Fix growth estimation for early-year calculations using weighted average

[nicklegan]

Problem

When running the tool early in the year (< 60 days), growth estimates
displayed zero delta for all future years, making projections unusable.

Root cause

The estimation logic copied previous year's stats as the prediction,
then calculated delta = predicted - previous, which was always 0.

Solution

Implement weighted average of recent years' growth deltas:

History available	Weights
4+ years	50% / 30% / 20%
3 years	60% / 40%
2 years	100% (single year)

Tested against short-running repos (< 2 years) - correctly shows
"Growth estimation unavailable" message.

[steffen]

@nicklegan Interesting idea on the weights! ✨ The generated code looks a little bit bloated to be honest. 😅 I kind of also like this simpler approach where it just takes last year's delta. I feel like this is easier to reason about and makes the previous year essentially the dominant factor for how this year's growth is estimated, which generally makes sense I think and is easy to reason about?

Maybe @TheZoran has some thoughts who worked in that area.

For reference, this pull request is to resolve #113.