GUI-ReRank Dataset: Annotations and Embeddings for the Rico GUI Dataset

Rico Annotation and Embedding Dataset for GUI-ReRank:

This dataset contains all annotation and embedding files required to import the Rico dataset into GUI-ReRank. It includes multi-dimensional annotations and precomputed embeddings for each GUI in the Rico dataset, enabling advanced search and reranking functionality within the GUI-ReRank system.

Contents:

annotations.json: Main annotation file mapping Rico GUI IDs to their multi-dimensional text annotations.

design.npy, domain.npy, functionality.npy, general_description.npy, gui_components.npy, text.npy: Numpy arrays containing embeddings for each search dimension.

failed_files.json: List of Rico GUIs that could not be processed.

index_mapping.json, inverse_index_mapping.json: Mappings between dataset indices and Rico GUI IDs.

meta.json: Metadata about the dataset and annotation process.

Usage:

Download and extract all files into the /rico/dataset directory of your GUI-ReRank installation. This dataset is required for importing the Rico dataset and enabling multi-dimensional GUI search and reranking.