This directory contains examples of Jupyter notebooks, Python code, and R code that can be used to perform the following SAS Model Manager tasks:
- Overview
- Calculate Fit Statistics, ROC, and Lift, and then Generate JSON Files
- Get the Number of Published Models
- Build and Import a Trained Python Model
- Fit a Scoring Script for Python Model Containerization
- Fit a Scoring Script for R Model Containerization
- Upload Data to the SAS Cloud Analytics Services (CAS)
- Delete Model Content Logs
- Score Open-Source Models with CAS in SAS Viya 3.5
- Score Python containers at runtime
- Create container images in an OpenShift environment
- Additional Resources
Note: Contributions from users other than the SAS Model Manager support team can be added to the /Python/external-examples and /R/external-examples subdirectories.
When you compare models, the model comparison output includes model properties, user-defined properties, and variables. The model comparison output might also include fit statistics, and lift and ROC plots for the models if the required model files are available. The fit statistics, as well as plots for lift and ROC, can be produced using Python packages that generate JSON files. These JSON files are used to show the fit statistics and plots when comparing models in SAS Model Manager.
To calculate fit statistics, ROC, and lift information when you build and import a trained Python model, see ImportPythonModel.ipynb in the examples directory.
You can return the number of projects with published models and the total number of published models by destination type.
To return a count for published models, see GetPublishedModelsCount.ipynb in the examples directory.
A Python model can be built and trained before importing the model in SAS Model Manager as a ZIP file. The ZIP file contains model files that are associated with a specific model and stored within the ZIP file. The ZIP file can contain model folders at the same level or in a hierarchical folder structure. Each model folder within the ZIP file is imported as a separate model object that contains the contents of the model folder. When you import models from a ZIP file into a project version, the hierarchical folder structure is ignored.
To build and import a trained Python model, see ImportPythonModel.ipynb in the examples directory.
The web services in the container store the input CSV file and pass the input file name to your scoring script. Web services will search for the first script whose file name ends with 'score.py' for a Python model. The scoring script reads the input data from an input CSV file and then stores the output data in the CSV file. The scoring script must follow the below pattern:
- The pickle file must be specified in the command-line arguments to be read by the script
- Input variables must be in the inputVar.json file, and the output variables in the outputVar.json file
To fit a scoring script for a Python model containerization, see PythonModel_ScoreFileTutorial.md in the examples directory .
The web services in the container store the input CSV file and pass the input file name to your scoring script. Web services will search for the first script whose file name ends with 'score.R' for an R model. The scoring script must follow the below pattern:
- Default R score code that helps score the R model with an RDA model file
- The RDA model file must be specified in the command-line arguments to be read by the script
To fit a scoring script for an R model containerization, see RModel_ScoreFileTutorial.md in the examples directory.
The SAS Scripting Wrapper for Analytics Transfer (SWAT) package can be used to upload local data to the CAS server.
To upload data to CAS, see UploadDataToCAS.ipynb in the examples directory.
When you publish a model to a destination, there are log and SAS code files that are generated within the contents of a model object. You can delete the files for a specific model, all models within a project, or all models within the common model repository. The model content is only deleted for the following file types: ScoreCodeGen{}.sas or ScoreCodeGen{}.log. The most recent revision of each file type is not deleted.
To delete model content logs, see DeleteModelContentLogs.ipynb in the examples directory.
SAS Model Manager in SAS Viya 3.5 provides the ability to score Python and R models within CAS. For Python models, the user must modify their Python model files. However, for R models, a user must modify their model files and verify details about the SAS Viya server.
To score open-source models within CAS, see ScoringOpenSourceModelsWithCAS.md in the examples directory.
SAS Model Manager can deploy Python models into a Python container. To score data using a Python container deployment, see PythonContainerRuntimeScoring.ipynb in the examples directory.
SAS Viya platform uses Kaniko to create container images for models and decisions when they are published to a container publishing destination. Because Kaniko requires root access to run, container images cannot be created when the SAS Viya platform is deployed in a Red Hat OpenShift environment. In order to create container images within an OpenShift environment you must instead publish your models and decisions to a Git destination and then use Docker to create the container images.
To create container images in an OpenShift Environment, see OpenShiftContainerImages.md
This project is licensed under the Apache 2.0 License.