Ilia Arutiunian 

This should be fairly straightforward.  But I think you'd benefit from reading through this article to get an idea of how to work with xml files: https://multifarious.filkin.com/2014/06/01/custom-xml/  It's nearly a decade old now but the principles are still good and you should learn how to handle this.

I created a sample file (which is always good to provide if you want help) based on the two lines you shared:

<?xml version="1.0" encoding="UTF-8"?>
<Instructions>
    <Text ClassID="1001" Text="When you open an XML, it's like a book; each tag a chapter, just take a look." />
    <Text ClassID="1002" Text="Attributes in XML are quite keen, like adjectives, they describe the scene." />
    <Text ClassID="1003" Text="Elements are friends, not food, nested neatly in a hierarchical mood." />
    <Text ClassID="1004" Text="Remember to close every tag you write, or the XML tree won't grow up right." />
    <Text ClassID="1005" Text="In the land of XML, 'well-formed' is king, ensuring your markup doesn't lose its zing." />
    <Text ClassID="1006" Text="XML is like a recipe, ingredients and steps in perfect harmony." />
    <Text ClassID="1007" Text="Do not fret if XML looks complex; with a bit of practice, you'll be coding in reflex." />
    <Text ClassID="1008" Text="Like a secret code without a Rosetta, XML seems tough until you getta." />
    <Text ClassID="1009" Text="With XML, you can craft a tale, of data and nodes that never go stale." />
    <Text ClassID="1010" Text="Schemas and DTDs are the maps, guiding your tags to avoid mishaps." />
</Instructions>

I created two rules in my custom XML that look like this:

The image shows two parser rules:

1. /Text/@Text which is set to "Always translatable".  This rule indicates that the content of the Text attribute within the Text element is always to be considered for translation.
2. //* which is set to "Not translatable".  This is a catch-all XPath expression that selects all elements in the XML document and marks them as not translatable.  This rule is processed first and ensures nothing else is parsed for translation, only the rules above it.

That gets me this:

Note that the Document Structure column on the right says ATT.  This is because these are all attribute values and better localization practice would be to not use attributes for translatable content... but nobody told the content producers!

Thank. I tried, but, unfortunately, I've had this.

Settings

Thank. I tried, but, unfortunately, I've had this.

Settings