HISE Docs
Support

Pin SDK Versions with CRAN and Conda (Tutorial)—DRAFT (this feature is inactive)

Last updated 2026-05-13

At a Glance

At a Glance

HISE now publishes R and Python SDKs as standard packages through CRAN and Conda channels, instead of baking SDK versions into IDE images. This tutorial shows you how to work in an IDE so you’re using a specific SDK version in your environment, and how to reuse that version across IDEs when you need consistent behavior.

When to Use This Feature

Use this process in the following circumstances:

  • Your analysis depends on a specific SDK version or behavior, and you don’t want updates to change your results unexpectedly.

  • You need to reuse the same SDK version across multiple IDEs or zones, so the IDEs behave consistently.

  • You want SDK packages to carry their own dependency rules instead of relying on dependency logic housed in the image.

SDK Packages and Channels

Previously, SDK versions and some dependencies were part of the IDE image: when you created an IDE, you got whatever SDK and libraries were included in the container, with key dependencies handled by the IDE management service. Small changes in SDK versions, IDE images, or base libraries could break your analysis because version and dependency logic were housed in the image, not in your own environment.

Now, however, SDKs are published as packages. R SDKs are released as R packages to CRAN, the standard repository R uses when you call install.packages(). Python and R SDK builds are also published to Conda channels like conda-forge, which Conda searches when you run conda install.
 Each SDK version carries dependency metadata. IDE workflows that read tabular data and HDF5 files, for example, might require pandas ≥ 2.0 and hdf5 ≥ 1.12. CRAN and Conda use dependency metadata to select specific package versions that satisfy all version constraints during install (such as choosing SDK, pandas, and HDF5 builds that all agree on required versions). This moves compatibility and dependency handling into the package ecosystem, where you already manage other research libraries, and out of the IDE image.

NOTE

Lorem ipsum

Instructions

Open an IDE Notebook

  1. Log in to HISE, go to RESEARCH > IDEs, and use the IDE creation flow to create a new IDE or open an existing one.


  2. Choose the Conda (or Pixi) environment you plan to use for this analysis. This environment is where SDK packages and their dependencies will be installed.

If you’re unsure which environment to choose, see existing environment tutorials such as Use the Conda Environment or Use the SDK to Save a Custom Conda Environment for guidance.


  3. After the IDE finishes provisioning, open a Python or R notebook from the IDE UI, or open a terminal if you prefer command-line installs.

Install a specific SDK package version

  1. Depending on whether the SDK is exposed through CRAN or a Conda channel, use the appropriate installer from inside the IDE:



    A. To use an R SDK from CRAN, install the package by name:



     
    install.packages("hise")

    R fetches the SDK package from CRAN into the R library for this environment.



     EXAMPLE
    conda install -c conda-forge r-hise=3.5.9

    Conda pulls the specified SDK version and its declared dependencies into the active environment.

    B. To use a Python SDK from a Conda channel, install the Python SDK as a Conda package:


     
    conda install <CHANNEL>::<PACKAGE>=<VERSION> OR conda install -c
     EXAMPLE
    conda-forge hise-sdk=1.0.0

    This pins the Python SDK version for the environment used by this IDE.

Pin your SDK version

  1. To confirm your SDK version, import the SDK, check its version, and make a note of it.

     
    import hisepy as hp hp.__version__

    This step makes the SDK version explicit in your workflow instead of implicit in the IDE image. Step 5 Reuse and pin the same SDK version in other IDEs. When you need the same SDK version elsewhere:

  2. Create another IDE using the same environment, or export and recreate the environment with the SDK package pinned (for example, via Conda environment export and rebuild).


  3. If you’re starting from a fresh environment, repeat the install step with the same CRAN or Conda package version (for example, r-hise=3.5.9 or hise-sdk=1.0.0).


  4. Optionally, capture the exact environment state (SDK + dependency versions) in an environment file or lockfile, and store it with your project so IDEs created later can reproduce the environment exactly.


Related Resources

Work with Studies (Tutorial)

Use the Upload Files SDK Method (Tutorial)

Use HISE SDK Methods and Get Help in the IDE