Please come tell the ampache project what you need in the API

[lachlan-00]

I'm making a call out to all the xml-api client software devs to come over the the repo and make their issues known.

I'm working on updating Ampache to version 4 and would like feedback on those who use the API to actually tell me what they need to make it better.

So far i've added smart playlists to the playlist calls as well as recording/flagging songs.

Please come by when you can and I hope to make all ampache-related software better for everyone! :)

[matiasdelellis]

Hi @lachlan-00 I arrived here looking for a good client for android, but I can not speak for this project. Maybe for @pragha-music-player, who uses this api. -

• The first thing is to use json, so that the results should be smaller.
• Implement a query to see the changes in the collection. A query of the complete collection can easily occupy 50mb, and it would be interesting to add a parameter '&since=UNIXTIME' to obtain only the changes that would be a very small query, and allow to easily maintain the collection on the client side.
  • It would be necessary to add a column 'archive' to the database. In case a file is deleted, just mark this song as 'archived', and update 'last-modified' -MM.. -I do not remember if the database has this column.-
  • In case a file changes the tags, only update the column 'last-modified'
  • So, If the URL call has a parameter '&since=UNIXTIME', in the sql query we must add 'AND last-changed > UnixTime'
  • Well, In normal SQL queries, you have to add an 'AND archived = FALSE'
  • So that the database does not grow eternally, you have to add a command to delete the songs 'archived'.
    • If the client consults for a longer time than the last time they were deleted, it will be necessary to send the complete collection.

Well .. Thank you for your work with Ampache !! Regards, Matias.

[lachlan-00]

thank you @matiasdelellis

• Json is a long term goal but in the short term i'm going to be working with XML. (The open pull for JSON is where we'll be working from after 4 is out)
• I do want to add smaller queries, since i've been working on this i've found the API to be very heavy in pretty much all song/playlist queries
• It would also be easy to maintain a max(id) for songs to know what the last import found so updates which would be easier for clients to implement with a little bit of extension.

These are all good points and thanks for sharing!