[MRG] provide "protocol" tests for Index, CollectionManifest, and LCA_Database classes

[ctb]

Use pytest fixtures and parametrized tests to start building standardized test suites for the Index, CollectionManifest, and LCA_Database classes.

This PR creates three new test files, test_index_protocol.py, test_manifest_protocol.py, and test_lca_db_protocol.py, that use fixtures and parametrized tests to evaluate the various Index, CollectionManifest, and LCA_Database classes. Part of the motivation is to provide a clearer and more uniform test suite for the various Sqlite* classes being introduced in #1808 and #1933.

While Index had a fairly well defined API already, this PR ends up defining (and refactoring) the CollectionManifest and LCA_Database APIs to better test and support core APIs moving forward.
Both LCA_Database and CollectionManifest received a lot of refactoring as a result, but no functionality was changed.

In addition to creating these new tests, this PR fixes a few problems caught by the new tests:

• fixes len(...) on SBT to only count the number of leaves, not the number of internal nodes;
• adding a test for sourmash lca rankinfo --minimum-num;
• fixes a column-off-by-one bug in the LCA tests/GTDB spreadsheet;
• implements __len__ for LazyLoadedIndex, and adds some correctness checks to constructor;

Remaining issues:

• RevIndex doesn't obey the full Index API; punted to index.RevIndex is not a fully implemented Index class #1939
• might want to clean up the basic lca_db and manifest creation code in those protocol tests 🤔

TODO:

• evaluate the sqlite manifest API in light of the sigs-to-manifest script and the updating code there;

Creates problem, and then fixes it;
Fixes #1949

[ctb]

@sourmash-bio/devs curious if there is a better approach within the pytest world.

[codecov]

Codecov Report

Merging #1936 (cb54d7d) into latest (6f7eb06) will increase coverage by 0.07%.
The diff coverage is 98.27%.

@@            Coverage Diff             @@
##           latest    #1936      +/-   ##
==========================================
+ Coverage   83.17%   83.25%   +0.07%     
==========================================
  Files         126      126              
  Lines       13954    13989      +35     
  Branches     1910     1913       +3     
==========================================
+ Hits        11606    11646      +40     
+ Misses       2075     2071       -4     
+ Partials      273      272       -1     

Flag	Coverage Δ
python	91.22% <98.27%> (+0.08%)	⬆️
rust	65.16% <ø> (ø)

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

Impacted Files	Coverage Δ
src/sourmash/lca/lca_db.py	91.80% <96.29%> (+0.50%)	⬆️
src/sourmash/index/__init__.py	97.10% <100.00%> (+0.37%)	⬆️
src/sourmash/index/revindex.py	63.46% <100.00%> (ø)
src/sourmash/lca/command_index.py	89.50% <100.00%> (ø)
src/sourmash/lca/command_rankinfo.py	84.61% <100.00%> (+3.22%)	⬆️
src/sourmash/manifest.py	95.09% <100.00%> (+0.84%)	⬆️
src/sourmash/sbt.py	85.37% <100.00%> (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 6f7eb06...cb54d7d. Read the comment docs.

[ctb]

Ready for review & merge, @sourmash-bio/devs !

[bluegenes]

this lgtm! ...but I can't speak to whether there are any better pytest strategies 🤷🏻‍♀️