After detecting bird species in an audio recording, extract and save a smaller .wav file snippet for each individual detection.
Each snippet should correspond to the start_time and end_time of the detected sound.

Requirements:

• Use an audio processing library like pydub or soundfile to handle audio slicing.
• Handle edge cases:
  • Start or end times outside the actual audio length.
  • Overlapping detections.
• Normalize audio snippets if needed
• For MVP, store snippets locally (e.g., under /snippets/ folder).
• Later, uploading snippets to cloud storage (like S3) can be added.