SimCATS logo

SimCATS-Datasets

SimCATS-Datasets is a Python package that simplifies the creation and loading of SimCATS datasets. Please have a look at this repository regarding SimCATS itself.

Installation

The framework supports Python versions 3.7 - 3.11 and installs via pip:

pip install simcats-datasets

Alternatively, the SimCATS-Datasets package can be installed by cloning the GitHub repository, navigating to the folder containing the setup.py file, and executing

pip install .

For installation in development/editable mode, use the option -e.

Loading Datasets

Datasets created with SimCATS-Datasets are stored in HDF5 files. These datasets can be loaded using the function load_dataset from simcats_datasets.loading.

The return value of the function is a named tuple. The fields can be accessed by their name or index. As with normal tuples, it is also possible to unpack the returned fields directly into separate variables. The available fields depend on which data was specified to be loaded. Please look at the docstring for further information.

Additionally, SimCATS-Datasets offers a pytorch dataset (see torch.utils.data.Dataset) implementation called SimcatsDataset. It allows the direct use of SimCATS datasets for machine learning purposes with Torch and can be imported from simcats_datasets.loading.pytorch.

Creating Datasets

To create a simulated dataset, import create_simulated_dataset from simcats_datasets.generation. This function allows the creation of simulated CSDs with ground truth very easily. It is also possible to add further CSDs to already existing datasets. The function will detect the existing dataset automatically. For the function’s usage, please have a look at its docstring.

:warning: WARNING

The functionalities for creating and extending simulated datasets using SimCATS expect that the SimCATS simulation uses the IdealCSDInterface implementation called IdealCSDGeometric. Other implementations might cause problems because the expected information for creating labeled lines etc. might be unavailable.

Alternatively, to using create_simulated_dataset and directly simulating a dataset with SimCATS, it is also possible to create a SimCATS-Dataset compatible dataset with existing data (for example, experimentally measured data or data simulated with other frameworks). This can be done using create_dataset from simcats_datasets.generation.

Citations

@article{hader2024simcats,
  author={Hader, Fabian and Fleitmann, Sarah and Vogelbruch, Jan and Geck, Lotte and Waasen, Stefan van},
  journal={IEEE Transactions on Quantum Engineering},
  title={Simulation of Charge Stability Diagrams for Automated Tuning Solutions (SimCATS)},
  year={2024},
  volume={5},
  pages={1-14},
  doi={10.1109/TQE.2024.3445967}
}

Examples

API Reference