Use latest pre-release of the SSSOM schema. (#625)

* Use latest pre-release of the SSSOM schema.

Update the sssom_schema dependency to use the latest 1.1.0 pre-release,
so that we can start using the new slots and enum values in the upcoming
version of the specification and ensure that SSSOM-Py is ready to work
with them.

* Use SSSOMSchemaView() when needing to get the list of mapping slots.

The `inject_metadata_into_df` function (whose entire purpose is very
dubious but that is another matter) needs a list of the slots available
in the mapping class. It gets that list by doing its own loading of the
SSSOM schema. It does so without specifying any encoding for the schema
file, and apparently the default YAML parser, when not provided with an
explicit encoding on Windows, assumes the encoding is CP-1252. This
results in a failure to load the schema if it happens to contain
UTF-8-encoded non-ASCII characters (which in version 1.1.0a2 it does,
because we are no longer in the 1980s).

This commit updates the `inject_metadata_into_df` function so that it
reuses the SSSOMSchemaView class, which is precisely intended to give
access to the contents of the schema.

Author: gouttegd

poetry.lock

@@ -18,7 +18,7 @@ dependencies = [
     "linkml-runtime>=1.7.5",
     "linkml>1.7.10",
     "pandas>1.0.3",
-    "sssom-schema==1.0.0",
+    "sssom-schema==1.1.0a2",
     "sparqlwrapper>=2.0.0",
     "validators>=0.20.0",
     "deprecation>=2.1.0",

pyproject.toml

@@ -30,7 +30,6 @@
 import numpy as np
 import pandas as pd
 import validators
-import yaml
 from curies import Converter, ReferenceTuple
 from jsonschema import ValidationError
 from linkml_runtime.linkml_model.types import Uriorcurie
@@ -61,7 +60,6 @@
     PREDICATE_MODIFIER,
     PREDICATE_MODIFIER_NOT,
     RDFS_SUBCLASS_OF,
-    SCHEMA_YAML,
     SEMAPV,
     SKOS_BROAD_MATCH,
     SKOS_CLOSE_MATCH,
@@ -1209,9 +1207,7 @@ def inject_metadata_into_df(msdf: MappingSetDataFrame) -> MappingSetDataFrame:
     # TODO add this into the "standardize" function introduced in
     #  https://github.com/mapping-commons/sssom-py/pull/438
     # TODO Check if 'k' is a valid 'slot' for 'mapping' [sssom.yaml]
-    with SCHEMA_YAML.open() as file:
-        schema = yaml.safe_load(file)
-    slots = schema["classes"]["mapping"]["slots"]
+    slots = SSSOMSchemaView().mapping_slots
     for k, v in msdf.metadata.items():
         if k not in msdf.df.columns and k in slots:
             if k == MAPPING_SET_ID: