Install Italian "Zingarelli" dictionary.

[xcube]

Hi,

Is it possible to install the italian dictionary "Zingarelli" in a pocketbook lux 2?

Regards.

[EastEriq]

Where do you get it from and in which format? I have a 7.2MB VocabolarioItaliano.dic, but I forgot where I got it from.

[xcube]

Originally Posted by EastEriq

Where do you get it from and in which format? I have a 7.2MB VocabolarioItaliano.dic, but I forgot where I got it from.

Hi,

Could you pass me the "VocabolarioItaliano.dic"?

I try to get the Italian dictionary from the kindle.

[EastEriq]

What I stored is probably this, from this post of 2012. The OP there says it is converted from a freely available one, at a link which is by now dead.

Digging in the fora I think I understand that at some point in time one firmware upgrade for some pocketbook model included as a bonus the Zingarelli. I don't know if that file has been circulated since.

OTOH, it may be possible to convert the vocabulary you have in .dic format. See e.g. https://www.mobileread.com/forums/sh...d.php?t=325203 . I myself would like to get it in StarDict format for KOreader if possible (I'm not aware of a toolchain for converting .dic to StarDict, only the opposite).

Lastly, I found on my disk also a smaller Zanichelli Italian Dictionary.ld2 (2.9MB which may still be a respectable size). I completely forgot what format it was and where I got it from.

ETA: oh, the zip from https://www.mobileread.com/forums/at...5&d=1347238978 contains also a plain txt file. Converting to StarDict may be easy.

[EastEriq]

Mentioned as bonus coming with a firmware for Pb 912 here

[Markismus]

I think my script will take it if you rename the txt-file to a csv-extension, e.g vocabolario_Italiano.txt.csv, and add the " " (3 spaces) as delimiter. Something like

perl pocketbookdic.pl vocabolario_Italiano.txt.csv it '   '.

The script and further explanation are on github.

[xcube]

Originally Posted by EastEriq

What I stored is probably this, from this post of 2012. The OP there says it is converted from a freely available one, at a link which is by now dead.

Digging in the fora I think I understand that at some point in time one firmware upgrade for some pocketbook model included as a bonus the Zingarelli. I don't know if that file has been circulated since.

OTOH, it may be possible to convert the vocabulary you have in .dic format. See e.g. https://www.mobileread.com/forums/sh...d.php?t=325203 . I myself would like to get it in StarDict format for KOreader if possible (I'm not aware of a toolchain for converting .dic to StarDict, only the opposite).

Lastly, I found on my disk also a smaller Zanichelli Italian Dictionary.ld2 (2.9MB which may still be a respectable size). I completely forgot what format it was and where I got it from.

ETA: oh, the zip from https://www.mobileread.com/forums/at...5&d=1347238978 contains also a plain txt file. Converting to StarDict may be easy.

Thank you very much for all the information and the dictionary.
I will try to find out about everything.

[EastEriq]

Originally Posted by Markismus

I think my script will take it if you rename the txt-file to a csv-extension, e.g vocabolario_Italiano.txt.csv, and add the " " (3 spaces) as delimiter. Something like

perl pocketbookdic.pl vocabolario_Italiano.txt.csv it '   '.

The script and further explanation are on github.

Yes right. At first look though there are some irregularities in the use of two and three spaces as field delimiters, maybe due to the previous conversion process. I'll see if I can find a pattern and an useful regexp to massage the source.

[EastEriq]

I've spent some time in regularizing this Vocabolario_Italiano.txt. Looking into, there were a number of blank lines, repeated keywords, repeated entries, and the use of spacing was irregular. Moreover, I saw no capital in proper names. Possibly all this comes from an earlier conversion attempt.
I corrected what I caught, and uniformized three spaces as delimiter. I attach here my result.

With that I'm able to run Markismus' tool and generate an xdxf (in fact two, _reconstructed and _unbloated, I don't know what would be better) and the .dic again for Pocketbook.

What I was not able at all, is to generate the StarDict version from the xdxf. PocketBookDict didn't, despite my setting isCreateStardictDictionary = 1 in DicControls.pm and DicGlobals.pm (I don't know perl); the new version of pyglossary requires python3.9 while I am at 3.8, and with an older pyglossary generates me an empty .dict.

If someone wants to take over from here....

PS. the attached file contains a single & at line 69831 which may have to be substituted with & depending on the workchain.

[nezih]

Originally Posted by EastEriq

What I stored is probably this, from this post of 2012. The OP there says it is converted from a freely available one, at a link which is by now dead.

Digging in the fora I think I understand that at some point in time one firmware upgrade for some pocketbook model included as a bonus the Zingarelli. I don't know if that file has been circulated since.

OTOH, it may be possible to convert the vocabulary you have in .dic format. See e.g. https://www.mobileread.com/forums/sh...d.php?t=325203 . I myself would like to get it in StarDict format for KOreader if possible (I'm not aware of a toolchain for converting .dic to StarDict, only the opposite).

Lastly, I found on my disk also a smaller Zanichelli Italian Dictionary.ld2 (2.9MB which may still be a respectable size). I completely forgot what format it was and where I got it from.

ETA: oh, the zip from https://www.mobileread.com/forums/at...5&d=1347238978 contains also a plain txt file. Converting to StarDict may be easy.

I unpacked the .dic file with this script I wrote: https://gist.github.com/anezih/b20e2...861efeff3f6072
When unpacked, I got a TSV file of 20K entries, which is way lower than the 80K of txt file. Anyway, I was able to convert the TSV file to the StarDict via PyGlossary, see the attached screenshot.

If you want to convert it yourself run: python pocketbook2tsv.py VocabolarioItaliano.dic

EDIT: There was a bug it seems, now it finds 80449 entries.

[EastEriq]

Good to know that Pocketbook .dic format has been decoded to some extent.

I also understood that your gist works correctly only in python 3.9, not in 3.8 nor in 2.7.

I quickly compared the results of your gist on either VocabolarioItaliano.dic of the original post, or on VocabolarioItaliano_reconstructed.dic which is the result of Markismus' PocketBookDict, and I see that the results differ mostly for spacing (which I corrected), and ordering of words with accented vowels (the result from VocabolarioItaliano_reconstructed.dic has them right imho, alphabetizing accented vowels as if they were not accented, as customary in italian, rather than listing them in code order at the after all unaccented letters). But I also spotted cases in which one file includes the acception (1) of some term, where the other the acception (2) (e.g: abbaglio, abbonare, acapnia, accapponare). We don't know how the original VocabolarioItaliano.dic was really generated, but I remark that these multiple acceptions are all present in the VocabolarioItaliano.txt of the original attachment. @Markismus, bug in your program? As for abbozzare, I have 3 entries in the csv I attached, wheras your result holds only (1).

Not sure what to do with _info.txt , which contain as Wordlist only a few of the terms, with some missing the initial (e.g. "ccozzaticcio", "d usum delphini" "pparenza"), but whatever...

[EastEriq]

Btw, @Markismus, why

abat-jour s.m. invar. paralume | lampada con paralume.

in the csv becomes

<head><k>abat-jour</k></head><def>abat-jour s.m. invar. paralume | lampada con paralume.</def>

in the xdxf? (repteated head term in the <def>, only case I spotted)

[nezih]

Wordlist is composed of every first entry in the each zlib stream. Dictionary program may use it to jump at relevant stream based on sorting. Other than that, the txt file contains keyboards, collates and other stuff that were required by the converter program.

About python version, type hints I added raises the min. version to 3.9, I guess.

For spacing differences, at line 58, I convert every 2+ space to one. If you don't want that add # before that line to comment it out. Finally, at line 59 I try to remove the control characters which I thought were the results of decoding errors but they may denote markup for styling like bold, italic etc. I don't have access to a Pocketbook device so I don't know how those are rendered in lookup results.

[EastEriq]

Originally Posted by nezih

Wordlist is composed of every first entry in the each zlib stream.

Strange that a few of them are off of one character, though. Possibly a bug in your script? But alas, not really a concern. And frankly, it is quite likely that the closed converter.exe, which is what generated those .dics, is much more bug ridden than the sources we can look into. That could explain the omission of multiple acception entries, alphabetic sorting, and what not.

[Markismus]

Originally Posted by nezih

Other than that, the txt file contains keyboards, collates and other stuff that were required by the converter program.

About python version, type hints I added raises the min. version to 3.9, I guess.

If you found new keyboard/collates/morphems txt-files, would you be willing to give a pull request to the github repository for the language files?