feat: Added implementation for ugly numbers

[SpiderMath]

This pull request aims to add the implementation for Ugly Numbers. Lmk if there is any issue in this pull request

[appgurueu]

Could you add comments to the code?

I'm also not sure whether this should necessarily be a function giving you only the n-th number of the series. An iterator or returning the list of ugly numbers up to the n-th one might be a better design.

[SpiderMath]

Could you add comments to the code?

Okay, I'll add some comments to the code (though I thought adding the bits above the file was enough but I can add some more if required)

I'm also not sure whether this should necessarily be a function giving you only the n-th number of the series. An iterator or returning the list of ugly numbers up to the n-th one might be a better design.

But the accompanying implementations in python, java and dart (though dart implementation also made a function isUgly) on the website do the same thing of returning the Nth ugly number, which is why I decided to code it in this way.

[appgurueu]

But the accompanying implementations in python, java and dart (though dart implementation also made a function isUgly) on the website do the same thing of returning the Nth ugly number, which is why I decided to code it in this way.

I think you should choose the best design for this implementation, not mirror the poor design decisions of the other implementations. If I want the first ten ugly numbers, an iterator is much preferable over such a "getter".

[raklaptudirm]

@appgurueu I would argue that getting the nth number of the series is also good design, because you can simply define a wrapper to convert it into an iterator or return the first n numbers. On the other hand, doing it the opposite way (nth number from iterator or range) is simply doing more work than necessary. Let me know what you think.

[appgurueu]

This algorithm generates a "stream" of numbers, since it needs the previous numbers to calculate the next ones. I think it is only reasonable to use an array or a generator to map this to code. The array may be wasteful though if only, say, the last number is needed, so I think a generator is better. Getting the n-th number from an array is easy; getting it from the generator would be slightly more work, but is pretty standard, at least in non-javascript spheres (Rust's iterators have an nth method for example).

I would argue that getting the nth number of the series is also good design, because you can simply define a wrapper to convert it into an iterator or return the first n numbers.

This is fine from a pure "design"/"correctness" perspective, but it makes an otherwise linear time operation quadratic time.

[SpiderMath]

Sorry for the long hiatus (for personal reasons), I'm finally back.
I read the points, and understood your suggestions.
I'll restructure the code as a generator, since it seems to be a better idea from a design perspective. I'll push the code in a day or so.

[SpiderMath]

I've made the required changes (I'm not sure on whether the writing of the tests was optimal however, since I don't know how exactly I'm supposed to write tests for generators)

[appgurueu]

I don't know how exactly I'm supposed to write tests for generators

I rewrote your test to just generate the first 11 numbers and compare them against the expected numbers.

(The failing CI is due to something unrelated.)

[SpiderMath]

I don't know how exactly I'm supposed to write tests for generators

I rewrote your test to just generate the first 11 numbers and compare them against the expected numbers.

That looks a lot cleaner than what I did, thanks!

(The failing CI is due to something unrelated.)

Ah, understood. Thanks for the clarification! 😄