I'm trying to blog every day, or at least keep notes. But there are so many interrupts. A constant twitter stream that I feel I have to react to as the issues change by the hour. I'm not going to be able to add links...
I recloned the Git repo and I think that's solved the large file problem. Git has so many ways of doing things so when you look them up on StackOverflow you get comments like "only use this method if you fwant huge pain". The commands are opaque iconsistent and really only useful for experts.
I've now hacked the automatically created dictionaries and they work!
Really excited - starting a project with colleagues in chemistry to mine the literature for new Battery materials. Great team of 3 early
career researchers , we built the site today:
https://github.com/the-grey-group/open-battery/
The relevance is that tech developed there helps us here and vice versa.
So I ran a simple search on three dictionaries. Worked.
### Collaboration on CORD-19 Nice approach from Sysrev over parsing the CORD-19 JSON (see below). We might annotate this.
We are allowed to go for a walk and I did so - about 1 hour and never nearer that 2 m to others.
The JSON has been created by parsing PDF. The pubklishers have XML but they haven't made this available, So it's corrupted. That's unacceptable. Also the material in CORD19 is a ragbag of papers and some non-papers. It's very difficult to navigate
I am FURIOUS with the scholarly publishing industry. The CORD19 was a drop in the ocean and probably of marginal value. what we need is TEN MILLION articles, not 20 K. We need articles on
- ventilators (90% of Taylor and Francis articles are paywalled). Paywalls kill people.
- plagues in history
- rapid prototyping
- mathematical modelling
- drugs
- zoonotic disease
and many many more.