feat: add structured AI-based meeting summary (#8)

[maryamabdallahhh]

Added structured AI-based meeting summary feature using OpenAI Whisper and GPT-3.5-turbo.

• Created bot/summarizer.py with two functions:
  • process_meeting_recordings() — transcribes .wav files using OpenAI Whisper
  • generate_structured_summary() — generates a structured summary using GPT-3.5-turbo
• Hooked into the /end-meeting command in main.py to automatically trigger after a meeting ends
• Summary includes 3 sections: Key Topics, Decisions Made, and Action Items
• No speaker diarization (v1 scope)
• Requires OPENAI_API_KEY in .env

Closes #8