CPS-461 Align DataNode for Get and Post/Put endpoints in CPS Core

CPS-461 - Getting issue details... STATUS

Introduction

Yang module used for testing

bookstore.yang

module stores {
    yang-version 1.1;
    namespace "org:onap:ccsdk:sample";

    prefix book-store;

    revision "2020-09-15" {
        description
        "Sample Model";
    }

    typedef year {
        type uint16 {
            range "1000..9999";
        }
    }

    container bookstore {

        leaf bookstore-name {
            type string;
        }

    list categories {

        key "code";

        leaf code {
            type string;
        }

        leaf name {
            type string;
        }

        list books {
            key title;

            leaf title {
                type string;
            }
            leaf lang {
                type string;
            }
            leaf-list authors {
                type string;
            }
            leaf pub_year {
                type year;
            }
            leaf price {
                type uint64;
            }
        }
    }
    }
}

It has been noted that the json content for read and write requests in CPS-Core is inconsistent

Description	Response for getDataNode	body when writing (root) DataNode
Operation	GET	POST/PUT
xPath	/	/
Response /Body	`{` `"bookstore-name":"Chapters", "categories":[ ... ] }`	`{ "bookstore":{` `"bookstore-name":"Chapters", "categories":[ ... ] } }`
Notes	The response includes the 'value' of the bookstore node ie. the leaves and children	the body needs to include the whole container object.

Solution

#	Solutions	Issues
1	Obtain the container from the xpath of the queried node. Node xpath will be /bookstore/categories[@code='01'] where bookstore is the container name given in the module info.	This should be reasonably straightforward as the top level node is named after the container name given by the module yang file.
2	Create a query to obtain the schema node directly from the database using and propagate through persistence and service layers.	Would need to create a new query which would parse the json data from the yang_resource content and find the container name which is very memory intensive for its use case. Also would need to pass a new schema node object to the toDataMap method which would not be used every time which it is called.

For the JSON output of Get DataNode we need to alter the DataMapUtils class. The function toDataMap translates a datanode object to a JSON output:

public static Map<String, Object> toDataMap(final DataNode dataNode) {
    final boolean isTopLevelNode = dataNode.getXpath().lastIndexOf('/') == 0;
    if (isTopLevelNode) {
        String containerName = dataNode.getXpath().substring(1);
        return ImmutableMap.<String, Object>builder().put(containerName,
            ImmutableMap.<String, Object>builder()
                .putAll(dataNode.getLeaves())
                .putAll(listElementsAsMap(dataNode.getChildDataNodes()))
                .putAll(containerElementsAsMap(dataNode.getChildDataNodes()))
                .build()
            ).build();
    } else {
        return ImmutableMap.<String, Object>builder()
            .putAll(dataNode.getLeaves())
            .putAll(listElementsAsMap(dataNode.getChildDataNodes()))
            .putAll(containerElementsAsMap(dataNode.getChildDataNodes()))
            .build();
    }
}

We only want the container name on the top level of the JSON output and as such have to distinguish between a top level node and non top level node. We create an outer map to wrap the inner map which creates the appropriate levels in the JSON output.

This produces the following response in the Post Request output:

{
    "bookstore": {
        "bookstore-name": "Easons",
        "categories": [
            {
                "code": "01",
                "name": "SciFi"
            }
        ]
    }
}

Tests would need to be updated to accept the new JSON output which is returned. Examples would need to be updated in openapi