Sharing translation cache

[mspykerez]

First let me ask, if a match is found on the translation cache does it prevent Textractor from sending a translation request to the machine translator altogether?

Now to the important stuff, can we choose to use the cache generated from other translators and select which ones to use even if we don’t have any of the selected translators enabled on the extensions window?

It would also be awesome if Textractor had some sort of database where people could share and grab each other’s cache. I know that machine translations aren’t perfect but DeepL can be quite a good option when there aren’t any hand made translations for a game which is pretty much all the new games and 90%+ of the old ones.
In my experience getting DeepL to work all the time without being subject to a ban is becoming more of a roulette so it would be nice if we were to have some alternative, hence I thought of a shared cache.

[Artikash]

To your first question, yes.
Second question, yes although it's unintuitive (I didn't think anyone would want to do this) you can rename your cache file from another translator to the cache file your current translator uses.
The sharing might be nice but it would involve me setting up and maintaining a server which is a pain.

[mspykerez]

It would be nice to always use the DeepL cache while using other translators in situations where DeepL online translator stops working and we are forced to use other translator instead.
Could also be useful when we are disconnected from the internet and wish to use all available caches at once by giving priority to the DeepL cache and from there going down a user made priority-list of other caches to consider if nothing is found on the previous cache.

I guess renaming the cache should work but it is not practical and could get messy to always have to manually rename and maybe copy-paste the text from one cache file to another if we intend to keep both caches translations at the same time. There should be an automatic way to do this inside the app itself...

I assume that the cache is not per-game but rather a universal thing, right? Does it have some logic attached to it, like trying to match phrases inside a bigger sentenced or it only works with complete sentences that it grabbed from the game as a single unit? Also does it lookup for 'dupe' stuff and avoids logging it again if detected? i imagine that it will not be good to have a huge file with multiple entries of the same thing logged...

It would indeed be nice to have a server connected to Textractor to automatically post/download/sync each others cache but if you think that it would be to much work I think that having some sort of link inside Textractor pointing to a repository/database/forum or something like that where people can manually share their caches should suffice.

[Artikash]

Reminded of this when going through PRs. I was thinking of adding these suggestions if someone else requested them, but they're a bit complicated to implement for something only one person seems to want and can be done manually if truly desired.

Regarding your questions, yes the cache is universal (for that translation provider) and it makes sure not to write dupes. It has no other notable logic, it just looks for exact sentence matches.

[mspykerez]

Regarding dupes, does it only search for matches on the RAW text?
To elaborate on that, there could be the case where the translator used got improved so I think it would be best if dupes were only consider as dupes if the translation matches and on another hand have Textractor replace the existing translation if a different translation is found for the same RAW text (because more recent translations are likely better).
It would be like if RAW+TRANSLATION=MATCH > do not write anything, but if RAW=MATCH+TRANSLATION=NOT MATCH then replace the existing translation with the new one.

[Artikash]

Yeah it just matches raw. If you want to update translations then uncheck Use translation cache. That option only stops Textractor from reading translations from the cache, it will still write and overwrite. The whole idea of using the cache is that Textractor can skip making a request if it already has a translation of that sentence.

[mspykerez]

I see... it's like to know if the translation is a dupe you've to make a request which defeats the purpose of not making requests.
The cache situation is truly complicated it seems.
I think some sort of profile system for it would be better.

[SRC267]

@mspykerez off-topic, but I saw your locked thread about SoRR Vita. You can run the game on Switch, via lineageOS (android on Switch), not ideal as it's not native on the Switch OS, but it's the only solution as it seems the devs don't like the idea of porting it to Switch hence the locked post. BTW lineageOS would dual boot, so it's not replacing the Switch OS.

[mspykerez]

cba dual booting... anyway looks like this also has a psp port so i will see if i can emulate that port.

[SRC267]

I've emulated it on my phone with the PSP port, it has some background glitches but it's playable. The most ideal way is playing the PC version though on my phone with Winlator emulator, no ports needed, just emulate it. No FPS drops, or slow downs like the Vita has (known bug)

Winlator could also work on Switch via LineageOS. But there's no other solution to have to dual boot. As there's no way to replace the Switch OS.