aard2 slob to stardict conversion is missing entries

[NgrNxk]

I've created the following dictionary:

╰─❯ slob info perrypedia-2025-11-19.slob                                                                                                                                                             ─╯


perrypedia-2025-11-19.slob
--------------------------
         id: 560cf0ea67b1453c94c22f8d9fbb1de9
   encoding: utf-8
compression: lzma2
 blob count: 64311
  ref count: 100451

Stuff like this works just fine:

╰─❯ slob find perrypedia-2025-11-19.slob Arkoniden                                                                                                                                                   ─╯
6094873 text/html;charset=utf-8 Arkoniden
6160385 text/html;charset=utf-8 Arkonidenanzug
73859074 text/html;charset=utf-8 Arkonidenbengel
6160386 text/html;charset=utf-8 Arkonidenliebe
6160384 text/html;charset=utf-8 Arkoniden (PR Neo)
6160387 text/html;charset=utf-8 Arkonidenzoo (PR Neo)

I then converted this aard2 to stardict via pyglossary -v4 perrypedia-2025-11-19.slob pp-sd --read-format=Aard2Slob --write-format=Stardict and although it looks nice from the outside:

╰─❯ sdcv -2l .                                                                                                                                                                                       ─╯
Dictionary's name   Word count
Perrypedia (de)    64223

The search is strange somehow:

╰─❯ sdcv -2x . Arkoniden                                                                                                                                                                             ─╯
Found 3 items, similar to Arkoniden.
0)Perrypedia (de)-->Arkonidizin
1)Perrypedia (de)-->Arkons Ende
2)Perrypedia (de)-->Koniden
Your choice[-1 to abort]: -1

I don't know what's going on here. The issue also shows up in my KOReader in exactly the same form -- it's not only an issue with sdcv.

[ilius]

If you want to use sdcv or KOReader, use this format instead:
https://github.com/ilius/pyglossary/blob/master/doc/p/stardict_merge_syns.md

If still have problems, please upload your slob file.

[NgrNxk]

Thanks for your response. The merge syns format didn't help: the issue remains the same. How can I upload my 80MB slob file here?

[ilius]

You can upload to google drive or mega.nz

[NgrNxk]

OK, so here it is: https://drive.google.com/file/d/1dVf9QGMf51_zTN63le6T64YtFsF4X87y/view?usp=drive_link

[ilius]

No access.
I requested access.

[NgrNxk]

Oh, sorry. You should have access now.

[ilius]

Looks like slob find does something like a substring ("contains") match (probably more complicated). But sdcv does not do that by default! (and KOReader uses sdcv)

These are the article keys that I found containing "Arkoniden":

• Arkonidenanzug
• Arkonidenliebe
• Arkonidenzoo (PR Neo)
• Das Doppelspiel des Arkoniden
• Der Arkonidenjäger
• Der Meisterplan des Arkoniden
• Dorgon 10 - Duell der Arkoniden
• Double (Arkoniden)
• Großkampfschiff (Neu-Arkoniden)
• Leichter Aufklärer der Arkoniden
• Stayn (Arkoniden)
• TRÄGER-Klasse (Arkoniden)

which matches the result of sdcv '*Arkoniden*' command:

Found 12 items, similar to *Arkoniden*.
0)Perrypedia (de)-->Arkonidenanzug
1)Perrypedia (de)-->Arkonidenliebe
2)Perrypedia (de)-->Arkonidenzoo (PR Neo)
3)Perrypedia (de)-->Das Doppelspiel des Arkoniden
4)Perrypedia (de)-->Der Arkonidenjäger
5)Perrypedia (de)-->Der Meisterplan des Arkoniden
6)Perrypedia (de)-->Dorgon 10 - Duell der Arkoniden
7)Perrypedia (de)-->Double (Arkoniden)
8)Perrypedia (de)-->Großkampfschiff (Neu-Arkoniden)
9)Perrypedia (de)-->Leichter Aufklärer der Arkoniden
10)Perrypedia (de)-->Stayn (Arkoniden)
11)Perrypedia (de)-->TRÄGER-Klasse (Arkoniden)
Your choice[-1 to abort]: 

[ilius]

Okay so the problem is pyglossary does not support slob aliases and ignores them.

[ilius]

The slob.Slob class does not expose aliases.
And time slob aliases perrypedia-2025-11-19.slob Arkoniden command takes about 1 second to finish!

How did you create your slob file BTW?

[ilius]

If you created this and want to use StarDict format (sdcv, KOReader, GoldenDict), I would suggest you convert it to StarDict Textual File (.xml)

pyglossary perrypedia-2025-11-19.slob perrypedia-2025-11-19.xml --write-format StardictTextual

Then modify and maintain this xml (add alternates, modify content etc), and convert to StarDict to use it.

[NgrNxk]

The slob.Slob class does not expose aliases. And time slob aliases perrypedia-2025-11-19.slob Arkoniden command takes about 1 second to finish!

How did you create your slob file BTW?

I first scraped the perrypedia with mwscrape and then used the mwslob command.

[NgrNxk]

If you created this and want to use StarDict format (sdcv, KOReader, GoldenDict), I would suggest you convert it to StarDict Textual File (.xml)

pyglossary perrypedia-2025-11-19.slob perrypedia-2025-11-19.xml --write-format StardictTextual

Then modify and maintain this xml (add alternates, modify content etc), and convert to StarDict to use it.

But here the aliases are also missing. I wouldn't know how to add them manually. I wanted to do as little manual work as possible. The mwscrape -> mw2slob path works beautifully easy. I wouldn't mind mwscrape -> stardict; but I haven't found a tool to do that...

[ilius]

How big is your couchdb - the output of mwscrape after you put it in an archive? (prefer 7zip)

[NgrNxk]

─❯ du -sh .couchdb/ perrypedia-couchdb.7z
545M    .couchdb/
175M    perrypedia-couchdb.7z

Do you want me to upload it somewhere?

[ilius]

Yes