feat(models): add support for generic platform resources (datahub-pro…

…ject#11531)

metadata-models/src/main/pegasus/com/linkedin/common/SerializedValue.pdl

@@ -0,0 +1,42 @@
+namespace com.linkedin.common
+
+/**
+ * Captures the serialized value of a (usually) schema-d blob.
+ */
+record SerializedValue {
+    /**
+     * The serialized blob value.
+     */
+    blob: bytes
+
+    /**
+     * The content-type of the serialized blob value.
+     */
+    contentType: enum SerializedValueContentType {
+        JSON,
+        BINARY
+    } = "JSON"
+
+    /**
+     * The schema type for the schema that models the object that was serialized
+       into the blob.
+     * Absence of this field indicates that the schema is not known.
+     * If the schema is known, the value should be set to the appropriate schema
+     * type.
+     * Use the NONE value if the existing schema categories do not apply.
+     */
+    schemaType: optional enum SerializedValueSchemaType {
+        AVRO
+        PROTOBUF
+        PEGASUS
+        THRIFT
+        JSON
+        NONE
+    }
+
+    /**
+     * An optional reference to the schema that models the object.
+     * e.g., 'com.linkedin.platformresource.slack.SlackConversation'
+     */
+    schemaRef: optional string
+}

metadata-models/src/main/pegasus/com/linkedin/platformresource/PlatformResourceInfo.pdl

@@ -0,0 +1,57 @@
+namespace com.linkedin.platformresource
+
+import com.linkedin.common.SerializedValue
+
+/**
+ * Platform Resource Info.
+ * These entities are for miscelaneous data that is used in non-core parts of the system.
+ * For instance, if we want to persist & retrieve data from auxiliary integrations such as Slack or Microsoft Teams.
+ */
+@Aspect = {
+  "name": "platformResourceInfo"
+}
+record PlatformResourceInfo {
+   /**
+    * The type of the resource. 
+    * Intended as a loose specifier of the generic type of the resource.
+    * Producer is not forced to conform to a specific set of symbols for
+    * resource types.
+    * The @PlatformResourceType enumeration offers a paved path for agreed upon
+    * common terms, but is not required to be followed.
+    * Example values could be: conversation, user, grant, etc.
+    * Resource types are indexed for ease of access. 
+    * e.g. Get me all platform resources of type user for the platform looker
+   */
+   @Searchable = {
+     "fieldType": "KEYWORD"
+   }
+   resourceType: string
+
+  /**
+   * The primary key for this platform resource.
+   * e.g. for a slack member this would be the memberID.
+   * primary keys specified here don't need to include any additional specificity for the
+     dataPlatform
+   * The @PlatformResourceKey is supposed to represent that
+   */
+  @Searchable = {
+      "fieldType": "KEYWORD"
+  }
+  primaryKey: string
+
+  /**
+   * The secondary keys this platform resource can be located by.
+   * I.e., for a slack member this would be email or phone.
+   */
+  @Searchable = {
+      "/*": {
+        "fieldType": "KEYWORD"
+      }
+  }
+  secondaryKeys: optional array[string]
+
+  /**
+   * The serialized value of this platform resource item.
+   */
+  value: optional SerializedValue
+}

metadata-models/src/main/pegasus/com/linkedin/platformresource/PlatformResourceKey.pdl

@@ -0,0 +1,28 @@
+namespace com.linkedin.platformresource
+
+/**
+ * Key for a Platform Resource.
+ * Platform Resources are assets that are not part of the core data model.
+ * They are stored in DataHub primarily to help with application-specific
+ * use-cases that are not sufficiently generalized to move into the core data model.
+ * For instance, if we want to persist & retrieve additional user profile data 
+ * from auxiliary integrations such as Slack or Microsoft Teams for resolving details later.
+ */
+@Aspect = {
+  "name": "platformResourceKey"
+}
+record PlatformResourceKey {
+  /**
+   * A unique id for this entity.
+   * There are no constraints on the format of this id, but most implementations
+   * will choose to use a UUID.
+   * This id should be globally unique for the entire DataHub instance and
+     uniquely identify the resource that is being stored, so most
+     implementations
+   * will combine logical attributes like platform name, platform instance,
+   * platform-specific-id and the resource type to create the unique id.
+   * e.g. slack:slack-instance:slack-user-id:user-info 
+   * or guid(slack, slack-instance, slack-user-id, user-info) etc.
+   */
+  id: string
+}

metadata-models/src/main/pegasus/com/linkedin/platformresource/PlatformResourceType.pdl

@@ -0,0 +1,17 @@
+namespace com.linkedin.platformresource
+
+/**
+* A set of symbols for loose agreements between producers and consumers of
+  platform resources
+  See @PlatformResourceInfo.resourceType for where this can be populated into
+  **/
+enum PlatformResourceType {
+    /**
+     * e.g. a Slack member resource, Looker user resource, etc.
+     */
+    USER_INFO,
+    /**
+     * e.g. a Slack channel
+     */
+    CONVERSATION
+}

metadata-models/src/main/resources/entity-registry.yml

@@ -615,6 +615,17 @@ entities:
     aspects:
       - dataHubConnectionDetails
       - dataPlatformInstance
+  - name: platformResource
+    doc: >-
+      Platform Resources are assets that are unmodeled and stored outside of
+      the core data model. They are stored in DataHub primarily to help with
+      application-specific use-cases that are not sufficiently generalized to move into the core data model.
+    category: core
+    keyAspect: platformResourceKey
+    aspects:
+      - dataPlatformInstance
+      - platformResourceInfo
+      - status
 events:
 plugins:
   aspectPayloadValidators: