Refactor some of the algorithms

[lostella]

Major changes:

• Forward-backward splitting is now split (a-ha) into fast and non-fast implementations.
• Fast FBS now takes the (optional) convexity modulus (default = 0) which is used in computing the extrapolation coefficients.
• The explicit linear mapping composed with smooth terms has been removed by algorithms that don’t exploit this structure (essentially all non-line-search methods); this simplifies there code quite a bit.
• Code for initial estimation of Lipschitz constants and step size backtracking is now in its own functions and shared by multiple algorithms.

Minor changes:

• The ProximalGradient and FastProximalGradient aliases have been added.
• Some source files have been renamed.

[codecov]

Codecov Report

Merging #51 (85324d2) into master (f945684) will increase coverage by 0.87%.
The diff coverage is 96.52%.

@@            Coverage Diff             @@
##           master      #51      +/-   ##
==========================================
+ Coverage   87.91%   88.79%   +0.87%     
==========================================
  Files          19       20       +1     
  Lines         786      794       +8     
==========================================
+ Hits          691      705      +14     
+ Misses         95       89       -6     

Impacted Files	Coverage Δ
src/ProximalAlgorithms.jl	100.00% <ø> (ø)
src/algorithms/douglas_rachford.jl	95.65% <ø> (ø)
src/algorithms/fista.jl	86.84% <ø> (ø)
src/algorithms/primal_dual.jl	64.40% <ø> (ø)
src/utilities/iteration_tools.jl	83.78% <ø> (ø)
src/algorithms/li_lin.jl	78.84% <71.42%> (ø)
src/algorithms/forward_backward.jl	97.14% <97.14%> (ø)
src/utilities/fb_tools.jl	97.14% <97.14%> (ø)
src/algorithms/fast_forward_backward.jl	97.77% <97.77%> (ø)
src/algorithms/davis_yin.jl	93.93% <100.00%> (ø)
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f945684...85324d2. Read the comment docs.

[lostella]

README needs to be updated

[lostella]

Benchmark shows what seems to be a performance regression:

2021-10-16T01:58:04.7149585Z 
2021-10-16T01:58:04.7150052Z                                           ID  time ratio memory ratio
2021-10-16T01:58:04.7150880Z   –––––––––––––––––––––––––––––––––––––––––– ––––––––––– ––––––––––––
2021-10-16T01:58:04.7151586Z   ["Lasso (Float64)", "FastForwardBackward"] 1.33 (5%) ❌  0.79 (1%) ✅
2021-10-16T01:58:04.7152369Z       ["Lasso (Float64)", "ForwardBackward"] 1.61 (5%) ❌  1.08 (1%) ❌
2021-10-16T01:58:04.7153060Z                 ["Lasso (Float64)", "PANOC"] 1.49 (5%) ❌  1.25 (1%) ❌
2021-10-16T01:58:04.7153687Z               ["Lasso (Float64)", "ZeroFPR"] 1.58 (5%) ❌  1.40 (1%) ❌