Refactor audio generation in NotebookMg to write directly to output file, eliminating the need for intermediate AudioSegment. This change improves performance and simplifies the audio processing workflow.

Refactor podcast segment regeneration to remove pydub dependency. Directly concatenate MP3 files for improved performance and simplicity.

cookies removed

remove ffmpeg installation from Dockerfile and add ffmpeg-python to requirements

readme updated

first commit

initial commit