Make bracket matching more robust for typeahead

[dingfeli]

Currently, the mechanism in typeahead for bracket matching is rather rudimentary:

amazon-q-eclipse/plugin/src/software/aws/toolkits/eclipse/amazonq/util/IQInlineSuggestionSegmentFactory.java

Lines 95 to 110 in e07722f

	private static boolean isCloseBracket(final String input, final int idx,
	final Stack<QInlineSuggestionOpenBracketSegment> unresolvedBrackets) {
	char c = input.charAt(idx);
	boolean isBracket = c == ')' || c == ']' || c == '}' || c == '>' || c == '"' || c == '\'';
	if (!isBracket) {
	return false;
	}
	if (c == '"' || c == '\'') {
	return !unresolvedBrackets.isEmpty() && unresolvedBrackets.peek().getSymbol() == c;
	}
	// TODO: enrich this check to eliminate false positives
	if (idx > 0 && Character.isWhitespace(input.charAt(idx - 1)) && c == '>') {
	return false;
	}
	return true;
	}

As a result, the typeahead session is susceptible to having erroneously assess whether or not a character is a part of a bracket pair.

We need a more robust mechanism to parse the suggestion text to provide a more accurate typeahead experience.