update the response result format

[mirsella]

this is no longer correct:

{
  url: String,
  title: String,
  description: String,
  image: String,
  author: String,
  favicon: String,
  content: String,
  published: Date String,
  type: String, // page type
  source: String, // original publisher
  links: Array, // list of alternative links
  ttr: Number, // time to read in second, 0 = unknown
}

latest broke my application as the result for example author, is now something like

  "author": {
    "@type": "Person",
    "name": "Chris Baraniuk"
  },

and published if not found, is now null instead of a empty string

[ndaidong]

@mirsella yes, the latest update relates to published. It should be empty string instead of null when no date value detected.
However it has nothing to do with author property unless the target page contains jsonld data.
Could you share your article url?

[mirsella]

hey !
for example this url https://www.bbcearth.com/news/the-winged-giant-that-was-bigger-than-t-rex

return as author:

  "author": {
    "@type": "Person",
    "name": "Chris Baraniuk"
  },

and as published, instead of the usual empty string when not finding a date, return null, on the latest version of this library, tested with the bun and deno version:

  "published": null,

[mirsella]

on your demo site https://extractor-demos.pages.dev/article-extractor this is still a empty string, so i guess its on a recent version that wasn't yet put on your site

[ndaidong]

@mirsella yes, we should avoid the inconsistency. I will release a new version.

[mirsella]

personally i've adapted my code to match the api return so i dont "care" anymore, but updating the schema for new user would be nice, and avoid breaking change like this on a minor version bump (both author and published are breaking change)

thanks, have a good day

[ndaidong]

@mirsella your are right, I didn't carefully check the recent pr!

[mirsella]

awesome !

just missing the update response format in the readme, i would have done a PR but i don't know enough the project to says what exactly can be assigned to the field author like the @type

[ndaidong]

@mirsella I try to keep the output structure as before. published and author must be string. So there is no change.

[mirsella]

oh ok author is also getting it's previous comportement, nice