Complicated database programming for searching and getting information

You have to realize that you are being hired for a part of the project and not the complete project so don’t worry about the other details not relevant to you You can suggest your own frameworks and the language which you feel are relevant Project Background: We are developing a service that allows developers to access a repository with thousands of radio station streams from around the world. This service has an API formatted in JSON to allow other developers to access our information and build their own applications. Some radio stations provide meta data that can include the currently playing song on the radio station. This metadata is not properly formatted. This metadata is at most 2 strings separated by a hyphen. One for song name, the other for album name. We have already built a database of over 250,000 streams . Now we have a data dump of 150 million songs, roughly 7 million albums and 5 million artists in separate tables. This dump is not on our server, we have this dump in xml format. It needs to be uploaded in the form of tables. In the discogs data dump there are 3 separate tables containing song names, artists, and albums of millions of songs. Your job is to fetch the metadata from 250,000 radio stations. Separate the metadata string into two strings. Then match the two strings against 3 tables (from the data dump) in order to identify which string is the artist and which is the song name. We don't know whether the format is "song name - artist" or "artist - song name" it could also use the album name instead of the artist or song name. This has to be completed within 10 secs for all the radio stations combined. The matching algorithm has to take into account that accents can be different which should not affect the match. Here is an example of a few accents of different letters.