Normalize Recruiter Transcript Text

Problem

Given a list of transcript strings, write a Python function clean_transcripts(lines) that returns a new list of cleaned strings. For each line, convert text to lowercase, trim leading and trailing whitespace, collapse repeated internal whitespace into a single space, remove filler tokens "um", "uh", and "like" when they appear as whole words, remove non-alphanumeric characters except spaces, and discard empty results. After cleaning, remove duplicate lines while preserving the first occurrence order.

Constraints

• 1 <= len(lines) <= 10^4
• 0 <= len(lines[i]) <= 10^3
• Each input element is a string
• Filler removal is case-insensitive and applies only to whole words