Added compressToCustom and decompressFromCustom function

[HelloLudger]

as discussed in / (maybe) useful for

• Is there a way to allow custom dictionaries? #169
• Can you add base91 support to lz-string? #90
• Implement IRI Encoding #165
• Feature request: compress to source-compatible character set #180
• Can utf-8 codes be avoided in the output? #182

It basically works similar to ...Uint8Array(), first compressing everything and then figuring out, how the UTF-16 characters can be stored with the provided dictionary(-string).

I couldn't make the _compress/_decompress route work, because the custom dictionary length is often not 2^something, so _decompress wouldn't work (or _compress would only use a subset of the provided characters).

[Rycochet]

Can this be updated to the new split file version - should be relatively easy to do - testing is getting improvements but the basic pattern is in there!

[HelloLudger]

Yes, looking into it. Hopefully this weekend.

[HelloLudger]

Hey, I hope this will work for you.

[Rycochet]

Only thing blocking this is getting the tests to use the same boilerplate as elsewhere - trying to figure out how to add this to the commandline code (still working on that so don't worry about adding it yourself!)

[HelloLudger]

I didn't use runTestSet because I don't know how to deal with the second parameter for the custom function (the custom dictionary). The "lazy" way would be to give them a default custom dictionary, then runTestSet could be used as is.

[Rycochet]

I didn't use runTestSet because I don't know how to deal with the second parameter for the custom function (the custom dictionary). The "lazy" way would be to give them a default custom dictionary, then runTestSet could be used as is.

I've replied on the thread, hopefully it makes more sense now :-)

[Rycochet]

Also note that we've got a consistent license header on all other files now, can you add that to your new ones :-)

[HelloLudger]

Next try. Hopefully it works now. Found a missing typing for compressToCustom. Seems like TypeScript and testing is useful after all.

I tried to create a test case with an 13-character heterogram (unproblematic), but then the tests "Repeating String" and "Long String" fail, because the "compressed" string is longer than the original string. The custom dictionary has to have at least 16 chars to pass these tests.

Just FYI, for the testing it is not important.

Edit: And another missing typing... don't know why the error wasn't triggered eartlier...

[Rycochet]

This is awesome!!

[Rycochet]

The size test is probably not important - I'm moving over to only testing known output and making sure that the raw data is hashed and known correct - should have that merged later tonight, which will give us a better testing base :-)

[marawannwh]

Sorry, I am a little bit confused. How to use this modification? I thought it added compressToCustom and decompressFromCustom to the lib, but I just installed it, and these functions don't exist.

I want it mainly for this issue #182 (I mean if this modification solves it)

[Rycochet]

@marwan-nwh How did you install it? v2 (that this is part of) is not currently released to npm so you need to install manually.

[marawannwh]

@marwan-nwh How did you install it? v2 (that this is part of) is not currently released to npm so you need to install manually.

Got it, thanks!

[HelloLudger]

Can you elaborate on the found problem? Anything I can do to help?

[Rycochet]

@HelloLudger I've created a draft PR with more detail and an easier way to test it - #200 - hope that helps!