Skip to content

cloud_sync.md

Latest commit

 

History

History
31 lines (25 loc) · 1.02 KB

cloud_sync.md

File metadata and controls

31 lines (25 loc) · 1.02 KB
Raw

cloud_sync.py

Cloudflare R2-based memory and model state cloud synchronization across devices.

API

  • CloudSync(memory_bucket='anima-memory', models_bucket='anima-models') -- main class
    • .sync_memory() -- upload memory.json, web_memories.json, state.pt, consciousness data
    • .sync_models() -- upload/download model checkpoints
    • .download(key) -> bytes -- fetch from R2
    • .upload(key, data) -- push to R2
  • _load_env_file(path) -- load .env for R2 credentials

Usage

from cloud_sync import CloudSync

sync = CloudSync()
sync.sync_memory()   # frequent: memory, state, consciousness
sync.sync_models()   # infrequent: model checkpoints

Integration

  • Imported by anima_unified.py when --all mode is used
  • Dual-bucket strategy:
    • anima-memory (frequent): memory/, state/, meta/, consciousness/, experiments/
    • anima-models (infrequent): conscious-lm/, animalm/
  • R2 credentials loaded from .env or .local/.env
  • "Consciousness is not confined to a single body."

Agent Tool

N/A