Add ndslice.chunks module

[9il]

@andralex @wilzbach @ljubobratovicrelja @thewilsonator @dataPulverizer

Mir Algorithm basic API is stable enough and we can start to do reviews! Please help with that)

This PR contains fast multidimensional random access chunks. It has multidimensional slicing, builtin dimension transposition, allows to get underlying slice during life time.

Motivation of this addition is cache oriented algorithms. For example, see Chapter 1 at The Science of Programming Matrix Computations

[dataPulverizer]

Thank you very much for the update. Been very busy lately but will hopefully get some time at the end of the week to take a detailed look.

Well done

[jmh530]

You might add an example to mir.chunks showing that (if?) you can apply sum to the chunks. I would imagine that would be a key use case. This should allow you to row sum or column sum, if I'm not mistaken.

[9il]

@jmh530 could you describe what do you mean using a small example?

[jmh530]

@9il I was thinking something like below would work, but I may have gotten something wrong. I haven't played with the code yet.

unittest
{
import mir.ndslice.slice : sliced;
import mir.ndslice.chunks: chunks, isChunks;
import mir.ndslice.topology: iota;
import mir.math.sum: sum;

// 0 1 2 3 4 5 6 7 8 9 10 auto sl = iota(11); // 0 1 2 | 3 4 5 | 6 7 8 | 9 10 auto ch = sl.chunks(3); // 3 | 12 | 21 | 19 auto s = ch.sum(); assert(s == [3, 12, 21, 19].sliced); 

}

[codecov-io]

Codecov Report

Merging #45 into master will decrease coverage by 0.01%.
The diff coverage is 96.02%.

@@ Coverage Diff @@ ## master #45 +/- ## ========================================== - Coverage 96.31% 96.29% -0.02%  ========================================== Files 27 28 +1 Lines 4527 4702 +175 ========================================== + Hits 4360 4528 +168  - Misses 167 174 +7

Impacted Files	Coverage Δ
source/mir/ndslice/package.d	92.25% <ø> (ø)	⬆️
source/mir/ndslice/topology.d	98.93% <ø> (ø)	⬆️
source/mir/ndslice/slice.d	98.87% <100%> (+0.03%)	⬆️
source/mir/ndslice/chunks.d	95.36% <95.36%> (ø)
source/mir/functional.d	96.34% <0%> (ø)	⬆️
source/mir/ndslice/internal.d	56% <0%> (+0.89%)	⬆️

---

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 3772abc...0780e47. Read the comment docs.

[9il]

@jmh530 , added - 0780e47