Refactor render logic

[bazzuser]

Description:
The current HTML/Jinja2 template uses multiple nested if/elif blocks to render different question types (text, radio, checkbox, candidate_id) and media (image, audio, video). This approach goes against our goal of a clean, object-oriented architecture and results in a template that is difficult to maintain.

Refactoring Goal:
Move the rendering logic out of the template and into the respective Python classes. Both Question subclasses and MediaItem subclasses should implement a to_html() method that is directly invoked from the template.

Example Jinja2 usage after refactoring:

{{ question.to_html() | safe }}

Acceptance Criteria:

• The template file (exam.html) no longer contains any if q.type or if m.type logic.

• All supported question types are rendered via their respective to_html() methods.

• Optional media items are rendered via to_media_html().

• The visual output is semantically identical to the current implementation.

• All to_html() methods follow a consistent interface and can be tested individually.

Notes:

• to_html() returns HTML strings (not Jinja snippets).

• If a NullMediaItem is used, to_media_html() simply returns an empty string.

• The goal is a "Smalltalk-inspired" design: objects are fully responsible for their own rendering.

[enno]

fixed

[bazzuser]

See source:/master/flexograder/builder/exam.py#L120