This section describes the format parsed by learn-ocaml for exercises, tutorial and lessons.

Exercices and groups format

Groups

A group is either a set of exercices or other groups. As such exercices are classified into a tree representation, where groups are nodes and leafs are exercises. This tree (or index of exercises) can be described by a file index.json at the root of the exercises directory from the repository. If no index.json is found, the subdirectories are scanned and all exercises found are returned.

The file index.json has the following format (in version "1"):

<group name> = <string>
<directory> = <string>

<exercices> = "exercises : [ <directory>, .., <directory> ]
<groups> =
  "groups" :
    { ( <group name> : { "title" : <string>, (<groups> | <exercices> ) } )+ }

{ "learnocaml_version": <string>, <groups> }

For example, this index.json describes two groups of exercices, one with only one exercise, two in the other:

{ "learnocaml_version": "1",
  "groups":
  { "imperative-traits":
    { "title": "Imperative traits",
      "exercises": [ "mooc-simple-input-output" ] },
    "data-structures":
    { "title": "Data Structures",
      "exercises": [ "mooc-binary-search-trees",
                     "mooc-trie-data-structure" ] } } }

This format reflects how exercises should be classified into the application. If this file does not exists, the exercises are simply put at the root of the index.

Exercise

An exercise is described by a directory containing at most the following files: - meta.json - descr.html - prelude.ml - prepare.ml - template.ml - solution.ml - test.ml - max_score.txt

meta.json

Json description with the following format (in version "1"):

{ "learnocaml_version" : "1",
  "kind"               : "exercise" | "problem" | "project",
  "stars"              : [1 .. 5]
}

Json description in version "2" contains more metadata:

{
  "learnocaml_version" : "2",
  "kind"               : "exercise" | "problem" | "project",
  "stars"              : [1 .. 5],
  "title"              : "Title of the exercise",
  /* In an exercise repository, each exercise must have a unique identifier. */
  "identifier"         : "some_unique_identifier",
  /* Authors with their emails. */
  "authors"            : [["Xavier Leroy", "some@email"], ["Damien Doligez", "someother@email"]],
  /* The skills and concepts that are practiced by this exercise. */
  "focus"              : ["skill1", ..., "skillN", ..., "concept1", ..., "conceptM"],
  /* The skills and concepts that are required to do this exercise. */
  "requirements"       : ["skill1", ..., "skillN", ..., "concept1", ..., "conceptM"],
  /* The suggested exercises in case of success */
  "forward_exercises"  : [ "exercise1", "exercise2", ... ],
  /* The suggested exercises in case of difficulty */
  "backward_exercises" : [ "exercise1", "exercise2", ... ],
  /* Maximum score for the exercise. */
  "max_score" : [ 0 .. n ],
}

descr.html

Description of the exercise in HTML format.

prelude.ml

OCaml file containing the definitions that are known by the user. In the web-application, the prelude is shown after the exercise description. The prelude is loaded in the environment before any other .ml files from the exercise directory.

prepare.ml

OCaml file containing some definitions that are unknown by the user. As such, it can be used to to redefine some functions, or execute some code. For example, it is useful to redefine some standard library functions to track their usage for algorithmic purpose. prepare.ml is loaded directly after prelude.ml in the environment.

template.ml

OCaml file containing a template of code for the exercise, and loaded into the editor by default if the user does not have a saved code for the current exercise.

solution.ml

OCaml file containing the solution for the exercise. It can be used for testing the user's code by comparing its results with a code reference.

test.ml

OCaml file containing the test program that is used to check and grade the user's code and the solution. It has access to many functions allowing the introspection of the code, which will be described and detailed in another section.

max_score.txt

Maximum score that is possible to get for this exercise, even if the grader grades more. Overridden by the field max_score, if present in meta.json.

Metadata

When building the corpus and extracting the metadatas of all exercises, the builder will generate two files requirements.json and focus.json, which aggregates every skill present in their respecting fields in every meta.json. It associates a skill to the list of exercises that requires it (or focuses on it). These files behave as some sort of static database for this information and should not be written by hand.

Their format is the following:

[{ "<skill name>" : [ "<exercise_id>" ; .. ; "<exercise_id>" ] }; .. ]