Add Streamlit waveform annotation workbench with DNN pre-labeling

[cangyeone]

Summary

• introduce a Streamlit-based waveform annotation workbench that loads demo or uploaded NPZ events, sorts traces by distance, and provides interactive filtering, visualization, and export tools
• implement reusable modules for event data models, filtering utilities, annotation state handling, waveform plotting, and synthetic demo generation
• bundle a lightweight convolutional neural network pre-labeler to auto-suggest P/S picks for rapid analyst review
• extend the data loader, sidebar workflow, and documentation to support HDF5 uploads with event metadata sourced from file attributes

Testing

• python -m compileall waveform-annotation-app

---

https://chatgpt.com/codex/tasks/task_e_68df74257d588323a64faa39d21ed34e