Add results of manual calls of 500 DOIs from on- and off-campus

[jglev]

This PR contains the completed 500-DOI sample, from on- and off-campus, with no "invalid" cells.

See this comment forward re: DOIs which possibly need to be manually edited for the off-campus column, depending on the outcome of that dicussion.

[dhimmel]

How risky / hard would it be to manually port the penn_access_date and open_access_date from the previous dataset. I think the rows are in the same order, so you may be able to copy and paste? This will make manual-doi-checks-500.tsv a bit nicer.

Otherwise, looks good to me once we resolve the login-wall determination.

[jglev]

How risky / hard would it be to manually port the penn_access_date and open_access_date from the previous dataset.

Not hard, and not risky, I think (I'll be careful, and post my code for doing so here using a JOIN approach rather than a cut-and-paste). I'll take care of it now.

[jglev]

This latest commit looks ok / not dangerous to me; @dhimmel, would you mind looking over the code, too, to confirm that it looks ok to you, as well?

[dhimmel]

The changes to manual-doi-checks-500.tsv in https://github.com/greenelab/library-access/pull/20/commits/5befc95a6af85e2ac39f896ea9c60fc837052f9d look good.

Can you confirm that login-walled articles are not considered access?

[jglev]

Following discussion in #15, I can confirm that I've manually change off-campus access to 0 for the following DOIs, as of afba7a1 (there are two commits because I amended the commit log to include an extra DOI and then git push --forced):

• https://doi.org/10.2307/3795274
• https://doi.org/10.2307/3726111
• https://doi.org/10.1039/tf9656100383
• https://doi.org/10.2307/3736092
• https://doi.org/10.2307/317747
• https://doi.org/10.2307/1122632
• https://doi.org/10.1001/jama.1965.03080200059025