lingomat / olac-ng Goto Github PK

Right now it's based on age of record. But specifically for dynamic archives, we can check identify and listrecords first hit to see if anything has changed, and if not, avoid reharvesting at all.

This was axed from OLAC because it was broken, so I don't know what it looked like, or what it should look like.

It would be reasonable to search for language, and free text in the title and so on. Searching for content in the metadata block is probably not going to be very good given the variation in quality, with some exceptions like date and linguistic type (which could be a drop down since old only has three...).

not really any point to this unless we're extracting the item rather than the page (or entire static archive!)

... saves embedding the svelte component in pages.

This might not actually be possible because at first glance there seems to be no way to request specific pages.

But check anyway because it's not ideal that an error in a single api request from thousands will end up discarding everything. The harverster does not even keep disk copies and network cached hits are no good because resumption tokens must be freshly generated.