EET - ESP-ESM Translator

[Epervier 666]

Hi
I just tested it and I do not have this problem.
Have you the ability to send me your database file, maybe your setting file, and giving me the exact mod name ?

[ElSolDolLo]

Sure.
First was Tamriel_Data.esm from Tamriel Rebuilt, but I reproduced the case on a smaller one
TR_Factions.esp
Check please 'Sadrith Mora, Telvanni Council House'

Settings + TextOnly DB + ESP

[ElSolDolLo]

I also have a feature request.
The topic phraseforms.
I think you know what they are but, just in case:

► Afficher le texte

Morrowind core has a support for different topic phraseforms. They ship separately for each esm/esp being translated (if any phraseform exist) in files with .top extension.
It is crucial for slavic languages, for example.
When there is a need to use a phraseform instead of a topic name, it is enclosed in @# symbols.
I will try to explain with official Morrowind.esm translation.
For example, if "Morrowind" is translated to "Морроувинд" and we need to use a phraseform "Морроувинде", then we do as you can see on a screenshot:

Then the game knows it should lookup "Морроувинде" inside a phrasephorm file and get original topic name.
The corresponding link between "Морроувинде" and "Морроувинд" will look like this:

One peculiar thing you may notice here - the asterisk symbol (*).
The reason why * need to be added to a phareform - is to avoid conflict when different topics can have same phraseforms.
The number of * can be any, and Morrowind will silentry strip them upon rendering.


But there is a very interesting concern. In data files each asterisk is a 0x7F byte, it only looks as * in .top files. For readability I guess.
Somehow in Tes Construction set it is not translated to *, I don't know why.
You can see here @Шестом Доме*# rendered with trailing 0x7F

But Construction Set is not important.

It is important that 0x7F byte should be translated to * symbol in every INFO text and translated back to 0x7F byte upon saving data files.
In short, currently 0x7F bytes are not rendered, but contained in database.



'@Шестом доме#' here should be actually displayed as '@Шестом доме*#', because there is a 0x7F byte before #

[ElSolDolLo]

The another feature request is dialogue INFO validation. It would be a huge helper.
As topic names are already highligthed in both original and translated texts, it would be great to have a validation tool.
The simpliest way would be to to check that all topics from original text present in translation.
The more complex way would be taking into account all possible topic phraseforms.

[Epervier 666]

Hi,

Thx for your files, concerning the bug, I found the cause and fixed it. No beta right now, but this bug is only for database Only Text, with xml format. If you convert the database in the format eet (in the tool menu),you won't have this bug. (it's just a typo in the code ) I'll have to generate a new beta/new version, but I want to wait for the other things below.
I know absolutly nothing about topix phraseform
Have you any example of mods that use it ? I tried to search for it in original morrowind, but no luck.
Concerning the validation, Maybe I can put together some kind of filter. But know that the coloration of the topic in INFO is just a simple regular expression (so maybe some of them are colored but not appear as such in the game. I never found a sure way to check this)

Cheers

[ElSolDolLo]

I know absolutly nothing about topix phraseform
Have you any example of mods that use it ? I tried to search for it in original morrowind, but no luck.

Sure, check this. Here is ESM with an NPC that provides a greeting with all possible phrasephorm use cases. There's also a game save right in front of him.
Please check INFO id 210029342322891805

This is the text expected to be seen in EET, notice the asterisk:

These will be highlighted without enclosing: phraseform topic one and phraseform topic two.
They have aliases: @topic one alias# and @topic two alias#. They must be enclosed and stored to PhraseFormsExample.top, otherwise will not be highlighted.
You can press @press here# to access first topic, or @press here*# to access second.

As I said, 0x7F pseudo asterisk byte should be replaced to * on load.
On ESP/ESM save it should be replaced back to 0x7F, but only if it is enclosed with @...#
This is how it is seen currently in EET:


This is how text is seen Vanilla Morrowind:

[ElSolDolLo]

Hi,

Thx for your files, concerning the bug, I found the cause and fixed it. No beta right now, but this bug is only for database Only Text, with xml format. If you convert the database in the format eet (in the tool menu),you won't have this bug. (it's just a typo in the code ) I'll have to generate a new beta/new version, but I want to wait for the other things below.

I very appreciate this, thanks!

Concerning the validation, Maybe I can put together some kind of filter. But know that the coloration of the topic in INFO is just a simple regular expression (so maybe some of them are colored but not appear as such in the game. I never found a sure way to check this)
Cheers

Yea, I guessed that.
This is exactly how OpenMW works at least, guess vanilla Morrowind too.
They first collect hyperlinks from info text, having all topics and topic phraseforms and paying attention to @#.
Then hyperlinks are filtered based on if the NPC cannot have an associated topic.

I've just now realized that implementing validation will be more complex than I initially thought.
It would be really desirable to have a validator and an editor for phraseforms somewhere in the interface.
But then we'd also have to somehow integrate them with the translation databases, etc.
Let's leave it as is for now. This is too huge and too Morrowind-specific. I think I will make a standalone tool for this.
The main issue is the problem with 0x7F.

[Epervier 666]

Hi,

Ok, I think I got everything working.
There is a new beta (link in my signature). Bug is fixed, Phraseform is taken in account and there is, in the menu verification (or check ? I don't remember the name in english) an option to check if topic numbers are coherent. It checks only the number of topic found in the info, not the real correspondance, but it can help.
Please tell me if everything works as intented

Pour les anglophobes, nouvelle beta qui corrige la base OnlyText si celle ci était un fichier en xml, prend en compte un truc appelé phraseform dont je ne connaissait pas l'existence, et ajoute une option pour afficher les INFO dont le nombre de topic (surligné) dans ORIGINAL et TRADUIT est différent.

Voilou

[Carlos_G]

I'd like to suggest the possibility of creating databases in a language that doesn't exist in the game, for example Skyrim Special Edition, which doesn't have Brazilian Portuguese as its native language. Therefore, I can't create a database in that language to store what I've done or made and use it to create new mods. I don't know if it has anything to do with the current game, but I've only found this possibility when the game has our language, like Baldur's Gate 3.

[ElSolDolLo]

Hi,

Ok, I think I got everything working.
There is a new beta (link in my signature). Bug is fixed, Phraseform is taken in account and there is, in the menu verification (or check ? I don't remember the name in english) an option to check if topic numbers are coherent. It checks only the number of topic found in the info, not the real correspondance, but it can help.
Please tell me if everything works as intented

Pour les anglophobes, nouvelle beta qui corrige la base OnlyText si celle ci était un fichier en xml, prend en compte un truc appelé phraseform dont je ne connaissait pas l'existence, et ajoute une option pour afficher les INFO dont le nombre de topic (surligné) dans ORIGINAL et TRADUIT est différent.

Voilou

Thank you!

What works:
* Topic count validation works
* Opening the original language file with 0x7F pseudo-asterisks - they are correctly converted to *.
* When saving the translated file - * are correctly converted to 0x7F.

What doesn't work:
* When doing a Fast Import: in the preview, in both Original and Translated, 0x7F is not replaced with *.
* After running a comparison with a Custom database - in applied translation - 0x7F is not replaced with *

What works strangely:
* A Custom database was created from the original and translated esm. Doing a Custom db comparion with the database in xml format - the substituted lines become GREEN. When comparing with the same database in eet format - the lines are ORANGE
* The INFO line with psedudo asterisks: when comparing with Custom xml database - the translation is applied and it becomes GRAY. When comparing with eet - it stays empty.

Also: not sure, but I think everywhere in xml and in eet 0x7F should stay 0x7F, as it is now.

https://drive.google.com/drive/folders/ ... uT-iskTpu8

xml comparison:


eet comparison:


Fast import: