Upcoming change on kaikki.org: Switching over to using .jsonl
file suffixes where appropriate
#603
Labels
enhancement
New feature or request
kaikki.org
Issues and requests regarding the kaikki.org site and frontend
Tatu has come around to the thought that we should change certain file suffixes on kaikki.org to use the
.jsonl
suffix where appropriate..jsonl files are like normal .json files, except they contain one or more json objects which are separated by newlines; you can read more at https://jsonlines.org/
The issue only concerns the file suffix used, and not the content of the files. It just makes it more explicit that we're technically breaking the json standard.
This message is just a heads-up that this change is coming at some point in the future, possibly quite fast, so that people are aware when the switch happens. Old links will break. My assumption is that no one actually uses the single-word and single-article .json files on word pages through any sort of automated process, because the whole big extract output download is much simpler to get than crawling through the website to get bits and pieces, but those should also change.
This isn't exactly a Wiktextract issue because the output is named by the user and not the program itself. The
--human-readable
option for output is also not valid .jsonl because it contains newlines inside the objects.On kaikki.org, the change should mostly (only?) happen with files with word data output; some other .json files we provide, like categories and error data are valid single-object json and won't need changing.
The text was updated successfully, but these errors were encountered: