Breaking down the text into semantic parts?

Question

Algorithms

Is there a service or algorithm that can break down a sentence into semantic parts?

Answer 1

The only true option here: building a "tree" of connections between parts of speech (within one sentence), collected from many different texts with the help of a person using specific reference examples.
That is, the algorithm needs to clearly know: where which part of speech can be located inside the sentence (with the highest degree of probability) relative to other parts of speech (located around, before, after, or absent altogether).
Submitting a text for analysis: "The server did not turn on because of a broken power supply."
We send to the input: "Why didn't the server turn on?"
You should receive: "The power supply was broken."
or
Send to input: "
Should receive: "Yes."
"This is the house that Jack built":
1. "The house" ("here" - we delete it, because it is a particle)
2. "The house was built by Jack" ("which" is replaced by the control noun "House")
PS: By links from Mikhail have only a simple template / dictionary search using rules and regular expressions - this is not a solution to the problem in the form in which it is possible to achieve the correct work and the correct result.

Answer 2

Before breaking it down into semantic parts ("semantic" parsing), it makes sense to learn how to parse it, as a result of which the sentence will be presented in a more "computer-readable" form. There are a lot of algorithms for this and that, implementations too, literature has its roots deep in the last century. You can start looking here:
https://habrahabr.ru/post/148124/
https://nlpub.ru/Text
processing otipl.philol.msu.ru/~soiza/testsynt/files/info-par...
https:/ /tech.yandex.ru/tomita/doc/dg/concept/inter...
Programs for extracting information (facts) from text
https://dmkpress.com/catalog/computer/databases/97...

Answer 3

To the breakdown into semantic parts of the task
has nothing to do
We count the number of syllables (vowels)
And for each person we calculate the maximum number of vowels per breath
We break the text
In the general case, the task of breaking the text into "semantic parts" is not feasible