SortedSet.remove() having trouble with SortedSet.binary_search()

[smurp]

The Nameless set isn't being drained properly as names arrive because SortedSet.remove() is failing thanks to a problem with SortedSet.binary_search().

This same little bug is likely at the root of remaining situations where nodes "stick* to the shelf, such as Jewish (occasionally) in the Jewish novelist dataset.

Possible Solution

• debug the dang binary_search
• or is add() misbehaving?
• or is the hash function for literals producing collisions, which is confusing SortedSet?

Steps to Reproduce (for bugs)

• http://beta.huviz.dev.nooron.com/#load+https://raw.githubusercontent.com/cwrc/testData/master/sparqlOutputs/jewish_novelist_biographies.ttl+with+http://sparql.cwrc.ca/ontology/cwrc.ttl

[smurp]

See #259

[smurp]

This was not what I thought it was! SortedSet was working fine... it was an issue with the virtual nodes for literals.

Fixed with https://github.com/smurp/huviz/commit/511a80fea6abc53f24eecba2f46343ccb89a23a2 and https://github.com/smurp/huviz/commit/e2c4004e55402e9c8dda640ff89a58b63809c4d5

[smurp]

http://localhost:9997/#load+https://raw.githubusercontent.com/cwrc/testData/master/qa/AmyLevy/LevyBioTriples.ttl+with+http://sparql.cwrc.ca/ontology/cwrc.ttl

has data which triggers:

remove failed at idx 88 to splice http://cwrc.ca/cwrcdata/levyam_SpatialContext_travelsTo_2_identifying out of Shelved because levyam_SpatialContext_inhabits_1_identifying was removed instead of levyam_SpatialContext_travelsTo_2_identifying