Make strict schema

docs/history/changelog.md

@@ -332,6 +332,8 @@ Per the [normative and non-normative content and changes policy](../governance/n
 * [#1450](https://github.com/open-contracting/standard/pull/1450) Replace a repeated example in schema/merging/ with a link to guidance/build/merging/.
 * [#1665](https://github.com/open-contracting/standard/pull/1665) Abandon in-file translations.
 
+* [#1480](https://github.com/open-contracting/standard/pull/1480) Add `UnitValue` definition.
+
 ### Documentation
 
 * [#1115](https://github.com/open-contracting/standard/pull/1115) Add guidance on when having multiple suppliers per award.

manage.py

@@ -23,7 +23,7 @@
 from babel.messages.pofile import read_po
 from docutils.utils import relative_path
 from lxml import etree
-from ocdskit.schema import get_schema_fields
+from ocdskit.schema import add_validation_properties, get_schema_fields
 
 basedir = Path(__file__).resolve().parent
 schemadir = basedir / "schema"
@@ -392,6 +392,126 @@ def get_versioned_release_schema(schema):
     return schema
 
 
+def add_key_based_validation_properties(schema):
+    """
+    Add validation properties based on key names.
+
+    * "format": "email" if the key is "email"
+    * "minimum": 0 to "quantity", "durationInDays" and "numberOfTenderers fields
+    * "required": ["id", "name"] to "Organization" and "OrganizationReference"
+    * "required": ["id"] to "Amendment" and "RelatedProcess"
+    * "$ref": "#/definitions/UnitValue" to "Unit.value"
+
+    Removes "integer" type from "id" and "projectID" fields.
+
+    :param dict schema: a JSON schema
+    """
+    if isinstance(schema, list):
+        for item in schema:
+            add_key_based_validation_properties(item)
+    elif isinstance(schema, dict):
+        for key, value in schema.items():
+            if key == "email":
+                value["format"] = "email"
+            elif key in ["quantity", "durationInDays", "numberOfTenderers"]:
+                value["minimum"] = 0
+            elif key in ["Organization", "OrganizationReference"]:
+                value["required"] = ["id", "name"]
+                value["properties"]["name"]["type"] = "string"
+                value["properties"]["id"]["type"] = "string"
+            elif key in ["Amendment", "RelatedProcess"]:
+                value["required"] = ["id"]
+                value["properties"]["id"]["type"] = "string"
+            elif key in ["id", "projectID"]:
+                if "type" in value and "integer" in value["type"]:
+                    value["type"].remove("integer")
+            elif key == "Unit":
+                value["properties"]["value"]["$ref"] = "#/definitions/UnitValue"
+
+            add_key_based_validation_properties(value)
+
+
+def get_strict_schema(schema):
+    """
+    Return the strict version of the schema.
+    """
+    # Update schema metadata.
+    identifier = schema["id"].split("/")
+    identifier[-1] = f"strict/{identifier[-1]}"
+    schema["id"] = "/".join(identifier)
+    schema["title"] = f'Strict {schema["title"][0].lower()}{schema["title"][1:]}'
+    schema["description"] = (
+        f'{schema["description"]} The strict schema adds additional validation rules planned for inclusion in OCDS 2.0. Use of the strict schema is a voluntary opportunity to improve data quality.'  # noqa: E501
+    )
+
+    # Update references to other schemas.
+    reference_strict_schemas(schema)
+
+    # Reference compiled release schema.
+    if schema["id"].endswith("record-schema.json"):
+        schema["properties"]["compiledRelease"]["$ref"] = schema["properties"]["compiledRelease"]["$ref"].replace(
+            "release-schema.json", "compiled-release-schema.json"
+        )
+
+    # Add validation properties
+    add_validation_properties(schema)
+
+    # Add key-based validation properties
+    add_key_based_validation_properties(schema)
+
+    # Remove null types from package schemas
+    if "package" in schema["id"]:
+        return remove_nulls(schema)
+
+    return schema
+
+
+def get_compiled_release_schema(schema):
+    """
+    Return the compiled release schema.
+    """
+    # Update schema metadata.
+    schema["id"] = schema["id"].replace("release-schema", "compiled-release-schema")
+    schema["title"] = schema["title"].replace("Open Contracting Release", "Open Contracting Compiled Release")
+
+    # Remove null types and enum values.
+    return remove_nulls(schema)
+
+
+def remove_nulls(schema):
+    """
+    Remove null types and enum values.
+    """
+    if isinstance(schema, dict):
+        for key, value in schema.items():
+            if key == "type" and isinstance(value, list) and "null" in value:
+                value.remove("null")
+            if key == "enum" and isinstance(value, list) and None in value:
+                value.remove(None)
+
+            remove_nulls(value)
+
+    return schema
+
+
+def reference_strict_schemas(schema):
+    """
+    Update $refs to reference strict schemas.
+    """
+    if isinstance(schema, dict):
+        for key, value in schema.items():
+            if key == "$ref" and value.startswith("https://standard.open-contracting.org/schema/"):
+                reference = value.split("/")
+                reference[-1] = f"strict/{reference[-1]}"
+                schema[key] = "/".join(reference)
+
+            reference_strict_schemas(value)
+
+    if isinstance(schema, list):
+        for subschema in schema:
+            reference_strict_schemas(subschema)
+
+
 @click.group()
 def cli():
     pass
@@ -494,6 +614,12 @@ def pre_commit():
     - meta-schema.json
     - dereferenced-release-schema.json
     - versioned-release-validation-schema.json
+    - strict/release-schema.json
+    - strict/release-package.json
+    - strict/record-package.json
+    - strict/dereferenced-release-schema.json
+    - strict/versioned-release-validation-schema.json
+    - strict/compiled-release-schema.json
     """
     nonmultilingual = {
         # Identifiers.
@@ -516,6 +642,7 @@ def pre_commit():
     }
 
     release_schema = json_load("release-schema.json")
+    strict_release_schema = get_strict_schema(deepcopy(release_schema))
     jsonref_release_schema = json_load("release-schema.json", jsonref, merge_props=True)
 
     counts = defaultdict(list)
@@ -568,6 +695,24 @@ def pre_commit():
     json_dump("dereferenced-release-schema.json", jsonref_release_schema)
     json_dump("versioned-release-validation-schema.json", get_versioned_release_schema(release_schema))
 
+    # Strict schemas.
+    directory = Path("strict")
+    json_dump(directory / "release-schema.json", strict_release_schema)
+    json_dump(directory / "record-schema.json", get_strict_schema(json_load("record-schema.json")))
+
+    strict_dereferenced_release_schema = json_load(directory / "release-schema.json", jsonref, merge_props=True)
+    json_dump(directory / "dereferenced-release-schema.json", strict_dereferenced_release_schema)
+    json_dump(
+        directory / "versioned-release-validation-schema.json", get_versioned_release_schema(strict_release_schema)
+    )
+
+    json_dump(directory / "release-package-schema.json", get_strict_schema(json_load("release-package-schema.json")))
+    json_dump(directory / "record-package-schema.json", get_strict_schema(json_load("record-package-schema.json")))
+    json_dump(
+        directory / "compiled-release-schema.json",
+        remove_nulls(get_compiled_release_schema(json_load(directory / "release-schema.json"))),
+    )
+
 
 @cli.command()
 @click.argument("file", type=click.File())