
upload - build and upload to cms (needs a CMS install)
clean - clean up build files

YAML Schema

Folder structure

cmsAOI expects each task to be in its own folder. The file task.yaml in this folder is where you can configure the task.

Example folder structure

$ tree .
.
├── attachment
│   ├── cpp
│   │   ├── grader.cpp
│   │   ├── mvv.cpp
│   │   └── mvv.h
│   └── typescript
│       ├── grader.ts
│       └── mvv.ts
├── private
│   ├── cpp
│   │   ├── grader.cpp
│   │   └── mvv.h
│   └── typescript
│       └── grader.ts
├── solution
│   ├── n2.cpp
│   └── nlogn.ts
├── statement
│   ├── mvv.md
│   ├── mvv.tex
│   └── sample.png
├── task.yaml
└── tc
    └── gen.py
$ cmsAOI build .

Basics

The task is configured in the task.yaml file in the YAML format. Before a task is built/uploaded, the YAML file is parsed and validated (so extra/misspelled keys etc will turn into an error).

Some options (such as name or author) just expect a string as the value. Others, such as statements[:][value] accept files. For example for a statement the following would cause the file statement/mvv.pdf to be uploaded.

statements:
  de: statement/mvv.pdf

YAML Schema

# extends (optional, path): specify a YAML file here to use all options
# specified there as the base. Any changes made in the task.yaml 
# file will overwrite the options in the "extended" file.
extends: ../base-task.yaml

# name (REQUIRED, string): specify the short codename of the task
name: MVV

# long_name (REQUIRED, string): specify the long name of the task
long_name: Sitzdesign

# author (optional, string): Add an author tag for bookkeeping, not used
# for anything in cmsAOI.
author: 'Flo'

# attribution (optional, string): Add an attribution tag for bookkeeping, 
# not used for anything in cmsAOI.
attribution: 'Flo'

# uses (optional, list of strings): A list of times this task was used.
# Not used for anything in cmsAOI.
uses:
  - AOI 2020 1. Wien Qualifikation

# statements (required): A mapping of language code to a statement PDF file.
statements:
  de: !latexcompile statement/mvv.tex

# statement_html (optional, file): An optional HTML file to display in
# the left description panel in frontendv2.
statement_html: !mdcompile statement/mvv.md

# default_input (optional, file): An optional file to prepulate the
# stdin field in test mode with frontendv2.
default_input: !raw |
   3
   GBR

# attachments (optional): A mapping of filename to attachments to add
# to the task (!zip is useful here).
attachments:
  mvv.zip: !zip attachment/*

# feedback_level (optional, string): The CMS feedback level (i.e. if all testcase
# outcomes are shown, or only up to the last successful one in a subtask).
# Either `RESTRICTED` or `FULL`. Defaults to `restricted`.
feedback_level: FULL

# score_options: additional options for CMS scoring
score_options:
  # decimal_places (optional, int): The number of decimal places to round
  # the score for this task to. Defaults to 0.
  decimal_places: 0

  # mode (optional, string): The CMS score mode (how subtasks scores are combined), options:
  #  - MAX_TOKENED_LAST: best of the tokened submissions and the last one
  #  - SUM_SUBTASK_BEST: Sum of the best result for each subtask
  #  - MAX: Plain max of all submissions
  # Defaults to SUM_SUBTASK_BEST
  mode: SUM_SUBTASK_BEST

  # type (optional, string): The CMS score type (how testcase outcomes
  # are combined to the score of a subtask), options:
  #  - GROUP_MIN: subtask score = min(testcase outcomes)
  #  - GROUP_MUL: subtask score = prod(testcase outcomes)
  #  - SUM: subtask score = sum(testcase outcomes)
  # Defaults to GROUP_MIN.
  type: GROUP_MIN

# time_limit (required, string): The time limit for this task, with unit s
time_limit: 1.5s

# memory_limit (required, string): The memory limit for this task, with unit MiB
memory_limit: 512MiB

# sample_solution (optional, file): A file to execute to automatically compute testcase outputs (in case they're not already given).
sample_solution: !cppcompile samplesol.cpp

# grader (optional, list of files): A list of files to upload as graders/managers
# in CMS. These will (depending on the extension), be included in the compilation
# environment
grader:
  - private/cpp/grader.cpp
  - private/cpp/mvv.h

# task_type (required): settings for the task type in CMS
task_type:
  # There are three types: batch, output_only and communication
  
  # ======== BATCH TYPE ========
  type: BATCH
  # stdin_filename (optional, string): If given, puts the input file at that location. Otherwise just pipes the input into the process via stdin.
  stdin_filename: ""
  # stdout_filename (optional, string): If given, puts the output file at that location. Otherwise just takes the stdout of the process as output.
  stdout_filename: ""
  # for checker, see top-level checker option

  # ======== OUTPUT_ONLY TYPE ========
  type: OUTPUT_ONLY
  # no further options
  # for checker, see top-level checker option

  # ======== COMMUNICATION TYPE ========
  type: COMMUNICATION
  # manager (required, file): A binary to coordinate the child processes.
  # Receives as arguments the child process input and output FIFOs.
  manager: !cppcompile private/manager.cpp
  # num_processes (optional, int): The number of user processes to start,
  # defaults to 1.
  num_processes: 2
  # user_io (optional, string): Whether to use stdio or files (given as arguments
  # to the graders) as input/output to the manager. 
  # Options: `std_io` or `fifo_io`. Defaults to `std_io`
  user_io: std_io

# subtasks (required, list): A list of subtasks (for no subtasks just 
# give a single subtask here)
subtasks:
    # points (required, int): The amount of points to award for this subtask
    # Note: for SUM score type this is the amount to give _per_ testcase
  - points: 30
  
    # public (optional, boolean): Whether the subtask outcome is public.
    # Defaults to true
    public: true

    # testcases (required, list): A list of testcases for this subtask
    testcases:
        # input (optional, file): The input file for this testcase.
      - input: tc/1-01.in
        # output (optional, file): The output file for this testcase.
        # If not given, the sample_solution is used to automatically compute this
        output: tc/1-01.out

        # public (optional, boolean): Whether this testcase's outcome is public (visible to user). Defaults to true.
        public: True

        # codename (optional, string): A codename for this testcase, defaults to an automatically generated one like 1-01.

        # Notes:
        #  - if an input/output filename ends with .gz, it's automatically extracted
        #  - you can use wildcards to match multiple input/output files
        input: tc/1-*.in
        output: tc/1-*.out

# checker (optional, file): An executable to run the user's output against to determine
# how many points they get (and a message). See CMS docs for more information.
checker: !cppcompile checker.cpp

# testcase_checker (optional, file): An executable to run each testcase against
# before uploading (to do some sanity checks for the testcases).
# The executable gets as the only argv argument the subtask number
# The testcase input is piped in via stdin.
testcase_checker: testcase_checker.py

# test_submissions (optional, mapping): A mapping of file to expected number of points.
# After the task has been uploaded, each of these submissions is uploaded to the server.
# If any test submission does not receive the expected number of points, it's shown here.
test_submissions:
  solution/nlogn.cpp: 100

# editor_templates (optional, list of files): What the prepopulate the
# editor screen with in frontendv2 depending on the language.
editor_templates:
  - attachment/cpp/mvv.cpp
  - attachment/csharp/MVV.cs

# test_grader (optional, list of files): Like grader, but used for test mode
# in frontendv2. These are separate from the grader option because some graders
# use secret strings to validate the output authenticity.
test_grader:
  - attachment/cpp/grader.cpp
  - attachment/cpp/mvv.h

austrian-olympiad-informatics/cms-aoi-import

CMS AOI

Installation

YAML Schema

Folder structure

Basics

Tags

YAML Schema