Calculate additional codelist values for schema using anyOf or oneOf,…

… like OCDS record packages

#125
open-contracting/lib-cove-ocds#106

CHANGELOG.md

@@ -17,6 +17,10 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - Restore jsonschema's type validator, as its performance has improved in recent Python versions https://github.com/OpenDataServices/lib-cove/pull/127
 - Allow `SchemaJsonMixin` classes to define a `validator` method, that accepts lib-cove's JSON Schema draft 4 validator class and its format checker, and returns a validator instance. https://github.com/OpenDataServices/lib-cove/pull/128
 
+### Fixed
+
+- Calculate additional codelist values for schema using `anyOf` or `oneOf`, like OCDS record packages https://github.com/open-contracting/lib-cove-ocds/issues/106
+
 ## [0.31.0] - 2023-07-06
 
 ### Changed

libcove/lib/common.py

@@ -456,17 +456,28 @@ def get_schema_codelist_paths(
         if "codelist" in value and path not in codelist_paths:
             codelist_paths[path] = (value["codelist"], value.get("openCodelist", False))
 
-        if value.get("type") == "object":
-            get_schema_codelist_paths(None, value, path, codelist_paths)
-        elif value.get("type") == "array" and isinstance(value.get("items"), dict):
-            if value.get("items").get("type") == "string":
-                if "codelist" in value["items"] and path not in codelist_paths:
-                    codelist_paths[path] = (
-                        value["items"]["codelist"],
-                        value["items"].get("openCodelist", False),
+        descendants = []
+        if "oneOf" in value and isinstance(value["oneOf"], list):
+            descendants = value["oneOf"]
+        elif "anyOf" in value and isinstance(value["anyOf"], list):
+            descendants = value["anyOf"]
+        else:
+            descendants = [value]
+
+        for value in descendants:
+            if value.get("type") == "object":
+                get_schema_codelist_paths(None, value, path, codelist_paths)
+            elif value.get("type") == "array" and isinstance(value.get("items"), dict):
+                if value.get("items").get("type") == "string":
+                    if "codelist" in value["items"] and path not in codelist_paths:
+                        codelist_paths[path] = (
+                            value["items"]["codelist"],
+                            value["items"].get("openCodelist", False),
+                        )
+                elif value.get("items").get("properties"):
+                    get_schema_codelist_paths(
+                        None, value["items"], path, codelist_paths
                     )
-            elif value.get("items").get("properties"):
-                get_schema_codelist_paths(None, value["items"], path, codelist_paths)
 
     return codelist_paths
 

tests/lib/fixtures/common/schema_with_oneof_codelists.json

@@ -0,0 +1,34 @@
+{
+  "$schema": "http://json-schema.org/draft-04/schema#",
+  "type": "object",
+  "properties": {
+    "data": {
+      "oneOf": [
+        {
+          "type": "object",
+          "properties": {
+            "entityType": {
+              "openCodelist": false,
+              "title": "Type",
+              "type": "string",
+              "codelist": "currency.csv"
+            }
+          }
+        },
+        {
+          "type": "object",
+          "properties": {
+            "entityType": {
+              "enum": [
+                "otherValue"
+              ],
+              "type": "string"
+            }
+          }
+        }
+      ]
+    }
+  },
+  "type": "array",
+  "version": "0.1"
+}

tests/lib/test_common.py

@@ -1347,3 +1347,35 @@ def test_get_additional_codelist_values():
             "extension_codelist": False,
         },
     }
+
+
+def test_get_additional_codelist_values_anyOf():
+    json_data = {"data": {"entityType": "additional"}}
+
+    schema_obj = SchemaJsonMixin()
+    schema_obj.schema_host = os.path.join(
+        os.path.dirname(os.path.realpath(__file__)), "fixtures", "common/"
+    )
+    schema_obj.release_pkg_schema_name = "schema_with_oneof_codelists.json"
+    schema_obj.pkg_schema_url = os.path.join(
+        schema_obj.schema_host, schema_obj.release_pkg_schema_name
+    )
+    # Ideally we wouldn't rely on getting codelists from an external source.
+    # (Makes for faster tests with less dependencies)
+    # But load_codelist always uses requests lib, so we have to.
+    schema_obj.codelists = "https://raw.githubusercontent.com/open-contracting/standard/1.1/schema/codelists/"
+
+    additional_codelist_values = get_additional_codelist_values(schema_obj, json_data)
+
+    assert additional_codelist_values == {
+        "data/entityType": {
+            "path": "data",
+            "field": "entityType",
+            "codelist": "currency.csv",
+            "codelist_url": "https://raw.githubusercontent.com/open-contracting/standard/1.1/schema/codelists/currency.csv",
+            "codelist_amend_urls": [],
+            "isopen": False,
+            "values": ["additional"],
+            "extension_codelist": False,
+        }
+    }