jpdsousa / rookit-parser Goto Github PK

Parsing Error:

• String: 'A R I Z O N A - Electric Touch (Lyrics _ Lyric Video)'
• Format: UNKNOWN

When a STAB is created through a format parser, its score should be affected by the track format's score.
Notes:

• Make sure that scores are added and not overwritten

Use an alternative tagging library that supports other formats:

• JAudioTagger

Ignore fields are not being stored in the database.

Expected behaviour: return the base result

Description

When a track is parsed with a certain format, store such information in the database, in the form of:
{value: String, occurrences: Integer}

• value: the format's string value;
• occurrences: the number of times the format has been picked for parsing

Score

Similar to the field, formats affect the parsing score, based on the number of occurrences.

In format parser (AbstractArtistField), a default key is hardcoded as a base key to be used when creating artists from field values.

The problem is that we can only extract the artist name from the field value, whereas there are more required values that cannot be inferred at this stage. This leads to the ArtistKey being "filled" with incorrect values, which goes against the idea of having those fields required.

The only approach here (AFAIK) is to give as much power to the user as we can:

• Let the user define if a default key will be used. If not, the artist inference will always fail, as long as we do not implement a fallback mechanism;
• Let the user define the default values for each key.

Both options will be defined through a Config-based type. Also, note that this is not a muttex situation, as we can perfectly combine both options.

When parsing a certain input, the mongodb log shows multiple connections to the database.

• Does this make sense?
• Is it mandatory?
• Can it be improved?

Create an insertMultiparser() method that receives a multiparser an returns a parser pipeline which output is a list of the current output (might need to creat an additional interface MultiParserPipeline.

Use Thread pools to split multi parsers, allowing for concurrent parsing.

Food for Through

• Consider changing the return collection to another more "concurrently-suitable" collection (e.g. queue): SO