Vincenzo Piro

You've got to love these localization friendly clients ;-)

Your rule is far too generic.  For example, this is what you capture with your opening expression:

This with your closing:

You are going to end up with a lot of overlap and probably error messages as well.  To create a rule specifically for this part you could try something like this:

Opening:

<img[^"]+""

Closing:

(?<=<img[^"]+"".*)""

A bit messy perhaps but it should do this:

Opening:

Closing:

The closing search on its own picks up more than needed but as this is intended to be used with the opening tag I think it'll be ok.  You could probably refine it if you really wanted to but hopefully this will give you the idea what what it means to add a rule for every single type of tag.

So I did quick test..

Also make sure you set the segmentation for the rule to exclude, and it looks like this:

So does the trick.  Now I'd just need the other rules... which I'm sure you have sorted.

Vincenzo Piro

In fact I added a few more rules and changed the one I gave you:

Looks a lot cleaner apart from the last bit which I deliberately left in:

I left it in because this simply looks wrong and if it was me I'd go back to my client and check:

 [DO">www.lotofwater.com/.../p>

I totally get your first comment about how generic my rule was, I must admit that I'm just a newbie in this.

However, I can't be more specific, as the file has a lot of lines with different <img alt=""[...]> tags.

Using the rules you gave me in first comment, this is my preview:

And these are the rules I had

So, the problem stays: the opening is ok, but the closing leaves parts that I must not translate.

Here's my idea for the closing:

The rule starts from the opening, ignores the opening rule and the text to translate, and states: "from "" to > everything must be part of the tag"

By the way, can you suggest me a good beginner manual about regex rules? :D

Edit:

About this:

I left it in because this simply looks wrong and if it was me I'd go back to my client and check:

 [DO">www.lotofwater.com/.../p>

Something went wrong during copy, I updated it in my test file.

Vincenzo Piro

Vincenzo Piro said:

Using the rules you gave me in first comment, this is my preview:

So it looks as though that rule I gave you works as it does exactly what was expected... apart from the additional bits which I think is due to a bug in Studio.  However, if you were to use the more complete rules I created in the second comment, and make sure the exclude is set, then it is much cleaner.

Vincenzo Piro said:

By the way, can you suggest me a good beginner manual about regex rules? :D

You can find a few article here and there are some tips in each one that may point you to an appropriate resource:

https://multifarious.filkin.com/tag/regex-tutorial/

You also have this manual written by Anthony Rudd which may be helpful.  I have not seen it myself but I know Anthony is very capable as he answers a lot in this community and delivered a webinar for us on the topic this Autumn:

https://appstore.sdl.com/language/app/practical-usage-of-regular-expressions-for-translators/892/

And of course the bible!!

https://www.regular-expressions.info/cookbook.html

In addition to Paul's regex reading suggestions,
I have also prepared a short introductory document for basic regex concepts:
www.facebook.com/.../4098648660192286

Dear Paul,

Sorry for the late reply.

I tried your rules: the one about  a href works, but I still have problems with the img alt tag.

I also tried a double placeholder combination (instead of the tag pair), but it doesn't seem to work.

About the first part I found that:

<img\salt="

works well, the problem is always in the second part.

May I send you via PM the original csv and xlsx files?

Thank you Anthony, very appreciated!

Vincenzo Piro

Vincenzo Piro said:

May I send you via PM the original csv and xlsx files?

Sure... pfilkin@rws.com

Vincenzo Piro

Apologies for the delayed response.  I have put an updated settings file here:

VV XLSX HUB.zip

It seems to do a fairly good job.

One thing I did note is that you are corrupting some characters in getting the CSV into Excel, for example:

I can't be sure since you gave me an XLSX that has different content to the CSV, but the problem is most likely because you need to add a BOM to the CSV before importing into Excel and then the encoding should be respected.  If the chars are already corrupted in the CSV then this won't help of course!