Files
gbrain/scripts/run-verify-parallel.sh
T
f702ec053b v0.41.16.0 feat: conversation parser cathedral + progressive-batch primitive (closes #1461) (#1510)
* v0.41.15.0 feat: conversation parser cathedral + progressive-batch primitive (closes #1461)

Replaces PR #1461's single-format Telegram regex with a 12-pattern
built-in registry covering iMessage/Slack, Telegram (×2), Discord
(×2), WhatsApp (×2 locales), Signal, Matrix/Element, IRC (×2), Teams.
Each pattern is hand-vetted from public format docs (signal-cli,
DiscordChatExporter, Telegram Desktop, WhatsApp export docs, Element
matrix-archive, irssi/weechat defaults); module-load validation runs
test_positive[] + test_negative[] for every pattern at startup so a
typo makes gbrain refuse to start.

PR #1461 contributor's BRACKET_TIME_RX + cleanSpeaker survive verbatim
as the `telegram-bracket` built-in pattern + DEFAULT_SPEAKER_CLEAN
export. All 33 of their test cases pass against the new orchestrator.

Three layers per page (orchestrator chooses):
  1. Built-in pattern registry (zero-cost, deterministic)
  2. User-declared simple_pattern via config (deferred to v0.42+)
  3. Opt-IN LLM polish + fallback (privacy-first; chat content goes
     to Anthropic only when user explicitly enables)

D18 priority scoring picks the highest-match-rate pattern across the
first 10 lines (not first-wins) so overlapping formats don't silently
mis-route. D5 multi_line per-pattern + D11 quick_reject prefix screen
+ D19 timezone_policy per-pattern complete the registry shape.

Companion: src/core/progressive-batch/ primitive (rule of three
satisfied across 12+ ad-hoc cost-prompt sites). Wintermute-inspired
ramp shape (trial 10 → 100 → 500 → full with verification at each
stage), productionized with verifier+policy injection (callers
describe HOW TO MEASURE SUCCESS, not WHEN TO WAIT FOR CTRL-C). D3
fail-closed budget gate: null tracker + null Policy.maxCostUsd →
abort_cost_cap reason='no_budget_safety_net'. D20 discriminated
Verifier union (output_count | idempotent_mutation | noop).
extract-conversation-facts is the one proven consumer in v0.41.15.0;
9-site retrofit deferred to v0.41.16.0+ per TODOS.md.

Codex outside-voice review absorbed 8 substantive findings:
  - Privacy posture (LLM polish/fallback flipped to opt-IN)
  - ReDoS theater (dropped arbitrary user regex; v0.42+ uses RE2)
  - LLM-inferred-regex persistence as silent-corruption machine
  - Pattern priority scoring across first 10 lines
  - Timezone policy on every PatternEntry
  - Verifier shape discriminated union
  - Behavior parity for sites that "jumped straight to full"
  - Real-corpus-redacted fixture gap (v0.42+ TODO)

CI gates:
  - bun run check:conversation-parser (13 fixtures, --no-llm, deterministic)
  - bun run check:fixture-privacy (banned-token grep)

Doctor surfaces 3 new checks: conversation_format_coverage,
progressive_batch_audit_health, conversation_parser_probe_health.

Tests: 198/198 across primitive + parser + LLM + nightly probe + eval
CLI + debug CLI + doctor checks + migration v97 round-trip + E2E
parser ↔ engine integration. Real bug caught + fixed during gap audit:
IdempotentMutationVerifier was comparing absolute mutated-count vs
per-stage expected (failed silently on stage 2+); now uses per-stage
delta semantics matching OutputCountVerifier.

Schema migration v97: conversation_parser_llm_cache table with
(content_sha256, model_id, call_shape) composite key. NO
inferred_patterns table (D17: silent-corruption machine).

Plan + 23 decisions + codex outside-voice absorption at
~/.claude/plans/system-instruction-you-are-working-cuddly-hollerith.md.

Co-Authored-By: garrytan-agents (PR #1461) <noreply@github.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* fix(check-privacy): allowlist scripts/check-fixture-privacy.sh

The new sibling privacy guard literally names the banned tokens in its
BANNED_TOKENS array — same meta-exception that check-privacy.sh itself
gets. Without this allowlist entry, bun run verify rejects the file
post-merge because the banned name appears in the rule-definition script.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* chore: renumber v0.41.15.0 → v0.41.16.0 (queue drift)

Mechanical rename across all surfaces: VERSION, package.json,
CHANGELOG (header + body refs), CLAUDE.md, TODOS.md, src/core/
migrate.ts (migration v98 comment), all src/core/conversation-parser/*
and src/core/progressive-batch/* file headers, all test/ headers,
scripts/check-privacy.sh allowlist comment, llms-full.txt regenerated.

Audit clean: VERSION + package.json + CHANGELOG header all show
0.41.16.0. verify 24/24, touched tests 179/179.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: garrytan-agents (PR #1461) <noreply@github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-26 17:31:48 -07:00

197 lines
7.3 KiB
Bash
Executable File

#!/usr/bin/env bash
# scripts/run-verify-parallel.sh — parallel verify dispatcher.
#
# Runs the 19+ verify checks (privacy, jsonb, source-id, … + typecheck +
# admin-build) as background jobs, waits for all, aggregates exit codes,
# surfaces failed-check name + tail of its log to stderr.
#
# Replaces the sequential `&&`-chain in package.json's `verify` script.
# Wallclock: 19 sequential checks (~15-25s on CI) → parallel (~3-5s).
#
# Usage:
# bash scripts/run-verify-parallel.sh # run every CHECK below
# bash scripts/run-verify-parallel.sh --dry-list # print check list, exit
#
# Env overrides:
# GBRAIN_VERIFY_TIMEOUT per-check wallclock cap, seconds (default 120)
# GBRAIN_VERIFY_LOG_DIR where to write per-check logs (default tempdir)
#
# Exit codes:
# 0 all checks passed
# 1 one or more checks failed (full details in stderr)
# 2 usage error / no checks defined
set -uo pipefail
cd "$(dirname "$0")/.."
# ──────────────────────────────────────────────────────────────────────────
# Checks to run. Order is irrelevant (parallel), but keep stable for log
# determinism + grep-ability. Each entry is a bun-script name (the
# `package.json` "scripts" key), invoked as `bun run <name>`.
#
# To add a check: append to this array. To skip in CI temporarily, comment
# the line — the parallel runner doesn't care about count.
# ──────────────────────────────────────────────────────────────────────────
CHECKS=(
"check:privacy"
"check:proposal-pii"
"check:test-names"
"check:jsonb"
"check:source-id-projection"
"check:source-config-leak"
"check:progress"
"check:test-isolation"
"check:wasm"
"check:admin-build"
"check:admin-scope-drift"
"check:cli-exec"
"check:system-of-record"
"check:eval-glossary"
"check:no-pii-agent-voice"
"check:synthetic-corpus-privacy"
"check:skill-brain-first"
"check:fuzz-purity"
"check:operations-filter-bypass"
"check:gateway-routed"
"check:fixture-privacy"
"check:conversation-parser"
"check:resolver"
"typecheck"
)
if [ "${#CHECKS[@]}" -eq 0 ]; then
echo "ERROR: no checks defined in run-verify-parallel.sh" >&2
exit 2
fi
# Dry-run path: list checks, exit. Used by tests + ops debugging.
if [ "${1:-}" = "--dry-list" ]; then
printf '%s\n' "${CHECKS[@]}"
exit 0
fi
if [ "$#" -gt 0 ] && [ "${1:-}" != "" ]; then
echo "ERROR: unknown arg: $1" >&2
echo "usage: bash scripts/run-verify-parallel.sh [--dry-list]" >&2
exit 2
fi
TIMEOUT="${GBRAIN_VERIFY_TIMEOUT:-120}"
# Per-check temp dir. Each check gets its own subdir so writes can't race
# on shared scratch state (the checks themselves are read-only — they grep
# the working tree — but defense-in-depth.)
if [ -n "${GBRAIN_VERIFY_LOG_DIR:-}" ]; then
LOG_DIR="$GBRAIN_VERIFY_LOG_DIR"
mkdir -p "$LOG_DIR" || { echo "ERROR: cannot create $LOG_DIR" >&2; exit 2; }
else
LOG_DIR="$(mktemp -d /tmp/gbrain-verify-XXXXXX)"
trap 'rm -rf "$LOG_DIR"' EXIT
fi
# Resolve `timeout` for per-check wallclock cap. macOS doesn't ship one;
# brew coreutils provides `gtimeout`. If neither is available, fall back to
# bg-pid + sleep-cap (slightly less reliable but still bounded).
TIMEOUT_BIN=""
if command -v gtimeout >/dev/null 2>&1; then TIMEOUT_BIN="gtimeout"
elif command -v timeout >/dev/null 2>&1; then TIMEOUT_BIN="timeout"
fi
START_TS=$(date +%s)
echo "[verify-parallel] running ${#CHECKS[@]} checks in parallel (timeout=${TIMEOUT}s, logs=$LOG_DIR)" >&2
# ──────────────────────────────────────────────────────────────────────────
# Spawn one background process per check. Each child captures its own exit
# code into a sentinel file under $LOG_DIR/<safe-name>.exit; the parent
# never trusts `wait`'s aggregate value because that maps to last-spawned.
#
# safe_name: turn `check:privacy` into `check_privacy` so it fits a filename
# without escaping.
# ──────────────────────────────────────────────────────────────────────────
PIDS=()
SAFE_NAMES=()
for c in "${CHECKS[@]}"; do
safe="${c//:/_}"
SAFE_NAMES+=("$safe")
LOG_FILE="$LOG_DIR/$safe.log"
EXIT_FILE="$LOG_DIR/$safe.exit"
(
if [ -n "$TIMEOUT_BIN" ]; then
"$TIMEOUT_BIN" "${TIMEOUT}s" bun run "$c" > "$LOG_FILE" 2>&1
else
bun run "$c" > "$LOG_FILE" 2>&1 &
pid=$!
( sleep "$TIMEOUT" && kill -TERM "$pid" 2>/dev/null && \
sleep 5 && kill -KILL "$pid" 2>/dev/null ) &
cap_pid=$!
wait "$pid" 2>/dev/null
kill "$cap_pid" 2>/dev/null
wait "$cap_pid" 2>/dev/null
fi
rc=$?
echo "$rc" > "$EXIT_FILE"
) &
PIDS+=($!)
done
# Wait for every background job. Ignore wait's aggregate exit — exit codes
# live in the sentinel files.
for pid in "${PIDS[@]}"; do wait "$pid" 2>/dev/null || true; done
END_TS=$(date +%s)
ELAPSED=$((END_TS - START_TS))
# ──────────────────────────────────────────────────────────────────────────
# Aggregate. For each check, read its exit file; on failure, append a
# labeled block (check name + tail of log) to the failure report. Surface
# one final summary line and the report to stderr if anything failed.
# ──────────────────────────────────────────────────────────────────────────
PASS=0
FAIL=0
FAIL_NAMES=()
FAIL_REPORT=""
for i in "${!CHECKS[@]}"; do
c="${CHECKS[$i]}"
safe="${SAFE_NAMES[$i]}"
EXIT_FILE="$LOG_DIR/$safe.exit"
LOG_FILE="$LOG_DIR/$safe.log"
rc=1
[ -f "$EXIT_FILE" ] && rc=$(cat "$EXIT_FILE" 2>/dev/null || echo 1)
if [ "$rc" = "0" ]; then
PASS=$((PASS + 1))
else
FAIL=$((FAIL + 1))
FAIL_NAMES+=("$c")
if [ "$rc" = "124" ]; then
FAIL_REPORT+=$'\n--- '"$c"' (TIMED OUT after '"${TIMEOUT}"'s) ---\n'
else
FAIL_REPORT+=$'\n--- '"$c"' (rc='"$rc"') ---\n'
fi
if [ -f "$LOG_FILE" ]; then
FAIL_REPORT+="$(tail -30 "$LOG_FILE")"
FAIL_REPORT+=$'\n'
fi
fi
done
if [ "$FAIL" -gt 0 ]; then
{
echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo "❌ verify failed: $FAIL/${#CHECKS[@]} checks did not pass"
echo "Failed: ${FAIL_NAMES[*]}"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
printf '%s' "$FAIL_REPORT"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo "[verify-parallel] elapsed=${ELAPSED}s | pass=$PASS fail=$FAIL"
} >&2
exit 1
fi
echo "[verify-parallel] elapsed=${ELAPSED}s | pass=$PASS fail=0 | all checks green" >&2
exit 0