- 1. Introduction
- 2. Installation
- 3. Uninstallation
- 4. Clara Administration
- 5. Core Concepts
- 6. How to Run a Reference Pipeline
- 6.1. Select a Reference Pipeline
- 6.2. Run Reference Pipelines by Sending DICOM Instances Over Network
- 6.2.1. External DICOM Sender and DICOM Receiver
- 6.2.2. Customize the Pipeline
- 6.2.3. Create the Pipeline
- 6.2.4. Configure the Clara DICOM Adapter
- 6.2.5. Trigger Pipeline
- 6.2.6. Check Job Status and Download/View Payloads in Clara Console/Dashboard
- 6.2.7. Verify That the External DICOM Receiver Received Your Images
- 6.2.1. External DICOM Sender and DICOM Receiver
- 6.3. Run Reference Pipelines using Local Input Files
- 7. Clara Pipeline Driver API
- 7.1. Clara Pipeline Driver Overview
- 7.2. Clara Pipeline Callbacks
- 7.3. Clara Pipeline Driver
- 7.4. NVIDIA Clara Pipeline Driver Guidance
- 7.5. Clara Pipeline Fast I/O Allocations
- 7.6. Clara Pipeline Fast I/O Context
- 7.7. Clara Pipeline Fast I/O Entries
- 7.7.1. Output Entries
- 7.7.2. Input Entries
- 7.7.3. FastIO Entry Automatic and Manual Allocation
- 7.7.4. FastIO Entry I/O Access
- 7.7.5. Name
- 7.7.6. Shape
- 7.7.7. Element Type
- 7.7.8. Size in Bytes
- 7.7.9. Dynamic Indices
- 7.7.10. Is Dynamic
- 7.7.11. Update Shape
- 7.7.12. Access
- 7.7.13. Allocate
- 7.7.14. Is Allocated
- 7.7.15. Map Allocation
- 7.7.16. Unmap Allocation
- 7.8. Clara Pipeline Payload
- 7.9. Clara Pipeline Payload Entry
- 7.10. Clara Pipeline Driver Callbacks
- 7.11. Clara Pipeline Fast I/O Allocations
- 7.12. Clara Pipeline Fast I/O Context
- 7.13. Clara Pipeline FastIO Entry
- 7.13.1. Pre-allocated FastIO Entries
- 7.13.2. Dynamic-size FastIO Entries
- 7.13.3. Input- and Output-Type FastIO Entries
- 7.13.3.1. FastIO Entry Accessors and Mutators
- 7.13.3.1.1. Name
- 7.13.3.1.2. Is Pre-Allocated
- 7.13.3.1.3. Access
- 7.13.3.1.4. Access
- 7.13.3.1.5. Shape
- 7.13.3.1.6. Element Type
- 7.13.3.1.7. Dynamic Indices of the Shape
- 7.13.3.1.8. Is Dynamic
- 7.13.3.1.9. Update Shape
- 7.13.3.1.10. Get Values of Allocation as Read-Only Bytes
- 7.13.3.1.11. Get Values of Allocation as Numpy Array
- 7.13.3.1.12. Set the Values in the Allocation
- 7.13.3.1. FastIO Entry Accessors and Mutators
- 7.14. Pipeline Stage Payload
- 7.15. Pipeline Payload Entry
- 7.16. Clara Pipeline Callbacks
- 7.17. Clara Pipeline Driver
- 7.18. Clara Pipeline Payload
- 7.19. Clara Pipeline Payload Entry
- 7.1. Clara Pipeline Driver Overview
- 8. Pipeline Definition Language
- 8.1. Clara Pipeline Definition Language Overview
- 8.2. Clara Deploy SDK Pipeline Glossary
- 8.3. Operators
api-version
0.7.0
api-version
0.5.0
- 8.6. The Anatomy of an Operator
- 8.7. Argo vs. Clara Orchestration
- 8.8. Pipelines
- 8.9. Services
- 8.10. Connections
- 8.11. Container
- 8.12. Document
- 8.13. Import
- 8.14. Input
- 8.15. Operator
- 8.16. Output
- 8.17. Pipeline
- 8.18. Requests
- 8.19. Service
- 8.20. Syntax & Grammar
- 8.21. Template
- 8.22. Clara Structures and Primitives
- 8.23. Variables
- 8.1. Clara Pipeline Definition Language Overview
- 9. Bundled Operators
- 9.1. Clara DICOM Reader
- 9.2. Clara DICOM Writer
- 9.3. Clara Deploy DICOM Parser Operator
- 9.4. Clara Deploy Base Inference Operator V2
- 9.5. Clara AI Liver Tumor Segmentation Operator
- 9.6. Clara Deploy SDK VNet Segmentation Operator
- 9.7. Clara AI Brain Tumor Segmentation Operator
- 9.8. Clara AI Hippocampus Segmentation Operator
- 9.9. Clara AI Spleen Segmentation
- 9.10. Malaria Microscopy Classification Operator
- 9.11. Clara AI Lung Tumor Segmentation Operator
- 9.12. Colon Tumor Segmentation Operator
- 9.13. Clara AI Pancreas Tumor Segmentation
- 9.14. Chest X-ray Classification Operator
- 9.15. Breast Microscopy Classification Operator
- 9.16. Clara AI Prostate Segmentation
- 9.17. CT Recon Operator
- 9.18. Clara Register Results
- 9.19. Clara Genomics Analysis
- 9.20. FastIO-ITK Operator
- 9.21. DICOM RTSTRUCT Writer Operator
- 9.22. Clara FastIO Variable Passthrough - Nifti to FastIO Variable Conveter Operator
- 9.23. Clara FastIO Variable Passthrough - Fastio to Npz Converter Operator
- 9.24. Clara FastIO Variable Passthrough - Compare Operator
- 9.25. DeepStream App Operator
- 9.26. Digital Pathology Image Processing Operator
- 9.26.1. Overview
- 9.26.2. Commands
- 9.27. Clara Deploy AI Lung Segmentation Operator
- 9.28. Clara Deploy AI COVID-19 Classification Operator
- 9.29. Clara Deploy AI COVID-19 Lesion Segmentation Operator
- 9.30. Clara Deploy Series Selector
- 9.31. Copy customized rules file overwriting the default one.
- 9.32. Digital Pathology Nuclei Segmentation Operator
- 9.33. Clara Deploy DICOM Segmentation Writer Operator
- 9.34. Clara Deploy DICOM Report Object Writer
- 9.1. Clara DICOM Reader
- 10. User Defined Operators
- 10.1. Clara Deploy Operator Development Guide
- 10.2. Developing Clara Operators Using the Operator Development Kit
- 10.2.1. Overview of the Operator Development Kit
- 10.2.2. Package Contents
- 10.2.3. Run the Example Liver Segmentation Model Locally
- 10.2.4. Customizing the Operator Development Kit Example Code
- 10.2.5. Deploy an Existing Container as a Clara Operator
- 10.2.6. Deploying a Single-Operator Pipeline to Existing Clara Cluster
- 10.1. Clara Deploy Operator Development Guide
- 11. Bundled Pipelines
- 11.1. Liver Segmentation Pipeline
- 11.2. CT Reconstruction Pipeline
- 11.3. Multi AI Pipeline
- 11.4. Vnet Segmentation pipeline
- 11.5. Clara AI Brain Tumor Segmentation Pipeline
- 11.6. Hippocampus Segmentation Pipeline
- 11.7. Spleen Segmentation Pipeline
- 11.8. Malaria Microscopy Classification Pipeline
- 11.9. Lung Tumor Segmentation Pipeline
- 11.10. Colon Tumor Segmentation Pipeline
- 11.11. Pancreas Tumor Segmentation Pipeline
- 11.12. Chest X-ray Classification Pipeline
- 11.13. Breast Pathology Image Classification Pipeline
- 11.14. Clara AI Prostate Segmentation Pipeline
- 11.15. De Novo Sequence Assembly - Clara Genomics Analysis
- 11.16. 3D Image Processing Pipeline using FastIO
- 11.17. FastIO Passthrough Pipeline Using FastIO Variable API
- 11.18. Suggested Reading
- 11.19. DeepStream Batch Pipeline
- 11.20. Digital Pathology Image Processing Pipeline
- 11.21. Clara Deploy AI COVID-19 Classification Pipeline
- 11.22. Clara Deploy DICOM Series Selection Pipeline
- 11.23. Digital Pathology Nuclei Segmentation Pipeline
- 11.1. Liver Segmentation Pipeline
- 12. User Defined Pipelines
- 12.1. Build Custom Operators
- 12.2. Place AI Model Files in the Right Directory
- 12.3. Start With a Reference Pipeline
- 12.4. Modify the Pipeline
- 12.5. External DICOM Sender and DICOM Receiver
- 12.6. Create the Pipeline
- 12.7. Configure the Clara DICOM Destination
- 12.8. Trigger Pipeline
- 12.9. Verify the Pipeline Execution in the Dashboard
- 12.10. Verify the External DICOM Receiver Received the Images
- 13. Bundled Services
- 13.1. Clara Results Service
- 13.1.1. Cleanup Service
- 13.1.2. Task States
- 13.1.3. Definitions
- 13.1.4. Requirements
- 13.1.5. Running the Results Service
- 13.1.6. API Usage
- 13.1.7. Database
- 13.1.8. License
- 13.2. Clara Dashboard Render Service
- 13.3. Clara Management Console
- 13.1. Clara Results Service
- 14. Connectivity With a PACS/DICOM Server
- 15. Clara Deploy GPU Profiler
- 16. Clara CLI
- 17. Load Generator
- 18. Platform API
- 18.1. Clara Deploy SDK Platform API Overview
- 18.1.1. Concepts
- 18.1.2. Clara Service
- 18.1.3. Jobs Service
- 18.1.4. Models Service
- 18.1.5. Payloads Service
- 18.1.6. Pipelines Service
- 18.1.7. Create a New Pipeline
- 18.1.8. Get a List of Existing Pipelines
- 18.1.9. Create and Start a Job
- 18.1.10. Get Status of a Job
- 18.1.11. Get a File Listing from a Payload
- 18.1.12. Download a File from a Payload
- 18.1.13. Introduction to GRPC and Protobuf
- 18.1.14. GRPC Examples
- 18.1.1. Concepts
- 18.2. GpuUtilization
- 18.3. ProcessDetails
- 18.4. Clara Stop RPC
- 18.5. ClaraStopRequest
- 18.6. ClaraStopResponse
- 18.7. Clara Stop RPC
- 18.8. ClaraUtilizationRequest
- 18.9. ClaraUtilizationResponse
- 18.10. Clara Version RPC
- 18.11. ClaraVersionRequest
- 18.12. ClaraVersionResponse
- 18.13. Pipelines Create RPC
- 18.14. PipelinesCreateRequest
- 18.15. PipelinesCreateResponse
- 18.16. Pipelines Details RPC
- 18.17. PipelinesDetailsRequest
- 18.18. PipelinesDetailsResponse
- 18.19. Pipelines List RPC
- 18.20. PipelinesListRequest
- 18.21. PipelinesListResponse
- 18.22. PipelineDetails
- 18.23. Pipelines Remove RPC
- 18.24. PipelinesRemoveRequest
- 18.25. PipelinesRemoveResponse
- 18.26. Pipelines Update RPC
- 18.27. PipelinesUpdateRequest
- 18.28. PipelinesUpdateResponse
- 18.29. PipelineDefinitionFile
- 18.30. Jobs Cancel RPC
- 18.31. JobsCancelRequest
- 18.32. JobsCancelResponse
- 18.33. Jobs Create RPC
- 18.34. JobsCreateRequest
- 18.35. JobsCreateResponse
- 18.36. Jobs List RPC
- 18.37. JobsListRequest
- 18.38. JobsListResponse
- 18.39. Jobs ReadLogs RPC
- 18.40. JobsReadLogsRequest
- 18.41. JobsReadLogsResponse
- 18.42. Jobs Start RPC
- 18.43. JobsStartRequest
- 18.44. JobsStartResponse
- 18.45. NamedValue
- 18.46. Jobs Status RPC
- 18.47. JobsStatusRequest
- 18.48. JobsStatusResponse
- 18.49. JobsDetails
- 18.50. JobFilter
- 18.51. JobOperatorDetails
- 18.52. JobOperatorStatus
- 18.53. JobPriority
- 18.54. JobState
- 18.55. JobStatus
- 18.56. Payloads Create RPC
- 18.57. PayloadsCreateRequest
- 18.58. Payloads Details RPC
- 18.59. PayloadsDetailsRequest
- 18.60. PayloadsDetailsResponse
- 18.61. Payloads Download RPC
- 18.62. PayloadsDownloadRequest
- 18.63. PayloadsDownloadResponse
- 18.64. Payloads Create RPC
- 18.65. PayloadsCreateRequest
- 18.66. Payloads Upload RPC
- 18.67. PayloadsUploadRequest
- 18.68. PayloadsUploadResponse
- 18.69. PayloadFileDetails Message
- 18.70. PayloadType
- 18.71. RPC Error Codes
- 18.72. Identifier
- 18.73. RequestHeader
- 18.74. ResponseHeader
- 18.75. Timestamp
- 18.76. Version
- 18.1. Clara Deploy SDK Platform API Overview
- 19. Third Party Tools
- 20. Release Notes
- 20.1. Version 0.7.4 (R7)
- 20.1.1. Introduction
- 20.1.1.1. What’s New
- 20.1.1.1.1. Clara GPU Profiler
- 20.1.1.1.2. Render Server - displaying orientation labels to slice view
- 20.1.1.1.3. Jupyter notebook based interactive rendering
- 20.1.1.1.4. Clara Management Console: Ability to view datasets which were not pre-registered
- 20.1.1.1.5. DICOM Adapter Open Source
- 20.1.1.1.6. Surface Mesh in STL Format in DICOM RTSTRUCT Instance
- 20.1.1.1. What’s New
- 20.1.2. Issues Fixed and Enhancements
- 20.1.3. Breaking Changes
- 20.1.4. Known Issues
- 20.1.4.1. Using different environment variables for the same pipeline service can make Clara Platform API Server confusing
- 20.1.4.2. Clara Management Console - Log Viewer may have problems visualizing very large logs
- 20.1.4.3. Clara Management Console - Jobs View filters don’t get reset on navigating to the other view
- 20.1.4.4. Multi AI Pipeline with shared memory
- 20.1.4.5. Digital Pathology Nuclei Segmentation Pipeline doesn’t work on A100
- 20.1.1. Introduction
- 20.2. Version 0.7.3 (R7)
- 20.2.1. Introduction
- 20.2.1.1. What’s New
- 20.2.1.1.1. Support for A100 platform
- 20.2.1.1.2. DICOM Adapter: Enable plug-in development for grouping data and triggering
- 20.2.1.1.3. AI Base Operator V2 Updated
- 20.2.1.1.4. COVID-19 Pipeline Updated with new AI operators
- 20.2.1.1.5. Render Server - Video Stream Encoding using Software Mode
- 20.2.1.1.6. GPU Profiling Tool
- 20.2.1.1.7. Clara Management Console - Job DAG View
- 20.2.1.1.8. Clara Management Console - Basic Authentication
- 20.2.1.1.9. Ansible Scripts
- 20.2.1.1.10. Platform Server: Enable metadata for jobs, models, pipelines and reusable payloads.
- 20.2.1.1.11. Platform Server - Clara::Utilization API
- 20.2.1.1. What’s New
- 20.2.2. Issues Fixed and Enhancements
- 20.2.3. Breaking Changes
- 20.2.4. Known Issues
- 20.2.1. Introduction
- 20.3. Version 0.7.2 (R7)
- 20.3.1. Introduction
- 20.3.1.1. What’s New
- 20.3.1.1.1. DICOM Adapter Now Supports Concurrent Associations
- 20.3.1.1.2. Exporting DICOM SR Object
- 20.3.1.1.3. COVID-19 Pipeline Support for DICOM Segmentation and Encapsulated PDF Output
- 20.3.1.1.4. Viewing payloads from Clara Console
- 20.3.1.1.5. Viewing Logs for Operators
- 20.3.1.1.6. GPU Profiling Tool
- 20.3.1.1.7. Model Analyzer
- 20.3.1.1. What’s New
- 20.3.2. Issues Fixed and Enhancements
- 20.3.3. Breaking Changes
- 20.3.4. Known Issues
- 20.3.1. Introduction
- 20.4. Version 0.7.1 (R7)
- 20.4.1. Introduction
- 20.4.1.1. What’s New
- 20.4.1.1.1. Segmentation Algorithm: Automatic Bone Segmentation for Multi-AI Pipeline
- 20.4.1.1.2. Reference Operator: Automatic detection of Lung ROI for Multi-AI Pipeline
- 20.4.1.1.3. Reference Pipeline: Nuclei Segmentation Pipeline for Digital Pathology
- 20.4.1.1.4. Reference Operator: DICOM Segmentation Object Writer
- 20.4.1.1.5. Clara Management Console: Downloading Payloads
- 20.4.1.1.6. Render Server: Shared Memory for Loading Datasets
- 20.4.1.1.7. Python Client
- 20.4.1.1.8. Platform Server Resource Management
- 20.4.1.1.9. Platform Server Job Operator Logs API
- 20.4.1.1. What’s New
- 20.4.2. Issues Fixed and Enhancements
- 20.4.3. Breaking Changes
- 20.4.4. Known Issues
- 20.4.4.1. Render Server fails to start on A100 GPU
nvidia-smi
- 20.4.4.3. Register results with multi AI + shared memory
- 20.4.4.4. Pipeline jobs which require more GPU resources than are available in the cluster are unable to be executed
- 20.4.4.5. Clara Orchestrator pipeline operator run times are incorrectly reported
- 20.4.4.6. Job status for jobs using Argo orchestrator is incorrectly reported
- 20.4.4.7. Job status for jobs using Clara Orchestrator without CPDriver is incorrectly reported
- 20.4.4.8. Failure to cancel downloading the operators/jobs payload
- 20.4.1. Introduction
- 20.5. Version 0.6.1 (R6)
- 20.5.1. Introduction
- 20.5.2. What’s New
- 20.5.2.1. Reference Pipeline: Implementation Update for the Digital Pathology Pipeline
- 20.5.2.2. Reference Pipeline: Implementation Update for the Multi-AI Pipeline
- 20.5.2.3. Clara Deploy GPU Profiling Tool
- 20.5.2.4. Reference Operator: New Base AI Inference Operator
- 20.5.2.5. Render Server: Area Measurement
- 20.5.2.6. Clara Console: Details for a Job
- 20.5.2.7. DICOM Adapter CLI
- 20.5.2.8. Clara Platform Server: Pod Cleaners and Job Controllers
- 20.5.3. Issues Fixed and Enhancements
- 20.5.4. Breaking Changes
- 20.6. Version 0.6.0 (R6)
- 20.6.1. Introduction
- 20.6.2. What’s New
- 20.6.2.1. Reference Pipeline: Detection of COVID-19 in CT datasets
- 20.6.2.2. Reference Pipeline: Usage of Shared Memory in Multi-AI CT Pipeline
- 20.6.2.3. Reference Operator: DICOM Parser Operator
- 20.6.2.4. Reference Pipeline: Digital Pathology
- 20.6.2.5. Reference Pipeline: DICOM Series Selection Pipeline
- 20.6.2.6. Render Server: Visualization of Multi-Resolution Data
- 20.6.2.7. Render Server: Supporting Color Data Type
- 20.6.2.8. Render Server: Static Scale
- 20.6.2.9. Render Server: Distance Measurement
- 20.6.2.10. Render Server: Picture-in-Picture
- 20.6.2.11. Configuring DICOM Adapter via REST APIs
- 20.6.2.12. Helm Upgrade
- 20.6.3. Breaking Changes
- 20.6.4. Known Issues
- 20.6.4.1. Jobs API Reports Incorrect Job State
- 20.6.4.2. Pipeline Services Conflict with Resource Manager
- 20.6.4.3. Deploying Triton via Pipeline Services Consumes All GPUs
- 20.6.4.4. Triton Inference Server Consumes Memory of Unassigned GPU(s)
- 20.6.4.5. Deleted Jobs Listed and/or Deleting Jobs with CLI Can Error
- 20.6.4.6. Corrupted Tiles in the Output of the Digital Pathology Image Processing Pipeline
- 20.7. Version 0.5.0 (R5)
- 20.7.1. Introduction
- 20.7.2. What’s New
- 20.7.2.1. Platform
- 20.7.2.2. Application
- 20.7.2.2.1. Fast I/O Integrated with Clara Platform Driver
- 20.7.2.2.2. Histopathology Pipeline
- 20.7.2.2.3. Prostate Segmentation Pipeline
- 20.7.2.2.4. Multi-AI Pipeline
- 20.7.2.2.5. De Novo Assembly Pipeline
- 20.7.2.2.6. 3D Cropping Pipeline using Shared Memory
- 20.7.2.2.7. Deep Stream Batch Pipeline
- 20.7.2.2.8. DICOM Parser Operator
- 20.7.2.3. Render Server
- 20.7.2.4. Management Console
- 20.7.2.1. Platform
- 20.7.3. Known Issues & Workaround
- 20.7.4. Deprecation Notice
- 20.8. Version 0.4.1 (hotfix)
- 20.9. Version 0.4.0 (R4)
- 20.9.1. Introduction
- 20.9.2. What’s New
- 20.9.2.1. Clara Pipeline Driver
- 20.9.2.2. Orchestration in two modes
- 20.9.2.3. File Adapter
- 20.9.2.4. CLI Improvements
- 20.9.2.5. Centralized Logging
- 20.9.2.6. Monitoring Performance
- 20.9.2.7. Shared Memory Context
- 20.9.2.8. Optimized CT Recon Pipeline
- 20.9.2.9. Optimized Liver Tumor Segmentation Pipelines
- 20.9.2.10. Render Server Improvements
- 20.9.2.11. Automatic Payload Cleanup
- 20.9.2.12. New Pipelines
- 20.10. Version 0.2.0 (R2)
- 20.11. Version 0.1.7 (Deprecated)
- 20.12. Version 0.1.6 (Deprecated)
- 20.13. Pre-alpha (Deprecated)
- 20.14. Known Issues
- 20.14.1. Installation
- 20.14.2. Installation error due insufficient Disk Space to Deploy Clara Container Images
- 20.14.3. Operators cannot access the internet or Clara Service IPs due to the institution-wide proxy server settings
- 20.14.4. Recon
- 20.14.5. Render Service
- 20.14.5.1. Minimum opacity value of the transfer function editor cannot be changed (since 0.4.0)
- 20.14.5.2. Intensity Range Selectors in Transfer Function Editor are not displayed properly (since 0.4.0)
- 20.14.5.3. White empty viewport
- 20.14.5.4. Session management
- 20.14.5.5. Changing datasets results in “Unable to reach RenderServer”
- 20.14.1. Installation
- 20.15. Support
- 20.1. Version 0.7.4 (R7)
- 21. License