[Koha-devel] Amazon Cover Images

[Frédéric Demians]

> The enhancement I want to see is allowing a mix of sources. It's on
> bugzilla but I can't remember the number.
> http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7187

My thoughts on designing such a system:

- On a Koha community server (or cluster of servers, or Koha library
  server), a repository of cached book covers URLs is maintained. It
  associates a book/resource cover per covers provider with resource ID
  (ISBN/EAN).

  Covers providers could be: Amazon (aws), Google Books (gb), Open
  Library (ol), Koha (kws), etc.

- On Koha pages, a JavaScript function grabs resource ID and send a
  request to the central server (kws) asking for ID cover URLS:

  http://kws.koha-community.org/cover/src=aws,gb,kws,ol&\
  isbn=a,b,c,d&icallback=myjscallback

  src: ordered list of targeted cover providers.
  isbn: list of ISBN (EAD?)
  callback: JS function to call back with ID URL

- kws server receive the request. Find directly URLs in its cache, send
  them immediately to JS call back function. Not already cached URL are
  searched in parallel, and send back if found. kws returns a JSON that
  looks like that:

  {
    "0596000278":{
        "id":"0596000278",
        "cover":"http://books.google.com/books?id=ezqe....",
    },
    "ISBN0765304368":{
        "id":"ISBN0765304368",
        "cover":"http://amazon.fr/thumnail/xx.jpg",
    },
  }

  A mechanism marks as failed, URLs not found from a provider at a time,
  in order not to fetch them again and again. But periodically, a batch
  task can retry fetching those URLs.

- On Koha pages, the JS call back function receive the result, get cover
  for each ID, and modify HTML DOM accordingly.