SDL Studio Terminjector Plugin - Reactivated
The original Terminjector plugin by Tommi Nieminen was repackaged for Studio 2017 (I guess the installer will also work for Studio 2015). Even though some of its functionalities, similar to match repair, are now also available natively in Studio, or through the Regex Autosuggest plugin, I believe it is still worth giving it a spin, because there’s always more than one way to skin a CAT.
Below I have copied my e-mail discussion with the plugin’s author from the end of 2012, when I was just beginning my adventure with SDL Trados Studio. I believe it may be useful not only for me, because it clarifies some things covered in the original instructions.
Regexp to replace hyphen with n-dash
45 messages |
| Piotr <piotrbienkowski@gmail.com> | Sat, Sep 29, 2012 at 2:00 PM |
To: tommi@tntranslations.com
|
Hi Tommi,
Sorry to bother you, maybe you do not have the time to provide support, but what is the regexp to replace a hyphen surrounded by spaces on both sides with an n-dash with spaces on both sides?
I tried to create something based on your examples, but failed.
Will appreciate,
Piotr
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sat, Sep 29, 2012 at 7:11 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi Piotr,
I don't mind doing support, it's good to know there are people giving TermInjector a try.
You can do this with two different rules. The simple rule is this:
- – -
(The middle character is n-dash.)
Since TermInjector applies each rule to every string surrounded by spaces (or other word boundary characters), all hyphens surrounded by spaces will be turned into n-dashes both in constructed proposal and fuzzy proposals. The drawback is that the rule is applied for every hyphen surrounded by word boundary characters, so some hyphens might be incorrectly changed (this is probably fairly rare, though).
The second rule is more complicated but only converts hyphens between two spaces:
([^ ]+) - ([^ ]+) \1 – \2
The regex in the first field matches strings of non-spaces followed by a space, hyphen a space and another string of non-spaces. In the second field the hyphen is converted to n-dash but the rest of match is copied as it is. This rule works only with new segments, if you want to convert hyphens in fuzzy matches, also add this rule:
\1 – \2 ([^ ]+) - ([^ ]+)
These rules should take care of most cases. One exception is a segment starting with a hyphen surrounded by spaces. In that case the complex regexp won't be used, as there is no string of non-spaces before the space. You might want to add another rule covering those cases as well, if it's necessary.
By the way, if you are using 2011 the Reload files button may not work. If that's the case, you can enter the rules in the editor (with concordance search key) or restart Studio (I think there might a bug in 2011, everything works fine in 2009).
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Sep 29, 2012 at 10:33 PM |
To: tommi@tntranslations.com
|
Thanks for the explanation. Yes I am using Studio 2011.
Regards,
Piotr
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Wed, Oct 3, 2012 at 5:56 AM |
To: tommi@tntranslations.com
|
Hi Tommi,
Your regex for new segment works, thanks. Did I understand you correctly that for it to work in fuzzy matches the pattern being replaced and the replacement need to be entered the other way round?
I also see that the third field need not be entered? In my attempts I was filling the third field as well and perhaps it was blocking it from working.
As for Studio 2011, disabling the plugin and then enabling it again also seems to work as a substitute for reloading files.
Regards,
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Wed, Oct 3, 2012 at 7:36 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi,
Your regex for new segment works, thanks. Did I understand you correctly
that for it to work in fuzzy matches the pattern being replaced and the
replacement need to be entered the other way round?
That's right, for fuzzy match replacement you specify the pattern to be replaced in the third field and the replacement in the second field. You can have both of the rules active at the same time, if you want to use the hyphen replacement in both new and fuzzy segments.
I also see that the third field need not be entered? In my attempts I was
filling the third field as well and perhaps it was blocking it from working.
In most rules filling two fields is enough to get the desired effect. If you fill all three fields, you get a rule that works with both new and fuzzy segments and that checks for the existence of the first field's pattern in the source segment before replacing the third field's content in the target segment.
-Tommi
|
|
| Piotr <piotrbienkowski@gmail.com> | Wed, Oct 3, 2012 at 11:35 AM |
To: tommi@tntranslations.com
|
Sounds a bit complicated now. Does it mean that for fuzzy segments the first field is left empty?
I guess I'll figure out this stuff with time.
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Wed, Oct 3, 2012 at 12:54 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
Yeah, leave the first field empty for a fuzzy segment rule, if you want it to work all the time.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Oct 6, 2012 at 9:29 AM |
To: tommi@tntranslations.com
|
Hi it's me again
I wanted sequences like this, including parentheses (AMP_MQSWP_06F_B) to be copied from source into fuzzy matches. So I created this regex rule:
\1 (\([A-Z0-9\_]+\))
first field being empty.
It does not seem to work.
What am I doing wrong?
Will appreciate.
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sat, Oct 6, 2012 at 10:46 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi,
This is one of the cases where you have to use the first field to get the desired effect. You want to copy the code from the source, so you should add the regex of the code to the source field inside parentheses. Then you need the same regex in the target field without parentheses (since you don't want to copy the code from the fuzzy match). Try this rule (the underscore doesn't need to be escaped, by the way):
(\([A-Z0-9_]+\)) \1 \([A-Z0-9_]+\)
With this rule TermInjector will go through the source segment looking for a string matching \([A-Z0-9_]+\). When it finds it, it will record the string in the first capture group. Then TermInjector will look for a string matching \([A-Z0-9_]+\) in the fuzzy target segment, and record the position of the found string. Finally it will replace the string at that position with the content of the first capture group (as \1 has been entered into the target field).
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Oct 6, 2012 at 12:26 PM |
To: tommi@tntranslations.com
|
Thanks, it works, but I must always remember to put the tab character before the end of line in my regex file. Took me a while to figure it out.
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Oct 6, 2012 at 1:20 PM |
To: tommi@tntranslations.com
|
I am sorry, I have to take back my words - ir DOES NOT work although it is entered according to your instructions and the regex is correct (checked in UltraEdit). There must be a bug in Studio 2011 or ... your plugin...
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sat, Oct 6, 2012 at 9:50 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi,
I tested the regex in 2009, and it seems to work, as you can see from the screenshot (first match is TermInjector output, second match is the same segment directly from the TM). What's the context in which you are trying to use the regex? Is the sequence separated from the rest of the text with a space or some other boundary character?
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
 | | test_for_piotr.PNG
12K |
|
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Oct 6, 2012 at 10:03 PM |
To: tommi@tntranslations.com
|
So in case of fuzzy matches I should look at the TM matches pane and not expect the changed fuzzy matches to pop in automatically in the editor?
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sat, Oct 6, 2012 at 10:19 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
If you have enabled the "Apply best match after successful lookup" option in Options -> Editor -> Automation, Studio should copy the best match from TermInjector automatically to the selected segment in the editor. This only works for completely untouched segments, though, so if you've previously edited the segment, you'll have to copy the changed fuzzy match yourself (you can use Alt+1 to get the best match).
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Oct 6, 2012 at 10:23 PM |
To: tommi@tntranslations.com
|
Tommi, I really appreciate that you are very helpful in trying to get your plugin to work, but it appears that Studio 2011 has some problems with it. But I will keep testing.
Regards,
Piotr
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Oct 6, 2012 at 10:34 PM |
To: tommi@tntranslations.com
|
Last question for today :-) What regex engine does the plugin support? I am most familiar with the Perl style regular expressions.
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sun, Oct 7, 2012 at 10:49 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Yeah, I'm starting to think there are more problems in Studio 2011 that are affecting the plugin. I'll have to do some testing on the newest version.
The regex engine supports POSIX extended regular expressions (but not POSIX character classes). I made my own engine because the amount of regular expressions can be very large (you might for instance want to load an entire glossary as regular expressions), and matching each regular expression separately on the segment might be too time-consuming (if you had ten thousand regular expressions, you would have to match the segment ten thousand times, which will probably cause visible slowdown). With my engine you can match all the regular expressions on the segment at once. I did make mistake with the | operator, though, it's currently non-standard. I'm going to change it back to the standard POSIX | in the next version.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Mon, Oct 15, 2012 at 6:36 AM |
To: tommi@tntranslations.com
|
Hi Tommi, thanks for the explanation.
What would be the regex to change 10A to 10 [A] in new segments? this string occurs at segment end. and 10 can be any number, also a single digit number..
Regards,
PIotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Mon, Oct 15, 2012 at 8:14 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi,
Try this one:
([0-9]+)A \1 [A]
The [0-9]+ expression matches all number strings with one or more numbers. You might want to add a replacement field in case you get memory matches later on:
([0-9]+)A \1 [A] [0-9]+ \[A\]
This will function the same as the first rule in new segments, and in fuzzy segments it will copy the number from the source to the fuzzy match.
Let me know if this works, I tested it out in 2009 but I've come to expect problems in 2011.
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sun, Dec 9, 2012 at 7:30 AM |
To: tommi@tntranslations.com
|
Hi Tommi,
I've just installed terminjector 2.1 . Does it work better in Studio 2011. I'm looking forward to trying it.
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sun, Dec 9, 2012 at 10:42 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi Piotr,
Yeah, the Reload Files button should now work in 2011. If you've still got the old plugin installed manually in 2011, you might have to remove that before installing the new one. Check if the version in the TermInjector settings screen is 2.1 to see if the new version is installed. If it still shows 2.0 after installing 2.1, close Studio, delete the TermInjector folder in c:\Users\<your_username>\AppData\Local\SDL\SDL Trados Studio\10\Unpacked and restart Studio.
Let me know if you have any problems, it would be good to know how the installer interacts with a previous manual installation.
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sun, Dec 9, 2012 at 12:56 PM |
To: tommi@tntranslations.com
|
I think I used the msi file to install TI 2.0, but I'll check anyway. Now I'm working on aother project of the same type so there is a lot of leverage from the TM, but it will be fun to hack some regex as a break from time to time :-)
Regards,
Piotr
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sun, Dec 9, 2012 at 3:12 PM |
To: tommi@tntranslations.com
|
It still showed ver. 2.0 so I did as you wrote, then removed and reinstalled ver. 2.1, and now studio says that terminjector translation provided could not be opened and has been disabled. So it does not work.
PB
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sun, Dec 9, 2012 at 3:36 PM |
To: tommi@tntranslations.com
|
False alarm. I had to remove the entry from project settings and then add it again, and then it worked.
Piotr
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sun, Dec 9, 2012 at 7:07 PM |
To: tommi@tntranslations.com
|
Hi Tommi,
I'm doing it just for the sake of checking whether it works in Studio 2011 because I can press CTRL+INS as well
this regex (\(SD\d+)(\-\d+\))
is sure to match text like: (SD971-11)
but no "translation" gets constructed. I wonder why. I have \1\2 in the 2nd field and the third field is empty.
I also have this rule where there is a single space in the 2nd field and [ ]{2,} in the third field.
The rule works and I'd expect it to collapse multiple spaces into a single space in fuzzy matches, but it leaves TWO spaces instead. Why is that?
Regards,
Piotr
Regards
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Mon, Dec 10, 2012 at 7:21 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi Piotr,
The first does not work because \d is not supported (it's part of the Perl regex scheme, but TermInjector only uses a subset of the POSIX regex scheme, you can check the supported operators from the TermInjector instructions). You can use [0-9] instead of \d.
The second one works the way it does because TermInjector applies the regular expressions only to parts of the segment that are after a word boundary character (like a space). In the case of consecutive spaces, this means that the first space is not included in the string that TermInjector captures, so only the later spaces will be collapsed into a single space. That's why you end up with two spaces. I think there is a good reason why this works the way it does (possibly to do with performance), but I can't remember it now.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
Hi Tommi,
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
TermInjector folder in c:\Users\<your_username>\**AppData\Local\SDL\SDL
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Mon, Dec 10, 2012 at 10:30 AM |
To: tommi@tntranslations.com
|
I see. I learned my regex with Perl, but I can pick up other dialects easily. :-) Does it also support references to character classes that have a colon at the start and end?
Thanks for your patience and explanations.
Regards,
Piotr Bieńkowski
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Mon, Dec 10, 2012 at 11:19 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
No, TermInjector doesn't support any character classes, you'll have to use bracket expressions to do the same thing (like [a-zA-Z0-9] for [:alnum:]). I guess character classes would be the next improvement, if I ever get around to making a new version.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
TermInjector folder in c:\Users\<your_username>\****
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Mon, Dec 10, 2012 at 4:46 PM |
To: tommi@tntranslations.com
|
Now it starts working for me, thanks. I was confused about the \d because it is supported in Studio itself in the view filter regexp box.
Does the regexp file in Terminjector support comments? I understand regular expressions (well, mostly) but sometimes I am a bit slow in figuring out what is what when I revisit the file only after some time.
Maybe you should consider asking a price for terminjector to keep you motivated for further development.
Regards,
Piotr Bieńkowski
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Tue, Dec 11, 2012 at 12:58 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
I think TermInjector only checks for the first three columns of the regex file, so you should be able to add an extra column for comments. Just make sure there are at least three tabs (or other column separators) on the line before the comments.
As for asking money for TermInjector, I think there needs to be a large userbase for the free version first. That might happen, but in the meantime I'll maintain TermInjector as a hobby project.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Wed, Dec 12, 2012 at 7:03 AM |
To: tommi@tntranslations.com
|
So here's an idea for the future. It'd be useful to be able to temporarily disable a regex without having to retype it again in the future. But wait, I think I can do it already by adapting your idea of adding more colums, and adding several empty tabs in front?
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Wed, Dec 12, 2012 at 8:29 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
I'm not sure how completely empty rules behave in TermInjector, so that might have some unintended consequences. Prefixing the contents of the first field with something like "###NOTINUSE### would probably make sure the rule would never fire.
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Wed, Dec 12, 2012 at 10:07 AM |
To: tommi@tntranslations.com
|
Now a completely different question. :-) How do you type the ¤ character suggested as a separator for adding terms through the editor interface. It isn't on the keyboard. Do you use the alt code (0164)?
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Wed, Dec 12, 2012 at 10:29 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
You can change it in the Advanced Settings. I only chose ¤ because I can type it with shift-4 on my keyboard and it isn't used in any texts, I think other keyboard layouts might have similar characters that are never used.
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Thu, Dec 20, 2012 at 6:16 PM |
To: tommi@tntranslations.com
|
Hi again Tommi
What programming language and what tools are required to develop plugins for Studio? Just asking out of curiosity, I won't start programming for Studio in the foreseeable future....
Regards,
Piotr
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Thu, Dec 20, 2012 at 6:55 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
I use C# and a free version of Visual Studio. I think you might be able to use other programming languages as well (like Visual Basic), but C# is really good, so I haven't tried the others.
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Thu, Dec 20, 2012 at 7:39 PM |
To: tommi@tntranslations.com
|
First I would need to learn some C# and have an idea how things work inside SDL Studio...
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Thu, Dec 20, 2012 at 8:54 PM |
To: tommi@tntranslations.com
|
Is it just C# or Visual C# .NET?
PB
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Fri, Dec 21, 2012 at 8:07 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
I think it's Visual C# .NET, that's probably the implementation they have in Visual Studio Express. Once you've signed up with the SDL developer program (free of charge), you can download project templates that work with Visual Studio Express (although you might have to import the files manually). Probably the easiest way to learn is just to make some adjustments to the templates. Just remember that for Studio plugins the target framework must be set to .Net Framework 3.5 (the default is 4).
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
[Quoted text hidden]
Piotr Bie?kowski
On Mon, Dec 10, 2012 at 11:19 AM, <tommi@tntranslations.com>
wrote:
No, TermInjector doesn't support any character classes, you'll
have
to
use
bracket expressions to do the same thing (like [a-zA-Z0-9] for
[:alnum:]).
I guess character classes would be the next improvement, if I ever
get
around to making a new version.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
I see. I learned my regex with Perl, but I can pick up other
dialects
easily. :-) Does it also support references to character classes
that
have
a colon at the start and end?
Thanks for your patience and explanations.
Regards,
Piotr Bie?kowski
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Dec 22, 2012 at 2:51 PM |
To: tommi@tntranslations.com
|
Hi Tommi,
Here I come again with my regular expressions...
I get sentences like these:
F41->Rear Combination Lamp (IN)
Where the -> stands for the tab character.
and I have this regex:
^([A-Z]+)([0-9]+)(\t) \1\2\3 ^([A-Z]+)([0-9]+)(\t)
This is my final attempt but it does not work without the ^ and without the (\t) either.
I want it to copy fragments like F41 to the beginning of a fuzzy match that does not have a similar fragment (the advanced option is enabled) or substitute it in a fuzzy match that has this fragment.
But nothing happens i.e. I don't see anything that could be the expected product of terminjector in the TM matches pane.
By the way, does terminjector support anchor characters like ^ and $ ?
Regards,
Piotr
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sun, Dec 23, 2012 at 11:40 AM |
To: tommi@tntranslations.com
|
Sorry to bother you again (nothing urgent, just to let you know)
I think that terminjector has a problem matching strings that contain a hyphen, especially if I put the hypen (escaped) in square brackets. I think it has to do with the fact that it is a POSIX regex engine.
PB
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Fri, Jan 4, 2013 at 10:49 AM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi Piotr,
Sorry about the late reply, just getting back from the holidays.
Your regular expression is good, it's just that the tab special character \t is not supported (it's treated as a literal character 't'). Another problem is that there is no boundary character between the tab and the "Rear Combination Lamp" part, so the regex won't be applied.
I got your regex to work by removing the \t special characters (and the \3 from the second field) and adding tab to the boundary characters (the only way to add it seems to be to copy it from a text editor or somewhere else). So the regex is like this:
^([A-Z]+)([0-9]+) \1\2 ^([A-Z]+)([0-9]+)
I'll probably add support for \t in the next version, it seems important.
Also, it seems it's not possible to copy the fragment to the start of a fuzzy match (that does not contain a matching fragment) with this kind of rule. The third field would have to be empty for the copying to happen, but then the replacing would not work. This is also something I should fix for the next version, although I'm not yet sure how the plugin should ideally behave.
TermInjector does support the ^ and $ characters, they are pretty much the only advanced POSIX functionality I've implemented. You can usually check if something is supported by doing a quick search for it in the TermInjector manual: if there's no references to some functionality, it most probably isn't supported.
I'll also make a note of that hyphen issue you mentioned in the other mail, it's probably a problem with the validation system I made.
-Tommi
Lainaus Piotr <piotrbienkowski@gmail.com>:
Hi Tommi,
Here I come again with my regular expressions...
I get sentences like these:
F41->Rear Combination Lamp (IN)
Where the -> stands for the tab character.
and I have this regex:
^([A-Z]+)([0-9]+)(\t) \1\2\3 ^([A-Z]+)([0-9]+)(\t)
This is my final attempt but it does not work without the ^ and without the
(\t) either.
I want it to copy fragments like F41 to the beginning of a fuzzy match that
does not have a similar fragment (the advanced option is enabled) or
substitute it in a fuzzy match that has this fragment.
But nothing happens i.e. I don't see anything that could be the expected
product of terminjector in the TM matches pane.
By the way, does terminjector support anchor characters like ^ and $ ?
Regards,
Piotr
|
|
| Piotr <piotrbienkowski@gmail.com> | Fri, Jan 4, 2013 at 10:59 AM |
To: tommi@tntranslations.com
|
Hi Tommi,
Thank you for your response. I wonder if I am the heaviest user of TI bothering you with the most questions... ;-)
I am finishing one automotive project and will be starting another one soon, so probably I will come up with more test cases.
PB
[Quoted text hidden]
|
|
| tommi@tntranslations.com <tommi@tntranslations.com> | Sat, Jan 5, 2013 at 1:24 PM |
To: Piotr <piotrbienkowski@gmail.com>
|
Hi Piotr,
I think you are the heaviest user, :). Questions are good, otherwise I'll never find the bugs or get new ideas. In the future I may make a more comprehensive plugin (with machine translation components and more versatile rules), and everything I learn now is going to help with that. Definitely the biggest problem is user-friendliness, that's one thing I've discovered so far.
[Quoted text hidden]
|
|
| Piotr <piotrbienkowski@gmail.com> | Sat, Jan 5, 2013 at 9:37 PM |
To: tommi@tntranslations.com
|
You know, there was a time when I was reading the regexp book (the one with the owl) at bedtime. :-)
You can expect another wave of feedback when I start another automotive project.
PB
[Quoted text hidden]
|
|
Comments
Post a Comment