@ctb @bluegenes
I have two questions for this PR.

1. I was wondering if it's better to include ANI calculation in multisearch, multigather, and pairwise in a single PR? or sub-PRs from this?
2. Should the ANI calculation be optional via a command line flag, or it can be a default column in the output files?

@bluegenes, Please let me know if you find any issues in the ANI code, so I can help fix them if needed.

2. Should the ANI calculation be optional via a command line flag, or it can be a default column in the output files?

As long as it doesn't degrade performance, I think adding it as a default column is good!

Reading list of query paths from: 'sig_paths.l'
Loaded 532 query signature(s)
Processed 0 comparisons
Processed 100000 comparisons
DONE. Processed 141246 comparisons

pairwise command before adding ANI

	Command being timed: "sourmash scripts pairwise sig_paths.l -c 16 -k 51 -s 10000 -o pairwise_result.csv -t 0.0"
	User time (seconds): 198.93
	System time (seconds): 4.22
	Percent of CPU this job got: 1471%
	Elapsed (wall clock) time (h:mm:ss or m:ss): 0:13.80
	Average shared text size (kbytes): 0
	Average unshared data size (kbytes): 0
	Average stack size (kbytes): 0
	Average total size (kbytes): 0
	Maximum resident set size (kbytes): 1781012
	Average resident set size (kbytes): 0
	Major (requiring I/O) page faults: 0
	Minor (reclaiming a frame) page faults: 623829
	Voluntary context switches: 142733
	Involuntary context switches: 150950
	Swaps: 0
	File system inputs: 0
	File system outputs: 50232
	Socket messages sent: 0
	Socket messages received: 0
	Signals delivered: 0
	Page size (bytes): 4096
	Exit status: 0

pairwise command after adding ANI

	Command being timed: "sourmash scripts pairwise sig_paths.l -c 16 -k 51 -s 10000 -o ani_pairwise_result.csv -t 0.0"
	User time (seconds): 286.95
	System time (seconds): 5.45
	Percent of CPU this job got: 1084%
	Elapsed (wall clock) time (h:mm:ss or m:ss): 0:26.96
	Average shared text size (kbytes): 0
	Average unshared data size (kbytes): 0
	Average stack size (kbytes): 0
	Average total size (kbytes): 0
	Maximum resident set size (kbytes): 1859196
	Average resident set size (kbytes): 0
	Major (requiring I/O) page faults: 3
	Minor (reclaiming a frame) page faults: 598861
	Voluntary context switches: 137504
	Involuntary context switches: 461133
	Swaps: 0
	File system inputs: 0
	File system outputs: 55432
	Socket messages sent: 0
	Socket messages received: 0
	Signals delivered: 0
	Page size (bytes): 4096
	Exit status: 0

@bluegenes Am I correctly calculating ANI from containment here? https://github.com/sourmash-bio/sourmash_plugin_branchwater/pull/188/files#diff-d62689bb23b648c8b52cd5f21c1bf1cf3974ef38477c5ca006c9b1741d380aa8R74-R107

is that really a 2x hit in time!?

@bluegenes Am I correctly calculating ANI from containment here? https://github.com/sourmash-bio/sourmash_plugin_branchwater/pull/188/files#diff-d62689bb23b648c8b52cd5f21c1bf1cf3974ef38477c5ca006c9b1741d380aa8R74-R107

This is an excellent opportunity for a test: calculate containment with regular sourmash code, then calculate it with your code, then compare the numbers...

is that really a 2x hit in time!?

Yes :/

This is an excellent opportunity for a test: calculate containment with regular sourmash code, then calculate it with your code, then compare the numbers...

I can load the comparisons CSV output and use the containment values (max, min) to calculate avg ani with sourmash then validate with the new rust ANI column. Is that a valid test?

This is an excellent opportunity for a test: calculate containment with regular sourmash code, then calculate it with your code, then compare the numbers...

I can load the comparisons CSV output and use the containment values (max, min) to calculate avg ani with sourmash then validate with the new rust ANI column. Is that a valid test?

Pretty much anything can be a valid test, but it sounds more roundabout than (a) running a comparison with sourmash, (b) saving the result, (c) writing a test that runs a comparison with branchwater and checks that you get approximately the same value? Hardcoding the known-good answer into a test is fine.

hi @mr-eyes --

I'm moving ANI calculation into rust in sourmash-bio/sourmash#2943.

Would you like to submit a PR with the equations file into that PR? I have some modifications I'll make, but I can do it over there.

If not, lmk and I'll just add the code directly

cc @ctb

hi @mr-eyes --

I'm moving ANI calculation into rust in sourmash-bio/sourmash#2943.

Would you like to submit a PR with the equations file into that PR? I have some modifications I'll make, but I can do it over there.

If not, lmk and I'll just add the code directly

cc @ctb

I don't mind copying or modifying the code. Although I would love to contribute to the amazing PR, I am not currently able to commit. Please go for it.

I don't mind copying or modifying the code. Although I would love to contribute to the amazing PR, I am not currently able to commit. Please go for it.

you just run git commit man. come on.

😁

I don't mind copying or modifying the code. Although I would love to contribute to the amazing PR, I am not currently able to commit. Please go for it.

you just run git commit man. come on.

😁

Hahaha, ok as long as my committed code will not be judged

@bluegenes @ctb
I thought I would do the full integration by opening a PR. That sourmash-bio/sourmash#3005 was fast :) Thank you!

Closing in favor of #236