Dataset for: Tree tensor network impurity solver based on Cayley-tree mapping

# Dataset for: Tree tensor network impurity solver based on Cayley-tree mapping ## Overview This repository contains the comprehensive numerical dataset necessary to reproduce the figures and results presented in the manuscript *"Tree tensor network impurity solver based on Cayley-tree mapping"*, published in Physical Review B (PRB). The simulations are based on Matrix Product States (MPS) and Tree Tensor Networks (TTN) applied to the Single Impurity Anderson Model and free-fermion models. To ensure maximum usability and adherence to FAIR data principles, the dataset separates lightweight plotting data (provided as individual files for quick web preview) from heavy raw computational trajectories (provided as a compressed archive). --- ## Repository Structure To optimize for both web preview and bulk downloading, the files are provided as follows: ```text . ├── README.md # This file ├── raw_data_h5.zip # Compressed archive of all raw HDF5 computational data ├── Fig3_... to Fig9_... .csv # Main text plotting data └── Fig10_... to Fig13_... .csv # Appendix benchmarking data ``` ### 1. Main Text & Appendix Plotting Data (.csv) Provided as individual `.csv` files for direct web preview and immediate plotting. Files are systematically named by their corresponding figure numbers (e.g., `Fig3_MPS_E.csv`). All necessary physics metadata, exact reference values, and scaling operations are included in the headers (lines starting with `#`). ### 2. Raw Computational Data (raw_data_h5.zip) Contains the heavy, unprocessed HDF5 files generated directly from the numerical solvers. These files represent the raw retarded Green's functions and full spatial profiles before spectral post-processing (like Linear Prediction). Please extract this `.zip` archive to access the `.h5` files. --- ## Usage Example (Python) The `.csv` files are designed to be easily read into any plotting software. Header rows containing critical metadata start with `#` and can be automatically handled as comments by tools like Pandas. ```python import pandas as pd import matplotlib.pyplot as plt # Since files are provided directly in the root directory, simply use the filename df = pd.read_csv('Fig6b_Aw.csv', comment='#') plt.plot(df['omega'], df['Aw_TTN_D30'], label='TTN') plt.plot(df['omega'], df['Aw_MPS_D90'], label='MPS', linestyle='--') plt.xlabel('Omega') plt.ylabel('A(omega)') plt.legend() plt.show() ``` --- ## CitationIf you utilize this dataset or find our methodology helpful in your research, please cite our Physical Review B paper: Bo Zhan, Jia-Lin Chen, Zhen Fan, and Tao Xiang. *"Tree tensor network impurity solver based on Cayley-tree mapping."* Physical Review B (Accepted, 2026). **DOI:** [10.1103/ycty-d5f9](https://doi.org/10.1103/ycty-d5f9)**URL:** https://link.aps.org/doi/10.1103/ycty-d5f9 ## Contact For any inquiries regarding the dataset, numerical methods, or code implementations, please contact Bo Zhan at [e-mail].