feat: integrate cellxgene-ontology-guide with cellxgene-schema CLI (#815

)

.github/workflows/push_tests.yml

@@ -36,10 +36,10 @@ jobs:
         uses: actions/checkout@v3
         with:
           ref: ${{ github.event.pull_request.head.ref }}
-      - name: Set up Python 3.8
+      - name: Set up Python 3.10
         uses: actions/setup-python@v1
         with:
-          python-version: 3.8
+          python-version: "3.10"
       - name: Python cache
         uses: actions/cache@v1
         with:

cellxgene_schema_cli/cellxgene_schema/cli.py

@@ -78,21 +78,6 @@ def remove_labels(input_file, output_file):
     anndata_label_remover.adata.write(output_file)
 
 
-@click.command(
-    name="convert",
-    short_help="Convert an h5ad from version 2.0.0 to version 3.0.0",
-    help="Convert an h5ad from version 2.0.0 to version 3.0.0. No validation will be performed on either"
-    "the input or the output file.",
-    deprecated=True,
-)
-@click.argument("input_file", nargs=1, type=click.Path(exists=True, dir_okay=False))
-@click.argument("output_file", nargs=1, type=click.Path(exists=False, dir_okay=False))
-def convert(input_file, output_file):
-    from .convert import convert
-
-    convert(input_file, output_file)
-
-
 @click.command(
     name="migrate",
     short_help="Convert an h5ad to the latest schema version.",
@@ -110,7 +95,6 @@ def migrate(input_file, output_file, collection_id, dataset_id):
 
 
 schema_cli.add_command(schema_validate)
-schema_cli.add_command(convert)
 schema_cli.add_command(migrate)
 schema_cli.add_command(remove_labels)
 

cellxgene_schema_cli/cellxgene_schema/migrate.py

@@ -369,10 +369,8 @@ def migrate(input_file, output_file, collection_id, dataset_id):
     # https://github.com/chanzuckerberg/single-cell-curation/blob/43f891005fb9439dbbb747fa0df8f0435ebf3f7c/cellxgene_schema_cli/cellxgene_schema/validate.py#L761-L762
     for key, value in list(dataset.uns.items()):
         if any(
-            [
-                isinstance(value, sparse_class)
-                for sparse_class in (scipy.sparse.csr_matrix, scipy.sparse.csc_matrix, scipy.sparse.coo_matrix)
-            ]
+            isinstance(value, sparse_class)
+            for sparse_class in (scipy.sparse.csr_matrix, scipy.sparse.csc_matrix, scipy.sparse.coo_matrix)
         ):
             if value.nnz == 0:  # number non-zero
                 del dataset.uns[key]

cellxgene_schema_cli/cellxgene_schema/ontology.py

@@ -1,8 +1,7 @@
 import enum
 import gzip
-import json
 import os
-from typing import List, Set, Union
+from typing import Union
 
 from . import env
 
@@ -108,164 +107,3 @@ def get_length(self, gene_id: str) -> int:
             return self.gene_dict[gene_id][1]
         else:
             raise ValueError(f"The id '{gene_id}' is not a valid ENSEMBL id for '{self.species}'")
-
-
-class OntologyChecker:
-    """Handles checking ontology term ids, retrieves ontology labels and ancestors"""
-
-    JSON_FILE = env.PARSED_ONTOLOGIES_FILE
-
-    def __init__(self):  # type: ignore
-        with gzip.open(self.JSON_FILE, "rt") as json_o:
-            self.ontology_dict = json.load(json_o)
-
-    def get_ontologies(self) -> List[str]:
-        """
-        rtype list[str]
-        return: a list of ontologies available in the checker
-        """
-
-        return list(self.ontology_dict.keys())
-
-    def get_term_dict(self, ontology: str, term_id: str) -> dict:  # type: ignore
-        """
-        Returns a dictionary with all the information from a given ontology and term_id
-
-        :param str ontology: the ontology id
-        :param str term_id: the ontology term id
-
-        :rtype dict
-        :return Dictionary with all the information for the given term id
-        """
-
-        self.assert_term_id(ontology, term_id)
-        return self.ontology_dict[ontology][term_id]  # type: ignore
-
-    def get_term_label(self, ontology: str, term_id: str) -> str:
-        """
-        Returns the label associated to an ontology term id
-
-        :param str ontology: the ontology id
-        :param str term_id: the ontology term id
-
-        :rtype str
-        :return Label associated to term id
-        """
-
-        self.assert_term_id(ontology, term_id)
-        return self.ontology_dict[ontology][term_id]["label"]  # type: ignore
-
-    def get_term_ancestors(self, ontology: str, term_id: str) -> Set[str]:
-        """
-        Returns the ancestors of an ontology id
-
-        :param str ontology: the ontology id
-        :param str term_id: the ontology term id
-
-        :rtype Set[str]
-        :return All term ids that are ancestors of the query term id.
-        """
-
-        self.assert_term_id(ontology, term_id)
-        return set(self.ontology_dict[ontology][term_id]["ancestors"])
-
-    def is_valid_ontology(self, ontology: str) -> bool:
-        """
-        Returns True if the ontology is present in the ontology dict
-
-        :param str ontology: the ontology id
-
-        :rtype bool
-        :return True if the ontology is present in the ontology dict, False otherwise
-        """
-
-        return ontology in self.ontology_dict
-
-    def is_term_id_deprecated(self, ontology: str, term_id: str) -> bool:
-        """
-        Returns True if the id has been deprecated (obsolete) in the ontology
-
-        :param str ontology: the ontology id
-        :param str term_id: the ontology term id
-
-        :rtype bool
-        :return True if id has been deprecated
-        """
-
-        self.assert_term_id(ontology, term_id)
-
-        return self.ontology_dict[ontology][term_id]["deprecated"]  # type: ignore
-
-    def is_valid_term_id(self, ontology: str, term_id: str) -> bool:
-        """
-        Returns True if term_id is a valid id  from ontology
-
-        :param str ontology: the ontology id
-        :param str term_id: the ontology term id
-
-        :rtype bool
-        :return True if term_id is a valid id  from ontology, False otherwise
-        """
-
-        self.assert_ontology(ontology)
-
-        return term_id in self.ontology_dict[ontology]
-
-    def is_descendent_of(self, ontology: str, query_term_id: str, target_term_id: str) -> bool:
-        """
-        Returns True if query_term_id is a descendent of target_term_id in a given ontology
-
-        :param str ontology: the ontology id
-        :param str query_term_id: the ontology term id
-        :param str target_term_id: the ontology term id
-
-        :rtype bool
-        :return True if query_term_id is a descendent of target_term_id in a given ontology, False otherwise
-        """
-
-        self.assert_term_id(ontology, query_term_id)
-        self.assert_term_id(ontology, target_term_id)
-
-        return target_term_id in self.get_term_ancestors(ontology, query_term_id)
-
-    def assert_ontology(self, ontology: str):  # type: ignore
-        """
-        Raises error if ontology is not present in the ontology dict
-
-        :param str ontology: the ontology id
-
-        :rtype None
-        """
-
-        if not self.is_valid_ontology(ontology):
-            raise ValueError(f"The ontology '{ontology}' is not present in the ontology checker")
-
-    def assert_term_id(self, ontology: str, term_id: str):  # type: ignore
-        """
-        Raises error if term_id is not present in ontology
-
-        :param str ontology: the ontology id
-        :param str term_id: the ontology term id
-
-        :rtype None
-        """
-
-        if not self.is_valid_term_id(ontology, term_id):
-            raise ValueError(f"The term id '{term_id}' is not present in the ontology '{ontology}'")
-
-    def assert_descendent_of(self, ontology: str, query_term_id: str, target_term_id: str):  # type: ignore
-        """
-        Raises error if query_term_id is not a descendent of target_term_id in a given ontology
-
-        :param str ontology: the ontology id
-        :param str query_term_id: the ontology term id
-        :param str target_term_id: the ontology term id
-
-        :rtype None
-        """
-
-        if not self.is_descendent_of(ontology, query_term_id, target_term_id):
-            raise ValueError(
-                f"The term id '{query_term_id}' is not a descendent of the term id '{target_term_id}'"
-                f" in the ontology '{ontology}'"
-            )