Data supporting "A generalized method for refining and selecting random crystal structures using graph theory"

The dataset contains all the structure files of the nine test systems in CIF and XYZ format. In each system folder, there are 12 folders (excluding C48): SBG-baseline: the initial structures QGBR_init: initial structures selected by QGBR QGBR: QGBR refined structrues discard: initial structures discarded by QGBR QGBR-B_init: initial structures selected by QGBR-B QGBR-B: QGBR-B refined structrues discard-b: initial structures discarded by QGBR-B refinement and their corresponding relaxed structures with "_relaxed" as the suffix. In each "_relaxed" folder in CIF format, there is an "energy.txt" that contains the energy (enthalpy) of the relaxed structures. Note that there are a few rare cases where the relaxation structure configuration is physically unreasonable but the energy (enthalpy) is lower than that of the ground state structure. These structures are not taken into account in our study.