Own address DB + cache improvements (reduce provider dependency) #2

Activity

1. coderabbitai commented on Dec 28, 2025

   

   coderabbitaibot

   on Dec 28, 2025 – with coderabbitai

   Last edited by coderabbitai

   More actions

   📝 CodeRabbit Plan Mode

   Generate an implementation plan and prompts that you can use with your favorite coding agent.

   • Create Plan

   Examples

   • Example 1
   • Example 2

   ---

   🔗 Similar Issues

   Related Issues

   • SDK: Core JS + React/Vue/Svelte + Web Component #1

   👤 Suggested Assignees

   • BleedingDev

   ---

   🧪 Issue enrichment is currently in open beta.

   You can configure auto-planning by selecting labels in the issue_enrichment configuration.

   To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

   issue_enrichment:
     auto_enrich:
       enabled: false

   💬 Have feedback or questions? Drop into our discord!
2. BleedingDev commented on Dec 28, 2025

   

   BleedingDev

   on Dec 28, 2025

   ContributorAuthor

   More actions

   Technical Notes

   Current Foundations

   • SQLite DB + migrations: apps/service-bun/src/sqlite.ts
   • L1/L2 cache + SWR: apps/service-bun/src/cache.ts
   • Search logging: apps/service-bun/src/search-log.ts

   Proposed Data Model

   • Canonical address table (addresses) + provider-specific table (address_sources) so one canonical address can have multiple sources.
   • Conservative dedupe key (“canonical key”) computed from normalized address parts; allow null when data is insufficient.
   • Optional full-text search (SQLite FTS5) over labels + address parts to serve suggestions from local DB.

   Suggested Runtime Flow

   1. L1 cache
   2. L2 cache (SQLite)
   3. Local address DB search (FTS/prefix search)
   4. External provider(s) as fallback
   5. Persist newly observed addresses asynchronously (do not block response)

   Acceptance Signal

   • Add an acceptance endpoint (POST /accept) to record which suggestion was selected.
   • Increment acceptance counters to improve ranking/confidence for future results.

   Ownership + Reuse

   • Add shared types + normalization helpers in packages/core (e.g. computeCanonicalKey, canonical address record type) so other services can reuse the model.

   Safety/Privacy

   • No user identifiers required.
   • Treat session tokens as short-lived correlation only (store only where needed for analytics, not in the canonical address DB).
3. BleedingDev

   added

   epicEpic (parent issue)Epic (parent issue)

   priority:p2Post-MVP / laterPost-MVP / later

   area:cache-dbCaching / address databaseCaching / address database

   on Dec 28, 2025