`RegexTextExtractor` - set `return_empty_on_no_match` default value to `False`

[anakin87]

In #10152, the return_empty_on_no_match init parameter was added to RegexTextExtractor with a default value of True.

We also added a DeprecationWarning:

haystack/haystack/components/extractors/regex_text_extractor.py

Lines 93 to 95 in 43e125d

	"Warning: In an upcoming release, the output when no matches are found will change from "
	"'{}' to {'captured_text': "
	"}"

While I would have preferred to handle #10105 with a small breaking change and always return {"captured_text": ...}, we now have to follow our deprecation policy and set return_empty_on_no_match default value to False on 2.22.0.
We should also remove the captured_texts=list[str] output edge, which is never populated.

[GunaPalanivel]

Hi @anakin87! I'd like to work on this issue. The changes seem straightforward:

1. Change the default value of return_empty_on_no_match from True to False
2. Remove the deprecation warning
3. Clean up the unused captured_texts output edge

I can have a PR ready soon. Would that be okay?

[anakin87]

@GunaPalanivel @majiayu000 please do not open PRs for this issue. I'd like to discuss this better internally first.

[anakin87]

Another cleaner idea I had:

• remove return_empty_on_no_match and always return {"captured_text": ...}: in case this parameter is used, we can find a way to ignore it and warn the user
• remove unused captured_texts=list[str] output edge

@sjrl WDYT?