[Fix] Truncate texts longer than the max_tokens

What does this MR do and why

Enable truncation when the text to detect is longer than the Detector max_tokens. Not hard-coded as the tokenizer enable to do this automatically.

MR acceptance checklist

Please evaluate this MR against this checklist :

• Linting and typing are OK,
• Licenses are OK (licensecheck + integration with reuse with the proper licenses),
• You have tested this MR locally,
• You have considered performance issues,
• You have considered availability and reliability risks,
• You have updated documentation if necessary.

How to set up and validate locally

Numbered steps to set up and validate the change.