Refactor audio generation prompt handling to remove system prompt from URL encoding

Refactor voice reply generation to include system prompt customization; add helper function to build system prompt and update encoded prompt formatting.

Remove unused pre_process import and commented-out system prompt logic from generate_voice_reply function

Refactor generate_voice_reply to use the original prompt instead of appending system prompt for audio generation

Refactor system prompt generation in generate_voice_reply function to use pre_process for improved clarity and maintainability.

Increase random seed range to enhance variability in image and trivia generation across multiple modules

Comment out example usage in main block for both FLUX.py and VoiceReply.py

Add random seed to voice reply generation and enhance image sending function with customizable caption

Update audio and image directory paths to use temporary directories

debuggings

flux & audio reply