v0.3.1: survival behaviour overhaul — storyline, biome-aware scout, wedge-relocate, food/perf fixes, monitor TUI (#28)

* docs(v0.3.1): PRD — LLM prompt cost optimization

Design-only commit; no runtime changes. Spec for the next patch iteration.

Goal: cut per-advise() input tokens from ~800 to ≤300, preserving the
LLM's ability to produce valid registered skill ids and useful rationale.

Five proposed changes ranked by impact:

  P1  Compact registry format (saves ~350t/call) — group by namespace,
      comma-list ids, drop human titles. Default mode for advisor;
      verbose mode kept for postmortem/reflect.
  P2  Need-scoped registry (~50t additional) — show LLM only skills
      relevant to the active Maslow need + always-available safety
      skills (survive.flee, pillar-up, recovery.tunnel-out, explore.*).
  P3  Snapshot pruning (~50t) — drop weather/experience/dimension/biome/
      players from the user prompt; the LLM doesn't consult them.
  P4  Prompt caching probe — check if TimeWeb passes through
      prompt_tokens_details.cached_tokens. If yes, restructure prefix
      to maximize cache hits (cached input is ~10x cheaper at OpenAI).
  P5  Per-trigger cost telemetry in scripts/list-improvements.js --stats:
      avg_in / avg_out / cost_₽ / share% per trigger_reason, using
      TIMEWEB_PRICE_IN_RUB_PER_M and TIMEWEB_PRICE_OUT_RUB_PER_M env.

Trigger: TimeWeb admin panel after first day of v0.3.0 live showed
34K tokens / day at low activity. At cap budget that projects to
~480₽/month (101₽/M in, 608₽/M out for gpt-5.4-mini). Manageable
but the savings are mostly free — repeated infra tokens, not signal.

All changes are additive; runtime behaviour stays the same. If the
LLM produces worse advice with the compact registry, flip back via a
single constant in fast-advisor.js.

Acceptance: re-run scripts/check-timeweb.js probe 3 — expect
tokens_in ≤ 300 (was ~800). Live for 1h, check --stats: avg_in ≤ 300
per trigger group. Existing 360 tests still green.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* feat(v0.3.1): storyline — canonical Minecraft survival quest

The bot has been stuck in a loop for two days:
  acquire-food (fail: no nearby food) → explore.far → pillar-up (fail) → repeat

Diagnosis: manifesto + LLM advisor both correctly identify "you need
food" but neither expresses *what concretely to do next*. Manifesto is
a priority ladder (need-detection), not a narrative arc.

This commit adds the missing narrative layer — an ordered list of
operational steps that mirror the vanilla Minecraft survival path:

  1. orient_self      — Понять где я
  2. first_wood       — Собрать 8 поленьев
  3. crafting_basics  — Сделать верстак и палки
  4. first_tools      — Деревянные орудия
  5. first_food       — Найти первую еду
  6. shelter_minimal  — Простой шелтер с кроватью
  7. stone_tier       — Каменные орудия
  8. food_security    — Запас еды на 16+
  9. iron_age         — Железо и печь
  10. settle_base     — Постоянная база
  11. village_grow    — Развивать деревню (ongoing)

Each step has:
  - completed(snapshot) → bool — detects achievement from snapshot
  - suggestSkill(snapshot) → { skillId, args? } — concrete next dispatch
  - emergencyPause(snapshot) → bool — defers to manifesto L0 alive
    emergencies (low HP near hostile, lava under foot, food = 0)
  - narration_ru — chat-friendly Russian one-liner spoken on entry

Components:

- runtime/goal/storyline.js — 11-step canonical quest catalogue
- runtime/goal/state.js — pickCurrentStep(snapshot) walks the list,
  returns first non-completed step + its suggestion. 3s cache.
  Validates suggestSkill's skillId against the live registry.
- runtime/reflex.js — curriculumReflex dispatch priority is now:
    1. manifesto (L0 alive emergencies always win)
    2. storyline (concrete operational subgoal)
    3. curriculum plan (legacy fallback)
  Tests pass ctx.disableStoryline=true for isolation.
- runtime/bot.js — snapshot.storyStep populated each tick so
  chatter/advisor/reflect observers see the same view.
- runtime/coach/fast-advisor.js — buildUserPrompt now embeds the
  current step + its suggested skill, so LLM advice is anchored
  ("step 5 first_food, storyline wants survive.acquire-food, but
  recent dispatches show it's failing — try explore.far + scout").
- runtime/coach/advisor-trigger.js — forwards ctx.storyStep into
  advise() and logs step id at trigger time.
- runtime/coach/reflect.js — reflection prompt includes storyline
  progress so 30-min self-assessment is anchored.
- runtime/persona/chatter.js — narrates step.narration_ru on
  transition. Rate-limited via existing maybeNarrateRaw().

New operator CLI:

- scripts/show-story.js — fetches the live snapshot via IPC sock and
  prints step progress with ✓/→/ markers, current skill, inventory.
  Falls back to --plain catalogue view when bot offline.

Token cost impact: ~+30 input tokens per advise() call (one extra
line in user prompt). Trivial vs the value of grounding LLM advice
in a concrete narrative.

Operator usage:

  node scripts/show-story.js          # live progress + which step + why
  node scripts/show-story.js --plain  # static catalogue of all 11 steps

Tests: 376 green (was 360, +16 storyline tests).

Also in this branch (already committed): dev/v0.3.1/PRD.md —
LLM prompt cost optimization design doc.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix(v0.3.1): storyline beats manifesto L1+ (only L0 alive emergencies override)

Found in live logs after the previous commit deployed:
  storyline: step 1/11: orient_self → explore.wander
  advisor-trigger: firing because wedged (planned=survive.acquire-food, ...)

Manifesto was still picking survive.acquire-food (L1 food) over the
storyline's orient_self → explore.wander. That's the wrong precedence —
storyline expresses a *concrete operational subgoal* and L1+ manifesto
needs are just "you'd benefit from food" priorities, not emergencies.

New dispatch precedence in curriculumReflex:
  1. manifesto L0 (alive emergencies: lava, low-HP+hostile, food=0)
  2. storyline (concrete narrative subgoal — beats L1+ manifesto)
  3. manifesto L1+ (fallback when storyline has no concrete suggestion)
  4. curriculum plan (legacy fallback)

This way the bot starts following the narrative arc even while
manifesto's L1 food is technically unsatisfied — orient_self runs to
completion before pursuing food explicitly. Storyline already handles
food as step 5 (first_food), so we're not skipping it.

Tests: 378 green (+2 priority-ordering tests):
- L0 manifesto emergency: upstream reflex (defend/modes) catches before
  curriculum dispatch
- storyline beats manifesto when both have suggestions: well-fed bot
  with logs → craft.planks (storyline crafting_basics), not gather.logs
  (manifesto L2)
- updated "manifesto fallback" test to require disableStoryline=true

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* feat(tui): fullscreen monitor-only TUI (opencode-style)

Replaces the old tui/tui.tsx hotkey-heavy dashboard with a read-only
observability screen. Operator actions live in scripts/* now —
TUI is for watching, not driving.

Layout (top to bottom, all auto-resizing to terminal):
  1. Header     — MC/IPC status, pos, HP, food, day/night, hostiles
  2. Storyline  — current step + 11-step quest map (✓/→/○)
  3. Activity   — last N skill dispatches (colour by outcome)
  4. MC Chat    — last N chat lines (cyan for bot, yellow for players)
  5. Advisor    — last N LLM recommendations (trigger + outcome + tokens)
  6. Improvements — open requests from knowledge.improvement_requests
  7. Footer     — 24h token usage + cost in ₽ + q-to-quit

Data sources:
  - IPC sock: snapshot frames, log frames, chat frames (push)
  - SQLite knowledge.db: advisor_recommendations + improvement_requests
    polled every 5s (pull)

Token cost displayed live using TIMEWEB_PRICE_IN_RUB_PER_M /
TIMEWEB_PRICE_OUT_RUB_PER_M env vars (defaults: 101 / 608 for
gpt-5.4-mini).

Switches:
  - npm run tui          → new monitor (this file)
  - npm run tui:legacy   → old action-driven tui/tui.tsx (kept for now)

Implementation notes:
  - Uses ink + alternate-screen-buffer ANSI for proper "opencode-feel"
    fullscreen behaviour; restores prior terminal contents on quit.
  - Skips alt-screen and useInput when stdin/stdout isn't a TTY
    (smoke tests, piped output) — both gracefully degrade.
  - Stable React keys via per-event uid counter, avoids reconciler
    duplicate-key warnings as logs/chat/dispatches stream in.
  - Resize handled via 1s stdout-dimension poll, NOT direct
    'resize' listener (which conflicts with ink's own listener and
    triggers MaxListenersExceededWarning).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* ui(tui): compact 4-section monitor (was 6) — fits 1080p without zoom

Operator reported the TUI overflowed the screen unless terminal was
zoomed way out. The 11-step storyline list alone was eating ~13
rows, and each advisor/improvement entry took 2-3 rows. Now:

- Header + storyline collapsed into one panel (2 lines):
    line 1: pepa · ●MC ●IPC · 1m50s · pepa_bot · (697,61,702) · HP 20 · food 5 · ☀ · ⚔60(creeper@58b)
    line 2: story ▓▒░░░░░░░░░ 1/11 orient_self · Понять где я → explore.wander
  The 11-step ladder is now a unicode progress bar (▓ done, ▒ current,
  ░ pending) — same info, fits in one row.

- Advisor entries: one line each instead of two.
    ✓ wedged_60s → survive.flee  802t 1900ms
    (outcome mark / trigger / target skill / tokens / latency)

- Improvements entries: one line each instead of two.
    #1 P2 ×3  Add craft.iron-pickaxe skill
  Description dropped from the row — use `node scripts/list-improvements.js`
  for full text.

- Sections: 4 (was 6).
    [header+story] · [activity | chat] · [advisor | improvements] · [footer]

Tested on a typical 1080p terminal — fits comfortably without zoom.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* feat(v0.3.1): real survival patterns — biome-aware scout, wedge-relocate, escape-pit-safe

Operator reported the bot wandered the same 50×50 patch for 2 hours
without making any progress toward food. Diagnosis showed three root
causes; this commit addresses all five open improvement_requests
the LLM (postmortem + tuner) flagged automatically.

Research basis (`Voyager`, `Plan4MC`, `GITM`, `Mindcraft`):
  - Coverage / commit-to-cardinal exploration when local scan fails
  - Biome-aware strategy switching using a static affordance table
  - Wedge detector above the skill layer that triggers RELOCATE not
    RETRY (per-skill stuck checks reset on re-entry — useless)
  - Time-in-region bbox heuristic + need-duration AND skill-cycle gate

Concrete changes:

1. `runtime/goal/storyline.js`
   - orient_self.completed: added timeout fallback (HP=full + session
     >120s → done) so barren biomes don't block the bot on step 1.
     Closes improvement #2 'Нет навыка оценки когда сменить район'.
   - first_food.suggestSkill: now picks survive.scout-food (new) when
     no passive mob is nearby; falls back to survive.acquire-food only
     when something is in immediate range.

2. `runtime/biome-affordances.js` (new)
   - Static table: 40+ biomes → {has_passive_mobs, has_trees,
     has_water, has_crops, livable}.
   - Unknown biomes return optimistic defaults to avoid regressions.
   - Closes improvement #1 'Нет навыка целевого поиска еды по биому'.

3. `runtime/skills/scout-food.js` (new — survive.scout-food)
   - Tiered strategy: biome check → scan 32 → scan 64 → commit a
     cardinal for 200 blocks rescanning every 16. On cardinal
     exhaustion, returns code:"exhausted" so the curriculum can
     escalate to village.relocate.
   - In barren biomes (desert/ocean/snowy_plains) the scan is
     SKIPPED — bot walks straight toward the nearest neighbour
     biome that affords passive mobs (8-direction biome probe at
     radius 64).

4. `runtime/awareness/wedge-detector.js` (new)
   - Rolling 10-min position bbox tracker. observe() called every
     tick; isWedged() returns true when bbox<50 AND active need
     unmet >5min AND skill cycles ≥3.
   - markRelocationStarted() suppresses further wedge firings
     until the bot has displaced ≥200b — prevents stack overflow
     of relocate calls.
   - Lives ABOVE the skill layer (in runtime/reflex.js), because
     any per-skill stuck check resets on re-entry.

5. `runtime/skills/relocate.js` (new — village.relocate)
   - 300-block walk in least-recently-used cardinal (per-incident
     memory in ctx.recentRelocations).
   - Re-paths every 32 blocks, soft-tolerates pathfinder failures
     (3 consecutive throws → exit with code:"stuck_in_place").
   - Closes improvement #2 + #4 ('low success rate trigger').

6. `runtime/skills/escape-pit-safe.js` (new — recovery.escape-pit-safe)
   - Surveys 4 cardinals AND ceiling height before committing.
     Picks the direction with most open blocks (≥3, no lava).
     Falls through to pillar-up only if ceiling clear ≥4b. Returns
     code:"no_strategy" if both blocked so curriculum can escalate
     to relocate.
   - Closes improvement #3 'Нет навыка для безопасного выхода'.

7. `runtime/reflex.js`
   - Wedge detector wired before manifesto/storyline. If wedge.wedged
     is true, dispatches village.relocate directly and returns —
     bypasses every other branch.
   - ctx.disableWedge flag for tests.

8. `runtime/coach/advisor-trigger.js`
   - LLM provider outage backoff: 3 consecutive http_400 / timeout /
     network_error → suppress advisor for 10 min. Today's TimeWeb
     gpt-5.4-mini was 400'ing for an hour straight; we were spending
     trigger budget on dead calls. Closes improvement implicit gap
     in #5.

9. `runtime/bot.js`
   - Tracks botSpawnedAt; snapshot._sessionMs exposed for storyline
     orient_self timeout fallback.

Tests: 396 green (was 378, +18):
  - runtime/biome-affordances.test.js — 8 tests
  - runtime/awareness/wedge-detector.test.js — 9 tests
  - runtime/goal/storyline.test.js — 1 new test (orient_self timeout)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix(coach/trigger-tuner): crash after ~1h — runOnce is sync, not a Promise

The live bot died overnight with:
  TypeError: runOnce(...).catch is not a function
  at trigger-tuner.js:42  →  [supervisor] child exited code=1

attach() wrapped the timer body as `runOnce().catch(...)` but
runOnce() returns a plain {ok, flagged, ...} object (pure SQL, no
await). The first tuner tick (60min after spawn) threw → killed the
whole bot process. Never surfaced before because the bot rarely ran
uninterrupted for a full hour during development.

Fix: guard the synchronous call with try/catch, matching how
persona/chatter.js already does its sync tick. (postmortem.drainOnce
and reflect.runOnce ARE async, so their .catch is correct — audited.)

Regression test added: captures the setInterval callback and invokes
it synchronously, asserting it does not throw.

Tests: 397 green.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix(v0.3.1): mechanical food/stuck fixes — bot reaches the chicken now

The wedge wasn't only in the manifesto layer; several mechanical bugs
kept the bot in a dead random-walk:

- storyline / manifesto / curriculum: "local food" now means an edible
  passive mob within <=32 blocks. A distant chicken or a cod no longer
  fools the bot into dispatching acquire-food (which then fails on
  no_path). Long-range food goes through scout-food instead.

- scout-food: partial approach to a target now counts as progress
  (approached_target, e.g. moved:14); a blocked heading is NOT counted
  as movement; added blind/tunnel fallback so it doesn't die when the
  pathfinder can't route cleanly.

- acquire-food: on no_path it now also tries a blind/tunnel approach to
  the animal; no_drop routes back into food scouting instead of giving
  up.

- explore.far / relocate / flee: fewer false "done" results (micro-steps
  no longer counted as success), more genuine escapes from stuck.

- scripts/show-story.js: live IPC now actually renders the current
  storyline step.

Verification: scripts/lint-patch.js clean; npm test 404/404 green; bot
relaunched in tmux `pepa`. Live logs show real progress — bot switched
to survive.scout-food, approached the chicken (approached_target
moved:14), then reached survive.acquire-food: hunting chicken. Food
isn't fully closed yet but the remaining issue is concrete pickup/drop,
not dead random-walk.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* fix(v0.3.1): sated bot stops chasing food + perf-leak + fuzzy improvement dedup

Third day of "bot just walks back and forth burning tokens". Root
causes were mechanical, not the manifesto:

1. SATED BOT CHASING FOOD (the big one)
   Bot had food=17 (nearly full) but storyline first_food + manifesto
   L1 required 2+ food ITEMS in inventory, so it looped scout-food /
   acquire-food for hours instead of working. Now both treat a hunger
   bar >= 14 (SATED_FOOD) as satisfied even with empty food inventory —
   a full bot chops wood / makes tools and grabs food opportunistically,
   only hard-pursuing food when actually hungry (< 14).
   manifesto/needs.js foodDetect + goal/storyline.js first_food.completed.

2. perf_hooks MEMORY LEAK (overnight OOM suspect)
   "MaxPerformanceEntryBufferExceededWarning: 1,000,001 measure entries".
   mineflayer/pathfinder emit perf marks we never consume. Added a
   60s reaper in bot.js (performance.clearMeasures/clearMarks). unref'd.

3. IMPROVEMENT QUEUE SELF-DUPLICATING
   The LLM re-filed closed gaps with reworded titles (#5/#8/#9 were
   dupes of implemented #1/#2/#3). Exact-title dedup missed them.
   Replaced with token-set fuzzy match (isDuplicateTitle): jaccard>=0.75
   OR >=3 shared meaningful tokens with jaccard>=0.5. Also: a re-filed
   gap that's already implemented/rejected is NOT resurrected as a new
   open row. Cleared all 5 open requests (now genuinely implemented).

Also confirmed (no change needed):
- canDig=true is a DELIBERATE codebase-wide choice ("without it the bot
  gets permanently stuck", actions.js). The stale memory recommending
  canDig=false is updated. ViaBackwards dig works partially (dug:1
  moved:1.8 observed); false would trap the bot in every pit.
- scout-food already has blind/tunnel fallback + 12s step timeout
  (operator's earlier edits) so trapped-pathfinder degrades instead of
  hanging 30s.

Tests: 407 green (was 404). Updated needs/state/storyline tests for the
SATED_FOOD threshold; added fuzzy-dedup + tokenize/jaccard tests.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

---------

Co-authored-by: Yuriy Mayatnikov <mayatnikov@me.com>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
This commit was merged in pull request #28.
This commit is contained in:
Yuriy Mayatnikov
2026-05-28 09:43:57 +03:00
committed by GitHub
co-authored by Claude Opus 4.7 mayatnikov
parent 7d96e44804
commit 15b6c11002
37 changed files with 3259 additions and 71 deletions
+487
View File
@@ -0,0 +1,487 @@
/**
* pepa monitor — fullscreen, read-only TUI.
*
* Replaces the old action-heavy tui/tui.tsx with a pure observability
* dashboard inspired by opencode's full-screen layout. No hotkeys to
* dispatch skills / send chat / approve proposals — operator does that
* via scripts/* or by writing to the IPC sock directly. This screen
* just shows what the bot is doing, in colour.
*
* Panels (top → bottom):
* 1. Header — connection / position / hp / food / time
* 2. Storyline — current step + the 11-step quest map
* 3. Activity — last N dispatches (colour-coded by outcome)
* + MC chat — last N lines
* 4. Advisor — last 6 LLM recommendations (trigger / outcome / tokens)
* + Improvements — open queue from improvement_requests
* 5. Footer — token usage today, advisor stats, q to quit
*
* Live data sources:
* - IPC sock (snapshot frames, log frames, chat frames)
* - SQLite (knowledge.db) polled every 5s for advisor + improvements
*/
import React, { useEffect, useReducer, useState } from "react";
import { render, Box, Text, useApp, useInput, useStdout } from "ink";
import { createIpcClient } from "./ipc-client.js";
import { EVENT_TYPES } from "../runtime/ipc-protocol.js";
import { stateDir } from "../runtime/config.js";
import { initKnowledge, isAvailable as knowledgeReady, recentRecommendations, listImprovements, recommendationStats } from "../runtime/knowledge/index.js";
// --- types ------------------------------------------------------------------
type LogEntry = { ts: string; level: string; source: string; text: string };
type ChatEntry = { uid: number; ts: string; from: string; text: string; kind: string };
type Snapshot = Record<string, any>;
type Dispatch = { uid: number; ts: number; kind: "start" | "end"; label: string; ok?: boolean; code?: string; detail?: string };
let _uidSeq = 0;
function nextUid() { _uidSeq += 1; return _uidSeq; }
type State = {
connectedIpc: boolean;
snapshot: Snapshot;
logs: LogEntry[];
chat: ChatEntry[];
dispatches: Dispatch[];
startedAt: number;
};
type Action =
| { type: "ipc-connected" }
| { type: "ipc-disconnected" }
| { type: "snapshot"; payload: Snapshot }
| { type: "log"; payload: LogEntry }
| { type: "chat"; payload: { from: string; text: string; kind: string }; ts: string };
const MAX_LOGS = 200;
const MAX_CHAT = 50;
const MAX_DISPATCHES = 30;
// --- reducer ----------------------------------------------------------------
function reducer(state: State, action: Action): State {
switch (action.type) {
case "ipc-connected":
return { ...state, connectedIpc: true };
case "ipc-disconnected":
return { ...state, connectedIpc: false };
case "snapshot":
return { ...state, snapshot: action.payload };
case "log": {
const logs = [...state.logs, action.payload].slice(-MAX_LOGS);
// also derive a dispatches view: lines like "→ label" or "← label ok/fail (...)"
const dispatches = extractDispatch(state.dispatches, action.payload);
return { ...state, logs, dispatches };
}
case "chat": {
const chat = [...state.chat, { uid: nextUid(), ts: action.ts, ...action.payload }].slice(-MAX_CHAT);
return { ...state, chat };
}
default:
return state;
}
}
function extractDispatch(prev: Dispatch[], log: LogEntry): Dispatch[] {
if (log.source !== "dispatch") return prev;
const ts = Date.parse(log.ts) || Date.now();
// "→ label" — skill starting
const startMatch = /^→\s+(\S+)/.exec(log.text);
if (startMatch) {
return [...prev, { uid: nextUid(), ts, kind: "start", label: startMatch[1] }].slice(-MAX_DISPATCHES);
}
// "← label ok/fail (...)" — skill ending
const endMatch = /^←\s+(\S+)\s+(ok|fail)(?:\s+\((.*)\))?/.exec(log.text);
if (endMatch) {
const [, label, outcome, detail] = endMatch;
return [...prev, { uid: nextUid(), ts, kind: "end", label, ok: outcome === "ok", detail }].slice(-MAX_DISPATCHES);
}
return prev;
}
// --- helpers ----------------------------------------------------------------
function formatAge(ts: number) {
const dt = Math.max(0, Date.now() - ts);
if (dt < 60_000) return `${Math.floor(dt / 1000)}s`;
if (dt < 3600_000) return `${Math.floor(dt / 60_000)}m`;
const h = Math.floor(dt / 3600_000);
const m = Math.floor((dt % 3600_000) / 60_000);
return `${h}h${m}m`;
}
function formatDuration(ms: number) {
const s = Math.floor(ms / 1000) % 60;
const m = Math.floor(ms / 60_000) % 60;
const h = Math.floor(ms / 3600_000);
if (h > 0) return `${h}h${m.toString().padStart(2, "0")}m`;
if (m > 0) return `${m}m${s.toString().padStart(2, "0")}s`;
return `${s}s`;
}
function shortTime(ts: number) {
const d = new Date(ts);
return `${d.getHours().toString().padStart(2, "0")}:${d.getMinutes().toString().padStart(2, "0")}:${d.getSeconds().toString().padStart(2, "0")}`;
}
function hpColor(hp: number | undefined) {
if (hp == null) return "gray";
if (hp <= 5) return "red";
if (hp <= 12) return "yellow";
return "green";
}
function foodColor(food: number | undefined) {
if (food == null) return "gray";
if (food <= 4) return "red";
if (food <= 10) return "yellow";
return "green";
}
// --- components -------------------------------------------------------------
function StatusHeader({ snapshot, connectedIpc, width, startedAt }: { snapshot: Snapshot; connectedIpc: boolean; width: number; startedAt: number }) {
const pos = snapshot.position
? `(${Math.round(snapshot.position.x)},${Math.round(snapshot.position.y)},${Math.round(snapshot.position.z)})`
: "?";
const hp = snapshot.health;
const food = snapshot.food;
const day = snapshot.isDay ? "☀" : "🌙";
const session = formatDuration(Date.now() - startedAt);
const mcOnline = snapshot.connected;
const story = snapshot.storyStep;
const idx = story?.index ?? 0;
const cur = story?.step;
const want = story?.suggestion?.skillId;
// 11-step progress bar in 11 cells
const bar = Array.from({ length: 11 }, (_, i) => {
if (i < idx) return "▓";
if (i === idx) return "▒";
return "░";
}).join("");
return (
<Box flexDirection="column" borderStyle="round" borderColor="cyan" paddingX={1} width={width}>
<Box>
<Text bold color="cyan">pepa</Text>
<Text dimColor> · </Text>
<Text color={mcOnline ? "green" : "red"}>{mcOnline ? "●MC" : "○MC"}</Text>
<Text dimColor> · </Text>
<Text color={connectedIpc ? "green" : "red"}>{connectedIpc ? "●IPC" : "○IPC"}</Text>
<Text dimColor> · </Text>
<Text>{session}</Text>
<Text dimColor> · </Text>
<Text bold>{snapshot.username ?? "?"}</Text>
<Text dimColor> · </Text>
<Text>{pos}</Text>
<Text dimColor> · </Text>
<Text>HP </Text><Text bold color={hpColor(hp)}>{hp ?? "?"}</Text>
<Text dimColor> · </Text>
<Text>food </Text><Text bold color={foodColor(food)}>{food ?? "?"}</Text>
<Text dimColor> · </Text>
<Text>{day}</Text>
{(snapshot.hostileCount ?? 0) > 0 ? (
<>
<Text dimColor> · </Text>
<Text color="red">{snapshot.hostileCount}</Text>
{snapshot.closestHostile ? <Text color="red">({snapshot.closestHostile.name}@{Math.round(snapshot.closestHostile.distance)}b)</Text> : null}
</>
) : null}
</Box>
<Box>
<Text color="magenta">story </Text>
<Text bold>{bar}</Text>
<Text dimColor> </Text>
<Text bold>{idx + 1}/11</Text>
<Text dimColor> </Text>
{cur ? (
<>
<Text color="white" bold>{cur.id}</Text>
<Text dimColor> · </Text>
<Text>{cur.title}</Text>
</>
) : <Text dimColor>(no story)</Text>}
{want ? (
<>
<Text dimColor> </Text>
<Text color="cyan">{want}</Text>
</>
) : null}
{story?.emergency ? <Text color="red" bold> [EMERGENCY]</Text> : null}
</Box>
</Box>
);
}
function ActivityPanel({ dispatches, width, height }: { dispatches: Dispatch[]; width: number; height: number }) {
const visible = dispatches.slice(-height);
return (
<Box flexDirection="column" borderStyle="round" borderColor="blue" paddingX={1} width={width}>
<Text bold color="blue">Activity (last {visible.length})</Text>
{visible.map((d) => {
if (d.kind === "start") {
return (
<Text key={d.uid} dimColor>
{shortTime(d.ts)} <Text color="white">{d.label}</Text>
</Text>
);
}
const color = d.ok ? "green" : "red";
const tail = d.detail ? ` (${String(d.detail).slice(0, 30)})` : "";
return (
<Text key={d.uid}>
<Text dimColor>{shortTime(d.ts)} </Text>
<Text color={color}> {d.label} {d.ok ? "ok" : "fail"}</Text>
<Text dimColor>{tail}</Text>
</Text>
);
})}
{visible.length === 0 ? <Text dimColor>(waiting for activity)</Text> : null}
</Box>
);
}
function ChatPanel({ chat, width, height }: { chat: ChatEntry[]; width: number; height: number }) {
const visible = chat.slice(-height);
return (
<Box flexDirection="column" borderStyle="round" borderColor="green" paddingX={1} width={width}>
<Text bold color="green">MC Chat (last {visible.length})</Text>
{visible.map((c) => (
<Text key={c.uid}>
<Text dimColor>{c.ts?.slice(11, 16) ?? ""} </Text>
<Text color={c.kind === "system" ? "gray" : c.from === "pepa_bot" ? "cyan" : "yellow"} bold={c.from !== "pepa_bot" && c.kind !== "system"}>
{c.from}:
</Text>
<Text> {c.text.slice(0, width - 12)}</Text>
</Text>
))}
{visible.length === 0 ? <Text dimColor>(no chat yet)</Text> : null}
</Box>
);
}
function AdvisorPanel({ recs, width, height }: { recs: any[]; width: number; height: number }) {
const visible = recs.slice(0, height);
return (
<Box flexDirection="column" borderStyle="round" borderColor="yellow" paddingX={1} width={width}>
<Text bold color="yellow">Advisor (last {visible.length})</Text>
{visible.map((r) => {
const ok = r.outcome_ok;
const outcomeMark = ok == null ? "·" : ok ? "✓" : "✗";
const outcomeColor = ok == null ? "gray" : ok ? "green" : "red";
const target = r.recommended_skill ?? r.action;
const trigger = String(r.trigger_reason ?? "?");
// One line per recommendation: " ✓ wedged_60s → survive.flee 802t 1900ms"
const lhs = `${outcomeMark} ${trigger}${target}`;
const rhs = `${r.tokens_in ?? "?"}t ${r.latency_ms ?? "?"}ms`;
const free = Math.max(20, width - rhs.length - 6);
return (
<Text key={`adv-${r.id}`}>
<Text color={outcomeColor} bold>{outcomeMark} </Text>
<Text color="white">{trigger}</Text>
<Text dimColor> </Text>
<Text color="cyan">{String(target).slice(0, Math.max(8, free - trigger.length - 6))}</Text>
<Text dimColor> {rhs}</Text>
</Text>
);
})}
{visible.length === 0 ? <Text dimColor>(no advisor calls yet)</Text> : null}
</Box>
);
}
function ImprovementsPanel({ items, width, height }: { items: any[]; width: number; height: number }) {
const visible = items.slice(0, height);
return (
<Box flexDirection="column" borderStyle="round" borderColor="magenta" paddingX={1} width={width}>
<Text bold color="magenta">Improvements (open {items.length})</Text>
{visible.map((r) => (
<Text key={`imp-${r.id}`}>
<Text color="gray">#{r.id} </Text>
<Text color="yellow">P{r.priority}</Text>
<Text dimColor> ×{r.votes} </Text>
<Text color="white">{String(r.title).slice(0, Math.max(20, width - 14))}</Text>
</Text>
))}
{visible.length === 0 ? <Text dimColor>(no improvement requests yet)</Text> : null}
</Box>
);
}
function Footer({ stats, width }: { stats: any[]; width: number }) {
const total = stats.reduce(
(acc, s) => ({
calls: acc.calls + (s.total ?? 0),
succ: acc.succ + (s.succeeded ?? 0),
fail: acc.fail + (s.failed ?? 0),
in: acc.in + (s.avg_in ?? 0) * (s.total ?? 0),
out: acc.out + (s.avg_out ?? 0) * (s.total ?? 0),
}),
{ calls: 0, succ: 0, fail: 0, in: 0, out: 0 },
);
const priceInRub = Number(process.env.TIMEWEB_PRICE_IN_RUB_PER_M) || 101;
const priceOutRub = Number(process.env.TIMEWEB_PRICE_OUT_RUB_PER_M) || 608;
const costRub = (total.in * priceInRub + total.out * priceOutRub) / 1_000_000;
return (
<Box flexDirection="column" borderStyle="round" borderColor="gray" paddingX={1} width={width}>
<Box>
<Text dimColor>Last 24h: </Text>
<Text bold>{total.calls}</Text>
<Text dimColor> advisor calls (</Text>
<Text color="green">{total.succ} ok</Text>
<Text dimColor> / </Text>
<Text color="red">{total.fail} fail</Text>
<Text dimColor>) tokens </Text>
<Text bold>{Math.round(total.in / 1000)}K</Text>
<Text dimColor> in / </Text>
<Text bold>{Math.round(total.out / 1000)}K</Text>
<Text dimColor> out </Text>
<Text bold color="cyan">{costRub.toFixed(2)} </Text>
</Box>
<Box>
<Text dimColor>q quit (bot keeps running)</Text>
</Box>
</Box>
);
}
// --- main app ---------------------------------------------------------------
function App() {
const { exit } = useApp();
const { stdout } = useStdout();
// Sample stdout dimensions periodically. Subscribing directly to
// stdout.on('resize') from a React hook conflicts with ink's own
// listener and produces "MaxListenersExceededWarning" / reconciler
// errors. A 1s poll is cheap and good enough — terminals don't
// resize often.
const [cols, setCols] = useState<number>((stdout as any)?.columns ?? 120);
const [rows, setRows] = useState<number>((stdout as any)?.rows ?? 30);
useEffect(() => {
const t = setInterval(() => {
setCols((stdout as any)?.columns ?? 120);
setRows((stdout as any)?.rows ?? 30);
}, 1000);
return () => clearInterval(t);
}, [stdout]);
const [state, dispatch] = useReducer(reducer, {
connectedIpc: false,
snapshot: {},
logs: [],
chat: [],
dispatches: [],
startedAt: Date.now(),
});
const [client] = useState(() => createIpcClient());
const [knowledgeOk, setKnowledgeOk] = useState(false);
const [recs, setRecs] = useState<any[]>([]);
const [improvements, setImprovements] = useState<any[]>([]);
const [stats, setStats] = useState<any[]>([]);
// IPC
useEffect(() => {
const onConnected = () => dispatch({ type: "ipc-connected" });
const onDisconnected = () => dispatch({ type: "ipc-disconnected" });
const onFrame = (frame: any) => {
switch (frame.type) {
case EVENT_TYPES.STATUS: dispatch({ type: "snapshot", payload: frame.payload }); break;
case EVENT_TYPES.LOG: dispatch({ type: "log", payload: frame.payload }); break;
case EVENT_TYPES.CHAT: dispatch({ type: "chat", payload: frame.payload, ts: frame.ts }); break;
case EVENT_TYPES.HELLO:
if (frame.payload?.snapshot) dispatch({ type: "snapshot", payload: frame.payload.snapshot });
if (frame.payload?.recentLogs) {
for (const lg of frame.payload.recentLogs) dispatch({ type: "log", payload: lg });
}
break;
}
};
(client as any).on("connected", onConnected);
(client as any).on("disconnected", onDisconnected);
(client as any).on("frame", onFrame);
return () => {
(client as any).off("connected", onConnected);
(client as any).off("disconnected", onDisconnected);
(client as any).off("frame", onFrame);
client.close();
};
}, [client]);
// Knowledge DB init + polling
useEffect(() => {
let mounted = true;
(async () => {
try {
await initKnowledge({ stateDir });
if (!mounted) return;
setKnowledgeOk(knowledgeReady());
} catch {}
})();
const poll = () => {
if (!knowledgeReady()) return;
try {
setRecs(recentRecommendations({ limit: 12 }));
setImprovements(listImprovements({ status: "open", limit: 12 }));
setStats(recommendationStats({ sinceHours: 24 }));
} catch {}
};
poll();
const t = setInterval(poll, 5000);
return () => { mounted = false; clearInterval(t); };
}, [knowledgeOk]);
// useInput requires TTY raw mode; skip it when stdin isn't a TTY
// (e.g. when piped during a smoke test). Real `npm run tui` is always TTY.
const isTty = !!process.stdin.isTTY;
if (isTty) {
// eslint-disable-next-line react-hooks/rules-of-hooks
useInput((input, key) => {
if (input === "q" || (key.ctrl && input === "c") || key.escape) {
client.close();
exit();
}
});
}
// Layout math: compact 4-section vertical stack.
// row budget:
// header (story + status) ≈ 4 rows
// middle activity/chat ≈ floor((rows - 4 - 4 - 4) / 2)
// bottom advisor/improvements ≈ same
// footer ≈ 4 rows
const totalWidth = Math.max(80, cols);
const halfWidth = Math.floor(totalWidth / 2);
const totalRows = Math.max(20, rows);
const middleRows = Math.max(5, Math.floor((totalRows - 4 - 4 - 4) / 2));
return (
<Box flexDirection="column" width={totalWidth}>
<StatusHeader snapshot={state.snapshot} connectedIpc={state.connectedIpc} width={totalWidth} startedAt={state.startedAt} />
<Box flexDirection="row" width={totalWidth}>
<ActivityPanel dispatches={state.dispatches} width={halfWidth} height={middleRows} />
<ChatPanel chat={state.chat} width={totalWidth - halfWidth} height={middleRows} />
</Box>
<Box flexDirection="row" width={totalWidth}>
<AdvisorPanel recs={recs} width={halfWidth} height={middleRows} />
<ImprovementsPanel items={improvements} width={totalWidth - halfWidth} height={middleRows} />
</Box>
<Footer stats={stats} width={totalWidth} />
</Box>
);
}
// --- entrypoint -------------------------------------------------------------
// Use the alternate screen buffer so the dashboard owns the entire viewport
// and the operator's prior terminal contents come back after quit.
// Skip when stdout isn't a TTY (e.g. piped output during smoke tests).
const useAltScreen = !!process.stdout.isTTY && process.env.PEPA_TUI_NO_ALT !== "1";
if (useAltScreen) {
process.stdout.write("\x1b[?1049h\x1b[2J\x1b[H");
const onExit = () => process.stdout.write("\x1b[?1049l");
process.on("exit", onExit);
process.on("SIGINT", () => { onExit(); process.exit(0); });
process.on("SIGTERM", () => { onExit(); process.exit(0); });
}
render(<App />);