Files
OmniRoute/open-sse/services/modelFamilyFallback.ts
Diego Rodrigues de Sa e Souza bf8b56b29f Release v3.8.20 (#3547)
* chore(release): open v3.8.20 development cycle

* fix(images): prefer bare combos over image aliases (#3527)

Integrated into release/v3.8.20

* fix(translator): map Codex local_shell tool (#3534)

Integrated into release/v3.8.20

* fix(usage): make opencode-go quota fetcher fail-open instead of throwing 500 (#3522)

Integrated into release/v3.8.20

* Fix Runtime page breaker state rendering (#3533)

Integrated into release/v3.8.20

* Expose provider breaker degradation threshold setting (#3535)

Integrated into release/v3.8.20

* fix(executor): strip provider prefix from versioned built-in tool model field (#3532)

Integrated into release/v3.8.20

* feat(providers): add Claude Fable 5 support (#3524)

Integrated into release/v3.8.20

* feat(resilience): add global provider cooldown tracking to prevent combo re-walking (#3556)

Integrated into release/v3.8.20 (default OFF, opt-in)

* fix(translator): scope thoughtSignature bypass to Antigravity/CLI only (#3560)

Integrated into release/v3.8.20. Co-authored-by: Six7Day <six7day@gmail.com>

* fix(routing): normalize thinking:disabled for combo-substituted models that reject it (#3554) (#3563)

Integrated into release/v3.8.20

* fix(usage): accept 0/empty budget limits so the dashboard can save and clear (#3537) (#3564)

Integrated into release/v3.8.20

* docs(changelog): credit @Six7Day for #3560 thoughtSignature fix (#3414)

The #3560 squash co-author trailer landed inline (unparsed by GitHub), so add
an explicit CHANGELOG credit ensuring @Six7Day (original #3414) and @oyi77 are
on the public record for the Gemini thoughtSignature fix.

* fix(gamification): dedup badge unlock via user_badges so events don't re-fire every request (#3472) (#3565)

Integrated into release/v3.8.20

* fix(routing): pass through 'auto' keyword on codex /v1/responses instead of rewriting to codex/auto (#3509) (#3566)

Integrated into release/v3.8.20

* fix(cli-tools): normalize apiKey null in guide-settings schema so cloud-mode config saves (#3552) (#3567)

Integrated into release/v3.8.20

* fix(catalog): reclassify PublicAI from keyless to one-time-initial (requires API key) (#3558) (#3568)

Integrated into release/v3.8.20

* fix(gemini-web): surface missing Playwright browser as actionable 503 + cooldown hint, not a retryable 500 loop (#3516) (#3570)

Integrated into release/v3.8.20

* fix(security): sanitize raw err.message in web executors + embeddings/search response bodies (Rule #12) (#3494, #3495) (#3573)

Integrated into release/v3.8.20

* fix(dashboard): point CustomHostsManager + FeatureFlagsGrid at real routes (#3486, #3487) (#3574)

Integrated into release/v3.8.20

* chore(providers): remove dead krutrim entry (#3483) + docs(api): fix agent-bridge per-agent state route (#3489) (#3575)

Integrated into release/v3.8.20

* docs(api): correct API_REFERENCE.md paths for skills/plugins/admin/cache/acp/system-info (#3497) (#3577)

Integrated into release/v3.8.20

* fix(proxy): drive SOCKS5 UI option from runtime ENABLE_SOCKS5_PROXY, not build-time NEXT_PUBLIC (#3508) (#3579)

Integrated into release/v3.8.20

* fix(playground): filter playground models by node prefix so custom-endpoint models appear (#3505) (#3581)

Integrated into release/v3.8.20

* fix(usage): show an informative message instead of a blank Kiro quota card when no usage breakdown (#3506) (#3582)

Integrated into release/v3.8.20

* docs(changelog): add the #3506 Kiro quota entry (missed in #3582 due to a stale-base CHANGELOG anchor) (#3583)

Integrated into release/v3.8.20

* fix(auto-update): use stable PROJECT_ROOT walker, not frozen process.cwd() (#3561)

Integrated into release/v3.8.20. Auto-update PROJECT_ROOT now uses a stable __dirname-anchored upward walker instead of the no-op process.cwd() resolver.

* fix: address PR #3518 review comments (lifecycle hooks, regex, indentation, route params) (#3562)

Integrated into release/v3.8.20. Addresses #3518 review: regex literals, logs/[id] route params (Next 16), indentation, and wires plugin lifecycle hooks (onInstall/onActivate/onDeactivate/onUninstall) in the loader so manager.ts can register them. Adds Rule #18 regression test.

* docs(changelog): credit @ViFigueiredo (#3423) for PROJECT_ROOT + log #3561/#3562 (v3.8.20)

* fix: openai to gemini incorrectly translates historical tool calls into text (#3569)

Integrated into release/v3.8.20. Standard Gemini direct path now maps historical tool calls to native functionCall/functionResponse parts (signaturelessToolCallMode: native) instead of inert text — validated against the real Gemini API (gemini-2.5-flash returns 200 for signatureless native functionCall, even with tools+thinking; Hard Rule #18). Eliminates the text-serialization leak. Antigravity/CLI sentinel path (#3560) untouched.

* docs(changelog)+test: reconcile standard-Gemini native mode (#3569) — update round-2 rationale comment + log VPS validation

* docs(changelog): reconcile v3.8.20 — add 9 missing bullets + move [Unreleased] to versioned section

* docs(changelog): complete v3.8.20 reconciliation — 27 bullets, 11 contributors

---------

Co-authored-by: Alexander Averyanov <alex@averyan.ru>
Co-authored-by: Hakan Kurşun <bykamaka@gmail.com>
Co-authored-by: Wilson <pedbookmed@gmail.com>
Co-authored-by: Randi <55005611+rdself@users.noreply.github.com>
Co-authored-by: Giorgos Giakoumettis <giorgos@yiakoumettis.gr>
Co-authored-by: PizzaV <103120356+pizzav-xyz@users.noreply.github.com>
Co-authored-by: Paijo <14921983+oyi77@users.noreply.github.com>
Co-authored-by: Markus Hartung <mail@hartmark.se>
2026-06-10 13:49:08 -03:00

234 lines
8.1 KiB
TypeScript

/**
* Model Family Fallback — Phase 2 Feature (T5)
*
* Implements two-phase model resolution:
* Phase 1 (static, pre-request): already done by model.ts alias resolution.
* Phase 2 (dynamic, post-error): when a provider returns a model-not-available
* error (400 with specific message or 404), we try sibling models within the
* same "family" before giving up.
*
* Inspired by Antigravity Manager's account-aware dynamic model remapping
* (commit 6cea566, Mar 8 2026).
*/
import { getModelContextLimit } from "../../src/lib/modelCapabilities";
import { parseModel } from "./model.ts";
import { CONTEXT_OVERFLOW_REGEX } from "./errorClassifier.ts";
import { getRegistryEntry } from "../config/providerRegistry.ts";
// ── Model Family Definitions ─────────────────────────────────────────────────
/**
* Ordered candidate lists per model family.
* First entry is the most preferred; fallback proceeds in order.
*/
const MODEL_FAMILIES: Record<string, string[]> = {
// Gemini 3 / 3.1 Pro family — ordered by preference
"gemini-3-pro": [
"gemini-3.1-pro-preview",
"gemini-3-pro-preview",
"gemini-3.1-pro-high",
"gemini-3-pro-high",
"gemini-3.1-pro-low",
"gemini-3-pro-low",
],
"gemini-3.1-pro": [
"gemini-3.1-pro-preview",
"gemini-3-pro-preview",
"gemini-3.1-pro-high",
"gemini-3-pro-high",
"gemini-3.1-pro-low",
"gemini-3-pro-low",
],
"gemini-3-pro-preview": [
"gemini-3.1-pro-preview",
"gemini-3-pro-high",
"gemini-3.1-pro-high",
"gemini-3-pro-low",
"gemini-3.1-pro-low",
],
"gemini-3.1-pro-preview": [
"gemini-3-pro-preview",
"gemini-3.1-pro-high",
"gemini-3-pro-high",
"gemini-3.1-pro-low",
"gemini-3-pro-low",
],
"gemini-3-pro-high": [
"gemini-3.1-pro-high",
"gemini-3-pro-preview",
"gemini-3.1-pro-preview",
"gemini-3-pro-low",
"gemini-3.1-pro-low",
],
"gemini-3.1-pro-high": [
"gemini-3-pro-high",
"gemini-3.1-pro-preview",
"gemini-3-pro-preview",
"gemini-3.1-pro-low",
"gemini-3-pro-low",
],
// Gemini 2.5 Pro family
"gemini-2.5-pro": ["gemini-2.5-pro-preview-06-05", "gemini-2.5-pro-exp-03-25"],
"gemini-2.5-pro-preview-06-05": ["gemini-2.5-pro", "gemini-2.5-pro-exp-03-25"],
// Claude Mythos family (Fable 5) — flagship falls to the next-best Opus
// tiers before the cheaper Sonnet, matching the Opus family ordering.
"claude-fable-5": ["claude-opus-4-8", "claude-opus-4-7", "claude-sonnet-4-6"],
// Claude Opus family
"claude-opus-4-8": ["claude-opus-4-7", "claude-opus-4-6", "claude-sonnet-4-6"],
"claude-opus-4-7": ["claude-opus-4-6", "claude-opus-4-5-20251101", "claude-sonnet-4-6"],
"claude-opus-4-6": ["claude-opus-4-6-thinking", "claude-opus-4-5-20251101", "claude-sonnet-4-6"],
"claude-opus-4-6-thinking": ["claude-opus-4-6", "claude-opus-4-5-20251101"],
// Claude Sonnet family
"claude-sonnet-4-6": ["claude-sonnet-4-5-20250929", "claude-sonnet-4-20250514"],
"claude-sonnet-4-5-20250929": ["claude-sonnet-4-6", "claude-sonnet-4-20250514"],
// GPT-5 family
"gpt-5": ["gpt-5-mini", "gpt-4o"],
"gpt-5.1": ["gpt-5.1-mini", "gpt-5", "gpt-4o"],
};
// ── Error Detection ──────────────────────────────────────────────────────────
/**
* Error message fragments that indicate the requested model is unavailable
* for the current account/provider, as opposed to a transient error.
*/
const MODEL_UNAVAILABLE_FRAGMENTS = [
"model not found",
"model_not_found",
"model not available",
"model is not available",
"no such model",
"unsupported model",
"unknown model",
"this model does not exist",
"invalid model",
"model not supported",
"does not support",
"not enabled for",
"access to model",
"improperly formed request", // Kiro 400 (model unavailable)
];
/**
* Returns true if the HTTP status + error message indicates the model
* itself is not available, not a transient server error.
*/
export function isModelUnavailableError(status: number, errorMessage: string): boolean {
if (status === 404) return true;
if (status !== 400 && status !== 403) return false;
const msg = errorMessage.toLowerCase();
return MODEL_UNAVAILABLE_FRAGMENTS.some((fragment) => msg.includes(fragment));
}
export function isContextOverflowError(status: number, errorMessage: string): boolean {
if (status !== 400) return false;
return CONTEXT_OVERFLOW_REGEX.test(errorMessage);
}
// ── Fallback Resolution ──────────────────────────────────────────────────────
/**
* Get the next fallback model from the same family.
*
* @param currentModel The model that just failed
* @param triedModels Set of model IDs already tried (to avoid cycles)
* @returns Next model to try, or null if family exhausted
*/
export function getNextFamilyFallback(
currentModel: string,
triedModels: Set<string>
): string | null {
const parsed = parseModel(currentModel);
const bareModel = parsed.model || currentModel;
const provider = parsed.provider || parsed.providerAlias || "";
const prefix = provider ? `${provider}/` : "";
// Normalize dots to hyphens for the lookup so kiro/claude-opus-4.8 finds the right entry
const lookupKey = bareModel.replace(/\./g, "-");
const family = MODEL_FAMILIES[lookupKey];
if (!family) return null;
// Resolve the provider's supported model IDs so we can match notation (dot vs hyphen)
const registryEntry = provider ? getRegistryEntry(provider) : null;
const supportedIds = registryEntry ? new Set(registryEntry.models.map((m) => m.id)) : null;
for (const candidate of family) {
let resolvedCandidate = candidate;
if (supportedIds && !supportedIds.has(candidate)) {
// Try dot-notation variants: claude-opus-4-8 → claude-opus-4.8
const dotVariant = candidate.replace(/-(\d+)-(\d+)$/, "-$1.$2");
const dotVariant2 = candidate.replace(/-(\d+)-(\d+)-/, "-$1.$2-");
if (supportedIds.has(dotVariant)) resolvedCandidate = dotVariant;
else if (supportedIds.has(dotVariant2)) resolvedCandidate = dotVariant2;
}
const fullCandidate = `${prefix}${resolvedCandidate}`;
if (!triedModels.has(fullCandidate)) {
return fullCandidate;
}
}
return null; // family exhausted
}
/**
* Check if a model belongs to any registered family.
*/
export function isInModelFamily(model: string): boolean {
const parsed = parseModel(model);
const bareModel = parsed.model || model;
return bareModel in MODEL_FAMILIES;
}
/**
* Get all members of a model's family (including itself).
*/
export function getModelFamily(model: string): string[] {
const parsed = parseModel(model);
const bareModel = parsed.model || model;
const prefix =
parsed.provider || parsed.providerAlias ? `${parsed.provider || parsed.providerAlias}/` : "";
const family = MODEL_FAMILIES[bareModel];
if (!family) return [model];
return [model, ...family.map((c) => `${prefix}${c}`)];
}
/**
* Find a model with larger context window from a list of candidate models.
* Uses models.dev synced capabilities to compare context limits.
*/
export function findLargerContextModel(
currentModel: string,
availableModels: string[]
): string | null {
const currentParsed = parseModel(currentModel);
const currentProvider = currentParsed.provider || currentParsed.providerAlias || "unknown";
const currentModelId = currentParsed.model || currentModel;
const currentLimit = getModelContextLimit(currentProvider, currentModelId) ?? 0;
let bestModel: string | null = null;
let bestLimit = currentLimit;
for (const candidate of availableModels) {
if (candidate === currentModel) continue;
const parsed = parseModel(candidate);
const provider = parsed.provider || parsed.providerAlias || "unknown";
const modelId = parsed.model || candidate;
const limit = getModelContextLimit(provider, modelId) ?? 0;
if (limit > bestLimit) {
bestLimit = limit;
bestModel = candidate;
}
}
return bestModel;
}