Development:Quiz Item Analysis of Multianswers Question Types

This page described a code proposal to access the embedded questions data in multianswers questiion types so that they can be displayed in the Item analysis report. See http://moodle.org/mod/forum/discuss.php?d=86598#387154 The code is experimental and need more testing and is not optimized.

All question types to be analyzed contain at least one embedded question

If we set this as a general framework, the code will be valid for all question types.

We need a question type function to retrieve the embedded questions

There is actually no question type function to know the number the embedded questions in a question. The default function number_of_embedded_questions() returns 1;

This is not the same as  actual_number_of_questions().

We need a convention of how to store the responses from the different embedded questions

The same convention should be used in the $session states and in the get_responses function. The actual responses convention used an array to contains the different answers and put in [''] element the actual response. an example for a numerical question with the tolerance limit

stdClass Object
(
   [id] => 2312
   [responses] => Array
       (
           [6837] => stdClass Object
               (
                   [answer] => 1822 (1817..1827)
                   [credit] => 1
               )
       )
)

In a multianswers qtype the different answers responses are stored in an array as.

stdClass Object
(
   [id] => 2815
   [responses] => Array
       (
            [1] => stdClass Object
               (
                   [id] => 2816
                   [responses] => Array
                       (
                           [7300] => stdClass Object
                               (
                                   [answer] => Wrong answer
                                   [credit] => 0
                               )
                           [7302] => stdClass Object
                               (
                                   [answer] => Correct answer
                                   [credit] => 1
                               )
                           [7305] => stdClass Object
                               (
                                   [answer] => Answer that gives half the credit
                                   [credit] => 0.5
                               )
                       )
               )
           [2] => stdClass Object
               (
                   [id] => 2817
                   [responses] => Array
                       (
                           [7303] => stdClass Object
                               (
                                   [answer] => Wrong answer
                                   [credit] => 0
                               )
                           [7304] => stdClass Object
                               (
                                   [answer] => Correct answer
                                   [credit] => 1
                               )
                       )
               )
           [3] => stdClass Object
               (
                   [id] => 2818
                   [responses] => Array
                       (
                           [7307] => stdClass Object
                               (
                                   [answer] => 65 (64.9..65.1)
                                   [credit] => 1
                               )
                       )
               )
       )
)

So the one question question types responses is tranformed in a similar structure The upper example becomes

stdClass Object
(
   [id] => 2312
   [responses] => Array
       (
           [0] => stdClass Object
               (
                   [id] => 2312
                   [responses] => Array
                       (
                           [6837] => stdClass Object
                               (
                                   [answer] => 1822 (1817..1827)
                                   [credit] => 1
                               )
                       )
               )
       )
)

To distinguish