feat: experiment library schema changes, ingestion report with errors, json schema warnings

chord_metadata_service/experiments/migrations/0009_v4_1_0.py

@@ -0,0 +1,17 @@
+from django.db import migrations
+
+
+def set_experiment_library_strategy(apps, _schema_editor):
+    Experiment = apps.get_model("experiments", "Experiment")
+    for exp in Experiment.objects.filter(library_strategy="WES"):
+        exp.library_strategy = "WXS"
+        exp.save()
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ('experiments', '0007_v4_0_0'),
+    ]
+
+    operations = [
+        migrations.RunPython(set_experiment_library_strategy)
+    ]

chord_metadata_service/experiments/schemas.py

@@ -1,10 +1,15 @@
 from .descriptions import EXPERIMENT, EXPERIMENT_RESULT, INSTRUMENT
 from chord_metadata_service.restapi.schemas import ONTOLOGY_CLASS_LIST, KEY_VALUE_OBJECT
 from chord_metadata_service.restapi.schema_utils import tag_ids_and_describe
+from chord_metadata_service.ontologies import readXsdSimpleTypeValues
 
 
 __all__ = ["EXPERIMENT_SCHEMA", "EXPERIMENT_RESULT_SCHEMA", "INSTRUMENT_SCHEMA"]
 
+LIBRARY_STRATEGIES = readXsdSimpleTypeValues(
+    'chord_metadata_service/ontologies/xsd/SRA.experiment.xsd.xml',
+    'typeLibraryStrategy',
+)
 
 EXPERIMENT_RESULT_SCHEMA = tag_ids_and_describe({
     "$schema": "http://json-schema.org/draft-07/schema#",
@@ -104,7 +109,7 @@
         "molecule_ontology": ONTOLOGY_CLASS_LIST,
         "library_strategy": {
             "type": "string",
-            "enum": ["Bisulfite-Seq", "RNA-Seq", "ChIP-Seq", "WES", "Other"]
+            "enum": LIBRARY_STRATEGIES
         },
         "library_source": {
             "type": "string",

chord_metadata_service/ontologies/__init__.py

@@ -0,0 +1,5 @@
+from .utils import readXsdSimpleTypeValues
+
+__all__ = [
+    "readXsdSimpleTypeValues",
+]

chord_metadata_service/ontologies/utils.py

@@ -0,0 +1,14 @@
+import xmltodict
+
+
+def readXsdSimpleTypeValues(xsd_file_path: str, type_name: str):
+    """Reads an XML Schema Definition (XSD) file and returns a type's values.
+    The XSD file is parsed using xmltodict following this spec:
+    https://www.xml.com/pub/a/2006/05/31/converting-between-xml-and-json.html
+    """
+    sra_file = open(xsd_file_path).read()
+    sra_experiment_data = xmltodict.parse(sra_file, namespaces={'xs': None})
+    simple_types = {sp["@name"]: sp for sp in sra_experiment_data["schema"]["simpleType"]}
+    target_type = simple_types[type_name]
+    values = [val['@value'] for val in target_type['restriction']['enumeration']]
+    return values