Implement initial SSSOM export - #124
Conversation
| **kwargs: Any, | ||
| ) -> None: | ||
| if converter is None: | ||
| converter = bioregistry.get_default_converter() |
There was a problem hiding this comment.
not clear if this package produces a coherent prefix map somewhere else, so Bioregistry is a good place to get a comprehensive one if you aren't sure what you'll get
|
|
||
|
|
||
| LOOKUPS = { | ||
| "=": v.exact_match, |
There was a problem hiding this comment.
question: what else can appear as a relation besides =?
There was a problem hiding this comment.
For now, I would keep it as it then we can fix this later.
|
@HamedBabaei is there a source of example matching objects? Ideally, I would want to include unit tests for the conversion into SSSOM objects |
Thanks @cthoyt for taking care of this. With what @abhishekreddyallu is doing, I think we can add more info to this module's SSSOM data. Sure, here are the samples: |
|
@HamedBabaei I added a test that uses the JSON file, but it does not contain any more exotic information (for now) I did notice that the scores are bigger than 1 - should these be interpreted as percentages out of 100? |
2 participants
Closes #123
This PR implements an initial SSSOM exporter.
Because the
matchingobject appears to be an arbitrary dictionary, I wasn't able to make any interesting additions to the SSSOM data model beyond the subject, predicate, object, and confidence beyond what I could understand from the Alignment API XML export.Ideally, there is a possibility for representing much more rich information about the methods used to produce mappings, which can be done in a follow-up.