memory: snapshot — CSM teardown, Chatterbox v2 + Fish v6 reworks, undefined-bug root cause, dia2 restart; archive 8 (2026-05-27) entries

2026-06-01 15:07:55 -07:00
parent 38d9e3b7aa
commit 7f9dc2dcf1
2 changed files with 224 additions and 220 deletions
@@ -174,6 +174,25 @@ _Entries moved out of persistent-memory.md to keep the active file scannable. Re
  to PATCH (which silently 2xx'd on the pre-fix Worldtree).
  _Archived 2026-05-31._
 - `[2026-05-27]` Worldtree v0.29.x landed full saga→echo refactor +
  config-validator hardening (v0.29.10 create_provider family-before-regex;
  v0.29.11 collect-then-raise echo startup validators; v0.29.12 ships
  providers.yaml.example/defaults.yaml.example canonical configs; v0.29.13
  reasoning_content extraction + catalog family lookup). Operator-asked,
  worldtree-dev-shipped, CI-deployed.
  _Archived 2026-06-01._
 - `[2026-05-27]` artemis-31b-v1i added to llama-swap + worldtree personal.
  BeaverAI Gemma 4 31B Q6_K (~28.6GB), 128K ctx, `--reasoning-format deepseek`
  (gemma format unsupported in deployed llama.cpp). Worldtree catalog
  `family: gemma` so GemmaProvider routes reasoning tokens.
  _Archived 2026-06-01._
 - `[2026-05-27]` Skaldsong streaming TTS v0.32.0→v0.32.2: chunked-batch SSE
  (one Kokoro POST per paragraph); defensive event_stream catch-all; NDJSON
  parsing for Kokoro /dev/captioned_speech multi-line responses.
  _Archived 2026-06-01._
 ## Tried and abandoned (archived)
 - `[2026-04-30]` task-board workflow with
@@ -383,3 +402,32 @@ _Entries moved out of persistent-memory.md to keep the active file scannable. Re
  substring replace on YAML top-level keys WILL match suffix-containing
  keys.
  _Archived 2026-05-31._
 - `[2026-05-27]` `docker compose up -d` inside the `docker:cli` sandbox:
  `${VAR:-./config}` defaults resolve `./config` to the sandbox CWD, but the
  Docker daemon interprets the path against the HOST filesystem → auto-creates
  an empty dir → entrypoint reseeded image-baked defaults (lost host-side
  providers.yaml patches). Fix: pass `-e WORLDTREE_CONFIG_DIR=/abs/path`.
  Folded into the docker-as-root convention note.
  _Archived 2026-06-01._
 - `[2026-05-27]` `:latest`-pinned compose + private gitea registry + sandboxed
  pull = recreate on ancient cached `:latest` (deploy pulls by SHA so the tag
  never advances; sandbox can't pull). Fix: retag SHA→`:latest` on host, then
  `--pull never`. Better: pin SHA in `.env`, advance in CI.
  _Archived 2026-06-01._
 - `[2026-05-27]` Container recreate during in-flight skaldsong gen kills the
  runner. With deploys every ~10min and stories >5min, structural not
  incidental. Roadmap (skaldsong-dev): pre-shutdown signal handler, per-scene
  resume-from-checkpoint, /api/admin/quiesce. None shipped.
  _Archived 2026-06-01._
 - `[2026-05-27]` `--reasoning-format gemma` on artemis-31b-v1i — unsupported in
  the deployed llama.cpp (accepts none|deepseek|deepseek-legacy). `deepseek`
  populates the `reasoning_content` SSE delta Worldtree GemmaProvider checks.
  _Archived 2026-06-01._
 - `[2026-05-27]` `head -c N` piped after a streaming curl SIGPIPEs the curl,
  killing the request early. Use file-write + separate read.
  _Archived 2026-06-01._
@@ -1,6 +1,6 @@
 # Persistent memory — eshpfi-management
-_Last updated: 2026-05-31_
+_Last updated: 2026-06-01_
 ## Repo purpose
@@ -77,84 +77,111 @@ Sister repos (separate gitea repos, deployed by playbooks here):
  resulting path against the HOST filesystem. Always pass `-e VAR=/abs/path`
  to the docker run invocation for any relative-default config dir.**
 - **`scripts/elway` sudo handling** — elway prompts for the sudo password
  ONCE via `getpass` before the first `sudo: true` step. That prompt is
  interactive → elway can't run unattended from a non-TTY tool if any step
  needs sudo. For sudo-free playbooks (no `sudo: true` steps) it runs fully
  non-interactive over key SSH. To create root-owned dirs WITHOUT host sudo,
  use the docker-daemon-root trick: `docker run --rm -v /worktank:/mnt alpine
  sh -c 'mkdir -p /mnt/<x> && chown -R 1000:1000 /mnt/<x>'`.
 ## Current state / in-flight
-_As of 2026-05-31:_
+_As of 2026-06-01:_
- **Dia2 LIVE as two instances on irv-ml1; zonos + csm still down.**
+- **TTS fleet on irv-ml1 reworked this session; asset-engine catalog now
-  `dia2-2b` (:8200, Dia2-2B) + `dia2-1b` (:8202, Dia2-1B) running +
+  17 services** (CSM removed). Canonical `docs/asset-engine/services.yaml`
-  healthy on GPU 0 from `local/dia:v2`; catalog `status: ready` (both
+  + vendored `vh/asset-engine` both pushed (eshpfi `38d9e3b`; asset-engine
-  exercised). Legacy Dia 1.6B retired. **zonos**: REST adapter built
+  `0b552bd`). Per-engine live state:
-  (`stacks/zonos/adapter/`, `local/zonos-api`) but the stack is DOWN —
+  - **dia2-2b (:8200) + dia2-1b (:8202)** — LIVE + healthy (restarted this
-  brought down to free 3090 VRAM for Dia2; catalog entry `status: down`,
+    session; `local/dia:v2`, voices + cache intact). `status: ready`.
-  adapter port 8203. **csm**: still PARKED (`status: down`, gated — needs
+  - **Chatterbox (:8196)** — LIVE. Catalog **v2**: routes to `/tts`, exposes
-  `CSM_HF_TOKEN` before first start; non-reproducible → `experimental`
+    emotion levers (exaggeration 0.5, cfg_weight 0.5, temperature 0.8), 28
-  when up). ComfyUI `:8188` still a catalog-deferred UI **link** (open
+    built-in predefined voices (default `Emily.wav`) + clone. Fast + competent
-  hand-off to asset-engine-dev, no form-field API).
+    cloner (it audibly cloned Imogen). Plain `Imogen.wav` uploaded to its clone
-
+    dropdown (`/get_reference_files` now: Imogen.wav + glados_25s.wav).
- **Dia2 voices staged — re-stage if host rebuilt.** The 43 curated
+  - **Fish-S2 (:8195)** — LIVE but **HEAVY**: ~240s torch.compile warmup on
-  voices baked at `/app/voices` aren't reachable from `/tts`'s clone path,
+    cold start, ~24 GB on the A6000; first call after any restart eats the
-  so they're copied into `/worktank/dia/reference_audio` (restic-included).
+    warmup. Catalog **v6**: full `/tts` ServeTTSRequest surface
-  This is a **one-time manual host step** (`docker exec dia2-2b sh -c 'cp
+    (temperature/top_p/repetition_penalty/latency/seed/format) + `reference_id`
-  -n /app/voices/* /app/reference_audio/'`), documented in
+    is a 32-voice static "Voice" dropdown (default Emily) + advanced base64
-  `stacks/dia/README.md`, NOT auto-run on recreate. Catalog voice default
+    clone. NOT OpenAI-compat (only `/v1/tts`, no `/openapi.json`, no `/voices`
-  is `voice_mode: clone` + `clone_reference_filename: Abigail.wav`. Open
+    API). Research/non-commercial license.
-  offer (not done): make staging auto-run via container entrypoint.
+  - **zonos (:8203)** down (adapter built, not deployed). **ComfyUI :8188**
-
+    catalog-deferred link.
- **asset-engine catalog fully synced + pushed.** `docs/asset-engine/
+- **BLOCKED on asset-engine-dev — P1 "undefined" select bug** (default voice
-  services.yaml` (canonical) and `vh/asset-engine`'s vendored
+  picks send the literal string `"undefined"` → Chatterbox 404s, Fish falls to
-  `data/services.yaml` (re-synced via `sync_catalog.py`, commit
+  default voice). ROOT-CAUSED: the form's Kokoro voice-blend widget
-  `58dadfe`) both on origin; catalog v1, 18 services; schema regenerated.
+  (`templates/partials/_fields.html`) runs `recompute()` at page load, reading
-  asset-engine deploy-on-push triggered. asset-engine-dev notified on
+  the Shoelace `<sl-select>.value` BEFORE hydration → writes "undefined" into
-  althing thread `01KSZM5MNQYP2DZ25Y2F2DXDBY` of the breaking dia→dia2
+  the submit field. Escalated with full root cause + 3-part fix (althing msg
-  swap (drop dia tile; add dia2-2b/dia2-1b/zonos), the `/tts` field
+  `01KT2K2SY9N7AY69R9V0B4RXSW` → asset-engine-dev). **Workaround until fixed:
-  shapes, and a UI-hardening ask (don't serialize empty optional selects
+  click the dropdown (fires change) or hit the API directly.** My half
-  as the string `"undefined"`; enforce clone_reference_filename
+  (`blendable: false` catalog flag) is queued — see Recent decisions.
-  required-when-clone).
+- **Fish cloning verdict UNRESOLVED — re-test parked.** My in-session "Fish is
-
+  a weak cloner" conclusion was confounded (bogus reference transcript + maybe
- **ttyd fleet driver seat operational on nh3-dev** (browser/iPad/Vision
+  `reference_id: "undefined"` from the form bug + weak resemblyzer encoder) —
-  Pro into the zellij `Claude` session): RO `:7681` / drive `:7682` via
+  see Tried and abandoned. Literature says Fish/OpenAudio is SOTA cloning.
-  a Caddy front, network-gated (no auth). Lives in `~/.config` on
+  Whisper has the real transcript ("If the red of a second bow…"). Clean
-  nh3-dev (out of this repo). Full wiring + gotchas → auto-memory
+  re-test (correct transcript + ECAPA scoring) ready when wanted. Future voice
-  `reference_ttyd_fleet_seat`.
+  clones are operator-handled (pitch-shift deepening abandoned).
-
+- **On-host consented voice library** — ~992 real-person clips cached in the
- **Worldtree all instances healthy on v0.29.13** (last-known, not
+  kyutai tts-voices repo (`/worktank/kyutai-tts/.../snapshots/.../`): VCTK
-  re-checked this session). Demo (:8080), personal (:8081), pinned
+  (CC BY 4.0, accent-tagged speaker IDs), Unmute voice-donations (CC0), EARS +
-  (:8082). v0.29.x = saga→echo refactor + selene-judgment validator +
+  expresso (CC BY-NC). Source for future clone voices. British-female Southern
-  canonical `*.yaml.example` configs.
+  England speakers p225/p228/p229 staged into Fish as Imogen/Eleanor/Beatrice.
-
+- **Parakeet ASR (:8765) is DOWN** — no container, but catalog says
- **Skaldsong on v0.32.2** (ana-docker:8300). Kokoro TTS streaming via
+  `status: ready` (drift). Use Whisper-in-a-container (`faster-whisper`) for
-  SSE working end-to-end; default voice `af_heart`. Last-known.
+  transcripts instead; it's self-contained (CTranslate2, no torch).
-
+- **Worldtree healthy v0.29.13** (last-known); **Skaldsong v0.32.2**
- **artemis-31b-v1i live on llama-swap + worldtree personal.** BeaverAI
+  (ana-docker:8300, Kokoro SSE streaming); **artemis-31b-v1i** live on
-  Gemma 4 31B Q6_K (~28.6GB, 128K ctx), reasoning via
+  llama-swap + worldtree personal; **ttyd fleet driver seat** on nh3-dev
-  `--reasoning-format deepseek`.
+  (RO :7681 / drive :7682 via Caddy) — all out of this repo, last-known.
-
+- **Disclosed-keys hygiene queue** — rotate at convenience: HF token `hf_HBl…`
- **All commits pushed (operator-authorized 2026-05-31).** The prior
+  (lkraven's HF account) leaked into BuildKit logs during the CSM build attempt
-  pending-push backlog + this session's TTS/catalog work are all on
+  (logs shredded, never committed — low urgency);
-  origin (eshpfi-management through `5c47843`; asset-engine through
+  `/tmp/wt-personal-skaldsong-prod.key` on nh3-dev; mead-hall's prior Worldtree
-  `58dadfe`). No local-only backlog. (Standing rule still holds: never
+  bearer (superseded by `a360822d`); Worldtree `Z_AI_API_KEY`/`ZAI_API_KEY`;
-  push without an explicit greenlight — this push was granted.)
+  chamber `forseti`/`agent_runner` api_keys (superseded by `50d85460`); Gitea
-
+  runner registration token (`a1135753…`).
- **VibeVoice catalogued for skaldsong #73 contract** (TTS alongside
+- **Still open from prior sessions:** rotate `MINIFLUX_PASSWORD` (leaked
-  Kokoro). Endpoint `http://10.100.79.3:8194`, OpenAI-compat. NO
+  twice); clean up legacy `news-digest` detritus on ana-docker; watch nh3-nas
-  per-word timings (diffusion); read-along is a v2 forced-alignment
+  `/volume1` (was 65%; recheck before ~80%); the `docker push 60s client-side
-  upgrade (Parakeet irv-ml1:8765 or Whisper).
+  ceiling` mystery remains uninstrumented.
 - **Disclosed-keys hygiene queue** — rotate at convenience:
  `/tmp/wt-personal-skaldsong-prod.key` on nh3-dev (shred after verify);
  mead-hall's prior Worldtree bearer (superseded by `a360822d`);
  Worldtree `Z_AI_API_KEY`/`ZAI_API_KEY` (2026-05-12 outage); chamber
  `forseti`/`agent_runner` api_keys (superseded by `50d85460`); Gitea
  runner registration token (`a1135753...`).
 - **Still open from prior sessions:** rotate `MINIFLUX_PASSWORD`
  (leaked twice); clean up legacy `news-digest` detritus on ana-docker;
  watch nh3-nas `/volume1` (was 65%; recheck before ~80%); the
  `docker push 60s client-side ceiling` mystery remains uninstrumented.
 ## Recent decisions
 - `[2026-06-01]` **CSM (Sesame csm-1b) torn down entirely** — removed from
  catalog, `stacks/csm/`, `playbooks/deploy-csm.yaml`, and host
  (`c54ab13`). Two reasons: (1) deep-research verdict — the acclaimed
  Maya/Miles demo is a fine-tuned, larger variant Sesame never open-sourced;
  the open 1B base ships no usable voices, can't generate text, English-only,
  no real-time streaming, absent from leaderboards, dominated by
  Kokoro/Dia2/Fish/IndexTTS. (2) The upstream image is rotted (see Tried and
  abandoned). asset-engine re-synced (`306aae9`).
 - `[2026-06-01]` **Chatterbox catalog v1→v2** (`93cedd6`) — re-point to the
  devnen wrapper's `/tts` (was the thin OpenAI `/v1/audio/speech` exposing
  none of the emotion knobs → poor output). Default `exaggeration: 0.5` per
  Resemble's README ("works well for most"); the server SHIPS demo-tuned 1.3
  which over-acts — deliberately NOT adopted. 28 predefined voices via
  `/get_predefined_voices` (default Emily.wav). Corrected seedable false→true.
 - `[2026-06-01]` **Fish-S2 catalog v1→v6** (`f3c260e`..`38d9e3b`) — was a
  blank 2-field form; now exposes the full fish-speech ServeTTSRequest surface
  + a 32-voice `reference_id` "Voice" dropdown (STATIC list — Fish has no
  /voices API). `reference_id`-by-name is the WORKING clone path; the
  path-form `references` shape 500s (see Tried and abandoned). Corrected vram
  4→24 GB, cold_start 8→240s, seedable false→true. Added 3 consenting
  British-female VCTK voices (Imogen=p225/Eleanor=p228/Beatrice=p229).
 - `[2026-06-01]` **`blendable: false` catalog flag — DEFERRED**, pending
  asset-engine-dev honoring it in the form template. Clone-voice selects
  (chatterbox/fish/dia2) must render as plain dropdowns, not the Kokoro blend
  widget (blending reference clones is nonsensical). I ship the catalog flag +
  schema field once they read it. Tracked: althing msg
  `01KT2K2SY9N7AY69R9V0B4RXSW` to asset-engine-dev.
 - `[2026-05-31]` Dia2 deployed as **two fixed-model instances**
  (`dia2-2b` :8200, `dia2-1b` :8202) from `local/dia:v2`, retiring
  legacy Dia 1.6B; catalog `dia` entry **removed → dia2-2b + dia2-1b**
@@ -185,171 +212,100 @@ _As of 2026-05-31:_
  Pydantic already supported both — schema file was just behind.)
 - `[2026-05-31]` TTS bench expanded with `stacks/{dia,zonos,csm}`
-  (`666f7f3` dia+zonos, `a4b8c2a` csm): dia = Nari dialogue TTS via
+  (`666f7f3` dia+zonos, `a4b8c2a` csm). The bench already had Fish
  devnen/Dia-TTS-Server (OpenAI-compat, :8200), zonos = Zyphra Gradio
  audition surface (:8199, no API), csm = Sesame conversational
  (gated model, :8201). dia + csm catalogued in
  `docs/asset-engine/services.yaml` `status: down` + repro audits
  (`8eac39b`); all three deployed PARKED to irv-ml1 (no GPU spin).
  Zonos + ComfyUI = catalog-deferred (no form-field API) → asset-engine
  UI **links**, handed to asset-engine-dev. The bench already had Fish
  S2-Pro / Chatterbox-Turbo / IndexTTS-2 / CosyVoice3 / Kokoro /
-  VibeVoice / Qwen3-TTS / Kyutai, so the research-agent's top-2 picks
+  VibeVoice / Qwen3-TTS / Kyutai. (csm since removed 2026-06-01.)
  were already deployed.
 - `[2026-05-31]` Remote browser/iPad/Vision-Pro driver seat for the
  agent-fleet zellij session `Claude` stood up on nh3-dev (ttyd behind
-  Caddy, network-gated). Brokkr-smithy-dev request; no Tailscale (WG at
+  Caddy, network-gated). Out of this repo — full architecture + the
-  firewall). Out of this repo — full architecture + the
+  HTTP2/OSC52/Safari-auth gotchas in auto-memory `reference_ttyd_fleet_seat`.
  HTTP2/OSC52/Safari-auth gotchas in auto-memory
  `reference_ttyd_fleet_seat`.
 - `[2026-05-30]` esh-docker-vm NFS boot-ordering fix:
  `playbooks/fix-esh-nfs-boot-ordering.yaml` (`c0458d9`, + `53157b1`
  drop-in filename-collision fix) adds `_netdev,nofail` to the four
  10.0.50.50 NFS mounts + a docker `After=remote-fs.target` drop-in —
-  resolves paperless `Exited(255)` on every reboot. traefik also gained
+  resolves paperless `Exited(255)` on reboot. traefik also gained
-  `restart: unless-stopped` (was Docker-default `no` → never
+  `restart: unless-stopped`. Full incident → auto-memory
  auto-started after a reboot; "homepage broken" was collateral via the
  traefik route). Full incident → auto-memory
  `incident_esh_docker_nfs_boot_race`.
- `[2026-05-27]` Worldtree v0.29.x landed full saga→echo refactor +
+_25 older entries archived to archival-memory.md._
  config-validator hardening. Net image:
  - v0.29.10 — `create_provider` checks model_catalog `family:`
    before regex auto-detect (artemis routing fix)
  - v0.29.11 — collect-then-raise echo startup validators (UX fix
    for cascading one-at-a-time crash cycle that bit demo + personal
    during the v0.29.0 → v0.29.9 deploy)
  - v0.29.12 — ships `providers.yaml.example` + `defaults.yaml.example`
    canonical configs (REQUIRED/OPTIONAL annotated, guaranteed-pass-
    validator). Replaces reverse-engineer-from-validator-errors as
    the operator's path.
  - v0.29.13 — `reasoning_content` extraction in base provider +
    catalog `family:` lookup fix (artemis reasoning route through
    worldtree GemmaProvider).
  Operator-asked, worldtree-dev-shipped, all auto-deployed via CI.
 - `[2026-05-27]` artemis-31b-v1i added to llama-swap + worldtree
  personal providers.yaml. BeaverAI Gemma 4 31B Q6_K (~28.6GB)
  pulled via `pull-hf-repo.yaml`, llama-swap config mirrors
  gemma4-31b-dense settings at 128K ctx, `--reasoning-format deepseek`
  (gemma format not supported in deployed llama.cpp). Worldtree
  catalog entry has `family: gemma` so GemmaProvider routes reasoning
  tokens correctly. Pinned locally — push pending operator greenlight.
 - `[2026-05-27]` Skaldsong streaming TTS v0.32.0 → v0.32.1 → v0.32.2.
  v0.32.0: chunked-batch via SSE on `/api/stories/{id}/audio?voice=X`,
  one Kokoro POST per paragraph. v0.32.1: defensive catch-all in
  event_stream after silent JSONDecodeError (made the real
  exception surface in an SSE error frame). v0.32.2: NDJSON parsing
  — Kokoro `/dev/captioned_speech` returns multi-line JSON for
  large inputs (one object per chunk); was using `resp.json()` which
  only parsed the first object. Verified end-to-end clean.
 _22 older entries archived to archival-memory.md._
 ## Tried and abandoned
- `[2026-05-31]` Building the dia2-capable image surfaced THREE upstream
+- `[2026-06-01]` **CSM bring-up** — upstream `phildougherty/sesame_csm_openai`
-  packaging quirks: (1) `pip install -e nari-labs/dia2` fails — its
+  Dockerfile pins NO `huggingface_hub` version, so it now resolves to 1.17.0
-  pyproject build backend lacks the PEP 660 `build_editable` hook; (2)
+  where `huggingface-cli` is fully removed (replaced by `hf`) → the build dies
-  plain `pip install` builds an empty `UNKNOWN-0.0.0` wheel because the
+  at `huggingface-cli login`/`download`. Passing the HF token as a build-arg
-  base image's setuptools (59.6) is older than dia2's required `>=70`, so
+  got the token in (the model downloads at BUILD time, not runtime — our
-  it ignores the `[project]` metadata; (3) `--no-deps` (which the wrapper's
+  compose only wired it as runtime env, the original failure) but then hit the
-  own auto-installer uses) leaves `transformers`/`sphn`/`whisper-timestamped`
+  rotted CLI. Fixing would need vendoring + patching the Dockerfile; not worth
-  missing. Fix that worked (`local/dia:v2`, `stacks/dia/dia2-image/`): copy
+  it for a weak model → CSM removed.
  the pure-python `dia2/` package straight into site-packages + install
  ONLY those 3 deps; torch 2.12 / numpy 2.2 in the base already satisfy
  Dia2, so no risky CUDA/torch churn.
- `[2026-05-31]` Dia2 predefined voices (43, baked at `/app/voices`) are
+- `[2026-06-01]` **Fish path-form references** (`{audio:"/app/references/x.wav"}`)
-  NOT reachable from the `/tts` clone path — it resolves
+  → HTTP 500 on this build, despite the stack README showing that shape. The
-  `clone_reference_filename` against the reference_audio dir ONLY
+  working clone paths are `reference_id` (by name → reads `<name>.wav` +
-  (`engine.py`). The OpenAI `/v1/audio/speech` `voice` param DOES
+  optional `<name>.txt`) or inline base64 `references`. The catalog now uses
-  auto-resolve them by name (separate code path), which masked the gap in
+  `reference_id`.
-  early testing. Fix: stage `/app/voices/*` into
+
-  `/worktank/dia/reference_audio` so `/tts` clone can use them. (So a
+- `[2026-06-01]` **Fish reference transcript as a provenance note** (not the
-  "verify on the endpoint the catalog actually targets" lesson — testing
+  actual spoken words) — Fish/fish-speech uses the reference transcript to
-  the OpenAI endpoint nearly shipped a catalog that 404s on `/tts`.)
+  disambiguate phonemes, and a misaligned transcript hurts cloning badly. I
  staged VCTK voices with provenance-note `.txt` sidecars, which CONFOUNDED my
  "Fish is a weak cloner" verdict (that verdict is retracted/unproven). Lesson:
  stage the REAL transcript (VCTK ships ground-truth; or ASR via Whisper) for
  any clone reference.
 - `[2026-06-01]` **Pitch-shift register control** (rubberband, to deepen Imogen
  to contralto/mezzo) — Fish ignores small reference shifts and overshoots
  large ones; `-3st` WITHOUT formant preservation → "baritone through a tuba"
  (formants collapsed); `-2/-3st` WITH `--formant` → too subtle / still bad.
  Abandoned at every depth; all variants deleted. Finer independent
  pitch/formant control needs praat (not installed). Future clones = operator's.
 - `[2026-06-01]` **resemblyzer for cloning-fidelity scoring** — weak/dated
  encoder (2019 LSTM, English-biased) + a synthetic-vs-natural domain gap that
  depresses cosine regardless of true fidelity → conclusions were soft. Use
  ECAPA-TDNN (speechbrain `spkrec-ecapa-voxceleb`) for rigorous
  speaker-verification scoring.
 - `[2026-05-31]` Building the dia2-capable image surfaced THREE upstream
  packaging quirks: (1) `pip install -e nari-labs/dia2` fails — no PEP 660
  `build_editable` hook; (2) plain `pip install` builds an empty
  `UNKNOWN-0.0.0` wheel (base setuptools 59.6 < dia2's required ≥70); (3)
  `--no-deps` leaves `transformers`/`sphn`/`whisper-timestamped` missing. Fix
  (`local/dia:v2`): copy the pure-python `dia2/` package into site-packages +
  install ONLY those 3 deps; base torch/numpy already satisfy Dia2.
 - `[2026-05-31]` Dia2 predefined voices (43, baked at `/app/voices`) are NOT
  reachable from the `/tts` clone path — it resolves `clone_reference_filename`
  against the reference_audio dir ONLY. The OpenAI `/v1/audio/speech` `voice`
  param auto-resolves them (separate code path), which masked the gap. Fix:
  stage `/app/voices/*` into `/worktank/dia/reference_audio`. Lesson: verify on
  the endpoint the catalog ACTUALLY targets.
 - `[2026-05-31]` `voice_mode=clone` with an empty `clone_reference_filename`
-  → asset-engine serializes it as the literal string `"undefined"` →
+  → asset-engine serializes it as the literal string `"undefined"` → `/tts`
-  `/tts` 404 `Reference audio file not found: undefined`. Reproduced
+  404. First observed on dia2; worked around in the catalog (default the field
-  directly. Worked around in the catalog (default the field to a real
+  to a real voice). **[2026-06-01] root cause found** — the Kokoro voice-blend
-  voice); proper fix is asset-engine-side (don't send `"undefined"` for
+  widget reading Shoelace `.value` before hydration (see Current state); the
-  empty optionals; enforce conditional-required) — flagged to
+  real fix is asset-engine-side and is escalated.
  asset-engine-dev.
- `[2026-05-31]` asset-engine `docs/asset-engine/services.schema.json`
+- `[2026-05-31]` asset-engine `services.schema.json` is DERIVED (regen from the
-  is a DERIVED artifact (regen from the Pydantic model via
+  Pydantic model via `dump_schema.py`) and had DRIFTED — rejected the
-  `dump_schema.py`) and had DRIFTED: it rejected the `lifecycle` field
+  `lifecycle` field 12/14 services use. RESOLVED: regenerated with
-  that 12/14 live services already use (13 pre-existing validation
+  `CatalogLifecycle` + `reproducibility.seed_field` (`b7b2130`). Lesson:
-  errors against the committed catalog).
+  hand-editing `services.yaml` shape without regenerating re-introduces drift.
  **[2026-05-31] RESOLVED:** schema regenerated with `CatalogLifecycle`
  + `reproducibility.seed_field` (`b7b2130`); catalog now validates clean.
  Standing lesson: the schema is generated from asset_engine's
  `catalog.py` — hand-editing `services.yaml` shape without regenerating
  re-introduces drift.
- `[2026-05-30]` esh-docker-vm: `hard` NFS mounts from 10.0.50.50 froze
+- `[2026-05-31]` ttyd-over-TLS forces HTTP/2 (kills ttyd's terminal WebSocket
-  a container worker in UNKILLABLE D-state when the NAS stalled —
+  → blank screen); Safari/WebKit never sends HTTP basic-auth on WS upgrades.
-  `docker restart` fails with "tried to kill container, but did not
+  Both solved for the fleet seat (Caddy forces HTTP/1.1; auth → network-gating)
-  receive an exit event"; only a host reboot clears it. Separately,
+  — detail in auto-memory `reference_ttyd_fleet_seat`.
-  `fstab defaults` (no `_netdev`) made NFS-bind containers `Exited(255)`
+
-  on reboot. Diagnosis + recovery steps → auto-memory
+- `[2026-05-30]` esh-docker-vm: `hard` NFS mounts from 10.0.50.50 froze a
  container worker in UNKILLABLE D-state when the NAS stalled — only a host
  reboot clears it. Separately, `fstab defaults` (no `_netdev`) made NFS-bind
  containers `Exited(255)` on reboot. → auto-memory
  `incident_esh_docker_nfs_boot_race`.
- `[2026-05-31]` ttyd-over-TLS forces HTTP/2 (no flag disables it) which
+_29 older entries archived to archival-memory.md._
  kills ttyd's terminal WebSocket (blank screen); Safari/WebKit never
  sends HTTP basic-auth on WS upgrades (→ ttyd "User code denied"). Both
  solved for the fleet seat (Caddy forces HTTP/1.1; auth dropped to
  network-gating) — detail in auto-memory `reference_ttyd_fleet_seat`.
 - `[2026-05-27]` `docker compose up -d` from within the `docker:cli`
  sandbox: `${VAR:-./config}` defaults resolve `./config` to the
  sandbox CWD (e.g. `/wt/config`). Docker daemon interprets the
  resulting path against the HOST filesystem, not the sandbox's.
  Auto-creates an empty dir on the host → container saw an empty
  bind target → entrypoint shim seeded image-baked defaults
  (lost all the host-side providers.yaml patches I'd applied earlier
  in the day). Root cause for ratatoskr-dev's Tier 3 sessions 500.
  Fix: pass `-e WORLDTREE_CONFIG_DIR=/opt/worldtree-personal/config`
  to the docker run invocation. Captured into the docker-as-root
  pattern note in Tools and conventions.
 - `[2026-05-27]` `:latest`-pinned compose + private gitea registry +
  sandboxed pull = recreate on ancient cached `:latest`. Local
  `:latest` tag never advanced because deploy.yml pulls by SHA
  (not by tag); sandbox can't `docker pull` (no creds for private
  registry); `compose up -d --pull never` used the 2-week-old
  cached `:latest`. Symptom: personal worldtree's matrix container
  crash-looped because the ancient image didn't understand the new
  config schema. Fix: retag SHA → `:latest` on host first
  (`docker tag <registry>/<repo>:<sha> <registry>/<repo>:latest`),
  then recreate with `--pull never`. Better long-term: pin SHA
  explicitly in `.env`, advance in CI.
 - `[2026-05-27]` Container recreate during in-flight skaldsong gen
  kills the runner. Pattern: my llama-swap restart at 23:57Z + a
  v0.31.2 `compose up -d` at 02:47Z both took out active dramatist
  passes. Confirmed by skaldsong-dev INV-006 ("partial state only
  flips on host restart"). With deploys landing every ~10min during
  active dev and stories taking >5min, runner-killed-by-recreate is
  structural, not incidental. Roadmap (skaldsong-dev): pre-shutdown
  signal handler (cleanup row state to failed/aborted), per-scene
  resume-from-checkpoint, `/api/admin/quiesce` endpoint. None shipped
  yet; operator's call which path.
 - `[2026-05-27]` `--reasoning-format gemma` on artemis-31b-v1i —
  not supported in the deployed llama.cpp build. Accepts only
  `none | deepseek | deepseek-legacy`. `deepseek` populates the
  `reasoning_content` SSE delta field which Worldtree GemmaProvider's
  fallback path checks. Was load-bearing for getting reasoning
  tokens through to ratatoskr's wire.
 - `[2026-05-27]` `head -c N` piped after a streaming curl SIGPIPEs
  the curl, killing the request early. Bit me twice while probing
  SSE-shape samples. Use file-write + separate read.
 _24 older entries archived to archival-memory.md._