Fix inversion of mappings containing subject_type or object_type slots. (#595)

* Add missing slots to COLUMN_INVERT_DICTIONARY.

Mapping inversion requires, among other things, to know which are the
slots whose value must be inverted. For example, knowing that
`subject_id` must be inverted with `object_id`, `subject_label` must be
inverted with `object_label`, and so on.

This knowledge is provided, in SSSOM-Py, in the
`COLUMN_INVERT_DICTIONARY` map in the `sssom.constants` module.

That map is missing the `subject_type` and `object_type` slots,
resulting in SSSOM-Py being unable to invert any mapping set that
contains one of those slots.

We add the missing entries to the `COLUMN_INVERT_DICTIONARY` map.

closes #554

* Add all "asymmetric" slots to the asymmetric test case.

The test supposed to ensure that "asymmetric" mapping sets (sets that
contain some `subject_*` columns without their `object_*` counterparts,
or the other way round) only contains a couple of asymmetric slots.

As a result, it didn't catch the problem caused by the absence of
`subject_type` and `object_type` in the COLUMN_INVERT_DICTIONARY map.

We fill the `asymmetric.tsv` set so that it contains _all_ the slots
that require switching when a mapping set is inverted.

Author: gouttegd

src/sssom/constants.py

@@ -158,13 +158,15 @@
     SUBJECT_SOURCE: OBJECT_SOURCE,
     SUBJECT_PREPROCESSING: OBJECT_PREPROCESSING,
     SUBJECT_SOURCE_VERSION: OBJECT_SOURCE_VERSION,
+    SUBJECT_TYPE: OBJECT_TYPE,
     OBJECT_ID: SUBJECT_ID,
     OBJECT_LABEL: SUBJECT_LABEL,
     OBJECT_CATEGORY: SUBJECT_CATEGORY,
     OBJECT_MATCH_FIELD: SUBJECT_MATCH_FIELD,
     OBJECT_SOURCE: SUBJECT_SOURCE,
     OBJECT_PREPROCESSING: SUBJECT_PREPROCESSING,
     OBJECT_SOURCE_VERSION: SUBJECT_SOURCE_VERSION,
+    OBJECT_TYPE: SUBJECT_TYPE,
 }
 
 

tests/data/asymmetric.tsv

@@ -9,9 +9,9 @@
 #  - orcid:5678
 #license: https://creativecommons.org/publicdomain/zero/1.0/
 #mapping_date: 2020-05-30
-subject_id	subject_label	predicate_id	object_id	mapping_justification	object_source	confidence
-x:appendage	appendage	owl:equivalentClass	y:appendage	semapv:ManualMappingCuration	y:example	0.841
-x:appendage	appendage	owl:equivalentClass	z:appendage	semapv:LexicalMatching	z:example	0.882
-x:appendage	appendage	owl:equivalentClass	z:appendage	semapv:ManualMappingCuration	z:example	0.841
-x:bone_element	bone element	owl:equivalentClass	y:bone	semapv:LexicalMatching	y:example	0.739
-x:bone_element	bone element	owl:equivalentClass	y:bone	semapv:ManualMappingCuration	y:example	0.535
+subject_id	subject_label	predicate_id	object_id	mapping_justification	object_source	confidence	subject_type	subject_category	object_source_version	subject_match_field	subject_preprocessing
+x:appendage	appendage	owl:equivalentClass	y:appendage	semapv:ManualMappingCuration	y:example	0.841	owl class	some category	http://example.org/y/1.0	rdfs:label	z:foo_processing
+x:appendage	appendage	owl:equivalentClass	z:appendage	semapv:LexicalMatching	z:example	0.882	owl class	some category	http://example.org/y/1.0	rdfs:label	z:foo_processing
+x:appendage	appendage	owl:equivalentClass	z:appendage	semapv:ManualMappingCuration	z:example	0.841	owl class	some category	http://example.org/y/1.0	rdfs:label	z:foo_processing
+x:bone_element	bone element	owl:equivalentClass	y:bone	semapv:LexicalMatching	y:example	0.739	owl class	some category	http://example.org/y/1.0	rdfs:label	z:foo_processing
+x:bone_element	bone element	owl:equivalentClass	y:bone	semapv:ManualMappingCuration	y:example	0.535	owl class	some category	http://example.org/y/1.0	rdfs:label	z:foo_processing