Requirements:

• Describe your strategy to solve the task and how you test your code to make sure it works as expected.
• Implement a function to take an array w[] as input and return a message version number that satisfies weighted random condition.
• Write tests to cover your code, make sure that the weighted random condition is satisfied and all edge cases are checked properly.
• Your code should be production ready, well organised and well tested.

Answer

Startegy to implement task

How this algorithm works:

• Add up all the weight
• Randomly pick a number from 1 to sum of weights
• Iterate over versions, subtract random weight by version weight
• Compare result to 0, if less than or equal to zero then get that version

The larger the weight the more likely to be less than zero

Strategy to test

To check the random algorithm is weighted correctly, we can do it both way: manual or auto test.

Manual

Because of limitation of manual test, in order to test this algorithm we have to wisely select input data

Example of test data:
* [1, 0, 1000] => expected result: most of time, the result will be message version 3
* [9999, 1] => expected result: most of time, the result will be message version 1
* ...

So the trategy is to set a very high weight for a specific version and expect that version will be retunred most of the time. OR set 0 weight for a version and expect that version will never be returned.

Automated

Automation can overcome the limitation of manual test that we can loop for thousand times and compare actual result with given weight. BUT we have to accept the fact that randomness is relative so the result of randomness may have deviation. In my test code i set the deviation = 10%/sum of weights