Optimize billing period - temperature merging

[jkwan2011]

The list of billing periods is built by looking up indices in the temperature input and creating a list of temperatures in a for loop. Assuming the temperature input is sorted by date, consider using a more optimal method of generating this list. @debajyotid2 has suggested using the bisect module.

Note that this in reference to lines 64-67 of engine.py for the latest commit on the summary-outputs branch.

[debajyotid2]

@jkwan2011 I made the changes. (The original lines are commented out) Please review here: https://github.com/codeforboston/home-energy-analysis-tool/commit/be7987a5f40e32f738b3bc4cf36a04f340a541b8 .

If the changes look good please close this issue as completed.

Edit: Reverted an unintended change that I accidentally left in. It has been fixed now here: https://github.com/codeforboston/home-energy-analysis-tool/commit/6a46074a1a0cf50584d273fd51a01bfb062e9753.

[jkwan2011]

Thanks for the contribution! It looks good, but for the end_idx, is there a reason to using bisect_left + 1, instead of bisect_right?

[debajyotid2]

Thanks for the contribution! It looks good, but for the end_idx, is there a reason to using bisect_left + 1, instead of bisect_right?

Thanks! The difference between bisect_left and bisect_right is that if there are repeated values of an element x that you want to insert in an array, bisect_left will return the left-most index where it can be inserted to keep the array sorted, while bisect_right will return the right-most index. In this case I assumed there would be no repeated dates, so there would be no difference in start_idx and end_idx if you used either of the two functions. The 1 is added as for the slice based indexing we want to be including temperatures from start_idx to end_idx inclusive of both ends. (Since the billing period start date and end date are both considered inclusive when searching for temperature values within that period).

[jkwan2011]

Thanks for the explanation!