Help in dictionary conversion

[Patrcio]

Hello. I have some problems with conversion to dic format. I need some extra dictionaries, because it will help me in studying languages. Mainly ENG-PL, ES-PL, DE-PL. After some research I found that obtaining dictionaries is difficult for pocketbook devices. I found on this forum simple advice how to convert everything.

https://www.mobileread.com/forums/sh...02&postcount=3

I started with preparations. I downloaded PHP and PERL software, KINDLEUNPACK, pocketbook converter ver3 from this forum, and scripts needed to convert between formats. I have got all maps for languages.

I downloaded dictionaries. Using kindleunpack (all settings on default) I converted .MOBI to .HTML. Later using cmd on my win10 I run PHP script to convert HTML to XDXF.

This is how is look my pocketbook converter directory:

Spoiler:

When I got XDXF I started conversion to dic. In PERL script I changed my home directory, I checked dic conversion on (from 0 to 1 in three configuration lines). I run the program. It asked me that languages are correct, I wrote in command line folder names. Rest I kept on default.

Spoiler:

After that it begins to transfer formats. I have got 2 extra xdxf files in directory, but at the end pocketbook converter won't to convert my file. I've got this message: This file is not in XDXF format

I noticed also that my definicions are messed up, there are wrong polish letters, and in every definitions I've got some </k>.

Spoiler:

How to fix this? What I doing wrong?

[Markismus]

No idea.
I am a bit surprized that it's working with full path d:/konwerter/d:/xdxf. I have always used dictionaries in a subdirectory of the one in which converter.exe resides.

That being said: Garbage in equals garbage out. So it's quite reasonable to assume that your mobi-2-xdxf conversion made things rather odd. If characters are wrong, you have probably edited a file and opened it as a windows character set, while is should be UTF-8.* No idea what php-script you're using. You do realize, that pocketbook.pl uses kindleunpack and converts from mobi? Of course, not from all mobi, because the dictionary format is rather loose.

Still, I suggest you put the dictionary in the subdirectory dict in your main converter directory and run pocketbook.pl either with the html-file from kindleunpack or directly with the mobi-file if you've correctly installed KindleUnpack.**

Why don't you upload the dictionary and post a link, than we can at least see whether it works on Linux. Nowadays, Windows 10 boasts a Linux subsystem, so if it works on Linux and you can't get it work on Windows, I won't be inclined to help you: You'll just have to use a Linux subsystem.

_____________
*Just a wild guess. However, it is safe to edit files with NotePad++ in Windows, because it allows for different character sets, Linux line endings and windows line endings, etc.
**If you have KindleUnpack working with the perl script under Windows, please leave the info in an issue on github, so that I can add it to the readme. Basically, it amounts to being able to run at the terminal in the KindleUnpack directory the code:

python kindleunpack.py -r -s --epub_version=A -i "$InputFile" "$OutputFolder"

[rcentros]

(Meant to start a new thread, instead hit "New Reply." Sorry, not the first time I've done this. I would delete this post if I could.)

[Markismus]

@rcentros Nice. Why are you posting in this thread, though?

[Patrcio]

Here is my dictionary. It's quite small. Could you check it please? I have also linux mint on other partition, so if it will work on your PC, I will try on linux. Thanks!

I tried also put dictionaries into pocketbook converter dictionary. Without luck, other dictionaries also are generating this error.

[rcentros]

Originally Posted by Markismus

@rcentros Nice. Why are you posting in this thread, though?

What happens is I think I've backed up to the main (in this case) PocketBook forum and I hit the "new post" button (or whatever it's called). I don't realize I'm still in a specific thread. I've done this about five times. Sorry for the interruption, it was not intentional.

[Markismus]

@Patricio KindleUnpack failed to convert the unpacked mobi-file to standard html-format.
The generated rawml-file was simple, though, so I added a subroutine for handling that format. The updated code can be found on its github page.

I saw that the final check for using KindlleUnpack is Linux. That's not really helpful, if you want to work under Windows. You can comment out this line, if you do want to test in on Windows:

else{ debug("Not Linux, so the script can't convert mobi-format. Quitting!"); die; }

The results are uploaded to the usual pcloud location.

[Markismus]

@rcentros

Maybe try clicking new thread instead of new post next time.

[rcentros]

Originally Posted by Markismus

@rcentros

Maybe try clicking new thread instead of new post next time.

Yep, I do "try" that. They look pretty much identical. I usually go a month or two before I fall into the "trap" again. It would be nice if we just delete mistakes.