Plan data handling & persistence

[ebowwa]

Proposed Issue: Data Handling & Persistence Strategy

Summary

• Capture decisions and open questions around persisting ring metrics beyond the current in-memory model.
• Document constraints from ring hardware (packet cadence, battery cost) and how they influence storage choices.
• Define milestones for introducing optional persistence once experimentation stabilises.

Background

• The SwiftUI app currently holds RingMetrics in memory only; no Core Data or external store is used.
• We now have an expanded “Sensor Capabilities & Application Opportunities” section describing what data we may want to keep.
• Legacy tooling (e.g., RingCLI, Home Assistant integrations) shows demand for historical logs and cross-device automations.

Goals

1. Evaluate lightweight persistence mechanisms (Core Data, SQLite, files, CloudKit) suitable for iOS-first scope.
2. Decide on data granularity to retain (raw packets vs. parsed RingMetricsUpdate events vs. aggregated summaries).
3. Outline migration path that keeps experimentation friction low while enabling opt-in storage later.

Non-Goals (for now)

• Shipping a full persistence layer in the current milestone.
• Implementing background sync, multi-device sharing, or cloud services.
• Guaranteeing compatibility with legacy firmware mods beyond our documented command set.

Proposed Tasks

• Audit metrics that need historical storage vs. session-only visibility.
• Spike Core Data model sketch for RingSession, RingSample, etc.; compare against alternatives.
• Draft data-retention policy (sample rates, max history, export formats).
• Identify testing strategy (unit tests for parsers, integration tests for persistence writes/reads).
• Capture battery and performance implications of realtime streaming + storage.

References

• COLMI_RINGS_COMPREHENSIVE_DOCUMENTATION.md — new sensor/application overview section.
• ring/ring/RingResponseParser.swift — source of parsed RingMetricsUpdate events.
• Legacy/ATC_RF03_Ring — demonstrates firmware tweaks that may affect data cadence.