slash grades should have integer scores

[johncalvinroberts]

Possible tentative fix for: #137

[musoke]

I suggest two changes here:

• round the top and bottom of the range in opposite directions
• add some tests

[johncalvinroberts]

@musoke updated according to your suggested implementation. Thank you so much for the guidance!

After updating, some tests (integration tests?) in readme.ts were failing due to hardcoded non-integer return values. This caused me to notice that French scale implementation was also suffering from the same defect, and could potentially return non-integer score values for slash grades. So, I updated with the same logic, assuming slash grades to be exclusive of adjacent grades in the french scale. Is that a correct?

Also, definitely LMK if there are any other issues with my implementation and I'll try to fix.

[musoke]

After updating, some tests (integration tests?) in readme.ts were failing due to hardcoded non-integer return values. This caused me to notice that French scale implementation was also suffering from the same defect, and could potentially return non-integer score values for slash grades. So, I updated with the same logic, assuming slash grades to be exclusive of adjacent grades in the french scale. Is that a correct?

Yes, now that you point it out, this this is going to be an issue with all the scales. Maybe you could define a function to do this check for each scales. Or I can add it to other scales in a followup.

[johncalvinroberts]

After updating, some tests (integration tests?) in readme.ts were failing due to hardcoded non-integer return values. This caused me to notice that French scale implementation was also suffering from the same defect, and could potentially return non-integer score values for slash grades. So, I updated with the same logic, assuming slash grades to be exclusive of adjacent grades in the french scale. Is that a correct?

Yes, now that you point it out, this this is going to be an issue with all the scales. Maybe you could define a function to do this check for each scales. Or I can add it to other scales in a followup.

Updated accordingly with a getRoundedScoreTuple utility func, and applied to other grade scales which could return a tuple with high/low based on adjacent scores. Is that what you had in mind? @musoke

It doesn't feel super DRY, but then again that seems to be the running style in this codebase? The enclosing getScore function is mostly the same across grade scales, so I also tried to follow roughly the same level of abstraction. Let me know if stylistically there is anything I should know 😄

[musoke]

That looks good to me.

Yes, there is a lot of repetition due to the adhoc evolution of sandbag. The special cases in methods like getScore do make this a little tricky. I would welcome refactors to improve this.