Constructor of custom subtitles for the user from the raw model response (in json)

[LebedevIV]

Constructor of custom subtitles for the user: to create the necessary fields for subtitles from the raw model response (in json), specifying the type template. Using an example to create Enhanced LRC format A2 (https://en.wikipedia.org/wiki/LRC_(file_format)#A2_extension:_word_time_tag):
If the raw answer is of the form:

{
"language": "en",
"text": "Hello World",
"segments": [{
"start": 0.0,
"end": 3.0,
"text": "Hello World",
"whole_word_timestamps":
[
{"word": "Hello", "start": 0.0, "end": 1.5, "timestamp": 1.5, "probability": 1.0},
{"word": " World", "start": 1.5, "end": 3.0, "timestamp": 3.0, "probability": 1.0},
]
}],
"diarization": [
{"startTime": 0.0, "stopTime": 3.0, "speaker": "SPEAKER_00"}
]
}

then the correspondence of the chunks will be approximately like this:
[mm:ss.xx] mm:ss.xx line 2 word 1 mm:ss.xx line 2 word 2 mm:ss.xx ... line 2 last word mm:ss.xx
[{segments.start}]{diarization.speaker}:<{segments.whole_word_timestamps.start}>{segments.whole_word_timestamps.word}

It is assumed that arrays of segments and words will be processed by sequentially outputting a given pattern for one line and word