How to define application specific error code in Swagger - swagger

I'd like to define the possible application error code, not HTTP status code, for example:
In the returned JSON
{
code: 2000,
message: "create success"
} // or
{
code: 2001
message: "user already exists"
}
....
I know you can define a schema for this structure, but schema is not enough here, since I need to define possible code and message.
Is it possible? If so, how to do it in Swagger?

Related

Serverless - Change the content before deploy

I'm using Serverless for working with our aws lambda / appsync.
For Error Handling, we are keep erro code with message in a json file. The Codes will be unique. Something like this:
//error-code.json
{
"1"": { code: 1, message: "Invalid User Input"},
"2"": { code: 2, message: "Invalid Input"},
//... so on
}
This wil lbe deploy as layer and all the lambda will use it. Issue is we cannot use it in the resolve template. There are some of the resolver will be only template file. These template files cannot access the json file nor can access the layer. How can I use the error-code.json here?
Solution 1:
Manually write the error code in templates and make sure there are alway unique. Something like this:
#set(#errorInfo = {
"erroCode": "1",
"errorMessage": "Invalid Input"
})
$util.error("Invalid Input", "errorType", $ctx.arguments,#errorInfo)
Rejected: Becasue we have to manually check everytime for the unique of error code. In case of lot of template file, we cannot rely on it.
Solution 2:
Create a table with error code (unique) and error message. Use this table to send error from template.
Rejected: Because we use multiple app sync instance and they all connect to dirferent database. So we have to make this table in all database, and thus unique across the app-sync is not maintained.
Solution 3:
Write the placeholder in vtl where we want to send the error. Before Deploy, replace the placeholder with the actual code using pre-hook script, but not in the actual vtl file but in the generated package that serverless deploy. Does Serverless even such thing?
if your errors are all static, there is one more option for consideration.
You create one more file that holds all errors defined in Velocity.
$util.qr( $ctx.stash.put("errors", {}) ) $util.qr(
$util.qr( $ctx.stash.errors.put("ONE", { "code": 1, "message": "Invalid User
Input"} )
...
$util.qr( $ctx.stash.errors.put("TWENTY", { "code": 20, "message": "20th error description"} )
For every velocity resolver that throws errors, you inject pre-defined errors at the beginning of its request mapping's file. Whenever you want to throw an error, it's done by retrieving a pre-defined error from $ctx.stash
$util.error ( $ctx.stash.errors.ONE.message, $ctx.stash.errors.ONE.code )
The error file is generated from error-code.json, or manually typed again for simplicity. $ctx.stash is used because stash is accessible from everywhere in a resolver, including pipeline ones.

Google reCAPTCHA Enterprise Unknown field error when Create Assessment

I did follow their documentation on https://googleapis.dev/ruby/google-cloud-recaptcha_enterprise-v1/latest/index.html
I first added the env variable RECAPTCHA_ENTERPRISE_CREDENTIALS then run the
client = ::Google::Cloud::RecaptchaEnterprise::V1::RecaptchaEnterpriseService::Client.new
documentation says response = client.create_assessment request but documentation didn't add sample request so I did assume that it will look like in this documentation https://cloud.google.com/recaptcha-enterprise/docs/create-assessment
{
"event": {
"token": "token",
"siteKey": "key"
}
}
But when I execute the client.create_assessment it is giving me error ArgumentError: Unknown field name 'event' in initialization map entry.
I also tried to follow this https://googleapis.dev/ruby/google-cloud-recaptcha_enterprise-v1/latest/Google/Cloud/RecaptchaEnterprise/V1/RecaptchaEnterpriseService/Client.html#create_assessment-instance_method
where the parameters is request and also gives me error ArgumentError: Unknown field name 'request' in initialization map entry.
What could be the correct parameter/request for creating assessment?
Gem I used is google-cloud-recaptcha_enterprise-v1 and google-cloud-recaptcha_enterprise
So I solve this issue by using client.create_assessment(parent: "projects/{your project-id}", assessment: event) my event value was event = Google::Cloud::RecaptchaEnterprise::V1::Event.new(token: "token from execute", site_key: "your site key")
It solves the error and answers the question (but another error after this one occurred about permission denied). I Will update once resolved.

How do I use the messages file to create my own messagesApi

In my play-framework application, there is a messages file which I can use to define my own messages. Eg
##############Application Error messages
error.incorrectBodyType=Incorrect body type. Body type must be JSON
error.incorrectBodyContentInLoginRequest=No Login info found in the request. Check format
error.incorrectUsernameOrPassword = Incorrect username or password
error.unregisteredUser = registration not complete
error.ghostUsername = No record found for the username
error.unauthenticated = You must first sign in
error.unauthorized = You are not permitted
####### Application Success Messages
success.signoutSuccessful = Signout Successful
The play frameworks makes these messages available to my Controller and I can use them like the messagesApi("error.unauthenticated")(langs.availables(0)).
I am unable to figure out how I can use the messages file in my unit tests. I am using compile time injection and am creating my own instances of the required classes. To create MessagesApi, there is a DefaultMessagesApi method but it take a map, not a File.
I am able to create the messages in my unit tests as follows but I have to duplicate the effort of copying messages from messages file to this map
val messagesApi = new DefaultMessagesApi( //takes map of maps. the first is the language file, the 2nd is the map between message title and description
Map("en" -> //the language file
Map("error.incorrectBodyType" -> "Incorrect body type. Body type must be JSON",
"error.incorrectUsernameOrPassword"->"Incorrect username or password",
"error.incorrectBodyContentInLoginRequest"->"No Login info found in the request. Check format", //map between message title and description
"error.unregisteredUser" -> "registration not complete",
"error.ghostUsername" -> "No record found for the username",
"success.signoutSuccessful" -> "Signout Successful",
"error.unauthenticated" -> "You must first sign in",
"error.unauthorized" -> "You are not permitted")
)
)
Is there a way I can read the messages file and use it in my test case?
My test case already extended OneAppPerSuiteWithComponents. I used components.messagesApi of OneAppPerSuiteWithComponents which was already available and reads from the messages file

How to add Error Code to validation attributes in MVC?

I'd like to be able to specify an error code to be returned from my API along with the error message in the event that validation fails. The aim is to specify a validation attribute on a property such as:
[Range(1, int.MaxValue, ErrorMessage = "Page must be 1 or greater", ErrorCode = 1234)]
And, in the event that someone requests page 0, return a 400 Bad Request with a JSON error object in the body like this:
{
"errorCode": 1234,
"errorMessage": "Page : Page must be 1 or greater"
}
I already have a custom ModelValidationFilter which returns an array of error messages from the ModelStateDictionary, so that's all working, but I can't see a straightforward way of getting the error code included in there too without overriding all of the MVC model validation classes.
This seems like a problem which should already have been solved by someone at some point, but I can't find anything to support that theory.

Stub Rails UJS/Ajax responses status to test returning message in Rspec/Capybara feature test

I have AJAX calls initiated by Rails UJS that I would like to test. specifically, I have used Rails UJS ajax events to provide for cases of errors.
I would like to test them but I don't know how to tell rspec/capybara to "stub" and assume the error code
$("button").
on('ajax:error',function(event,xhr, status, error){
if(status == "timeout") {
var msg;
msg = Messenger().post({
message: "This is taking too long"
});
} else {
var msg;
msg = Messenger().post({
message: "it seems there is a bug. Please try again."
});
};
});
I would like to do something like the following:
describe "test returning error message", js: true do
it "should be successful" do
visit deal_page_path(deal)
first('a.button').click
stub(:xhr.status) = "timeout"
expect(page).to have_content('This is taking too long')
end
end
How to do this?
Note: the ajax requests are internal they don't go to third party API or services (such as facebook for ex).
When testing with Capybara (JS enabled drivers) it has no access to the request or response except through the changes it creates in the browser. You could build a test mode into your relevant controllers that could be turned on and off to allow it to output the errors you want, but the cleanest way to do this is probably to use a programmable proxy like puffing-billy which will allow you to selectively return whatever you'd like for any given request from the browser. One thing to realize is that this isn't testing that app correctly returns errors, it's just testing that your front-end handles errors the way you expect.

Resources