how can I create any dictionary?

[moshpirit]

I miss a Spanish - Portuguese dictionary so I would like to make it but I don't know how.

[rdoeffinger]

Unfortunately the Spanish wiktionary uses a completely different (and IMHO messy) format, and is not supported so far. That means the only way to create the kind of dictionary you want is by using only the English one and its translation entries, which probably won't be great quality. Otherwise the basic process is:

• Get copy of Dictionary, DictionaryPC and Utils repos, go into DictionaryPC
• run data/downloadInputs.sh to get the data (you only need the English one probably)
• run ./compile.sh (probably needs quite a few things to be installed or paths in it configured)
• run ./WiktionarySplitter.sh
• normally: run ./generate_dictionaries.sh. But in your case, you need to run the specific command. For what you want that's something like ./run.sh --lang1=ES --lang2=PT --lang1Stoplist=data/inputs/stoplists/es.txt --dictOut=data/outputs/ES-PT.quickdic --dictInfo=(EN)Wiktionary-based ES-PT dictionary. --input1=data/inputs/wikiSplit/en/EN.data --input1Name=enwikitionary --input1Format=EnTranslationToTranslation --input1LangPattern1=es --input1LangPattern2=pt

Someone properly documenting this process is welcome.

[moshpirit]

┌─ alberto manjaro  ~/Otros/DictionaryPC 
└──┤|▶ ./compile.sh 
ICU4J needs to be installed

How can I get it?

[rdoeffinger]

Just install it via the package manager? Otherwise the project homepage (icu-project.org) also has downloads. You might need to adjust the paths in the scripts (compile.sh and run.sh).

[moshpirit]

If you mean

/usr/share/icu4j-55/lib/icu4j.jar

This is what I have and no "icu4j" in my repos :(

┌─ alberto manjaro  ~ 
└──┤|▶ /usr/share/i
 i18n/        icedtea-web/ icu/         imlib2/      inkscape/    itstool/     
 iana-etc/    icons/       idl/         info/        intltool/    
 ┌─ alberto manjaro ~ 
 └──┤|▶ ls /usr/share/icu/56.1
 config  install-sh  license.html  mkinstalldirs

[rdoeffinger]

Sorry for the delay in answering. Do you mean the file /usr/share/icu4j-55/lib/icu4j.jar exists on your computer? Is it also readable? compile.sh also checks read permissions. If yes to all these, I don't understand why you get that message. You can just comment out the "exit 1" to make it try to compile anyway.

[rdoeffinger]

I added a ES-PT dictionary in 369e8e489ead42680e9a06c818b1e351238b2280. Until a new version with that is released you can download it manually for the releases page. I still wouldn't mind ensuring the process is possible for people besides me though.

[moshpirit]

Oh, perfect!! thanks a lot!!

[papel]

Hello. I want to request to add a dictionary of Portuguese definitions in Portuguese. You can use data from the Portuguese wiktionary. The format is not a problem, because there was an app that used the Portuguese wiktionary. Thanks.

[rdoeffinger]

Single-language ES and PT dictionaries have been added in latest version (but please preferably open separate tickets for dictionary requests, this one was meant for discussing/documenting the process).

[strlcat]

Hi, It is possible to get RU-CS (Russian-Czech) and vice versa dictionaries directly too? I tried to build manually, but the java stuff is a bit messy for me. I stuck at WiktionarySplitter, which throws NullpointerException. I use English as a proxy currently, but it's inconvenient for me.