Final report

[dstansby]

This is a PR for the final report, in the form of the website. I intend for this to be pretty much the last thing we merge, but opening as a draft so everyone knows I'm working on this, and so I can use it to iterate on what graphs we want to put in here to tell the story.

Fixes #109
Fixes #108

[dstansby]

@ruaridhg @K-Meech could you review what's here? I still need to add sections for chunk size and different data (image/sparse/dense), but would be good to review what's here, and I'll open issues to remind myself to finish off the other sections after this PR.

[K-Meech]

Thanks @dstansby - looks good! I put some comments below.

Some more general points:

• We should give some details about the hardware the benchmarks were run on e.g. OS, RAM, number of cores, maybe type of storage (SSD type), as we are using the full time to read/write to disk?
• It would be good to mention that the points on the graphs are the mean values from 5 runs
• Exact numbers will have to be updated throughout once the benchmarks are run again with more iterations. I think a number of these plots aren't up to date with the latest example data?

[ruaridhg]

Looks good, just made a few minor suggestions.

[dstansby]

RIGHT - I think I have included all the graphs 🎉 @K-Meech & @ruaridhg I'd appreciate if you could give this another review when you have time to check that I haven't missed anything. If I haven't, we can merge and the benchmarking will be done!

[dstansby]

(actually, not quite done, we should address #110 before considering this finsihed)

[ruaridhg]

Just left a couple of comments for things that could/should be added.

[ruaridhg]

I've added alt text for all the plots as I think this is what @K-Meech meant

[K-Meech]

@dstansby I'm part way through reviewing the final report - should be able to finish this tomorrow. I wasn't sure if the plots / text had been updated to match the latest results?

E.g. running python src/zarr_benchmarks/create_plots.py --example_results on the main branch + looking at data/plots/heart/write/format_v2, seems to show a different plot to the first graph used in the report? The one in the report has multiple points below 1 second, while most in the main plot are above 1 second write time.

[K-Meech]

Thanks @dstansby - looks good! I've put some comments below:

[ruaridhg]

Added more detailed alt text descriptions if necessary

[dstansby]

Thanks all for comments/contributions to this!