[MRG] use picklist to exclude signatures

[bluegenes]

Use a picklist to exclude, rather than include, signatures.

Ref #1599

[ctb]

well that is way cleaner syntax than I was thinking about!

(I was noodling on about using ~ somewhere in the picklist args. Don't know why. This is nice!)

[codecov-commenter]

Codecov Report

Merging #1623 (2c2cd15) into latest (0814bcc) will increase coverage by 8.10%.
The diff coverage is 93.10%.

@@            Coverage Diff             @@
##           latest    #1623      +/-   ##
==========================================
+ Coverage   81.39%   89.49%   +8.10%     
==========================================
  Files         103       76      -27     
  Lines       10549     6864    -3685     
  Branches     1228     1236       +8     
==========================================
- Hits         8586     6143    -2443     
+ Misses       1754      510    -1244     
- Partials      209      211       +2     

Flag	Coverage Δ
python	89.49% <93.10%> (+<0.01%)	⬆️
rust	?

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

Impacted Files	Coverage Δ
src/sourmash/sourmash_args.py	93.94% <87.50%> (-0.18%)	⬇️
src/sourmash/picklist.py	91.57% <95.23%> (+0.55%)	⬆️
src/core/src/index/sbt/mod.rs
src/core/src/ffi/hyperloglog.rs
src/core/src/sketch/nodegraph.rs
src/core/src/signature.rs
src/core/src/wasm.rs
src/core/src/errors.rs
src/core/src/ffi/nodegraph.rs
src/core/src/ffi/cmd/compute.rs
... and 19 more

---

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 0814bcc...2c2cd15. Read the comment docs.

[bluegenes]

The reporting needs a bit of a rethink. For now I have:

sourmash/src/sourmash/sourmash_args.py

Lines 86 to 97 in f82375e

	def report_picklist(args, picklist):
	if picklist.pickstyle == PickStyle.INCLUDE:
	notify(f"for given picklist, found {len(picklist.found)} matches to {len(picklist.pickset)} distinct values")
	n_missing = len(picklist.pickset - picklist.found)
	elif picklist.pickstyle == PickStyle.EXCLUDE:
	notify(f"for given picklist, found {len(picklist.found)} matches by excluding {len(picklist.pickset)} distinct values")
	n_missing = len(picklist.pickset) - len(picklist.found)
	if n_missing:
	notify(f"WARNING: {n_missing} missing picklist values.")
	if args.picklist_require_all:
	error("ERROR: failing because --picklist-require-all was set")
	sys.exit(-1)

... but thinking about it, n_missing doesn't really make sense in this context. Plus, because we're excluding rather than including, n_missing can end up negative! However, we do need to make sure picklist_require_all is still considered.

@ctb @luizirber What might be a better reporting strategy for exclude? We're currently not keeping track of which signatures were excluded, just those that we want to keep.

[ctb]

hah! I have no informed idea... I was actually hesitant about the reporting in the first place for related reasons, but it seemed to work out OK and I left it in.

Maybe if it's exclude, omit any n_missing calculations?

[bluegenes]

This PR enables one way to handle #433: if trying to use gather on a signature in the database and want to ignore the exact match, can exclude any sigs matching the query sig.

[bluegenes]

ok, I think this is ready for review! @ctb @luizirber

[ctb]

I think maybe one additional test, for an explicit :include?

[ctb]

hah! added test in 2c2cd15. Nice!