Typing and refactoring of serializer.js

[DJ1TJOO]

Some context
I'm currently working on a project for my university to create a visual shacl editor, to accomplish this I'm using this library. Since the documentation is limited I'm looking a lot through the source code to find out how I can do certain operations. As a hobby project I would like to get better at refactoring code, so my idea to refactor the serializer.js.

The plan

• Fase 1: Convert serializer.js -> serializer.ts, by adding typings, trying to avoid any code changes unless absolutely nessesary (repo)
• Fase 2: Split up the code into multiple classes (repo):
  • There are different types of serializers in the serializer.js file: statementsToN3, statementsToNTriples, statementsToXML, and statementsToJsonld
  • All of these share code like namespaces/prefixes, base, flags, store, etc.
  • Some share other code
    • statementsToN3 and statementsToNTriples share atomicTermToN3
    • statementsToN3 and statementsToXML share rootSubjects
    • statementsToJsonld is just statementsToN3 with some extra code after it
  • To address this I have the following in mind: split up into multiple classes, but still do not change any code of the methods, just move

• Fase 3: Refactor code: like flags could become a object with clearer names
• Fase 4: Possible add new features, I'm currently thinking about giving options on which predicates should come first, or a way to add comments?

The question
Is rdflib.js interested in any of this, or should I just keep it to my self?

[bourgeoa]

Thanks for your interest in rdflib.js
Every help is appreciated.

Changes should not break the way functions are used.
Your step 1 should not be an issue.

The major problem is documentation. #656
I suggest you could also look at the high level functions fetcher and updateManager and to LiveStore a store with fetcher and updateManager

[DJ1TJOO]

Changes should not break the way functions are used.

In my implementation of fase 2 I have added a LegacyCompatibleSerializer with a deprecated tag, which is exported by the index as Serializer, the index then also exports the new serializers.

I have updated the internal uses of the Serializer to do use the new ones

Both fase 1 and fase 2 contain // @TODO(serializer-refactor): comments where I explaine some decisions, why things where done the way they are and/or ask question

[bourgeoa]

I created a refactor-serializer branch

[DJ1TJOO]

In my implementation of fase 2 I have also added many test cases for the serializer. I have covered
Stmts: 87.03%
Branch: 79.29%
Funcs: 89.38%
Lines: 88.2%

With which I'm pretty confident I can start changing code

[DJ1TJOO]

@bourgeoa could you explain to me what writeStore is used for in the serializer?