Developer Guides

• Simple Python Stage
  
  • Background
  • The Pass Through Stage
  • Putting the Stage Together
  • Testing our new Stage
  • Putting the Pipeline Together
    
    • Alternate Morpheus CLI example
• Real-World Application: Phishing Detection
  
  • Data Preprocessing
    
    • The Completed Preprocessing Stage
  • Predicting Fraudulent Emails with Accelerated Machine Learning
    
    • Launching Triton
    • Defining our Pipeline
    • The Completed Pipeline
    • Alternate Morpheus CLI example
  • Stage Constructors
  • Defining a New Source Stage
    
    • The Completed Source Stage
  • Defining a New Sink Stage
    
    • The Completed Sink Stage
  • Note
• Simple C++ Stage
  
  • A Simple Pass Through Stage
    
    • The Complete C++ Stage Header
    • Source Code Definition
    • Python Proxy and Interface
    • The Complete C++ Stage Implementation
    • Python Changes
• Creating a C++ Source Stage
  
  • Header Definition
  • Source Definition
  • A note on performance:
  • Python Proxy & Interface
  • Python Changes
• Digital Fingerprinting (DFP)
  
  • Overview
  • Training Sources
    
    • Defining a New Data Source
  • DFP Examples
    
    • The “Starter” Example
    • The “Production” Example
    • DFP Features
      
      • AWS CloudTrail
      • Azure Active Directory
        
        • Derived Features
      • Duo Authentication
        
        • Derived Features
  • High Level Architecture
    
    • Training Pipeline
    • Inference Pipeline
    • Monitoring
  • Runtime Environment Setup
    
    • Services
    • docker-compose
      • System requirements
      • Building the services
      • Downloading the example datasets
      • Running the services
        
        • Jupyter Server
        • Morpheus Pipeline
        • Output Fields
        • Optional MLflow Service
    • Running via Kubernetes​
      
      • System requirements
  • Customizing DFP
• Digital Fingerprinting (DFP) Reference
  
  • Morpheus Configuration
    
    • Pipeline Structure Configuration
    • Morpheus Config
    • Schema Definition
      
      • DataFrameInputSchema
      • ColumnInfo
      • CustomColumn
      • RenameColumn
      • BoolColumn
      • DateTimeColumn
      • StringJoinColumn
      • StringCatColumn
      • IncrementColumn
    • Input Stages
      
      • MultiFileSource
      • DFPFileBatcherStage
      • DFPFileToDataFrameStage
    • Output Stages
      
      • WriteToFileStage
      • WriteToS3Stage
    • Core Pipeline
      
      • DFPSplitUsersStage
      • DFPRollingWindowStage
      • DFPPreprocessingStage
  • Training Pipeline
    
    • Training Stages
      
      • DFPTraining
      • DFPMLFlowModelWriterStage
  • Inference Pipeline
    
    • Inference Stages
      
      • DFPInferenceStage
      • FilterDetectionsStage
      • DFPPostprocessingStage
• Python Morpheus Modules
  
  • Background
  • The Passthrough Module
    
    • Module Definition and Registration
    • Loading the Module
    • Module Chaining and Nesting
    • Wrapping Modules in Practice
• C++ Morpheus Modules
  
  • Background
  • The Passthrough Module
  • Advanced Topics
    
    • Dynamic Module Creation and Loading
• Morpheus Control Messages
  
  • Background
  • Anatomy of a Control Message
    
    • Working with Tasks
    • Managing Metadata
    • Handling Payloads
• Introduction to Modular Digital Fingerprinting Pipeline (Integrated Training)
  
  • Table of Contents
  • Introduction
    
    • Motivation
    • Overview
  • Setting up Morpheus
  • Morpheus Modules
  • DFP Deployment
    
    • fsspec Dataloader
  • DFP Training and Inference Pipelines
    
    • DFP Preprocessing
    • Control Message Filter
    • File Batcher
    • File to DF DataLoader
    • DFP Split Users
    • DFP Rolling Window
    • DFP Data Prep
  • DFP Training Pipeline
    
    • DFP Training
    • MLFlow Model Writer
  • DFP Inference Pipeline
    
    • DFP Inference
    • Filter Detections
    • DFP Post Processing
    • Serialize
    • Write to File