QuizQuestion

true_false: answered true or false. multiple_choice: exactly one correct option. multi_select: several options may be correct, scored all-or-nothing. ordering: the options are items whose authored order is the answer. short_answer: the options are the accepted spellings of a typed answer. numeric: a number within a tolerance of correct_number. matching: each option is a label/match_label pair the taker reunites. fill_in_blank: the prompt's runs of three or more underscores are blanks, and the options are the accepted answers, each naming its blank_index. long_answer: a free-text essay answer with no accepted spellings, always graded by hand. drawing: the taker draws on a canvas, over drawing_starting_image when one is attached, always graded by hand.

one of "true_false" | "multiple_choice" | "multi_select" | "ordering" | "short_answer" | "numeric" | "matching" | "fill_in_blank" | "long_answer" | "drawing"

plaintext: the prompt renders verbatim. markdown: the same prompt text renders as Markdown. pdf: an uploaded PDF carries the question; the prompt text is the instruction shown above it.

one of "plaintext" | "markdown" | "pdf"

Original name of the attached PDF, for pdf prompts; null otherwise. The file itself is fetched through the question's pdf endpoint.

Whether the attached PDF is drawn beside the question or offered as a download only (`00051`). A one-page scenario belongs on screen; a long reference the taker works from on paper only buries the answers beneath it. Ignored on prompts with no PDF.

Shown in answer review when the quiz reveals correct answers.

What a full-marks answer is worth, 1 to 1000; an attempt's score is the points earned as a percentage of the points presented.

minimum 1 · maximum 1000

Whether a marker awards the points by hand. Forced on for long_answer and drawing questions. A machine-gradable kind flagged this way is still checked at submit, and the verdict shown to the marker as a suggestion; nothing counts until the marker records points.

The marking scheme, as markdown: shown to markers on the grading page, and to takers only when rubric_visible_to_takers. Rendered through the sanitized pipeline question prompts use.

at most 50000 characters

Whether takers see the rubric too — beneath the prompt while answering, and again in answer review. Off, the rubric reaches markers only.

The image a drawing question's taker draws over, as a data URL; null for a blank canvas, and on every other kind.

at most 3000000 characters · matches ^data:image\/(png|jpeg|webp|gif);base64,[A-Za-z0-9+/]+=*$

The answer of a true_false question; null for every other kind.

The answer of a numeric question; null for every other kind.

How far a numeric response may stray and still count; null otherwise.

Choices, ordering items (authored order = correct order), accepted short answers, matching pairs or the accepted answers of each blank, in authored order; empty for true_false, numeric, long_answer and drawing questions.

When set, the record is retired; it stays readable but should not grow.