Deepak

Split output of jsonschemagen to one json per class

2

Not sure if this has been discussed before but have we thought about splitting the output of jsonschemagen to one json per class? This would, - make the individual classes...

Separate out unit and integration tests

1

Arrange unit and integration tests separately.

Make sure notebooks are run/tested to ensure that they don't break between releases.

Use proper paths for resources instead of relative paths in tests. For example, ```py cwd = os.path.abspath(os.path.dirname(__file__)) resource_dir = os.path.join(cwd, '../resources') target_dir = os.path.join(cwd, '../target') ```

Remove ontobio's usage of deprecated networkx methods

1

Networkx 2.4 has deprecated a few methods including `graph.node[]`. The recommendation is to use `graph.nodes[]`. All calls to `graph.node` --> `graph.nodes`

Handle associations from multiple relations by taking their union

1

Fixes #370

AssociationSetFactory does not handle associations from multiple relations for a subject

1

When fetching associations, say gene to phenotype associations, `AssociationSetFactory` populates an association map for each subject. But this code does not handle associations with multiple relations. For example, `HGNC:6764` has...

There are certain tests in the current test suite that fail due to certain assumptions which have changed over the last several months. As an example: https://github.com/biolink/ontobio/blob/master/tests/test_golr_associations.py#L137 where a check...

Currently there are three config files that define URLs for various remote calls made by Ontobio: - `conf/config.yaml` - `ontobio/config.yaml` - `test/resources/test-config.yaml` We should consolidate these configurations into one file...