如何为JSONSchema定制错误消息?
有没有办法根据给定的条件提供自定义错误消息?
我使用的是https://github.com/networknt/json-schema-validator版本1.0.43
这是我的JSON架构:
{
"$id": "https://configurations/provider.json",
"$schema": "http://json-schema.org/draft-07/schema#",
"title": "Configuration",
"type": "object",
"properties": {
"provider": {
"description": "Name of the provider.",
"enum": [
"Provider #1",
"Provider #2"
]
},
"configuration": {
"$schema": "json-schema/configurations/configuration.json"
}
},
"if": {
"properties": {
"parcsProvider": {
"const": "Provider #1"
}
}
},
"then": {
"required": [
"configuration"
]
},
"else": {
"not": {
"required": [
"configuration"
]
}
}
}
如果提供程序的值是";Provider#1";,则配置对象是必需的,如果它是";Provider#2";并且传递了配置,则会发生错误。我要自定义该错误,以便响应与现在相同,但使用像";Provider 2这样的自定义消息不能有配置。";
当前错误消息/响应:
{
"timestamp": "2020-11-23T12:50:56.20658+01:00",
"message": "invalid.json.input",
"validationErrors": [
{
"field": "$",
"message": "$: should not be valid to the schema "not" : {"required":["configuration"]}"
}
]
}
解决方案
我有一个类似的要求要在我的一个项目中实现。对于验证,我使用https://github.com/everit-org/json-schema。
以下是我所做的
- 已对验证器抛出的所有类型的错误[必须有特定关键字]进行分类
- 现在,一旦拥有了所有密钥,您就可以轻松处理错误并发送自定义错误/响应。
以下是我为不同情况收集的钥匙,这可能会对您有所帮助-
MIN_LENGTH_VIOLATION = "expected minLength"
MAX_LENGTH_VIOLATION = "expected maxLength"
PATTERN_VIOLATION = "does not match pattern"
DATA_TYPE_VIOLATION = "expected type"
DEPENDENCY_VIOLATION = "is required"
FORMAT_VIOLATION_OR_ENUM_VALIDATION_VIOLATION = "is not a valid"
MANDATORY_FIELD_VIOLATION_OR_CONDITIONAL_VIOLATION = "required key"
NUMBER_IS_LESS_THAN_VIOLATION = "is not greater or equal to"
NUMBER_IS_GREATER_THAN_VIOLATION = "is not less or equal"
EXCLUSIVE_NUMBER_IS_GREATER_THAN_VIOLATION = "is not less than"
EXCLUSIVE_NUMBER_IS_LESS_THAN_VIOLATION = "is not greater than"
MULTIPLE_OF_VIOLATION = "is not a multiple"
示例代码-
private static void validate(JSONObject request) {
try {
// Schema, that might be fetched dynamically from some data source
JSONObject schema = new JSONObject();
Schema loadedSchema = SchemaLoader.load(schema);
loadedSchema.validate(request);
} catch (ValidationException ve) {
List<String> allErrorMessages = ve.getAllMessages();
List<String> mandatoryFields = parseMandatoryField(allErrorMessages);
if (CollectionUtils.isNotEmpty(mandatoryFields)) {
throw new MandFieldMissingExp(mandatoryFields);
} else {
List<String> invalidFields = parseInvalids(allErrorMessages);
throw new InvalidFieldExp(invalidFields);
}
}
}
private static List<String> parseMandatoryField(List<String> validationExceptionMessages) {
Set<String> mandatoryListSet = new HashSet<>();
validationExceptionMessages.forEach(errorMessage -> {
if (StringUtils.containsAny(errorMessage, MANDATORY_FIELD_VIOLATION_OR_CONDITIONAL_VIOLATION, DEPENDENCY_VIOLATION)) {
mandatoryListSet.add(StringUtils.substring(errorMessage, errorMessage.indexOf('[') + 1, errorMessage.indexOf(']')));
}
});
return new ArrayList<>(mandatoryListSet);
}
private static List<String> parseInvalids(List<String> validationExceptionMessages) {
Set<String> invalidParamsSet = new HashSet<>();
validationExceptionMessages.forEach(errorMessage -> {
if (StringUtils.containsAny(errorMessage, MIN_LENGTH_VIOLATION, MAX_LENGTH_VIOLATION, PATTERN_VIOLATION,
FORMAT_VIOLATION_OR_ENUM_VALIDATION_VIOLATION, DATA_TYPE_VIOLATION, NUMBER_IS_LESS_THAN_VIOLATION,
MULTIPLE_OF_VIOLATION, EXCLUSIVE_NUMBER_IS_GREATER_THAN_VIOLATION, NUMBER_IS_GREATER_THAN_VIOLATION
, EXCLUSIVE_NUMBER_IS_LESS_THAN_VIOLATION)) {
invalidParamsSet.add(StringUtils.substring(errorMessage, errorMessage.indexOf('/') + 1, errorMessage.indexOf(':')));
}
});
return new ArrayList<>(invalidParamsSet);
}
希望能有所帮助
相关文章