• Stars
    star
    111
  • Rank 312,744 (Top 7 %)
  • Language
    C++
  • License
    Apache License 2.0
  • Created over 2 years ago
  • Updated 9 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Unreal Engine - VP Workflow Validation Framework

Supported Versions

UE4.27+: Supported v1.0.1 Only: βœ…
UE5.0: Testing: ❗
UE5.1: Supported v1.1.0+: βœ…
UE5.2: Supported v1.2.0+: βœ…

1. Intro

Validation UI

The Validation Framework is designed as an extendible and customizable framework to host and manage automated validation checks and fixes which can identify and help address problems within specific workflows.

Initially focused on virtual production workflows, and primarily ICVFX, the goal is to help production teams to catch and resolve common issues which can cause unexpected render results, data outputs and represent performance challenges when working with Unreal Engine.

Given that some issues are not obvious and some might not even be visible to the human eye, and with the pressure and complexity of operating on live production and virtual production stages, it is very easy for human error to occur and oversee some of these important settings.

Having a simple toolkit which can be used to run pre-flight checks, acts as a safety net to the operators and the production.

With this in mind, the system can easily be extended to encompass custom workflows, adding custom validations and fixes, allowing teams to check their own custom steps.

The Validation Framework is not solely for use on a stage, it can also be utilized during the prep and content build phases, ensuring issues are handled before moving down the production pipeline ahead of time.

The plugin is shipped with a set of validations for the most commonly required checks and fixes which cause challenges on a regular basis during ICVFX productions. This set of validations, comes as a result of a collaboration between the Virtual Production teams at Netflix and Epic Games, and their collected knowledge and experience of ICVFX productions.

The system can also be integrated into pipelines and additional tooling around CI/CD to generate validation reports.

Contents

  1. Intro
  2. Installation
    2.1 Binaries
    2.2 Source
  3. Quick Start
  4. UI
    4.1 Menu Bar
    4.2 Validation Framework UI
    4.3 Workflow
    4.4 Level Validation
    4.5 Project Validation
    4.6 Refresh
    4.7 Run All Validations
    4.8 Run All Fixes
    4.9 Name
    4.10 Description
    4.11 Run
    4.12 Validation Status
    4.13 Validation Message
    4.14 Fix
    4.15 Fix Description
  5. Validation
    5.1 Workflows
    5.2 Scopes
    5.3 Validation
    5.4 Validation Status
    5.5 Fixes
    5.6 Validation Fix Status
  6. Validation Project Settings
    6.1 Setting Up Project Settings
  7. CI/CD
    7.1 Reports
  8. Extending & Customizing
    8.1 Adding New Validation
    8.2 Adding New Workflows

2. Installation

The Validation Framework comes as a self contained plugin with no external dependencies aside from the Unreal Engine itself.

The plugin consists of a C++ set of validations, utilities, base classes and helpers.

It also contains Content in the form of blueprints and other .uassets.

The plugin can simply be added to the Unreal Project>Plugins folder as either a pre compiled binary or can be compiled from source code.

2.1 Binaries

  1. Download the relevant zip file for the given releases, unreal version and platform. Releases

  2. Unzip the plugin & place the folder inside the Plugins Folder within the Unreal Project

Plugin Folder

  1. Open the unreal project in Unreal and from the plugins menu ensure that the plugin is loaded.

Plugin Load

2.2 Source

  1. Clone the repository and either copy or symlink the ValidationFramework Folder into the UnrealProject>Plugins folder.

  2. Right Click The .uproject and select Generate Visual Studio project files

  3. Once the solution is generated, open the solution and compile the project.

  4. Once compiled, open the unreal project in unreal and from the plugins menu ensure that the plugin is loaded.

3. Quick Start

TODO INSERT LINK TO QUICKSTART VIDEO/S

4. UI

The UI is built entirely via Blueprints as an EditorUtilityWidget.

ValidationFramework>Content>Blueprints>VP_ValidationFramework.uasset

The blueprint contains no specific logic in regards to the validations, all of the logic is available via API, and any logic in the Blueprint is related to updating the UI itself.

4.1 Menu Bar

The main UI can be loaded from the menu item within the Netflix menu.

Validation Menu Bar

4.2 Validation Framework UI

The main UI for manual user interaction with the validation framework.

Validation UI

4.3 Workflow

Drop down box which allows the user to select a specific workflow we want to run validations for.

This filters the validations visible and runnable from the UI.

4.4 Level Validation

Toggle to select all of the Level-based validations for the selected workflow.

This filters the validations visible and runnable from the UI

4.5 Project Validation

Toggle to select all of the Project-based validations for the selected workflow

This filters the validations visible and runnable from the UI

4.6 Refresh

Refresh button acts as an overall reset/refresh of the UI. This clears and resets all the existing states of the validations. As some validations are implemented via Blueprints, these may be registered after the UI and the engine has loaded: We can’t know about Blueprints until they compile, unlike c++ which we know is compiled ahead of time. Refresh ensures any new or updated validations are all populated into the UI.

4.7 Run All Validations

Runs all of the validations for the selected Workflow and Scope, in the current Level and/or Project.

4.8 Run All Fixes

Runs all of the validation fixes for the selected Workflow and Scope, in the current Level and/or Project

4.9 Name

The short hand name of the validation

4.10 Description

A detailed description of what the validation is going to check for and why.

4.11 Run

Run buttons exist for each validation, this runs the validation check for this specific validation.

4.12 Validation Status

A simple graphical display of the validations result, this shows whether a validation was a Success, a Fail or has Warnings that the user should inspect further.

4.13 Validation Message

Description of the outcome from the validation which was run. Descriptions often identify offending actors, objects and their incompatible settings so a user could choose to resolve manually if required.

4.14 Fix

Fix buttons exist for each validation, this runs the validation fix, for the specific validation.

Not all issues can be automatically fixed, however the user will be notified of this from the validation itself.

4.15 Fix Description

Description of what applying the fix will do to the Level or Project.

5. Validations

Validations can be implemented via Blueprints or by C++, they all inherit from a ValidationBase class, which provides a few key features which we outline below.

5.1 Workflows

Validations are tagged/associated with specific workflows. This allows us to filter validations and only execute the ones which are appropriate.

Validations can be associated with multiple workflows.

Currently there are several in-built workflows

  • ICVFX
  • SIMULCAM
  • VAD
  • VR SCOUTING

Custom workflows can be added with ease see Adding New Workflows

5.2 Scopes

Validations can also be tagged/associated with a scope, this is another level of filtering to ensure we are running the correct set of validations.

Validations can only be part of a single scope.

There are 2 scopes Level & Project

5.2.1 Level

Level Scope indicates that the validation will run on the currently open level, (including sublevels).

Any fixes applied will be applied to the level.

NOTE: Changes are not auto saved, a user will still need to save any changes made to the levels for them to persist.

5.2.2 Project

Project Scope indicates that the validation will run on the current unreal project.

These are cross level validations which often relate to the Project Settings themselves.

5.3 Validation

Validations must all implement a check, blueprint validations override this function whilst c++ validations implement it directly.

This is the function/logic which is run when the user requests a validation to be run.

5.4 ValidationStatus

Validations all return a ValidationStatus, this comprises of user information regarding any issues which have been detected, as well as a status indicating the level of severity.

There are 3 status's - Success, Warning & Fail

5.4.1 Success

Validations which pass indicate that no issues where detected and report a success
Fail Icon

5.4.2 Warning

Validations sometimes return warnings. There are some situations which may or may not be valid, however it is not possible to tell without knowledge of the production. These require the user to investigate and decide if this is indeed an issue or not for their production and workflow.
Fail Icon

5.4.3 Fail

Validations return a failure when something has been detected which is known to critically break the workflow. It is important to note that some of these issues may not be visible to the naked eye, and only appear much later in the production workflow.

Users are encourage to either fix manually based on the description, or apply the automated fixes when applicable.
Fail Icon

5.5 Fix

Validations must all implement a fix, blueprint validations override this function whilst c++ validations implement it directly.

This is the function/logic which is run when the user requests a validation fix the detected issues.

This logic should apply any changes to the UObjects or UEProject as appropriate.

Sometimes its not possible to fix things automatically, in which case the fix still needs implementing but this is as simple as returning a ValidationFixStatus with status ManualFix.

5.6 ValidationFixStatus

Fix Status's are returned from the execution of any fix logic. These carry both informative messages which can be relayed back to the user, along with an overall status related to the success.

Status can be Fixed, ManualFix or NotFixed

5.6.1 Fixed

The fix or fixes were applied successfully.

5.6.2 ManualFix

There is no automated fix available or this requires more context relating to the production and requires a user to manually fix.

5.6.3 NotFixed

The fix was unable to be applied or failed and the issue still persists

6. Validation Project Settings

In order to keep the validations themselves flexible and reusable they sometimes need information about the overall workflow.

For instance validations may want to know the frame rate for the production.

Rather than have individual validations for each possible frame rate, we store this value as a project setting from which the validations can then refer too.

Should a different project have different requirements the validations then do not need to change project to project.

Similarly we also set a default OCIO Config which can be used to define the default source and destination color spaces

By setting up the validation project settings, you will greatly improve the accuracy and effectivness of the validations

6.1 Setting Up Project Settings

  1. Create a new blueprint class from the parent class VFProjectSettingsBase. Given these are project related settings, its recommended to keep this within the UnrealProject rather than a Content Plugin. (Unless you use the same settings over and over again)

Create Project Settings

  1. Within the blueprint we set the default inherited values for the project frame rate and the OCIO Config we expect to be using.

Project Settings Defaults

  1. If you are extending the framework with custom validations, as this is all based on blueprints its possible to add custom values to the VFProjectSettings blueprint.

Create Project Settings

  1. Once you are happy with the VFProjectSettings blueprint, compile and save, then these need to be set into the Unreal Project Settings

Create Project Settings

  1. With this set, within the validations we can now use the GetValidationFrameworkProjectSettings node to retrieve them. You can then either cast to the base type to access just the inherited values such as frame rate, or if you are building custom workflows you can cast to your custom settings class and retrieve your custom settings.

Create Project Settings

7. CI/CD

Whilst running validations via UI is advantageous for users during prep and live operation in productions.

It can be also useful to execute these validations as part of a larger pipeline as an automated step. This could be as part of an asset publish, or a CI/CD build pipeline etc.

Whilst this framework does not provide the entire pipeline and framework for establishing the CI/CD and asset pipelines themselves.

It does provide the API building blocks for teams to integrate into their systems.

static bool GenerateValidationReport(const FString LevelPath, const EValidationWorkflow Workflow, const FString ReportPath="");

Available as both blueprint node and c++ function, this api call allows the user or pipeline to specify a level, a workflow, and an output path in which to generate validation reports which can then be consumed by applications outside of unreal engine.

Validations of both Project & Level Scope are executed.

7.1 Reports

Reports are generated both via the API call above, but also via the UI when Run All Validations is executed.

This serializes the results which are often displayed in the UI into reports which are stored along side the project folder. (default, unless report path is specified via the API)

UnrealProject>ValidationReports>NameOfLevel

These are stored as both csv or json format so are easily consumed by other applications, or simply sent back as part of support requests to supervisors or technical help groups. Reports Folder

8. Extending & Customizing

The initial version of the validation framework aimed at providing a framework for virtual production workflows, along with a base set of validations mainly targeting ICVFX to ensure the integrity of the color pipeline and maintaining the image chain as a linear workflow.

There will be many more validations which are needed in future as new issues and or standards are discovered/implemented.

Many new workflows both within, and outside of virtual production field will emerge and teams, products and productions may also have their own set of standards for very bespoke workflows and applications.

With this in mind we ensured that the framework can be extended with both new workflows and new validations by both technical artists and developers alike.

8.1 Adding New Validations

Validations can be added either via C++ or via Blueprints.

Whilst the validations can live anywhere within an unreal project it is recommended to keep them within a custom content plugin.

This allows companies, teams and productions to store and transport their own validations between projects without tying themselves to a single Unreal Project or populating the validation framework with none generic workflows

8.1.1 Blueprints

Not everyone wants to or is able write c++, so we provide the ability to implement your own validations and fixes via blueprints.

  1. Create a new blueprint class from the right click menu in the content browser, as mentioned earlier these can live anywhere in the project, however it is recommended to keep them in a dedicated content plugin so they can be easily shared between projects

Create Blueprint

  1. Select the ValidationBase from the All Classes section to act as the parent class

Base Class Select

  1. Within the blueprint editor, select the class defaults and on the right we set the default values for the validation such as its name, a detailed description of the check and what it does, a detailed description of the fix and what it will do. Also set the workflows that this validation is applicable to, and the scope of the validation.

Set Defaults

  1. There are then two functions which need to be overridden one called Validation and one called Fix. We need to implement both.

Implement Overrides

  1. Within the Validation function we can implement whatever logic we need to determine whether something is valid or not. We need to return a ValidationResult from the function, with the status and any messages that should be presented to the user in either the UI or the reports.

Implement Validation

  1. Similarly within the Fix function we can implement whatever logic we need to apply any fixes that can be done. We need to return a ValidationFixResult from the function which holds the status, and any messages we want to pass back to the user

Implement Fix

  1. With the blueprint finished, save and compile the blueprint before returning to the ValidationFramework UI. Select the workflow and scope your new validation is applicable to and hit refresh. You will now see the new validation appear in the UI with the details supplied to the class defaults.

Refresh UI

  1. We can now hit run on this validation, which currently does not do anything but we can see is functioning as expected as the status in this case returns a Success and our message is displayed in the UI.

Run Validation

8.1.2 C++

Sometimes the logic to run the validation or fix is to complex to be described in a blueprint, more over as the complexity of the validation and fix increases the messier or more complex the blueprint becomes.

Often its just easier to implement a new validation in C++.

  1. Create a new class called CustomValidation which inherits from UValidationBase
#pragma once

#include "CoreMinimal.h"
#include "ValidationBase.h"
#include "CustomValidation.generated.h"

/**
* Example Validation We Want To Implement
*/
UCLASS()
class VALIDATIONFRAMEWORK_API UCustomValidation final : public UValidationBase
{
	GENERATED_BODY()

public:
	UCustomValidation();
	virtual FValidationResult Validation_Implementation() override;
	
	virtual FValidationFixResult Fix_Implementation() override;
};
  1. In the .cpp within the constructor we set the information we want to provide to the user such as name, description, fix description.

We also set which workflows the validation is going to be valid for and what scope it should be used for.

#include "CustomValidation.h"


UCustomValidation::UCustomValidation()
{
	ValidationName = "Custom Validation Name";
	ValidationDescription = "Useful Description Of What Im Checking For And Why";
	FixDescription = "Useful Description Of What Im Going To Do To Fix It";
	ValidationScope = EValidationScope::Project;
	ValidationApplicableWorkflows = {
		EValidationWorkflow::ICVFX
	};
}
  1. We then need to implement the validation, here we implement our logic to detect the problem we are trying to check for. We return a FValidationResult which contains any information we want along with the overall status.
FValidationResult UCustomValidation::Validation_Implementation()
{
    FValidationResult ValidationResult = FValidationResult(EValidationStatus::Pass, "");

    // Insert logic to check for problems
    //
    FValidationResult.Result = EValidationStatus::Fail;
    FValidationResult.Message = "I Deliberately Failed This Validation";
	return ValidationResult;
}
  1. Finally implement any logic which can be done to fix the issue. Care and consideration should be taken to consider if a fix can be done safely and cover all eventualities.
FValidationFixResult UCustomValidation::Fix_Implementation()
{
    FValidationFixResult ValidationFixResult = FValidationFixResult();

    // Insert logic to fix  problems
    //
    ValidationFixResult.Result = EValidationFixStatus::ManualFix;
    ValidationFixResult.Message = "Needs manual fix we don't even know what we are checking for";
    return ValidationFixResult;
}

8.2 Adding New Workflows

There are many workflows in the world, especially within an engine which is as open and flexible as Unreal.

With this in mind, users or teams may wish to extend the framework with their own workflows for their own uses.

This is as simple as extending a single enum and recompiling the plugin.

ValidationFramework/Source/ValidationFramework/Public/ValidationCommon.h

/**
* An enum representing all of the available workflows which validations can belong too.
* Simply extending this enum exposes new options in the workflow selection from the UI, as well as
* allowing validations to register to this new workflow option
*/
UENUM(BlueprintType)
enum class EValidationWorkflow : uint8
{
	ICVFX			UMETA(DisplayName = "ICVFX"),
	VRScouting		UMETA(DisplayName = "VR Scouting"),
	SimulCam		UMETA(DisplayName = "SimulCam"),
	VAD			UMETA(DisplayName = "VAD")
};

Adding an additional enum value such as MyWorkflow is as simple as

/**
* An enum representing all of the available workflows which validations can belong too.
* Simply extending this enum exposes new options in the workflow selection from the UI, as well as
* allowing validations to register to this new workflow option
*/
UENUM(BlueprintType)
enum class EValidationWorkflow : uint8
{
	ICVFX			UMETA(DisplayName = "ICVFX"),
	VRScouting		UMETA(DisplayName = "VR Scouting"),
	SimulCam		UMETA(DisplayName = "SimulCam"),
	VAD			UMETA(DisplayName = "VAD")
	MYWORKFLOW		UMETA(DisplayName = "MyWorkflow")
};

After recompile this option will display in the workflow selection dialog, and will be available to associate validations with.

More Repositories

1

Scumblr

Web framework that allows performing periodic syncs of data sources and performing analysis on the identified results
Ruby
2,643
star
2

stethoscope

Personalized, user-focused recommendations for employee information security.
Python
2,002
star
3

sleepy-puppy

Sleepy Puppy XSS Payload Management Framework
JavaScript
1,029
star
4

sketchy

A task based API for taking screenshots and scraping text from websites.
JavaScript
996
star
5

diffy

β›” (DEPRECATED) Diffy is a triage tool used during cloud-centric security incidents, to help digital forensics and incident response (DFIR) teams quickly identify suspicious hosts on which to focus their response.
Python
632
star
6

riskquant

Python
609
star
7

aardvark

Aardvark is a multi-account AWS IAM Access Advisor API
Python
470
star
8

stethoscope-app

A desktop application that checks security-related settings and makes recommendations for improvements without requiring central device management or automated reporting.
JavaScript
457
star
9

policyuniverse

Parse and Process AWS IAM Policies, Statements, ARNs, and wildcards.
Python
421
star
10

zerotodocker

Dockerfiles to be used to create Dockerhub trusted builds of NetflixOSS
Python
407
star
11

rewrite

Distributed code search and refactoring for Java
Java
291
star
12

gcviz

Garbage Collector Visualization Tool/Framework
Python
266
star
13

repulsive-grizzly

Application Layer DoS Testing Framework
Python
244
star
14

hystrix-dashboard

JavaScript
233
star
15

jvmquake

A JVMTI agent that attaches to your JVM and kills it when things go sideways
Python
154
star
16

zerotocloud

Scripts and instructions for Zero To Cloud With NetflixOSS
Groovy
147
star
17

bpftoolkit

Shell
128
star
18

aws-credential-compromise-detection

Example detection of compromise credentials in AWS
Python
118
star
19

WSPerfLab

Project for testing web-service implementations.
Java
116
star
20

cloudy-kraken

AWS Red Team Orchestration Framework
Python
102
star
21

historical

A serverless, event-driven AWS configuration collection service with configuration versioning.
Python
93
star
22

jenkins-cli

Simple Jenkins Command Line Interface
Perl
91
star
23

swag-client

Cloud multi-account metadata management tool.
Python
87
star
24

cloudtrail-anomaly

Python
82
star
25

cloudaux

Cloud Auxiliary is a python wrapper and orchestration module for interacting with cloud providers
Python
76
star
26

aws-metadata-proxy

AWS Metadata Proxy for protection against SSRF
Go
69
star
27

service-capacity-modeling

Python
61
star
28

titus-isolate

Python
55
star
29

skunky

Marking instances dirty since 2018
Python
47
star
30

raven-python-lambda

Sentry/Raven SDK Integration For AWS Lambda (python) and Serverless
Python
47
star
31

dynaslave-plugin

Jenkins DynaSlave plugin
Java
46
star
32

s3-flash-bootloader

A tool for flashing OS images onto stateful servers
Shell
45
star
33

rl_for_budget_constrained_recs

Jupyter Notebook
41
star
34

logstash-configs

Logstash Configs used by Netflix
31
star
35

spectatord

A high performance metrics daemon
C++
23
star
36

framerate-utils

Useful conversion utilities for working with video frame rate and display
TypeScript
17
star
37

qiro

The Qiro Project
Java
17
star
38

listening-test-app

C++
16
star
39

iep-apps

Example apps using Netflix Insight libraries from the Spectator, Atlas, and IEP projects.
Scala
15
star
40

zerotocloud-gradle

Gradle Plugin to Initialize the Cloud Environment and Utilize it for Continuous Delivery Purposes
Groovy
15
star
41

stethoscope-examples

Example Express application for collecting data from the Stethoscope app
HTML
14
star
42

bucketsnake

An AWS lambda function that grantsss S3 permissionsss at ssscale.
Python
14
star
43

causaltransportr

R package to generalize and transport causal effects.
R
12
star
44

Numerus

Counters, Percentiles, etc for in-memory metrics capture.
Java
12
star
45

mesos-on-pi

Shell
12
star
46

nfflink-connector-iceberg

Java
11
star
47

swag-api

REST API and UI for SWAG data
Python
10
star
48

repokid-extras

Python
10
star
49

raven-sqs-proxy

A Raven/Sentry SQS message proxy forwarder
Python
10
star
50

atlas-node-client

C++
10
star
51

post2crucible

Crucible code review uploader client
Java
8
star
52

StethoscopeMobile

JavaScript
8
star
53

netflixoss-dsl-seed

DSL Scripts to create build jobs for @NetflixOSS projects
Groovy
7
star
54

grails-jade

Grails plugin for rendering Jade templates with the spring-jade4j library
Groovy
6
star
55

spectator-js-nodejsmetrics

Generate node.js internal metrics using the nflx-spectator node module
JavaScript
6
star
56

historical-reports

Lambda functions to generate report artifacts from Historical
Python
6
star
57

cligraphy

Python
5
star
58

atlas-system-agent

Agent that reports system metrics through SpectatorD.
C++
5
star
59

node-pagerduty-netflix

pagerduty REST API interface in node.js
JavaScript
5
star
60

swag-functions

Lambda functions for SWAG management
Python
4
star
61

grails-context-param

Grails plugin to automatically add parameters specified as @ContextParam on a controller to redirect calls.
Groovy
4
star
62

ng-nflx

Miscellaneous utilities for AngularJS
JavaScript
4
star
63

hive2iceberg-migration

Scala
3
star
64

ec2blockdevcfg

Tools and configuration for Amazon EC2 NVMe block devices
Python
2
star
65

kmd

JavaScript
2
star
66

atlas-native-client

C++
2
star
67

qiro-logo

Code for generating the qiro logo
Java
2
star
68

corepipe

Rust
2
star
69

flagpole

Flag arg parser to build out a dictionary with optional keys.
Python
1
star
70

scumblr-spillguard

Python
1
star
71

adversarial_approach_to_recommender_systems

Python
1
star