Testing the plugins
In order to test the performance of the plugins, you need a list of DOIs with the associated names of domains
where the metadata of the papers are stored, for example: "10.1038/s41586-020-2649-2" and "www.nature.com".
Place this information in a .csv file with two columns: doi and domain. Then, assuming that this file is
called test.csv, issue the following command:
python -m benchmark.batch_download test.csv
This test command will generate information for every plugin that is called and will show the test results, for example:
============================================================
Total processed: 10
Successful downloads: 6
Success rate: 60.0%
============================================================
Generating benchmark report...
Report generated: benchmark/reports/performance_report.txt
Data exported to: benchmark/reports/performance_data.csv
✓ Reports generated in 'reports/' directory
The reports named above, contain more detailed information on the test run, for example: performance per plugin, performance per journal and combined performance. Note that for efficiency reasons not every plugin is called for every DOI. The plugins with names starting later in the alphabet will only be used for DOIs that the earlier called plugins could not process.