As the non-final element of Indic compounds, จักร- is enunciated tɕàk - krà -

The vowel lengths for เมตร and เพชร definitely need to be marked as exceptions.

I greatly appreciate your feedback. I will add those exceptions.

Did you mean falling tone for เมตร, not the vowel length?

As per จักร - yes, the rule insists that ไม้หันอากาศ is always followed by a final.

However, I think it's rather reduplication issue. And the biggest problem with reduplication is that it usually occurs in the middle of polysyllabic word and does not occur at the end (วัฏจักร vs จักรภพ). Boundaries of meaningful words can be detected only by dictionary-based tool, of course.

I cannot think of any effective algorithm, and I would be obliged if someone gave me an idea how to overcome this.

If one navigates away from the screen (at least, backwards), the tree has collapsed when one returns.

Yes, it's a common problem of stateless controls. I will try to find workaround, and I'm also thinking on simplifying the tree (it's redundant a lot).

Did you mean falling tone for เมตร, not the vowel length

Both เพชร and เมตร have short vowels and high tone. If you mark the vowel as short in your exception dictionaries, you will get the high tone automatically if you use the vowel length after exception look-up to determine tone. For example, would you take mai tho on a dead syllable with low class initial as indicating a long vowel if the vowel length is otherwise unmarked?

Boundaries of meaningful words can be detected only by dictionary-based tool, of course.

It depends on your use case, but there is some benefit from detecting the boundary at the end of the input!