These are some quick and dirty Python scripts for interfacing with the Openreview platform, which I wrote for the MIDL 2022 Conference. The scripts are somewhat hacky and should mainly be used as templates for creating your own scripts for interfacing with the Openreview's Python API.
The code was most recently tested with Python 3.10.6 and the packages in the requirements.txt file.
In order to run the scripts, you need to fill in your Openreview credentials in the call of openreview.Client of each script.
You can install all requirements using
pip install openreview-py pandas tqdm
or if you need to reproduce my exact environment use
pip install -r requirements.txt
Warning: There are two Openreview packages. The official one is openreview-py not openreview.
make-paper-selection-table.pycreates a wide table with all review information for each paper. This can be formatted in Excel (or similar) to filter, and visualise in a way to make it useful for the final decision-making.make-reviewer-rating-table.pycreates a table of all reviewer scores rated by the ACs. Useful for determining best reviewer awards.make-gender-stats-table.pycollects all review, and meta-review scores along with first and last author gender as specified in the authors' Openreview profile. Useful for checking for potential biases in the review process.