Allow customizing truncate! based on the algorithm

[mtfishman]

Closes #27.

The idea here is that this allows customizing the truncation based on the algorithm if needed, in addition to the truncation strategy. The reasoning is that certain algorithm backends might output the spectrum in a different format/ordering, which would require a different implementation of the truncation.

I'm open to suggestions on other ways of implementing this. Right now it is based on passing the TruncatedAlgorithm object into truncate!, which includes both the algorithm and the truncation strategy, and then by default it unwraps just the truncation strategy. An alternative could be to pass the truncation strategy and algorithm as separate arguments, though in general they might not be separate, for example in a sparse/low rank factorization like a randomized SVD or Krylov SVD.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Files with missing lines	Coverage Δ
src/implementations/eig.jl	100.00% <100.00%> (ø)
src/implementations/eigh.jl	98.41% <100.00%> (ø)
src/implementations/orthnull.jl	91.28% <100.00%> (ø)
src/implementations/svd.jl	94.65% <100.00%> (ø)
src/implementations/truncation.jl	87.77% <100.00%> (+1.70%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[mtfishman]

I talked to @lkdvos about this and we decide against this for now, we can bring it back if there is a compelling use case.