Calculate ANI based on a CLI flag

[mr-eyes]

I think incorporating ANI calculation in the branchwater plugin will be very useful. For performance, the code will need to be rewritten in Rust, though.

[mr-eyes]

Hi @ctb and @bluegenes

I wrote a Rust implementation of the sourmash ANI estimation code many weeks ago but never committed it 3b1c1b8

Just putting it here for now in case it's useful until I can continue the integration later, or maybe someone faster than me will do.

The Rust implementation is 712X faster than the original Python implementation.

[ctb]

The Rust implementation is 712X faster than the original Python implementation.

not good enough. Do better.

😆 that's awesome!

[bluegenes]

nice!!! I was just thinking I needed to write one so I could add ANI into fastmultigather results 😂

[mr-eyes]

nice!!! I was just thinking I needed to write one so I could add ANI into fastmultigather results 😂

Let's coauthor a PR to add this to all commands (I hope you do the testing 😂).

[ctb]

Let's coauthor a PR to add this to all commands (I hope you do the testing 😂).

🤔

[mr-eyes]

Let's coauthor a PR to add this to all commands (I hope you do the testing 😂).

🤔

I am not very skilled at writing tests in sourmash because it requires a broader context that I am not very familiar with.

[bluegenes]

Streamlined ANI from sourmash-bio/sourmash#2943 seems to be 2x faster than this rust implementation, as it doesn't cause a 2x slowdown compared to not calculating ANI :)

The benchmark here was a strong motivator for the streamlining work. So yay!

comparison benchmark (ANI vs not): #236 (comment)

streamlined ani_utils are here: https://github.com/sourmash-bio/sourmash/blob/latest/src/core/src/ani_utils.rs

[ctb]

Remaining functionality added in #298 and #361. ANI should now be available everywhere ;)