Version 3.1.4

Added support to make the file upload size configurable

Added support for multi-file uploads

In the Runtimes list page, added support to distinguish between REST and BATCH deployments in the Last Activity Column

Updated the error messages when the ModelOp runtime receives a malformed asset

For the ModelOp Monitoring package, added support to allow non-predictor columns to be monitored for drift

Added support to configure the client-registration-id to the Opaque query param Introspector. This update allows EndUsers to define the {{client-registration-id}} for the QueryParamIntrospector and AuthenticationManager, if added value is empty, then it will use the default one as {{query_param_introspector}}

Added support to be able to do a hot-update of the list of protected endpoints through the Gateway, so that an admin can add/remove endpoints without having to redeploy / restart GW.

Added additional optional access control for BPMN deployments to restrict to certain authorized groups

Added support to automatically distinguish PySpark models from regular python models

Addressed minor issue with the PowerBI plugin to handle Gateway link issues

Upgraded ModelOp Center to use Angular 14

Added a new Custom Metadata UX for adding/editing custom metadata via the UI

Optimized the ModelOp OOTB monitors to be more efficient in terms of memory usage

Added support for the ModelOp runtimes to resolve configuration values from SCCS

Added support to the ModelOp runtime to allow for more configuration tuning via configuration files

Added support to select the Runtime kafka credentials based on producer/consumer and/or topic

Added support to the ModelOp runtime for aws:kms encryption

Allow multiclass classification metrics in the Performance Monitor: Classification Metrics

Updated the MLC diagrams in the Snapshot and MLC pages to provide more detailed information when hovering over a step in the MLC

Updated the Job Details page to make the Model Test Result more prominent

Added a service alert banner if a core ModelOp Center service experiences degredation

Updated deployments label in the Model Snapshot page

Updated the labels in the Monitoring Scheduling tab

Added support for further filtering in the Jobs page

Added UI tabs for Associated Models in the business model and snapshot pages

Added support to include the full Jenkins job error message when a Jenkins job fails in any type of Jenkins service interaction

Added support to generate and verify schemas via a Jupyter notebook

Updated the default "deployment" MLC to automatically add a standard set of monitors to the model snapshot

Added a new Deployment details UI page to allow the user to see details of the deployment

Added criteria to search for associatedModelSnapshotId in the MTRSummary findByOptional Endpoint

Added the capability to add additionalAssets on a DeployedModel, such that a DMN (e.g. Dashboard dmn) can be added to the Deployment

Added a new asset role DASHBOARD_RESULT_COMPARATOR (to be used in deployedModel for Dashboard recognition

Updated the Jobs UI details page to handle jobs that do not contain a model

Added support for external credentials for Gitlab pipeline integration

Updated the Job details page to include more details of Jenkins and Gitlab pipeline job information

Added support to re-run Jenkins or Gitlab Pipeline jobs

Added a GitLab Service to talk to REST API client

Added a GitLab Job MLC Delegate to launch jobs

Added a GitLab Job Monitor to process Gitlab Pipelines

Added support for a GitLab job output 

Added support for Gitlab Job input variables

Created an updated ModelOp runtime image that includes support for loading CSV's directly into an R dataframe

Optimized handling of json input files for metrics jobs to have the python runner read the files directly via a Pandas call

Added support to store the output of Dashboard jobs in external storage (e.g. S3)

Optimized the default bpmns and delegates to use object IDs instead of the fully hydrated objects, allowing for more efficient memory and storage usage for the mlc service

Added support for sending custom variables when triggering an mlc signal (e.g. when sending a scheduler signal)

Created a new MLC to send an email instead of opening a ticket if a Monitoring job fails

Added new delegates for all ModelOp Center core objects that take a PATCH statement

Addressed issue with "next gen" Jira environments where there can be duplicate issue type names with different ids

Added support for mTLS configuration to the ModelOp runtime

Added support for AzureAD for the ModelOp Center Tableau plugin

Added support for AzureAD for the ModelOp Center PowerBI plugin

Added support for AzureAD for the ModelOp Center SageMaker integration

Added support for AzureAD for the ModelOp Center Jupyter integration

Added support for AzureAD for the ModelOp Center Spark integration

Added support for AzureAD for the ModelOp Center CLI

Added support for Okta for the ModelOp Center Tableau plugin

Added support for Okta  for the ModelOp Center PowerBI plugin

Added support to the standard MLC to configure how generic runtime matching should be done when group isolation is not required

For the ModelOp monitoring package, for models where input features are not provided, added ability to to run comprehensive Volumetrics on the score fields to allow for univariate analysis.

Added support for configuring the logging level of the ModelOp Center python packages. If the environment variable {{MODELOP_SDK_ENV_VAR_NAME}} is present then general LOG levels should be adjusted to the  env variable value

Added the ability to create generic bar, table, and line charts from a metrics model

Updated the ModelOp UI tags to trim any whitespace, thus avoiding any issues in matching tags in the MLC

Added support for a downloadable link to S3 assets for authorized users

Updated the UX of Model Test Results when the model has a large number of columns

Added business model name (reference model) in the main Job list page for each of the jobs

Updated the OOTB Stability monitor to be able to run even without a Score column

For the ModelOp monitoring package, added improve error messaging if the Identifier is not specified for the OOTB Volumetrics Comparison monitor

Added support for creating snapshots of SageMaker models directly in the ModelOp Center UI

Added Approval notification type and Approval section of the UI to distinguish specific model approvals throughout a model's life cycle

Added support to the ModelOp Runtime to send an access token when connecting to the web socket in secured mode

Addressed Vulnerability: <[https://nvd.nist.gov/vuln/detail/CVE-2019-17495|https://nvd.nist.gov/vuln/detail/CVE-2019-17495|smart-link] > - Critical - A Cascading Style Sheets (CSS) injection vulnerability in Swagger UI before 3.23.11 allows attackers to use the Relative Path Overwrite (RPO) technique to perform CSS-based input field value exfiltration, such as exfiltration of a CSRF token value

For the ModelOp monitoring package, added a new OOTB Monitoring Model that calculates Data Drift using Wasserstein Distance

Added UI support for custom fields in the Home Dashboard screen

Fixed several styling issues with the charts/graphs in UI dark mode

Updated the MLC tab of a Snapshot to sort the MLC's by start time

Addressed a minor UI issue with the highlighting of sub-menu items

Updated the Dashboard population approach to read the most recent model test results for a given model. This will allow more flexibility and improved performance for the Home Dashboard population

Updated the Spark runtime service to support AWS Cognito

Updated AWS Cognito support for the "refresh_token" mechanism

Improved end user error messages for non source code files when failing with a git import issue

Updated the Jobs UI page to allow for canceling Jenkins and Gitlab Pipeline jobs

Created an updated ModelOp runtime image that includes support for R-4.2.1

Updated the Compliance report to include the Approvers for all Approvals

Added support for AzureAD to feth the group-name in addition to the GroupID

Updated the Oauth2 implemenation process to create default generic OAuth2 clients for existing idPs, following these providers definitions:

• gateway-service

• internal-client

• go-cli

• external-integration-client

Added support for PingFederate integration to dynamically extract the user values from LDAP

For the ModelOp monitoring package, added a new OOTB monitor for calculting Linearity metrics via Box-Tidwell

Updated SparkSQL Support, including:
(1) centralizing the Spark configurations
(2) adding additional Spark Job options to the Job creation screen
(3) ability to write Spark job outputs as embedded assets in ModelOp Center
(4) ability to configure authorization options for Cloudera clusters
(5) limiting the log size for a Spark-submit job to 10MB
(6) updating the error messaging if a the HADOOP_CON_DIR is not present

Added the ModelOp runtime image name in the Platform information tab of the Runtime details UI page

Formerly added support to the ModelOp runtime for reading/writing data sets from Redshift 

Added /actuator/refresh endpoint to runtimes to support reloading configuration from SCCS without having to restart the runtime.

Updated how the MLC history is being managed to optimize overall memory and storage usage of the MLC service

Added support for connecting to RDS through an IAM DB Auth Token