#1594 Improve support for markup in DSL dictionaries

[Gabix]

In the DSL markup language, square brackets [ and ] are used to mark formatting tags. When you need to show square brackets in a dictionary article, you have to duplicate them, i. e. to put in [[ and ]]. However, OmegaT‘s DSL parser seemingly tries to process everything between duplicated square brackets as tags and just drops the content when showing an article in the dictionary pane.

Steps to reproduce:

1. Unpack and open the attached sample project in OmegaT.

It has only one segment matching an article in the dictionary supplied. The article contains two transcription areas: [t][[yī ge yàng]][/t] and [t]yī yàng[/t].

1. Open the dictionary pane to see that the first transcription area disappears with only a bare ] left, while the second transcription area is shown.

[Hiroshi Miura]

From the specification;

[s],[/s] – multimedia zone (used to add pictures or sound files into a dictionary entries ).

I've reported OmegaT dictionary pane does not show image but it ca be;

[bugs:#1069]

If you want to improve DSL that also show images, would you please mind to vote and raise a proposal for DSL?

Here is an example with EPWING driver and my proposal of the improvement;

https://sourceforge.net/p/omegat/bugs/_discuss/thread/57eca9ed0d/5c62/989a/attachment/omegat-chameleon.png

[Hiroshi Miura]

Here is an explanation of multimedia tag https://documentation.help/ABBYY-Lingvo8/PictureInDictEntry.htm

 [s]mypicture.bmp[/s].

Because a content of tag is filename of media file, so you may want to ignore not only tag also its content, isn't it?

[Gabix]

Yes, the contents can be ignored, unless you (or someone else) undertakes to implement a viewer for images. The same applies to preview and video tags. But, as I see, you already suggest a solution for images.

[Hiroshi Miura]

Does following make sense?

            reList.add(new RE("\\[s\\](.+?\\.wav)\\[/s\\]", "<a href=\"file://$1\"/>SOUND: $1</a>"));
            reList.add(new RE("\\[s\\](.+?\\.jpg)\\[/s\\]", "<img src=\"file://$1\"/>"));
            reList.add(new RE("\\[s\\](.+?\\.bmp)\\[/s\\]", "<img src=\"file://$1\"/>"));
            reList.add(new RE("\\[video\\](.+?)\\[/video\\]", "<a href=\"file://$1\">VIDEO: $1</a>"));
            reList.add(new RE("\\[s\\](.+?)\\[/s\\]", "UNSUPPORTED MEDIA: $1"));

[Gabix]

Looks like worth trying. I’d also suggest:

            reList.add(new RE("\\[s\\](.+?\\.tif)\\[/s\\]", "<img src=\"file://$1\"/>"));
            reList.add(new RE("\\[s\\](.+?\\.tiff)\\[/s\\]", "<img src=\"file://$1\"/>"));
            reList.add(new RE("\\[s\\](.+?\\.png)\\[/s\\]", "<img src=\"file://$1\"/>"));

[Hiroshi Miura]

https://documentation.help/ABBYY-Lingvo8/PictureInDictEntry.htm Does DSL support PNG?

[Hiroshi Miura]

Pushed to ask review for image support.
https://github.com/omegat-org/omegat/pull/153

[Gabix]

Unfortunately, this build does not load my test dictionary (the previous one does, although ignoring multimedia items). Attaching a modified test case.

[Hiroshi Miura]

Observed

Error load dictionary from 'DSL Test/dictionary/DSLTestForOmegaT.dsl': Cannot invoke "String.length()" because "replacement" is null 

[Hiroshi Miura]

A issue is an order of initialization.
It solve the issue but we can see another problem.

diff --git a/src/org/omegat/core/dictionaries/LingvoDSL.java b/src/org/omegat/core/dictionaries/LingvoDSL.java
index 2f3ad50d4..78b2c57e7 100644
--- a/src/org/omegat/core/dictionaries/LingvoDSL.java
+++ b/src/org/omegat/core/dictionaries/LingvoDSL.java
@@ -80,8 +80,8 @@ public class LingvoDSL implements IDictionaryFactory {

         LingvoDSLDict(File file, Language language) throws Exception {
             data = new DictionaryData<>(language);
-            readDslFile(file);
             dictionaryDir = file.getParent();
+            readDslFile(file);
         }

         private void readDslFile(File file) throws IOException {

You can see bmp and tiff files seems not to be supported on the pane in Java/Swing.

[Hiroshi Miura]

I've seen an another issue.
There is too much spaces between lines. It is caused by <p> tag replaced for [m?].

This can be fixed with followings:

diff --git a/src/org/omegat/core/dictionaries/LingvoDSL.java b/src/org/omegat/core/dictionaries/LingvoDSL.java
index 2f3ad50d4..a2c727fc6 100644
--- a/src/org/omegat/core/dictionaries/LingvoDSL.java
+++ b/src/org/omegat/core/dictionaries/LingvoDSL.java
@@ -80,8 +80,8 @@ public class LingvoDSL implements IDictionaryFactory {

         LingvoDSLDict(File file, Language language) throws Exception {
             data = new DictionaryData<>(language);
-            readDslFile(file);
             dictionaryDir = file.getParent();
+            readDslFile(file);
         }

         private void readDslFile(File file) throws IOException {
@@ -180,15 +180,15 @@ public class LingvoDSL implements IDictionaryFactory {
             reList.add(new RE("\\[lang\\]", ""));
             reList.add(new RE("\\[/lang\\]", ""));
             reList.add(new RE("\\[m\\](.+?)\\[/m\\]", "$1"));
-            reList.add(new RE("\\[m1\\](.+?)\\[/m\\]", "<p style=\"text-indent: 30px\">$1</p>"));
-            reList.add(new RE("\\[m2\\](.+?)\\[/m\\]", "<p style=\"text-indent: 60px\">$1</p>"));
-            reList.add(new RE("\\[m3\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
-            reList.add(new RE("\\[m4\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
-            reList.add(new RE("\\[m5\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
-            reList.add(new RE("\\[m6\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
-            reList.add(new RE("\\[m7\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
-            reList.add(new RE("\\[m8\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
-            reList.add(new RE("\\[m9\\](.+?)\\[/m\\]", "<p style=\"text-indent: 90px\">$1</p>"));
+            reList.add(new RE("\\[m1\\](.+?)\\[/m\\]", "<div style=\"text-indent: 30px\">$1</div>"));
+            reList.add(new RE("\\[m2\\](.+?)\\[/m\\]", "<div style=\"text-indent: 60px\">$1</div>"));
+            reList.add(new RE("\\[m3\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
+            reList.add(new RE("\\[m4\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
+            reList.add(new RE("\\[m5\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
+            reList.add(new RE("\\[m6\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
+            reList.add(new RE("\\[m7\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
+            reList.add(new RE("\\[m8\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
+            reList.add(new RE("\\[m9\\](.+?)\\[/m\\]", "<div style=\"text-indent: 90px\">$1</div>"));
             reList.add(new RE("\\[p\\]", ""));
             reList.add(new RE("\\[/p\\]", ""));
             reList.add(new RE("\\[preview\\]", ""));

[Gabix]

Unfortunately, trying to download the compiled JAR, I get an incomplete file of just 4.5 MiB.

[Hiroshi Miura]

I'd like to upload up-to-date one.

[Gabix]

I upvoted that suggestion.

[Gabix]

At least, GoldenDict renders PNG files correctly when they are referenced to in [s] tags. Unfortunately, I do not have ABBYY Lingvo at hand to test.

[Aaron Madlon-Kay]

Ticket moved from /p/omegat/bugs/1065/

[Hiroshi Miura]

@Gabix I'd like to add your name as copyright holder.
Could you tell me what name to be recorded? just Gabix is ok?

[Gabix]

I am Dmitri Gabinski. Thanks for adding me :)

[Aaron Madlon-Kay]

Implemented in master, [73ca09].

[Aaron Madlon-Kay]

Implemented in OmegaT 5.7.0.