Evaluation datasets for LuperQ

Published: 29 December 2021| Version 2 | DOI: 10.17632/g4tz8m5p6k.2
Rafael Corchuelo


README.TXT ========== This package provides everything you need to repeat the experimentation regarding the following article: A hybrid quantum approach to leveraging data from HTML tables, by Patricia Jiménez, Juan C. Roldán, and Rafael Corchuelo It has two folders that are explained below. FOLDER DATA ----------- This folder provides the full repository used to perform the experimentation. It has the following sub-folders: - nishida: this folder provides the experimental tables in a format that is amenable to be used with Nishida et. al.'s proposal. - original_tables: this folder provides the experimental tables in a format that is amenable to be used with the other proposals. - tomate: this folder has some meta-data that are used for validation purposes. The annotation tool is available at http://tomatera.tdg-seville.info. There are also some additional files: - Cambria2.ttf: this is the font used in the notebook; installing it is optional. - enwiki_20180420_100d.txt: this are the embeddings computed to experiment with Nishida et. al.'s proposal. - text-metadata-dict.pk: this is a dictionary with meta-data about the tables used in the experimentation. FOLDER LuperQ ------------- This folder provides the code that we prepared to perform the experimentation. Before you run it for the first time, we recommend that you should: - Create a conda environment called "luperq" in which you must install Python 3.7. - Activate the "luperq" environment. - Install the dependencies with "pip install -r requirements.txt". Now, you should be able to launch the experimentation script from the command line as follows: - Change the current directory using "cd LuperQ" - Execute the script with "python -u ./Experimental-Analysis.py" Alternatively, you can launch it as a Jupyter notebook, as long as you perform the following steps: - Install notebook conda kernels with "conda install -n luperq nb_conda_kernels" - Install jupyter kernel with "conda install -c anaconda ipykernel" - Install jupyter kernel into the "luperq" environment with "python -m ipykernel install --user --name=luperq" Now, you should be able to load the experimentation script in Jupyter. Just execute all of the cells in sequence. Whatever means you use to launch the experimentation, it will result in a number of CSV files with the experimental results in folder "workarea". NOTES ----- LuperQ requires access to the D-Wave D-2000 Quantum Processor. Please, make your browser for https://www.dwavesys.com/ and follow the instructions therein to create your account and to install the access key required to use their processor. To facilitate testing LuperQ, this implementation uses a Simulated Annealing Sampler provided by D-Wave. Please, change it if you have a D-Wave account.


Steps to reproduce

See the instructions above.


Universidad de Sevilla


Applied Computer Science