Update logic for choosing prompt and correct answer for pre- and post-tests.

[davidharting]

Update Please see my next comment that outlines a better approach for dealing with this.

---

For normal book reading, we decide if we are using a number or size prompt based on the "mode" (which is in turn based on the "treatment" that is set up at the beginning).

However, for pre- and post-tests there is only one prompt and correct answer per question.

So when choosing a prompt and when grading answers, the logic will need to be updated by passing in the "type" (pre-test, post-test, or regular book) so that we can update the logic.

Effected places:

• Getting the choices
  Currently, this based on difficulty. However, for pre- and post-tests, difficulty is not relevant. So we need to detect that we are in a pre- or post-test and simply select the one difficulty level that is present.
  src/models/Question.ts#getDifficulty
• Getting the prompt: src/models/Question.ts getPrompt
• Grading a selection
  src/models/Question.ts#isSelectionCorrect which in turn depends on src/models/Choice.ts#isSelectionCorrect.
  For a regular book, identifying the correct answer depends on the mode. However, for pre- and post-tests, we will make the assumption that only one of the fields is specified - either the correct number stimulus or the correct size stimulus. So we will need to pass along information about the book type. If a regular book, use mode as normal. If a pre- or post-test, then ignore the mode and identify which field is not null and use that one.

To identify more places that may be effected, in VS Code do a global search for "mode" - CTRL + SHIFT + F to open up the global search on windows.

[davidharting]

I may have had a bit of an epiphany on this task, so please read this (which conflicts with the description above) before working on this!

The above description is basically solving this problem by saying - anywhere we have to select something based on the mode, add a new branch of logic for pre- and post-tests.

It would be much simpler to solve the problem at the source and in one place - the logic where we pick the mode!

Check out readSlice.ts and find the function selectMode. Right now, the mode depends on the current treatment.

Current implementation:

export const selectMode = (state: RootState): Mode | null => {
  const treatment = selectTreatment(state);
  if (treatment === "mixed") {
    return state.read.randomMode;
  }
  return treatment;
};

We could introduce some logic here where we figure out what kind of book we are dealing with:

export const selectMode = (state: RootState): Mode | null => {
  const book = selectBook(state);
  const bookType = Book.getType(book);
  
 
  if (bookType === "normal") {
    const treatment = selectTreatment(state);
    if (treatment === "mixed") {
      return state.read.randomMode;
    }
    return treatment;
  }

  // New logic goes here - we are dealing with a pre- or post-test book
  const question = selectQuestion(state)
  return getModeForTestBook(question)
};

Then, the trick would be to implement getModeForTestBook. For test books we expect only one prompt - size or number. You could determine the mode based on the presence of the prompt.

[davidharting]

getModeForTestBook should probably be a static method on the question class inside of models/Question.ts.

Not in love with the name of that function but it would work if we can't think of anything better!