Add MetalLexer class with tokenization functionality

[huntermarchi]

Improves lexer performance by optimizing tokenization. This update refines the MetalLexer class and its token matching process for better efficiency.
This PR doesn’t directly fix an issue but addresses performance improvements.
closes #123

• Optimized token matching in the MetalLexer class by refining the regex patterns.
• Reduced unnecessary backtracking and improved processing speed for tokenization.
• Refined handling of whitespace characters to skip them during tokenization.
• Tested with a sample input string: "12 + 34 - 56", which produced correct tokenized output.
• Manually verified performance improvements with larger input files.
• No new automated tests were added, as the existing ones cover basic functionality.

[huntermarchi]

solve the issue please check it

[huntermarchi]

@NripeshN done and improved kindly see