shepherd-core 2026.2.1

Programming- and CLI-Interface for the h5-dataformat of the Shepherd-Testbed

Core Library

Main Documentation: https://nes-lab.github.io/shepherd

Source Code: https://github.com/nes-lab/shepherd-tools

Main Project: https://github.com/nes-lab/shepherd

---

shepherd-core is designed as a library and bundles data-models and file-access-routines for the shepherd-testbed, that are used by several codebases.

For postprocessing shepherds .h5-files usage of shepherd_data is recommended.

Features

• read and write shepherds hdf5-files
• create, read, write and convert experiments for the testbed
  • all required data-models are included
• simulate the virtual source, including virtual harvesters (and virtual converter as a whole)
• connect and query the testbed via a webclient (TestbedClient in alpha-stage)
  • offline usage defaults to static demo-fixtures loaded from yaml-files in the model-directories
• work with target-firmwares
  • embed, modify, verify, convert
  • Note: working with ELF-files requires external dependencies, see Installation-Chapter
• decode waveforms (gpio-state & timestamp) to UART
• create an inventory (for deployed versions of software, hardware)

See official documentation or example scripts for more details and usage. Most functionality is showcased in both. The extra-directory holds data-generators relevant for the testbed. Notably is a trafficbench-experiment that's used to derive the link-matrix of the testbed-nodes.

Config-Models in Detail

These pydantic data-models are used throughout all shepherd interfaces. Users can create an experiment, include their own content and feed it to the testbed.

• orchestration /data-models with focus on remote shepherd-testbed
• classes of sub-models
  • /base: base-classes, configuration and -functionality for all models
  • /testbed: meta-data representation of all testbed-components
  • /content: reusable user-defined meta-data for fw, h5 and vsrc-definitions
  • /experiment: configuration-models including sub-systems
  • /task: digestible configs for shepherd-herd or -sheep
  • behavior controlled by ShpModel and content-model
• a basic database is available as fixtures through a tb_client
  • fixtures selectable by name & ID
  • fixtures support inheritance
• the models support
  • auto-completion with neutral / sensible values
  • complex and custom datatypes (i.e. PositiveInt, lists-checks on length)
  • checking of inputs and type-casting
  • generate their own schema (for web-forms)
  • pre-validation
  • store to & load from yaml with typecheck through wrapper
  • documentation
• experiment-definition is designed securely
  • types are limited in size (str)
  • exposes no internal paths
• experiments can be transformed to task-sets (TestbedTasks.from_xp())

Compatibility

OS	PyVersion	Comment
Ubuntu	3.8 - 3.13
Windows	3.8 - 3.13	no support for elf and hex-conversions yet
MacOS	3.8 - 3.13	hex-conversion missing

Notes:

• hex-conversion needs a working and accessible objcopy
• elf-supports needs
  • shepherd-core[elf] installs pwntools-elf-only
  • most elf-features also still utilize hex-conversion

Installation

The Library is available via PyPI and can be installed with

pip install shepherd-core -U

# or for the full experience (includes core)
pip install shepherd-data -U

For bleeding-edge-features or dev-work it is possible to install directly from GitHub-Sources (here dev-branch):

pip install git+https://github.com/nes-lab/shepherd-tools.git@dev#subdirectory=shepherd_core -U
# and on sheep with newer debian
sudo pip install git+https://github.com/nes-lab/shepherd-tools.git@dev#subdirectory=shepherd_core -U --break-system-packages

If you are working with .elf-files (embedding into experiments) you make "objcopy" accessible to python. In Ubuntu, you can either install build-essential or binutils-$ARCH with arch being msp430 or arm-none-eabi for the nRF52.

sudo apt install build-essential

For more advanced work with .elf-files (modify value of symbols / target-ID) you should install

pip install shepherd-core[elf]

and also make sure the prereqs for the pwntools are met.

For creating an inventory of the host-system you should install

pip install shepherd-core[inventory]

Unittests

To run the testbench, follow these steps:

1. Navigate your host-shell into the package-folder and
2. install dependencies
3. run the testbench (~ 320 tests):

cd shepherd-tools/shepherd_core
pip3 install ./[tests]
pytest