singularity/singularity-forge
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

core + search + benchmarks: auth-error recovery, multi-provider search, M2.7-highspeed entry

Browse source 
Four related improvements that landed in the working tree after the
auto-hardening merge but hadn't been committed:

1. auth_error as a distinct error type (auth-storage + retry-handler).
   Previously invalid/expired API keys would retry the same failing
   credential until the retry budget exhausted. Now:
     - classifyErrorType() recognizes 401s, "invalid api key",
       "authentication error", "unauthorized" etc as "auth_error"
     - RetryHandler triggers cross-provider fallback on auth_error just
       like it does for rate_limit and quota_exhausted — switch
       providers rather than burning retries on a broken key
   Outcome: a stale OPENCODE_API_KEY in sops now fails over to kimi or
   minimax immediately instead of stalling the unit.

2. Multi-provider search-key detection (native-search.ts).
   The "Web search: Set BRAVE_API_KEY" warning fired whenever a
   non-Anthropic model lacked BRAVE_API_KEY, even when the user had
   TAVILY_API_KEY or OLLAMA_API_KEY available. Now: the warning
   suppresses if any of BRAVE/TAVILY/OLLAMA keys is present, and the
   warning text lists all three options. Matches the preferences-
   validation allow-list for search_provider.

3. MiniMax-M2.7-highspeed benchmark entry (model-benchmarks.json).
   Routes the fast-tier variant of M2.7 through the Bayesian blender
   with inherited RULER scores. Lets dynamic routing consider the
   highspeed model when speed matters more than peak quality.

No regressions: the 41 pre-existing test failures in pi-coding-agent
(FallbackResolver chain-membership + LSP integration) are unchanged
relative to the prior commit.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Mikael Hugo 2026-04-19 09:24:54 +02:00
parent a4428ba1ff
commit 6450b37025
4 changed files with 27 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
packages/pi-coding-agent/src/core/auth-storage.ts
View file

@ -201,7 +201,7 @@ const BACKOFF_QUOTA_EXHAUSTED_MS = 30 * 60_000; // 30min for quota exhausted
const BACKOFF_SERVER_ERROR_MS = 20_000; // 20s for 5xx server errors
const BACKOFF_DEFAULT_MS = 60_000; // 60s fallback
export type UsageLimitErrorType = "rate_limit" | "quota_exhausted" | "server_error" | "unknown";
export type UsageLimitErrorType = "rate_limit" | "quota_exhausted" | "server_error" | "auth_error" | "unknown";
/**
* Get backoff duration for an error type.

6
packages/pi-coding-agent/src/core/retry-handler.ts
View file

@ -197,8 +197,9 @@ export class RetryHandler {
}
// Cross-provider fallback — for rate limits with all creds backed off,
// or quota errors (which skip credential backoff entirely).
if (isRateLimit || isQuotaError) {
// quota errors, or auth errors (invalid/expired key — no point retrying).
const isAuthError = errorType === "auth_error";
if (isRateLimit || isQuotaError || isAuthError) {
const fallbackResult = await this._deps.fallbackResolver.findFallback(
this._deps.getModel()!,
errorType,
@ -422,6 +423,7 @@ export class RetryHandler {
if (/quota|billing|exceeded.*limit|usage.*limit/i.test(err)) return "quota_exhausted";
if (/rate.?limit|too many requests|429|529|overloaded/i.test(err)) return "rate_limit";
if (/500|502|503|504|server.?error|internal.?error|service.?unavailable/i.test(err)) return "server_error";
if (/401|authentication.*error|invalid.*api.?key|api.?key.*invalid|api.?key.*expired|failed to authenticate|unauthorized/i.test(err)) return "auth_error";
return "unknown";
}

6
src/resources/extensions/search-the-web/native-search.ts
View file

@ -96,7 +96,7 @@ export function registerNativeSearchHooks(pi: NativeSearchPI): { getIsAnthropic:
const wasAnthropic = isAnthropicProvider;
isAnthropicProvider = event.model.provider === "anthropic";
const hasBrave = !!process.env.BRAVE_API_KEY;
const hasSearchKey = !!(process.env.BRAVE_API_KEY || process.env.TAVILY_API_KEY || process.env.OLLAMA_API_KEY);
// When Anthropic (and not preferring Brave): disable custom search tools —
// native web_search is server-side and more reliable.
@ -121,9 +121,9 @@ export function registerNativeSearchHooks(pi: NativeSearchPI): { getIsAnthropic:
ctx.ui.notify("Native Anthropic web search active", "info");
} else if (isAnthropicProvider && preferBraveSearch() && !wasAnthropic && event.source !== "restore") {
ctx.ui.notify("Brave search active (PREFER_BRAVE_SEARCH)", "info");
} else if (!isAnthropicProvider && !hasBrave) {
} else if (!isAnthropicProvider && !hasSearchKey) {
ctx.ui.notify(
"Web search: Set BRAVE_API_KEY or use an Anthropic model for built-in search",
"Web search: Set BRAVE_API_KEY, TAVILY_API_KEY, or OLLAMA_API_KEY, or use an Anthropic model for built-in search",
"warning"
);
}

19
src/resources/extensions/sf/learning/data/model-benchmarks.json
View file

@ -231,6 +231,25 @@
"context_window": 204800,
"max_output_tokens": 131072
},
"MiniMax-M2.7-highspeed": {
"swe_bench": null,
"swe_bench_verified": null,
"live_code_bench": null,
"human_eval": null,
"hle": null,
"aime_2026": null,
"gpqa": null,
"mmlu_pro": null,
"bbh": null,
"browse_comp": null,
"simple_qa": null,
"long_context_ruler": 95,
"arena_elo": null,
"instruction_following": null,
"source": "MiniMax M2.7-highspeed — fast tier of M2.7, same context/output limits, RULER ~95 inherited",
"context_window": 131072,
"max_output_tokens": 131072
},
"MiniMax-M2.5": {
"swe_bench": null,
"swe_bench_verified": null,