learn-content-engine - v0.13.1
    Preparing search index...

    Interface PictureImage

    One image option in a PICTURE_CHOICE exercise.

    EXP-039: modeled explicitly (was an inline dict[str, str]) so the generated JSON-Schema / TS types carry the structured {src, label, is_correct?} shape instead of a loose string map. extra="forbid" + the two required fields replace the former key-subset / src+label-present checks; the "exactly one correct" rule stays in _validate_picture_choice_fields.

    is_correct stays a str ("true" marks the answer) for backward compatibility with authored content, not a bool.

    interface PictureImage {
        is_correct?: IsCorrect;
        label: string;
        src: string;
    }
    Index
    is_correct?: IsCorrect
    label: string
    src: string