Effortless Testing Solution for Algorithm Migration

Features

  • Single package definition file
  • Generate pipelines for all MATLAB projects
  • Leverage existing CI/CD platforms
  • Run jobs in parallel
  • Integrated with Nexus, and or Artifactory
  • Report errors, numerical differences and warnings

Standardize on pipelines for regression testing without writing additional unit tests

Why Choose Package Tester?

Upgrading algorithms to new releases can be a daunting task. The unpredictability of effort and the complexities involved often make the process challenging. Package Tester simplifies this journey by offering a seamless solution for dynamic pipeline generation and algorithm migration without having to write unit tests.

Experience the Benefits:

  • Package Tester ensures smooth transitions, eliminating the guesswork involved in predicting the effort required.
  • Dynamic pipelining made easy: once set up, dynamic pipelining becomes a breeze. Harness the power of automation without the headache of complex configurations.
  • Remove the burden from end-users to maintain testing pipelines.
  • Cost-effective solution: why hire additional personnel to build and maintain automation when you can buy a product that does it all? Package Tester saves you time, effort, and money.

Step 1

End-Users Create a Package with:

Package Definition, Source Code, Test Vectors and Dependencies

Step 2

Trigger Package Tester Dynamic Pipeline Generation

Step 3

CI Platform Runs Algorithms in Defined MATLAB Versions in Parallel

Step 4

Compare Results, Warnings, Errors, Performance

Step 5

Predict Migration Effort Based on Detailed Compatibility Report

How Package Tester Works:

  1. Package Algorithm with Test Vectors and Definition:
    • Seamlessly integrate your algorithms with Package Tester using comprehensive test vectors and configurations tailored to your needs, including MATLAB versions.
  2. Run Package Tester:
    • Package Tester is smart enough to only generate jobs in the pipeline for new combinations of algorithms, tests and versions.
    • Leverage your CI platform to run each configuration, conducting rigorous regression tests and providing detailed reports.
    • Results are stored in your artifacts repository.
  3. Compare Results
    • Pull from your artifacts repository previous results
    • Avoiding running the same code with the same test vectors on the same MATLAB versions

What does Package Tester answer?

When moving to a new version of MATLAB:

  • Are deprecated functions called?
  • Are there any new defaults that impact your algorithms?
  • Are there any numerical differences?
  • How many errors are there?
  • What errors are thrown?
  • Does your code run warning-free?
  • Is your algorithm also backward compatible?

What’s included?

Package Tester offers a comprehensive suite of utilities designed to streamline your algorithm migration process:

  • Estimate effort in migrating to a new version by analyzing numerical differences, errors and warnings
  • API Server that triggers Pipeline Generation that only runs what needs to be run.
  • Integration with Nexus or Artifactory for storing input vectors, code, outputs, and results.

Additional Services

Migration

VersionBay has a deep knowledge and experience in migrating code from one software environment to another, namely migrating to a newer version of MATLAB, Python, Julia or even migrating between them.

DevOps

We are a team of seasoned professionals specializing in DevOps practices and principles. Our comprehensive consulting services are designed to help businesses and organizations streamline their software development and deployment processes, enabling faster, more reliable, and efficient delivery of applications.

MATLAB Deployment

We specialize in helping businesses and individuals effectively deploy their MATLAB applications and optimising MATLAB deployment across enterprises. We have helped several companies with the deployment of MATLAB and maintaining internal developed toolboxes.