I’d like to share a status update on the Open Journal Matcher. As I’ve mentioned before, I have a grant to refactor and make the OJM more sustainable. I’m considering two different approaches to the rewritten OJM. The first relies entirely on PythonAnywhere, while the second still involves Google Cloud Platform. This post describes where things stand:

Option 1:

The first approach involves moving all the code over to PythonAnywhere, so that it no longer relies on GCP. The idea is that django-q would farm out the computational work to a bunch of “always-on tasks” (as PythonAnywhere calls them), thereby cutting out GCP. The upside is that everything will be on one platform. The downside is that (for me) there is a learning curve getting up to speed on django and django-q.

Option 2:

The second approach continues with GCP, but aims to make it cheaper. I’ll do this by pre-processing all of the journal abstracts that I’ve harvested from the Directory of Open Access Journals, and store them as binaries rather than as text. The upside is that this approach looks easier to implement; although it’s possible that it may not produce a dramatic improvement.

Conclusion:

I think I’m going to pursue Option 2, which looks easier, and I’ll see how it goes before attempting Option 1.

This entry is licensed under a Creative Commons Attribution 4.0 International license.