limit DOI intake

[maxheld83]

No description provided.

[maxheld83]

same as in #50 here, this is now (crudely) prevented in the shiny app, but we should test this downstream as well, ideally maybe using the same logic.

[maxheld83]

@michaelavoigt just suggested this again as a feature.

Some background: The principle place for use to limit the number of DOIs is inside the report, where as part of the ingest status, users will be informed how many DOIs were dropped because they exceeded the limit (see #60, and more details in #124).

There are two reasons for this:

1. We want to take on as few potentially invalid strings into the interactive R session as possible.
   For that reason, textInput currently only enforces a (generous) character limit, not a DOI limit.
   This is pretty cheap and may even be done in the client sometime soon (see https://github.com/subugoe/biblids/issues/40).
2. Limiting the number of DOIs is actually a bit more subtle than at first expected, because we'd first want to filter out invalid and duplicate DOIs and then apply the limit.

But all that said we should still absolutely advertise right in the entry box what our limit is #163.