Files

T

retoorandClaude Sonnet 5 9dbd1fe667 Add plexer subsystem, container test infra, and last-resort opencode box backend

- Plexer: always-active terminal multiplexing. NativePlexerController/
  TmuxPlexerController share one interface, chosen by which env var is
  already set (TAI_PLEXER_SOCKET vs TMUX_PANE); NativePlexerBackend is
  the pty.fork()-based wrapper (raw mode, SIGWINCH resize forwarding,
  a Unix-socket control protocol, continuous session logging, a
  Ctrl+A prefix key distinct from tmux's own Ctrl+B). Auto-wraps new
  interactive shells via a .bashrc hook mirroring the existing
  command-not-found install pattern, gated on $TMUX/$TAI_PLEXER_ACTIVE/
  a $HOME/.tai-no-plexer bypass. New plexer_send/read/type/status
  tools, env-gated on a detected controller.
- tests/container.py + docker-compose.yml + tests/docker/Dockerfile:
  a disposable podman/docker image for the tests that must never touch
  a real machine (installing the plexer hook into a real ~/.bashrc, a
  live tmux session). Required suite (test_seal.py, test_tai.py,
  tests/plexer.py, tests/bashrc.py) gates the result; tests/rag.py
  runs as best-effort since its real-LLM dependency can't be satisfied
  inside a disposable container with no nested engine.
- tests/bashrc.py: real end-to-end hook install + bash -i auto-wrap,
  anti-recursion, tmux-active skip, non-interactive skip, a live
  TmuxPlexerController test, and an install/uninstall round-trip.
- tests/plexer.py: NativePlexerBackend under a real pty, plus a
  dedicated robustness pass (file permissions, >64KB payload framing,
  a malformed control request, a genuine self-deadlock fix in the
  write-to-child path).
- tai_test.py: the shared async process-driving framework (pipe/pty
  transport, group-wide kill-and-reap, a StdinBridge for --interactive)
  both test files build on.
- ChatClient last-resort backend: when there's no local opencode CLI
  but the sandbox box is already running, install opencode into it
  (official Node tarball, not apt -- apt's nodejs/npm pull in
  webpack/babel/jest as transitive deps) and route through it instead
  of giving up. Only reachable after the first retry round already
  failed everything else; never creates the box itself. TmuxPlexerController.status()
  fixed to use `list-panes` instead of `display-message -p`, which
  silently fell back to the default pane instead of erroring against
  a dead target.
- Fix 4 pre-existing test_tai.py bugs surfaced by running it in a
  clean environment for the first time: two stale mock signatures
  (subprocess.run's stdout= kwarg, Chat.complete's deadline= kwarg),
  one $TMUX-dependent catalog assertion, one tty-dependent color
  assertion.
- research/: a sourced research archive (Source Dossier/Citation
  Card/Conclusion tiers) backing the above design decisions, with its
  own audit.py consistency checker.
- .gitignore: fix a blanket *.md rule that was about to silently drop
  the entire research/ archive from version control; add __pycache__/
  and the one-off debug transcript this session started from.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-10-10 04:36:05 +02:00

2.3 KiB

Raw Blame History

What Is Reciprocal Rank Fusion?

Metadata

Field	Value
Institute	Molodetz Institute — Research & Retrieval Division
Subject	sqlite-rag-hybrid-search
Document type	Source Dossier
Tier	1
Source title	What is Reciprocal Rank Fusion?
Source URL	https://www.paradedb.com/learn/search-concepts/reciprocal-rank-fusion
Source publisher / author	ParadeDB (vendor "learn" page)
Source type	vendor docs
Source date (published)	Unknown — not stated
Date researched	2026-10-09
Researched by	tai (agent), model: Claude Sonnet 5, on behalf of retoor
Research method	web search synthesis
Confidence	Medium
Status	Active

Summary

Covers the same RRF mechanism as the other two RRF sources in this subject, but is the specific source for RRF's main documented limitation: because RRF only uses rank position, it treats a document that is overwhelmingly the best match in one list only slightly better than the runner-up — the magnitude of the win is discarded, not just normalized.

Key findings

RRF's score-magnitude blindness is named explicitly as the technique's main limitation, with the suggestion that a score-aware fusion or a post-RRF reranker can help when score gaps genuinely matter.
Corroborates (independently of spiceai-reciprocal-rank-fusion.md) the standard formula and the k=60 convention.

Pros / strengths

The clearest, most direct statement among the three RRF sources of RRF's actual weakness, rather than treating RRF as a strictly-better free lunch.

Cons / limitations / caveats

Vendor "learn" page (ParadeDB, a Postgres-search company); the limitation is stated plainly but no quantitative example or benchmark accompanies it in what was gathered during this research pass.

Relevance to tai

Directly informed the final proposal's framing (in the separate agent-memory-recency-tracking subject, not this one) that if tai ever revisits combining signals — e.g. lexical BM25 plus some future recency/frequency signal — it should go in eyes-open about RRF discarding score magnitude, which matters more for a small, low-cardinality signal like recency than for two large ranked lists. See conclusions/no-vectors-bm25-chunked-rag.md.

Changelog

Date	Change
2026-10-09	Initial dossier created.

2.3 KiB Raw Blame History