@lawaty/lacode 0.7.1 → 0.8.0

package/dist/pool.js

@@ -34,12 +34,15 @@ var MAX_CLAIM_AGE_MS = 30 * 60 * 1e3;
 var HANG_COOLDOWN_MS = 10 * 60 * 1e3;
 var UNAVAILABLE_COOLDOWN_MS = 60 * 1e3;
 var MAX_WEIGHT = 100;
+var SLOT_CEILING_FACTOR = 2;
+var slotCeiling = (slot) => slot.weight * SLOT_CEILING_FACTOR;
 var MAX_SLOTS = 4;
 var COOLDOWN_BASE_MS = 60 * 1e3;
 var COOLDOWN_MAX_MS = 15 * 60 * 1e3;
 var QUOTA_COOLDOWN_BASE_MS = 15 * 60 * 1e3;
 var QUOTA_COOLDOWN_MAX_MS = 6 * 60 * 60 * 1e3;
 var STRIKE_DECAY_MS = 30 * 60 * 1e3;
+var STATED_COOLDOWN_MAX_MS = QUOTA_COOLDOWN_MAX_MS;
 var isQuotaReason = (reason) => /quota|limit|rate|too-many|overloaded|429/i.test(reason);
 var BASES = ["explore", "implement", "context"];
 var PRESETS_DIR = join2(PACKAGE_ROOT, "presets");
@@ -207,12 +210,13 @@ var isUnavailableText = (text) => UNAVAILABLE_PATTERN.test(text);
 var TRANSIENT_GATEWAY_PATTERNS = [/queue budget/i, /exceeded queue/i, /aborted due to timeout/i, /timeout/i];
 var isTransientGatewayText = (text) => TRANSIENT_GATEWAY_PATTERNS.some((re) => re.test(text));
 var statedCooldown = (text) => {
-  const hm = /resets?\s+in\s+(\d+)\s*(?:h|hr|hour)s?\s*(\d+)?\s*(?:m|min|minute)?s?/i.exec(text);
-  if (hm) return (Number(hm[1]) * 60 + (hm[2] ? Number(hm[2]) : 0)) * 60 * 1e3;
-  const m = /resets?\s+in\s+(\d+)\s*(?:m|min|minute)s?/i.exec(text);
-  if (m) return Number(m[1]) * 60 * 1e3;
-  const s = /retry[\s-]?in\s+(\d+)\s*(?:s|sec|second)s?/i.exec(text);
-  if (s) return Number(s[1]) * 1e3;
+  const bounded = (ms) => Number.isFinite(ms) && ms > 0 && ms <= STATED_COOLDOWN_MAX_MS ? ms : void 0;
+  const hm = /resets?\s+in\s+(\d+)\s*(?:hour|hr|h)s?\s*(\d+)?\s*(?:minute|min|m)?s?/i.exec(text);
+  if (hm) return bounded((Number(hm[1]) * 60 + (hm[2] ? Number(hm[2]) : 0)) * 60 * 1e3);
+  const m = /resets?\s+in\s+(\d+)\s*(?:minute|min|m)s?/i.exec(text);
+  if (m) return bounded(Number(m[1]) * 60 * 1e3);
+  const s = /retry[\s-]?in\s+(\d+)\s*(?:second|sec|s)s?/i.exec(text);
+  if (s) return bounded(Number(s[1]) * 1e3);
   return void 0;
 };
 var retryAfterMs = (headers) => {
@@ -246,10 +250,14 @@ function state(opts = {}) {
       cooling: /* @__PURE__ */ new Map(),
       strikes: /* @__PURE__ */ new Map(),
       activity: /* @__PURE__ */ new Map(),
+      probes: /* @__PURE__ */ new Set(),
+      probing: false,
+      lastProbe: 0,
       children: /* @__PURE__ */ new Map(),
       reaped: 0,
       reaping: false,
       lastSweep: 0,
+      lastDirSweep: 0,
       id,
       dir,
       modelsFile,
@@ -286,10 +294,88 @@ function init(s, onWarn) {
     onWarn(`claim dir ${s.dir} unusable; pool load tracked in memory only`);
   }
 }
+var STATE_SWEEP_INTERVAL_MS = 60 * 60 * 1e3;
+function sweepStateDir(s) {
+  const removed = [];
+  if (!s.dirReady) return removed;
+  let entries;
+  try {
+    entries = readdirSync(s.dir);
+  } catch {
+    return removed;
+  }
+  const now = Date.now();
+  const drop = (name) => {
+    try {
+      unlinkSync(join2(s.dir, name));
+      removed.push(name);
+    } catch {
+    }
+  };
+  for (const entry of entries) {
+    let age;
+    try {
+      age = now - statSync(join2(s.dir, entry)).mtimeMs;
+    } catch {
+      continue;
+    }
+    if (entry.endsWith(".tmp")) {
+      if (age > ORPHAN_TMP_MS) drop(entry);
+      continue;
+    }
+    const claim = /^claims\.(.+)\.json$/.exec(entry);
+    if (claim) {
+      if (claim[1] === s.id) continue;
+      if (/^\d+$/.test(claim[1]) && !isPidAlive(Number(claim[1]))) {
+        drop(entry);
+        continue;
+      }
+      if (age > DEAD_FILE_MS) drop(entry);
+      continue;
+    }
+    if (entry.startsWith("limit.") && entry.endsWith(".json")) {
+      try {
+        const strike = JSON.parse(readFileSync2(join2(s.dir, entry), "utf8"));
+        if (typeof strike.until !== "number") {
+          drop(entry);
+          continue;
+        }
+        if (strike.until > now) continue;
+        if (typeof strike.last === "number" && now - strike.last < STRIKE_DECAY_MS) continue;
+        drop(entry);
+      } catch {
+        drop(entry);
+      }
+    }
+  }
+  return removed;
+}
 function pruneOwn(s) {
   const now = Date.now();
   for (const [key, claim] of s.own) if (now - (claim.hb ?? claim.t) > CLAIM_TTL_MS) s.own.delete(key);
 }
+function pruneActivity(s, now) {
+  if (s.activity.size === 0) return;
+  const live = /* @__PURE__ */ new Set();
+  for (const claim of s.own.values()) if (claim.g) live.add(claim.g);
+  for (const [id, seen] of s.activity) {
+    if (live.has(id)) continue;
+    if (now - seen > CLAIM_TTL_MS) s.activity.delete(id);
+  }
+}
+function pruneChildren(s, now) {
+  if (s.children.size === 0) return;
+  for (const [parent, ring] of s.children) {
+    const fresh = ring.filter((child) => now - child.t <= CHILD_MEMORY_MS);
+    if (fresh.length === 0) s.children.delete(parent);
+    else if (fresh.length !== ring.length) s.children.set(parent, fresh);
+  }
+}
+function pruneSessions(s) {
+  const now = Date.now();
+  pruneActivity(s, now);
+  pruneChildren(s, now);
+}
 function publish(s) {
   if (!s.dirReady) return;
   const now = Date.now();
@@ -435,6 +521,50 @@ function coolUnavailable(s, model) {
   }
   return strike;
 }
+function clearCooldowns(s, models) {
+  const cleared = [];
+  const now = Date.now();
+  const wanted = models && models.length > 0 ? new Set(models) : void 0;
+  let entries;
+  try {
+    entries = readdirSync(s.dir);
+  } catch {
+    return cleared;
+  }
+  for (const entry of entries) {
+    if (!entry.startsWith("limit.") || !entry.endsWith(".json")) continue;
+    const path = join2(s.dir, entry);
+    let strike;
+    try {
+      const parsed = JSON.parse(readFileSync2(path, "utf8"));
+      if (parsed && typeof parsed.model === "string") strike = parsed;
+    } catch {
+      strike = void 0;
+    }
+    if (wanted && (strike === void 0 || !wanted.has(strike.model))) continue;
+    try {
+      unlinkSync(path);
+    } catch {
+      continue;
+    }
+    if (strike === void 0) {
+      cleared.push({ model: entry, reason: "unreadable cooldown file", strikes: 0, secondsLeft: 0 });
+      continue;
+    }
+    s.cooling.delete(strike.model);
+    s.strikes.delete(strike.model);
+    cleared.push({
+      model: strike.model,
+      reason: typeof strike.reason === "string" ? strike.reason : "unknown",
+      strikes: typeof strike.strikes === "number" ? strike.strikes : 0,
+      secondsLeft: typeof strike.until === "number" ? Math.max(0, Math.ceil((strike.until - now) / 1e3)) : 0
+    });
+  }
+  return cleared;
+}
+var markProbe = (s, sessionID) => s.probes.add(sessionID);
+var unmarkProbe = (s, sessionID) => s.probes.delete(sessionID);
+var isProbe = (s, sessionID) => s.probes.has(sessionID);
 function refreshCatalog(s, serverUrl) {
   if (!serverUrl || Date.now() - s.catalogAt < CATALOG_TTL_MS) return;
   s.catalogAt = Date.now();
@@ -457,6 +587,7 @@ function isAvailable(s, model, cooling) {
 }
 function snapshot(s) {
   pruneOwn(s);
+  pruneSessions(s);
   const siblings = readSiblings(s);
   const cooling = readLimits(s);
   const counts = /* @__PURE__ */ new Map();
@@ -466,7 +597,7 @@ function snapshot(s) {
   return { counts, cooling, siblings, own: s.own, dirReady: s.dirReady };
 }
 var loadOf = (counts, model) => counts.get(model) ?? 0;
-var underCeiling = (slot, counts) => loadOf(counts, slot.model) < slot.weight;
+var underCeiling = (slot, counts) => loadOf(counts, slot.model) < slotCeiling(slot);
 function roundRobinSequence(slots) {
   const seq = [];
   for (const slot of slots) for (let turn = 0; turn < slot.weight; turn++) seq.push(slot.index);
@@ -629,6 +760,9 @@ export {
   PRESETS_DIR,
   QUOTA_COOLDOWN_BASE_MS,
   QUOTA_COOLDOWN_MAX_MS,
+  SLOT_CEILING_FACTOR,
+  STATED_COOLDOWN_MAX_MS,
+  STATE_SWEEP_INTERVAL_MS,
   STUCK_IDLE_MS,
   STUCK_MIN_AGE_MS,
   UNAVAILABLE_COOLDOWN_MS,
@@ -640,6 +774,7 @@ export {
   backfillChild,
   bindChild,
   claimKey,
+  clearCooldowns,
   condemnReason,
   coolProvider,
   coolUnavailable,
@@ -650,10 +785,12 @@ export {
   init,
   isAvailable,
   isLimitText,
+  isProbe,
   isQuotaReason,
   isTransientGatewayText,
   isUnavailableText,
   manualVariant,
+  markProbe,
   noteActivity,
   parseSlotText,
   parseSlots,
@@ -661,6 +798,7 @@ export {
   pickSlot,
   providerOf,
   pruneOwn,
+  pruneSessions,
   publish,
   readLimits,
   readSiblings,
@@ -673,10 +811,13 @@ export {
   retryAfterMs,
   roundRobinSequence,
   selfFile,
+  slotCeiling,
   slotFor,
   snapshot,
   splitModel,
   state,
   statedCooldown,
-  underCeiling
+  sweepStateDir,
+  underCeiling,
+  unmarkProbe
 };

package/docs/OMNIROUTE.md

@@ -25,8 +25,10 @@ as an opencode provider for you. It is **off by default**: if you do not set
    `auto` is the zero-config default, but on a free tier it often resolves to a
    slow chain. [A tested fast combo](#a-tested-fast-combo-free-mix) below is
    ordered for latency and is worth applying on a new client.
-3. Optionally, to route pooled work through it, make `omniroute/auto 3` the
-   **first** line of `~/.config/lacode/pool.txt`. `pool.txt` is optional user
+3. Optionally, to route pooled work through it, add an `omniroute/<id>` slot to
+   `~/.config/lacode/pool.txt` exactly as you would any other provider. The
+   integration gives it no special precedence, priority or weight — its place and
+   weight in the pool are whatever you configure. `pool.txt` is optional user
    config; the bundled default preset does **not** include OmniRoute.
 4. **Restart opencode.** Nothing here hot-reloads.
 
@@ -64,7 +66,7 @@ The full option set, for reference:
 | `enabled` | `false` | Turns the integration on. |
 | `baseURL` | `http://localhost:20128/v1` | Gateway base URL, normalised to exactly one `/v1`. |
 | `apiKey` | — | Bearer key. **`OMNIROUTE_API_KEY` overrides this value.** The key is never logged. |
-| `combo` | — | An OmniRoute combo id (or `auto`) to expose as a selectable model even though the gateway does not list it. Exposed first. |
+| `combo` | — | An OmniRoute combo id (or `auto`) to expose as a selectable model even though the gateway does not list it. It leads the provider's own model list — an OmniRoute-side detail, unrelated to pool order. |
 | `models` | — | A **hard allowlist**: only these ids are ever exposed, even if the gateway lists more. |
 | `tiers` | — | `{ "primary": [...], "backup": [...] }` preference order over models the allowlist already permits. |
 | `stickyPerSession` | `true` | Keep a session on one model while that model is healthy. |
@@ -95,23 +97,23 @@ failed.
 
 **Using OmniRoute in the pool.** The integration only *registers* the provider —
 it never touches your pool, and it does not place a slot for you. To route pooled
-work (exploration, implementation, cartography) through the gateway, list it in
-your pool like any other model:
+work (exploration, implementation, cartography) through the gateway, add an
+`omniroute/<id>` slot to your pool like any other model:
 
 ```
 # ~/.config/lacode/pool.txt
-omniroute/auto             3
-opencode/space-bunny-free  2
-opencode-go/longcat-2.5-preview-free 1
-opencode/big-pickle        1
+omniroute/free-mix
+opencode/space-bunny-free
+opencode-go/longcat-2.5-preview-free
 ```
 
-`omniroute/auto` then takes slot 1 (weight 3) and drains pooled mechanical work
-before it reaches the bundled opencode/opencode-go ("zen") models. Any
-`omniroute/<id>` works, provided the id is exposed — by `combo`/`models`, or the
-whole catalog when neither is set. Because the plugin writes the pool nothing,
-the only thing that changes when the gateway is down is that `omniroute/<id>`
-fails; every other slot keeps working. See
+OmniRoute is a provider like any other: the pool decides its place and weight,
+and neither the pool nor this integration gives it any built-in precedence or
+priority — where it sits and how many turns it takes are whatever you configure.
+Any `omniroute/<id>` works, provided the id is exposed — by `combo`/`models`, or
+the whole catalog when neither is set. Because the plugin writes the pool
+nothing, the only thing that changes when the gateway is down is that
+`omniroute/<id>` fails; every other slot keeps working. See
 [docs/USER-STORIES.md](USER-STORIES.md) US-28.
 
 

package/docs/USER-STORIES.md

@@ -102,10 +102,10 @@ The reconciliation with US-7 is that **acceptance and endorsement are different
 **Story ID(s) for tests:** US-5
 **Acceptance criteria:**
 - [ ] Given a heavy slot and a light slot, When spawns are routed below both ceilings, Then the rotation walks the list in order with each slot taking as many consecutive turns as its weight says, rather than the heavy slot being filled and refilled first.
-- [ ] Given a slot is at its weight ceiling or cooling, When its turn comes up in the rotation, Then the turn is skipped and the rotation resumes at the following turn.
+- [ ] Given a slot is at its ceiling (twice its weight) or cooling, When its turn comes up in the rotation, Then the turn is skipped and the rotation resumes at the following turn.
 - [ ] Given every slot is at its ceiling, When another spawn is routed, Then it goes to the least-loaded slot as overflow.
 - [ ] Given all slots carry equal weight, When spawns are routed, Then the picks round-robin in declaration order.
-- [ ] Given per-slot load is read, When it is asserted, Then a slot's load never exceeds its weight ceiling except by overflow.
+- [ ] Given per-slot load is read, When it is asserted, Then a slot's load never exceeds its ceiling (twice its weight) except by overflow.
 
 ### US-6: Live pool visibility
 **As a** user, **I want** to see per-slot load and which model is cooling and why, **so that** I understand where my subagents are going.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lawaty/lacode",
-  "version": "0.7.1",
+  "version": "0.8.0",
   "description": "Ultimate opencode setup for Go users — cost-controlled pooled subagents, hang recovery, and notifications for opencode.",
   "license": "MIT",
   "type": "module",

package/presets/README.md

@@ -62,7 +62,7 @@ else means the flat form.
 | Field | Required | Meaning |
 |-------|----------|---------|
 | `model` | yes | `provider/model-id`, split on the **first** slash only |
-| `weight` | no | soft ceiling on concurrent claims, and turns taken per rotation. Absent = 1 |
+| `weight` | no | turns taken per rotation (the slot's share of the rotation); its concurrency ceiling is twice that. Absent = 1 |
 
 `$comment` is a JSON-legal array of strings and is ignored by the loader; the
 bundled preset uses it to explain itself in place.
@@ -91,9 +91,9 @@ because the pool requires it.
 
 ## Weight, precisely
 
-A weight is a count of concurrent sessions — a soft ceiling — and the number
-of turns a slot takes per rotation. Routing is **order-first weighted
-round-robin**:
+A weight is the number of turns a slot takes per rotation — its share of the
+work — and its concurrency ceiling is **twice that weight**: a weight-3 slot may
+hold 6 concurrent sessions. Routing is **order-first weighted round-robin**:
 
 1. the slot list is expanded into a run in which each slot appears once per
    unit of its weight — `[3,2,1,1]` becomes `1,1,1,2,2,3,4` — and the next

package/presets/free-tier.json

@@ -13,8 +13,9 @@
     "AT MOST 4 SLOTS: one hidden variant per base per slot is declared, so a fifth entry would",
     "route at an agent that does not exist. Extra entries are dropped with a warning.",
     "",
-    "WEIGHT IS THE ONLY KNOWN KNOB, and it is a SOFT CEILING on how many concurrent claims a",
-    "slot holds -- not a share of the work and not a priority. Routing is ORDER-FIRST",
+    "WEIGHT IS THE ONLY KNOWN KNOB: one position per unit of weight in the rotation run,",
+    "and a concurrency ceiling of TWICE that weight -- a share of the work, not a priority.",
+    "Routing is ORDER-FIRST",
     "WEIGHTED ROUND-ROBIN: the slot list is walked IN ORDER and each slot takes as many",
     "consecutive turns as its weight says, so weights 3,2,1,1 produce the run 1,1,1,2,2,3,4,",
     "which then repeats. The rule is:",