Extract non-string value from response object - azure-api-management

In APIM I want to get an OAuth token before calling the back-end and store it in the APIM cache.
I manage to get the OAuth token.
I wish to store my token in the cache for this I need a duration.
Extracting the duration from the response body is my issue.
First I extract the token and duration to save these in variables.
Then the variables are used to store a value in the cache.
When extracting the integer value of "expires_in" it fails.
Response object:
{
"access_token": "accesstokenresponse",
"token_type": "bearer",
"expires_in": 3600
}
extracting data and put in variables:
<set-variable name="bearer-token"
value="#(((IResponse)context.Variables["response"]).Body.As<JObject>()["access_token"].ToString())" />
<set-variable name="expires-in"
value="#(((IResponse)context.Variables["response"]).Body.As<JObject>(preserveContent: true)["expires_in"].ToString())" />
storing in cache:
<cache-store-value key="OAuthKey"
value="#((string)context.Variables["bearer-token"])"
duration="#((int)context.Variables["expires-in"])" />
Extracting the "expires_in" results in an Object reference not set to an instance of an object.
set-variable (153.612 ms)
{
"messages": [
{
"message": "Expression evaluation failed.",
"expression": "((IResponse)context.Variables[\"response\"]).Body.As<JObject>(preserveContent: true)[\"expires_in\"].ToString()",
"details": "Object reference not set to an instance of an object."
},
"Expression evaluation failed. Object reference not set to an instance of an object.",
"Object reference not set to an instance of an object."
]
}

Try to maintain the value of Objects using (preserveContent: true) on first line as well, thus:
value="#(((IResponse)context.Variables["response"]).Body.As<JObject>(preserveContent:true)["access_token"].ToString())" />
Let me know if it's solved.

Related

Azure Logic App - Parse JSON with dynamic key/name

just want to know if and how I can parse a HTTP response with a dynamic name in a JSON?
I used the Azure Management API to receive the managed identities (system- and user assigned managed identities) to receive all managed identities.
With a foreach I am iterating the results.
If a resource has a system assigned managed identity and user assigned managed identity, the response looks like this:
{
"principalId": "<principalId1>",
"tenantId": "<tenantId>",
"type": "SystemAssigned, UserAssigned",
"userAssignedIdentities": {
"/subscriptions/<subscriptionId>/resourcegroups/<resourceGroupName>/providers/Microsoft.ManagedIdentity/userAssignedIdentities/<userAssignedIdentitiesName>": {
"principalId": "<principalId2>",
"clientId": "<clientId>"
}
}
}
Now, I would like to get the <principalId2>.
Unfortunately, the Name of the object is dynamic related to the scope of the resource /subscriptions/<subscriptionId>/resourcegroups/<resourceGroupName>/providers/Microsoft.ManagedIdentity/userAssignedIdentities/<userAssignedIdentitiesName>.
How can I parse the JSON to receive the needed <principalId2>?
For all other responses I can easily use the Data operations Parse JSON with the payload I inserted from the HTTP response.
Is there a way to use a wildcard? Otherwise, could I somehow just select the first object of userAssignedIdentities to receive the needed value?
Ok, this should work for you. This is the flow I tested with ...
Initialise JSON
Your JSON as a string, how you do that in your solution may differ slightly.
Initialize XPath Result
Defined as an Array and the expression is ...
xpath(xml(json(concat('{ root: ', replace(variables('JSON'), 'PrincipalId', 'principalId'), '}'))), '(//principalId)[2]')
Initialize Result
A bit more work again but defined as a String and the expression is ...
array(xpath(xml(base64ToString(variables('XPath Result')[0]?['$content'])), '//text()'))[0]
The end result should be your value ...

How to parse nested JSON, within a string, using Kusto

I have a Python Azure Function that produces custom logging messages when the Function executes. I need to pull out some of the JSON values nested in the logging message.
How can I use Kusto to access the nested JSON within the logging message string?
Example logging message:
Desired values marked with <----------
####### EventGrid trigger processing an event:
{
"id": "long-guid",
"data": {
"api": "FlushWithClose",
"requestId": "long-guid",
"eTag": "long-guid",
"contentType": "application/octet-stream",
"contentLength": 16264, <----------------------
"contentOffset": 0,
"blobType": "BlockBlob",
"blobUrl": "https://function.blob.core.windows.net/parentdir/childdir/file.name",
"url": "https://function.dfs.core.windows.net/parentdir/childdir/file.name", <---- JUST FILE.NAME here
"sequencer": "long-guid",
"identity": "long-guid",
"storageDiagnostics": {
"batchId": "long-guid"
}
},
"topic": "/subscriptions/long-guid/resourceGroups/resourceGroup/providers/Microsoft.Storage/storageAccounts/accountName",
"subject": "/blobServices/default/containers/containerName/blobs/childDir/file.name",
"event_type": "Microsoft.Storage.BlobCreated"
} #######
I imagine it has something to do with the Kusto extend function, but piping in...
| extend parsedMessage = todynamic(message)
| project timestamp, test = parsedMessage["id"]
...yields only an empty test column
message in your specific case isn't a valid JSON payload - as it has the ###... EventGrid trigger processing an event: prefix (and a somewhat similar suffix).
That is why todynamic() isn't able to process it and why you're not able to reference properties in the JSON payload that's included in it.
Ideally, you would change the payload you ingest to be a valid JSON payload, and re-type the target column to dynamic instead of string.
If you can't do that, you can use the substring() function or parse operator to get everything but the aforementioned prefix/suffix, and parse the output of that using todynamic()
though note that doing that each time you query the data bears runtime overhead that could be avoided by following the advice above.

AWS Step function string/json concatenation

I have orchestrated a data pipe line using AWS Step function.
In last state I want to send a custom notification. I'm using an Intrinsic function States.Format to format my message and subject. It works fine for Context object element. Here, I have tested that in Message parameter.
But it doesn't work with input JSON. This is my input JSON
{
"job-param":{
"pipe-line-name":"My pipe line name", "other-keys":"other values"
}
}
"Success State": {
"Type": "Task",
"Resource": "arn:aws:states:::sns:publish",
"Parameters": {
"Message.$": "States.Format('Execution Id:{}, completed successfully!', $$.Execution.Id)",
"Subject.$": "States.Format('[INFO] {} completed successfully!', $.job-param.pipe-line-name)",
"TopicArn": "arn:aws:sns:us-east-1:************:sns-topic"
},
"End": true
}
While saving this state machine, it gives me following error message:
The value for the field 'Subject.$' must be a valid JSON Path
I checked Input and Result path. They have this value. I can directly use this value as parameter. This is working fine. But I can't format with other string.
"Subject.$": "$.job-param.pipe-line-name"
Alternate approach would be to call lambda to customize and trigger SNS. But I want to avoid that.
Can I request some suggestions to fix this error?
Thanks in advance!
If you want to use any name with - in your JSON then you can write your JSON Path like this:
"Subject.$": "States.Format('[INFO] {} completed successfully!', $['job-param']['pipe-line-name'])",
But it would be easier if you change your input JSON and replace - with _:
"Subject.$": "States.Format('[INFO] {} completed successfully!', $.job_param.pipe_line_name)",

creating RAML file for dynamic key/values for get/post

I was using a Mule application to get data from another system which is invoked by an HTTP endpoint, used java with jersey api to get the rest component.
Sample input data is given below, here keys and values are not fixed, it may vary based on user request. Array size will increase may be 2 to n entries. It is working fine with Mule and Java rest based component.
Input JSON data:
[
{
"Company": "BEG1",
"Account": "10011",
"Deptid": "111",
"Location": "SM1",
"Transaction Date": "2014-07-15",
"Description": "Invoice1",
"Debit": 0,
"Credit": 13.46,
"Invoice Nbr": "16824321"
},
{
"Company": "BEG92",
"Account": "10092",
"Deptid": "222",
"Location": "SL2",
"Transaction Date": "2014-07-19",
"Description": "Invoice End2",
"Debit": 13.46,
"Credit": 0,
"Invoice Nbr": "168243292"
}
]
Planning to migrate to APIkit with RAML: how can I make a RAML template for above case? Since keys are dynamic, this doesn't seem straightforward.
2) With same approach for GET, I will get the data for my get request, I am not sure what is the key and its corresponding values, only I was doing is get the data, parse it and send it to the user. How do I create RAML template on this situation.Will mule APIkit with RAML will work here?
My existing code:
#POST
#Path("/post")
#Consumes(MediaType.APPLICATION_JSON)
#Produces(MediaType.APPLICATION_JSON)
public List<GLExport> postOperation(#Payload String content) throws ParseException {
JSONParser jsonParser = new JSONParser();
Object jsonObjectInstance =null;
jsonObjectInstance = jsonParser.parse(new StringReader(content));
...
return glExportList;
<http:inbound-endpoint exchange-pattern="request-response" host="${hostname}" port="${glport}" path="QBJournalExport/QBGLRest" doc:name="HTTP"/>
<jersey:resources doc:name="REST">
<component class="com.qb.rest.GLExportService"/>
</jersey:resources>
It seems your question is more about JSON Schema than RAML. In essence, you are asking how to support unknown fields in a JSON Schema.
The answer is by adding:
"additionalProperties": true
to your JSON Schema object definition.
I strongly suggest that if there is set of fields are known, you declare them explicitly in your schema: users of your API will thank you for that.
Also, if some of these fields are guaranteed to be present, mark them as required as well.
APIkit should have no trouble dealing with additional properties.

Rally JSON I/O error creating a test case result

I am trying to create a test case result using a REST client, but get this error:
"Errors": ["Cannot parse input stream due to I/O error as JSON document: Parse error: expected '{' but saw '\uFFFF' [ chars read = >>>\uFFFF<<< ]"]
I get the same error when the name of the object, testcaseresult is not specified in the request body. Here are the steps to create a test case result using a browser REST client:
a) Generate the authorize key using "GET" method and the following URL:
https://rally1.rallydev.com/slm/webservice/v2.0/security/authorize
This is the response that I get back, with the security token: "123abc..."
{"OperationResult": {"_rallyAPIMajor": "2", "_rallyAPIMinor": "0", "Errors": [], "Warnings": [], "SecurityToken": "abc123..."}}
b) Use "POST" method, and the following URL:
https://rally1.rallydev.com/slm/webservice/v2.0/testcaseresult/create?key=abc123...
notice the security token in the end.
c) here is an example of a request body:
{
"testcaseresult":
{
"Build":"1",
"Tester":"/user/777",
"Date":"2010-09-04T19:56:05.000Z",
"TestCase":"/testcase/1111",
"Verdict":"Pass"
}
}
Only the required fields and the fields you want to set need to be referenced. Notice the outer key/value pair
{
"testcaseresult":{}
}
The fields that point to a full object, like "Tester" (points to User object) and "TestCase" (points to a TestCase object that owns the result) have to be referenced by their ObjectIDs:
"Tester":"/user/777",
"TestCase":"/testcase/1111",