Literature Updating: Citation counts sometimes very wrong from semantic scholar

[keelinm]

Sometimes semantic scholar gives citation counts that are much smaller than the existing stored ones. This should be investigated to determine why it happens and if anything can be done to fix it. Currently the code prints a warning if (previous_cits > (1.5 * n_cits)) and (previous_cits - n_cits > 10) ... these are the cases of interest.
On a related note, doi.org returns citations which are stored in the field "citation-count" - should investigate if these are accurate, although at one time I think we decided they were not. (Could check what doi.org gives for the cases of interest mentioned above, for example)