@lawaty/lacode 1.1.0 → 1.2.0

package/dist/pool.js

@@ -35,9 +35,9 @@ var HANG_COOLDOWN_MS = 10 * 60 * 1e3;
 var UNAVAILABLE_COOLDOWN_MS = 60 * 1e3;
 var UNAVAILABLE_REASON = "endpoint-unavailable";
 var MAX_WEIGHT = 100;
-var SLOT_CEILING_FACTOR = 2;
-var slotCeiling = (slot) => slot.weight * SLOT_CEILING_FACTOR;
-var MAX_SLOTS = 4;
+var slotCeiling = (slot) => slot.weight;
+var slotBucket = (slot) => slot.bucket ?? slot.model;
+var SLOTS_WARN_AT = 64;
 var COOLDOWN_BASE_MS = 60 * 1e3;
 var COOLDOWN_MAX_MS = 15 * 60 * 1e3;
 var QUOTA_COOLDOWN_BASE_MS = 15 * 60 * 1e3;
@@ -47,8 +47,12 @@ var STATED_COOLDOWN_MAX_MS = QUOTA_COOLDOWN_MAX_MS;
 var isQuotaReason = (reason) => /quota|limit|rate|too-many|overloaded|429/i.test(reason);
 var MODEL_MISSING_REASON = "model-missing";
 var MODEL_MISSING_COOLDOWN_MS = QUOTA_COOLDOWN_MAX_MS;
+var AUTH_FAILURE_REASON = "auth-failed";
+var AUTH_FAILURE_COOLDOWN_MS = QUOTA_COOLDOWN_MAX_MS;
 var MODEL_MISSING_PATTERN = /model\s+(?:not\s+found|does\s+not\s+exist|is\s+not\s+available)|(?:unknown|no\s+such)\s+model/i;
 var isModelMissingText = (text) => MODEL_MISSING_PATTERN.test(text);
+var AUTH_FAILURE_PATTERN = /\b401\b|\b403\b|unauthori[sz]ed|not\s+authori[sz]ed|invalid[\s_-]+(?:api[\s_-]?(?:key|token)|token|credential|bearer)|api[\s_-]?key\s+(?:is\s+)?(?:invalid|missing|expired|not\s+valid)|authentication\s+(?:failed|error|required)|incorrect[\s_-]+api[\s_-]?key/i;
+var isAuthFailureText = (text) => AUTH_FAILURE_PATTERN.test(text);
 var BASES = ["explore", "implement", "context"];
 var PRESETS_DIR = join2(PACKAGE_ROOT, "presets");
 var DEFAULT_PRESET = join2(PRESETS_DIR, "free-tier.json");
@@ -71,15 +75,18 @@ function parseSlots(parsed, origin) {
   entries.forEach((entry, i) => {
     const model = typeof entry?.model === "string" ? entry.model.trim() : "";
     const at = `slot ${i + 1}`;
-    if (slots.length >= MAX_SLOTS) {
-      problems.push(`${at} (${model || "?"}): the pool is capped at ${MAX_SLOTS} slots (one variant agent per base per slot is declared); entry ignored`);
-      return;
-    }
     if (!model.includes("/")) return problems.push(`${at}: model must be written provider/model-id, got "${model}"`);
     const weight = readWeight(entry, at, model, problems);
-    if (slots.some((s) => s.model === model)) return problems.push(`${at}: ${model} is already used by another slot; slots must be distinct models`);
-    slots.push({ index: slots.length + 1, model, weight });
+    const rawBucket = entry?.bucket;
+    const bucket = typeof rawBucket === "string" ? rawBucket.trim() : void 0;
+    if (rawBucket !== void 0 && !bucket) {
+      problems.push(`${at} (${model}): bucket must be a non-empty name; using the model id`);
+    }
+    slots.push({ index: slots.length + 1, model, weight, ...bucket ? { bucket } : {} });
   });
+  if (slots.length > SLOTS_WARN_AT) {
+    problems.push(`${origin}: ${slots.length} slots is unusually large, but every entry is kept (the pool has no slot cap)`);
+  }
   if (slots.length === 0) {
     problems.push(`${origin} yielded no usable slot; using built-in defaults`);
     return { slots: FALLBACK_SLOTS, problems };
@@ -91,8 +98,17 @@ function parseTextSlots(text, origin) {
   for (const raw of text.split(/\r?\n/)) {
     const line = raw.split("#")[0].trim();
     if (!line) continue;
-    const [model, weight] = line.split(/\s+/);
-    entries.push(weight === void 0 ? { model } : { model, weight });
+    const [model, ...rest] = line.split(/\s+/);
+    let weight;
+    let bucket;
+    for (const token of rest) {
+      if (token.startsWith("b:")) {
+        if (bucket === void 0) bucket = token.slice(2);
+      } else if (weight === void 0) {
+        weight = token;
+      }
+    }
+    entries.push({ model, ...weight === void 0 ? {} : { weight }, ...bucket === void 0 ? {} : { bucket } });
   }
   return parseSlots({ slots: entries }, origin);
 }
@@ -236,9 +252,17 @@ var retryAfterMs = (headers) => {
   for (const [key, value] of Object.entries(headers)) {
     if (key.toLowerCase() !== "retry-after") continue;
     const seconds = Number(value);
-    if (Number.isFinite(seconds) && seconds > 0) return Math.min(seconds * 1e3, COOLDOWN_MAX_MS);
+    if (Number.isFinite(seconds) && seconds > 0) {
+      const ms = seconds * 1e3;
+      if (ms > STATED_COOLDOWN_MAX_MS) return void 0;
+      return ms;
+    }
     const when = Date.parse(value);
-    if (Number.isFinite(when)) return Math.min(Math.max(when - Date.now(), 1e3), COOLDOWN_MAX_MS);
+    if (Number.isFinite(when)) {
+      const delay = when - Date.now();
+      if (delay > STATED_COOLDOWN_MAX_MS) return void 0;
+      return Math.max(delay, 1e3);
+    }
   }
   return void 0;
 };
@@ -288,7 +312,7 @@ function ensureSlots(s, onWarn) {
   for (const problem of problems) onWarn(problem);
   return slots;
 }
-var slotSignature = (origin, slots) => `${origin}\0${slots.map((slot) => `${slot.index}:${slot.model}:${slot.weight}`).join("|")}`;
+var slotSignature = (origin, slots) => `${origin}\0${slots.map((slot) => `${slot.index}:${slot.model}:${slot.weight}:${slot.bucket ?? ""}`).join("|")}`;
 function reloadSlots(s, onWarn) {
   const { slots, problems, origin } = resolveModels({ models: s.inlineModels, modelsFile: s.modelsFile });
   const changed = slotSignature(origin, slots) !== slotSignature(s.origin, s.slots);
@@ -466,7 +490,7 @@ function readSiblings(s) {
     try {
       const parsed = JSON.parse(readFileSync2(path, "utf8"));
       const claims = Object.values(parsed).filter(
-        (c) => c && typeof c.v === "string" && typeof c.m === "string" && now - (c.hb ?? c.t) <= CLAIM_TTL_MS
+        (c) => c && typeof c.v === "string" && typeof c.m === "string" && typeof c.k === "number" && Number.isFinite(c.k) && now - (c.hb ?? c.t) <= CLAIM_TTL_MS
       );
       siblings.push({ id: match[1], claims });
     } catch {
@@ -516,20 +540,32 @@ function recordStrike(s, model, reason, stated) {
   if (!s.dirReady) return;
   if (reason === UNAVAILABLE_REASON) return coolUnavailable(s, model);
   const now = Date.now();
-  const prior = s.strikes.get(model);
+  const slotsForModel = s.slots.filter((candidate) => candidate.model === model);
+  const buckets = [...new Set(slotsForModel.map((slot) => slotBucket(slot)))];
+  const targets = buckets.length === 0 ? [model] : [...new Set(
+    s.slots.filter((candidate) => buckets.includes(slotBucket(candidate))).map((candidate) => candidate.model)
+  )];
+  if (targets.length === 0) targets.push(model);
+  const prior = targets.map((target) => s.strikes.get(target)).filter((candidate) => Boolean(candidate)).reduce((worst, candidate) => !worst || candidate.strikes > worst.strikes ? candidate : worst, void 0);
   const strikes = prior && now - prior.last < STRIKE_DECAY_MS ? prior.strikes + 1 : 1;
   const quota = isQuotaReason(reason);
   const base = quota ? QUOTA_COOLDOWN_BASE_MS : COOLDOWN_BASE_MS;
   const cap = quota ? QUOTA_COOLDOWN_MAX_MS : COOLDOWN_MAX_MS;
   const backoff = Math.min(cap, base * 2 ** (strikes - 1));
-  const strike = { model, until: now + (stated ?? backoff), strikes, reason, last: now };
-  s.cooling.set(model, strike);
-  s.strikes.set(model, strike);
-  const path = join2(s.dir, limitFile(model));
-  try {
-    writeFileSync(`${path}.tmp`, JSON.stringify(strike));
-    renameSync(`${path}.tmp`, path);
-  } catch {
+  const statedOk = typeof stated === "number" && Number.isFinite(stated) && stated > 0 && stated <= STATED_COOLDOWN_MAX_MS;
+  const until = now + (statedOk ? stated : backoff);
+  let strike;
+  for (const target of targets) {
+    const record = { model: target, until, strikes, reason, last: now };
+    s.cooling.set(target, record);
+    s.strikes.set(target, record);
+    const path = join2(s.dir, limitFile(target));
+    try {
+      writeFileSync(`${path}.tmp`, JSON.stringify(record));
+      renameSync(`${path}.tmp`, path);
+    } catch {
+    }
+    if (target === model) strike = record;
   }
   return strike;
 }
@@ -619,30 +655,26 @@ function snapshot(s) {
   const siblings = readSiblings(s);
   const cooling = readLimits(s);
   const counts = /* @__PURE__ */ new Map();
-  const bump = (m) => counts.set(m, (counts.get(m) ?? 0) + 1);
-  for (const claim of s.own.values()) bump(claim.m);
-  for (const sibling of siblings) for (const claim of sibling.claims) bump(claim.m);
+  const bump = (k) => counts.set(k, (counts.get(k) ?? 0) + 1);
+  for (const claim of s.own.values()) bump(claim.k);
+  for (const sibling of siblings) for (const claim of sibling.claims) bump(claim.k);
   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) < slotCeiling(slot);
+var loadOf = (counts, slot) => counts.get(slot) ?? 0;
+var underCeiling = (slot, counts) => loadOf(counts, slot.index) < slotCeiling(slot);
 function pickSlot(s, snap) {
   const slots = s.slots.length > 0 ? s.slots : FALLBACK_SLOTS;
   const usable = slots.filter((slot) => isAvailable(s, slot.model, snap.cooling));
   const pool = usable.length > 0 ? usable : slots;
   for (const slot of usable) {
-    if (loadOf(snap.counts, slot.model) < slot.weight) return slot;
-  }
-  for (const slot of usable) {
-    if (loadOf(snap.counts, slot.model) < slotCeiling(slot)) return slot;
+    if (loadOf(snap.counts, slot.index) < slot.weight) return slot;
   }
-  return pool.reduce((best, slot) => loadOf(snap.counts, slot.model) < loadOf(snap.counts, best.model) ? slot : best, pool[0]);
+  return pool.reduce((best, slot) => loadOf(snap.counts, slot.index) < loadOf(snap.counts, best.index) ? slot : best, pool[0]);
 }
 function poolPass(s, snap) {
   const slots = s.slots.length > 0 ? s.slots : FALLBACK_SLOTS;
   const usable = slots.filter((slot) => isAvailable(s, slot.model, snap.cooling));
-  if (usable.some((slot) => loadOf(snap.counts, slot.model) < slot.weight)) return 1;
-  if (usable.some((slot) => loadOf(snap.counts, slot.model) < slotCeiling(slot))) return 2;
+  if (usable.some((slot) => loadOf(snap.counts, slot.index) < slot.weight)) return 1;
   return "overflow";
 }
 function agentFor(base, slot) {
@@ -749,6 +781,8 @@ function coolProvider(s, provider, until, reason) {
   return affected;
 }
 export {
+  AUTH_FAILURE_COOLDOWN_MS,
+  AUTH_FAILURE_REASON,
   BASES,
   CATALOG_TTL_MS,
   CLAIM_TTL_MS,
@@ -757,7 +791,6 @@ export {
   FALLBACK_SLOTS,
   HANG_COOLDOWN_MS,
   MAX_CLAIM_AGE_MS,
-  MAX_SLOTS,
   MAX_WEIGHT,
   MODEL_MISSING_COOLDOWN_MS,
   MODEL_MISSING_REASON,
@@ -766,7 +799,7 @@ export {
   PRESETS_DIR,
   QUOTA_COOLDOWN_BASE_MS,
   QUOTA_COOLDOWN_MAX_MS,
-  SLOT_CEILING_FACTOR,
+  SLOTS_WARN_AT,
   STATED_COOLDOWN_MAX_MS,
   STATE_SWEEP_INTERVAL_MS,
   STUCK_IDLE_MS,
@@ -790,6 +823,7 @@ export {
   findStuck,
   idleFor,
   init,
+  isAuthFailureText,
   isAvailable,
   isLimitText,
   isModelMissingText,
@@ -821,6 +855,7 @@ export {
   resolveModels,
   retryAfterMs,
   selfFile,
+  slotBucket,
   slotCeiling,
   slotFor,
   slotSources,