[wip][rfc] New readme from Fig1; fix #42

[mkborregaard]

This adds a readme that recreates Fig1 of the NLMpy paper as suggested in #42.

This is their Fig1:

This is the one of this PR:

The code here ports this file https://besjournals.onlinelibrary.wiley.com/action/downloadSupplement?doi=10.1111%2F2041-210X.12308&file=mee312308-sup-0002-dataS2.py to our julia format. As you can see there are some discrepancies that we could revisit to make our syntax more intuitive.

I also don't really know about Fig1s, and of course I just need to update the colors on the classified ones.

[codecov-io]

Codecov Report

Merging #45 (7784ee8) into main (efab0af) will decrease coverage by 0.04%. The diff coverage is 0.00%.

@@           Coverage Diff            @@
##            main     #45      +/-   ##
========================================
- Coverage   6.80%   6.75%   -0.05%     
========================================
  Files         13      13              
  Lines        294     296       +2     
========================================
  Hits          20      20              
- Misses       274     276       +2     

Flag	Coverage Δ
unittests	6.75% <0.00%> (-0.05%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/algorithms/nnelement.jl	0.00% <0.00%> (ø)
src/classify.jl	0.00% <0.00%> (ø)

---

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 efab0af...7784ee8. Read the comment docs.

[mkborregaard]

Now with updated colors

[tpoisot]

AMAZING 👀

[tpoisot]

I'm going to be my usual self and say: can we update the discrete color palette with something that is accessible, like Wong or IBM?

[mkborregaard]

I agree that would be nice, but isn't there something to be said for choosing the exact same one as the original paper?

[tpoisot]

I'm all for reproducibility, but I think it's better to make a point about accessible color schemes - it wouldn't change the message, and it would make the figure accessible to ~ 10% of people with different color perception.

[rafaqz]

Looks great. I'm just wondering why the Y axis seems to be flipped in d and e in our version?

[mkborregaard]

@tpoisot in fact this is one of the high-quality color schemes from RColorBrewer, which are all vetted for accessibility and scientific accuracy. I'm a 100% in agreement with making a statement for using high-quality accessible color schemes, but don't think we should support the notion that this restricts us to a very small portion of schemes, like wong. Even for full color deficiency it should be possibly to tell these apart, with possibly a slight issue for 100% deuteranopics:

if you like, maybe we can use a different color scheme for the docs?

@rafaqz because Plots and matplotlib interpret matrix orientation differently in heatmap. "Correct" matrix orientation is a real religion among plotting packages.

[tpoisot]

Interesting, the simulation plugin I used made the color much less distinguishable than your screenshot - let's go with this then.