Claude
17c7f5033f
Decision (Q15 resolved): start with text-only via TEI + bge-m3 in Phase 3.1, prepare data shapes so Phase 3.2 (visual collection via infinity + jina-clip-v2) is a pure additive operation — no rename, no schema migration, no n8n rewiring. Concretely: - Qdrant collection renamed nexa_knowledge → nexa_knowledge_text (1024-dim for bge-m3) with modality-aware payload (modality, source_type, media_uri, graph_iri, content_hash, context). Visual placeholder schema committed alongside (qdrant_schema_visual.json, 768-dim, jina-clip-v2). - Image attachments captured in 3.1 are recorded in GraphDB as nexa:Note with nexa:modality "image" + nexa:pendingVisualIndex true; the 3.2 backfill workflow picks them up and embeds. No data lost between phases — the queue is the GraphDB itself. - RDF schema (docs/08) gains nexa:modality, nexa:mediaUri, nexa:vectorCollection, nexa:pendingVisualIndex from day one. - docs/02 roadmap split: 3.1 = text RAG (Path A), 3.2 = visual collection (Path C), 3.4 = Ontotext GraphDB. - docs/09 grows a "Phase add-on: visual collection (Phase 3.2)" section with the TEI→infinity swap, second collection create, LiteLLM second model registration, and the SPARQL-driven backfill query. - New open questions: Q16 (queue ergonomics + does SAIA already proxy an embed model?), Q17 (reuse Immich's CLIP for photo-library queries?). - docs/03 + CLAUDE.md updated so future runs use the new collection names and don't re-decide the staging.
4.0 KiB
4.0 KiB
03 — Architecture Overview
A one-page mental model. For details follow the cross-links.
Topology
┌────────────────┐
voice / typing ──────▶│ Memos │◀──── Nexa replies as comments
│ (interface) │
└───────┬────────┘
│ webhook (- [ ] / #nexa:*)
▼
┌────────────────┐
IMAP / RSS / NC ────▶│ n8n │◀──── workflows live in
Karakeep / Bluesky │ (logic) │ ./nexa-core/n8n-workflows
└───┬────────┬───┘
classify ▲ │ │ ▲ retrieve
│ ▼ ▼ │
┌──────────┐ ┌──────────┐
│ SAIA │ │ Qdrant │
│ LiteLLM │ │ vector │
└──────────┘ └────┬─────┘
│
▼
┌──────────┐
│ GraphDB │ Ontotext, SPARQL
│ (RDF) │ Phase 3.4
└──────────┘
│
▼ writes
┌──────────────────────────────┐
│ Nextcloud (Tasks, Calendar, │
│ Mail, Files / Obsidian) │
└──────────────────────────────┘
Components
| Component | Role | Where it runs (today) |
|---|---|---|
| Memos | Interface, voice input, webhook source | docker host LXC 104 → memos.nuclide.systems |
| n8n | Workflow / logic engine | docker host LXC 104 → n8n.nuclide.systems |
| SAIA / LiteLLM | Model gateway, embeddings, classification | docker host LXC 104 → ai.nuclide.systems (LiteLLM internal :4000) |
| Qdrant | Vector memory (semantic recall) | docker host LXC 104 — reuse qdrant_scientific. Collections: nexa_knowledge_text (Phase 3.1, 1024-dim) and nexa_knowledge_visual (Phase 3.2, 768-dim) |
| Ontotext GraphDB | Structural memory via SPARQL (Phase 3.4) | not yet deployed; see 09-deployment |
| TEI (HF text-embeddings-inference) | Self-hosted text embeddings, BAAI/bge-m3, Phase 3.1 |
docker host LXC 104, CPU only — swapped for infinity in Phase 3.2 to add jina-clip-v2 |
| Nextcloud | Tasks, calendar, mail, files | dedicated LXC 105 → nc.nuclide.systems |
| ntfy | Push channel for system alerts | docker host → ntfy.nuclide.systems |
| Backrest | Backup orchestration | LXC 103 |
| Zoraxy | Reverse proxy + TLS | LXC 108 (192.168.1.4:8000) |
| AdGuard DNS | Internal name resolution | LXC 102 |
| Home Assistant | Voice + house automation | VM 100 (HAOS) |
Two-pillar memory
- Qdrant answers "what is similar / relevant?" (cosine search over embeddings).
- Ontotext GraphDB answers "who, what depends on what, how is it structured?" (SPARQL over RDF).
Both pillars are queried in parallel for #nexa:ask and merged before SAIA generates the final answer. See 08 — GraphRAG architecture.
Dual-context routing
Every input is classified work or personal before any side effect (task creation, calendar write). See 04 — Integration matrix and 06 — Classification logic.