Restarted testdatadb service (new template live), canary-pushed 1 cert (updated,
0 errors), then re-pushed all PASS certs for the 68 Final-Test-content-clean
models via uploadBySerialNumbers from a fresh node process.
Result over 30,423 PASS certs: updated=26022 unchanged=2738 created=0 errors=0
skipped=1663. The 26,022 updates replace the old defective DSCA renders on the
live site with the rebuilt, byte-content-validated ones. The 1,663 skips are the
count-guard correctly refusing any individual cert whose value count != its
template row count (ambiguous) — never publishing misaligned data.
Artifacts: push-clean68.js (driver), dsca-clean68-models.json (the published
set). NOT yet published: the 26 last-digit-diff models, the ~32 ambiguous/null
layout families, and the 231 untemplated models.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
validate-dsca-stage3.js: read-only harness that, for every staged DSCA original
we have ground truth for (2806 across 126 models), looks up the DB record,
renders it through the live path, and content-compares. GATE = the FINAL TEST
RESULTS section (rule lines canonicalized, whitespace collapsed — so the deferred
column-spacing cosmetic doesn't register); accuracy-section diffs reported
separately as informational.
First run verdict (report attached):
- 68 models FINAL-TEST CONTENT-CLEAN (0 mismatch over compared certs).
- 2123/2316 certs match the Final-Test content exactly (91.7%).
- 26 models show measured-value last-digit diffs only — structure (names, specs,
row alignment, statuses) is correct. Two root causes, neither structural:
* rounding-mode: JS double toFixed vs QB single-precision half-up
(e.g. raw 9.9995 code3 -> "9.999" here, "10.000" in golden). Fixable but
float-precision-sensitive; risks regressing currently-clean values.
* data vintage: staged .TXT is a different test run than the DB latest-wins
record (Fix 3) — e.g. Supply Current 19.6 vs 20.3, 0.7 apart. Not a render
bug; can't be reconciled against an older staged sheet.
- ~32 models render null (count-guard): DSCA33-*, DSCA45-*, DSCA49-* families
whose raw_data carries load points the template omits -> need per-subtype slot
mapping (the canonical-slot approach) before they can render.
Still NOT published: service not restarted, nothing re-pushed to Hoffman.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Deployed file is C:\Shares\testdatadb\templates\datasheet-exact.js; this
reconciles the repo copy + adds dsca-templates.json (STAGE 1 output).
What changed in generateExactDatasheet (DSCA family only; 5B/8B/7B/DSCT/SCMVAS
paths byte-unchanged):
- Load dsca-templates.json once at module top (126 per-model layouts).
- DSCA Final-Test now renders names + specs from the staged template rows, not
the single hardcoded DATA_LINES['DSCA'] + buildTSpecs DSCA branch.
- Value-bearing raw_data STATUS groups map positionally onto the spec-bearing
template rows; empty-spec rows (240VAC Withstand / Hi-Pot) render blank+PASS.
Removed the duplicate hardcoded 240VAC/Hi-Pot footer for DSCA (now rows).
- ACCURACY header uses the template accOut ("Output (V)"/"Output (mA)") with '-'
rule separators instead of "Vout (V)" + '='.
- Header/columns match the staged originals (Measured Value*, 25/15/19/6 rule).
Two real bugs fixed (both are the handoff's "lines drop" / wrong-value defect):
- formatMeasuredExact reads the value from index 4 so negative signs survive
("PASS-4.24..." -> "-4", not "4"); also decimal-code N -> toFixed(N) exactly
(DSCA differs from 5B/8B where code 2 means 1 decimal).
- parseRawData no longer consumes the first DSCA STATUS group as a bare
step-response line when that line is absent (dropped 3 rows on e.g. DSCA39-01).
Safety: when value count != spec-row count the positional zip is ambiguous
(subtype measures load points the template omits, e.g. DSCA49 5mA pair), so the
cert is SKIPPED (null) and left for STAGE 3 per-subtype mapping rather than
emitting misaligned data.
Validation: DSCA38-05 (SN 180224-1) Final-Test block byte-identical to its
staged original. 92/126 templated models render cleanly; 7 ambiguous + 27
no-spec skip. Remaining ACCURACY-block spacing diffs are the deferred cosmetic
gap. NOT YET LIVE: testdatadb service not restarted, nothing re-pushed to
Hoffman (STAGE 3 gate).
Coverage gap to resolve before publish: only 126/357 DSCA models in the DB have
a staged template (56,074 certs, 70.1%); 231 models / 23,866 certs have none and
now render null — needs a STAGE 1 extension (more staged originals).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Plain-language summary for John (EE, non-developer) covering all 4 problems
(RTD label, DSCA table, same-day run, missing units), framed as ACG-owned fixes.
Cross-links the four supporting technical docs. Draft - not yet sent.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Per Mike: import runs every 15 min, so routine timing isn't the cause. The 379
absent units are confined to 2025-10..2026-01 (stop after Jan 2026) on TS-4L/4R/1R
- fingerprint of a one-time overwrite during the incident/recovery (fresh DOS logs
overwrote accumulated appended server-side logs for ~2 weeks). One-time, not
recurring; backfill from the surviving staged .TXT.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
608 staged datasheets absent from DB. Two causes: (1) 229 units with encoded/
non-standard serials the importer's leading-digit regex silently skips - data is
in the .DAT, recoverable; full blind spot is 840 serials / 9,510 records / 141
models dropped fleet-wide. (2) 379 units whose per-model .DAT was overwritten by a
later work order - recoverable only from the staged .TXT or a log backup. Adds
John-facing report, raw data, and the chase-missing-units.js tool.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Whole-source sweep (981,716 records / 406,549 serials): 6,515 same-day multi-run
events; DB holds a NON-latest run for 311 (the strictly-greater-date conflict rule
freezes on an arbitrary same-day run). Corrects the verdict doc to flag same-day
retests as a latest-wins faithfulness violation (not benign). Adds the proposed
>= -with-data-differs conflict-rule fix (diagnose-only) and the sweep tool.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Validated all 11,922 staged original .TXT datasheets against test_records.
0 genuine parse faults across 11,239 comparable records; mismatches all explained
(retests, reused serials, VAS format, legacy out-of-scope units). Adds the
validate-parsing.js tool, raw report, and verdict. Two follow-ups (NOT parse bugs):
608 staged units absent from DB (ingestion completeness), and same-day retests keep
the first run (ON CONFLICT strictly-greater-date).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Repo copy only (review before deploying to C:\Shares\testdatadb). Folds RTD
(sensorNum 7) into the temperature path so the ACCURACY input column shows
'Temp. (C)' with signed temperature values, matching the original DOS-generated
datasheets and thermocouples (3-6). raw_data stimulus is already in deg C; no
conversion. getSensorNum and the i==13 ohm/ohm unit override are untouched.
Verified read-only against deployed env: 8B35 SN 179553-13 now shows Temp (C);
regression over 184 5B/8B renders -> 15 RTD changed (intended), 0 non-RTD changed.
Does NOT address Defect B (DSCA template). See DATASHEET-RTD-BUG-DIAGNOSIS-2026-06-17.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Relocated verbatim from .claude/CLAUDE.md so the ad2 fork stops editing the
shared fleet harness doc. After rebase onto main, .claude/CLAUDE.md = the lean
fleet version; Dataforth ops context lives here. Keeps future ad2 syncs clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Vanilla single-file (no build/CDN) command-center redesign of the testdatadb search
UI: omni-search with serial/model/text routing + auto-select fast path, dense
monospace results table with PASS/FAIL pills + web-published indicator, persistent
split-pane datasheet inspector (iframe to /api/datasheet), left filter rail
(result/date/model/station/log), server pagination, CSV export, URL state, keyboard
nav (/ ↑↓ ↵ Esc), clinical light theme. Hits the existing API; deployed to AD2 as
public/index.redesign.html (preview at :3000/index.redesign.html). Synthesized from
Grok + Gemini concepts (both converged on this command-center design).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Remote SSH/VPN to AD2 keeps flapping; hand the remaining datasheet fixes to the
local AD2 session. Includes the per-subtype approach (DSCA_TEMPLATES from staged
originals — STAGE 1 done, dsca-templates.json on AD2 = 126 models), the render-wiring
+ per-subtype byte-validation gate, Fix 5 (379 backfill via legacy_cert_text), the
discipline (backup/save-state/validate-before-publish), and the derive-dsca-templates
tool. Ref ticket #32441.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
While using the new 3-retry gemini path for live VPN research, two bugs surfaced:
- emit_or_fail checked auth_failed INSIDE the retry loop; a benign mid-run token-refresh line
matched the over-broad auth regex (bare login|credential|authenticat|oauth|401) and aborted the
retries with a false "auth error" - even though `gemini -p` auth tested fine. Moved auth-classify
to AFTER the retries (it only picks the final error message now) and tightened auth_failed to real
signatures (invalid_grant, not authenticated, login with google, token expired, ...).
- Added quota_exhausted() + a QUOTA FALLBACK: the pinned strong model (gemini-3.1-pro-preview) hit
"exhausted your capacity on this model" mid-session; emit_or_fail now retries once on the default
(lighter) model by stripping -m (separate quota). Validated: capped pro run -> fell back -> 2.9KB answer.
CT_THOUGHTS Thought 2 Resolution updated with both. (Search-bot reliability hardening continues.)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Mike's must-fix. Diagnosed from RAW output of failing queries (not guessed):
- grok xsearch = TIMEOUT: grok-4.20-multi-agent web_search runs past budget on multi-part queries
(286s/280s, rc=124, still searching - 183 thoughts, only progress-noise text); buffered json => total loss.
- gemini search = INTERMITTENT empty turn (a clean re-run gave a real 2.6KB answer in 122s); the wrapper
retried only once, so two empties in a row failed spuriously.
Fixes:
- ask-gemini.sh emit_or_fail: retry up to 3x with 3s/6s backoff (was 1).
- ask-grok.sh xsearch: --output-format streaming-json (salvage partials) + AUTO-FALLBACK to
ask-gemini.sh search when grok doesn't finish (rc!=0 or empty). Validated e2e: grok timed out
(rc=124) -> fell back -> gemini returned a real sourced answer (UniFi Teleport invite-link API).
grok's own multi-agent timeout is an xAI-side limitation; the fallback makes xsearch reliable regardless.
Docs: grok SKILL.md xsearch row + CT_THOUGHTS Thought 2 Resolution.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Mike's correction: web search (grok xsearch + gemini search) carries at least as much weight as
live API probing - the searches gave the real leads this session (connector proxy, teleport setting
path); blind endpoint-probing is "highly suspect" (mostly 404s). And the search bots MUST be properly
fixed - both returned empty repeatedly on UniFi research despite the same-day partial grok fix.
- docs/CT_THOUGHTS.md: Thought 2 (HIGH PRIORITY) - web-search reliability must-fix, with the observed
failures + a proper-fix investigation plan (capture failing-query JSON; max-turns/streaming-json/
retry; cross-fallback grok<->gemini; 5/5 acceptance).
- memory feedback_web_search_over_probing: lead with web search/docs; probe only to CONFIRM a
hypothesis, never as primary discovery. Reading our own config is fine; guessing paths is not.
- errorlog correction logged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Earlier "no usable Teleport API" was wrong (probed /rest/teleport, /stat/teleport, /v1/teleport).
Gemini research + live verification: Teleport config lives at /api/s/<site>/rest/setting/teleport
(GET/PUT, also under /get/setting key 'teleport') - reachable via the connector. Brooklyn confirmed
enabled, subnet 192.168.1.1/24. Invite generate/revoke is reportedly POST /api/s/<site>/cmd/teleport
{"cmd":"generate-invite"|"revoke-invite"} (untested - it creates a live VPN access link; gate as a
write). Invites are WiFiman-app-only. Proxy path is /v1/connector/consoles/{id}/proxy/... (Gemini's
/v1/hosts/{id}/proxy form 404s). Doc updated.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
neighbor-collect.sh: add `--console <name> [--site <short>]` so the AP name/BSSID/IP map can come
from the cloud connector (/v1/connector/.../stat/device) instead of a UOS direct-login -- lets the
disable-analysis collector run against ANY console we have AP-VLAN reach to (the AP SSH harvest of
/proc/ui_neighbor is unchanged and still needs L3 reach). UOS path untouched. Validated against
Cascades via connector: source=CONNECTOR, built 77-mac + 450-bssid map for the 75 online APs.
This completes the hybrid (don't-lose-functionality): connector for airtime everywhere + neighbor-
collect (any source) for the SNR matrix -> NEIGHBOR_JSON -> optimize-radios disables on remote sites.
Documented (references/site-manager-api.md): the neighbor-collect --console flow, and the gateway
VPN/Teleport reach -- connector reaches /rest/networkconf (VPN servers: wireguard-server/openvpn-
server, site-to-site) read+writable in principle (gate writes like gw-control); Teleport has no
usable API (v1/ea/teleport 404, per-console /teleport 403).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Validated the cloud-connector analysis against a KNOWN entity (Cascades, normally UOS-Mongo).
The connector reaches the self-hosted "UOS Server" host; Cascades is its site `va6iba3v`.
Two fixes from the validation:
- rf-analyze.py: pass macs:[<all uap macs>] to /stat/report/*.ap. The UniFi report endpoint
returns only a small DEFAULT subset otherwise -- Cascades came back as 10 of 77 APs until the
MAC list was supplied. Now profiles all 75 (uaps with 2.4 radios), matching the UOS path.
- model-rank.sh / optimize-radios.sh: --console now accepts --site <name> (internal short name
from /api/self/sites) for multi-site controllers like the UOS Server (Cascades = va6iba3v).
Result lines up with the known UOS-Mongo figures: 75 APs, 2.4GHz util 65-90% / interf 53-78% /
~1 client each, all power-down, 0 disables (roam graph absent via connector -> same coverage-safe
degradation; disables still need NEIGHBOR_JSON). Apples-to-apples confirmed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
