Add script for unlinking some bad agents. #1153
Conversation
jannistsiroyannis
requested review from
niklasl,
andersju,
klngwll,
kaipoykio,
lrosenstrom and
kwahlin
|
Is LXL-3919 the correct issue for this? |
Urgh.. nope :P It's LXL-3913 |
|
I also have a sneaking suspicion that maybe we shouldn't just remove them, but replace them with local entities? |
|
Ah, check. Yes, its only the linkage that's to be removed; "some" local entity data should be kept (presumably only names and possibly lifeSpan (albeit that might be wrong if added on the linked entity after original match)). Ask for clarification/spec of fields. |
| boolean isABadLink(String candidate) { | ||
| String s = candidate.substring(0, candidate.length()-3) // Trim off the #it | ||
| System.err.println(candidate + " -> " + s) | ||
| return s.endsWith("53hlt8kp5swj700") || |
There was a problem hiding this comment.
Nit-pick: safer (and slightly faster) to put these in a Set and use that here and to build the select string.
There was a problem hiding this comment.
To use a Set I'd instead have to shave of the https://libris{-qa,-stg}/ (which varies in length), which could go wrong. I think I'll keep this as is!
There was a problem hiding this comment.
Or use candidate.substring(candidate.lastIndexOf('/') + 1, candidate.lastIndexOf('#'))?
There was a problem hiding this comment.
candidate.split('[#/]')[-2] should work too!
There was a problem hiding this comment.
Yes, that is certainly much more succinct (albeit somewhat less performant, as it would parse a regexp (unless cached under the hood), splitting on all /:s and building a new list of those substrings; though I bet any performance loss is dwarfed by the I/O going on here).
No description provided.