Merge remote-tracking branch 'upstream/master'

# Conflicts:
#	.gitignore
#	open-sse/handlers/chatCore.js
This commit is contained in:
decolua
2026-07-16 11:59:46 +07:00
162 changed files with 9368 additions and 1287 deletions
@@ -1,3 +1,5 @@
import { getCapabilitiesForModel } from "../../providers/capabilities.js";
// Strip request params a given provider/model rejects upstream (e.g. HTTP 400).
// Config-driven: add a rule instead of scattering `delete body.x` across executors.
@@ -12,6 +14,13 @@ const STRIP_RULES = [
{ provider: "github", match: (m) => /claude/i.test(m) && !/claude.*(opus|sonnet).*4\.6/i.test(m), drop: ["thinking", "reasoning_effort"] },
// Cloudflare Workers AI: content must be plain string, rejects OpenAI content-part array (#1926)
{ provider: "cloudflare-ai", flattenContent: true },
{ provider: "volcengine-ark", match: /glm-5/i, clampToModelMaxOutput: true },
// VolcEngine Ark caps the Kimi family at max_tokens <= 32768, but the model's
// advertised ceiling is far higher (Kimi-K2.7-Code resolves to maxOutput 262144),
// so clampToModelMaxOutput alone leaves it uncapped and the request 400s with
// "integer above maximum value, expected <= 32768". Pin an explicit endpoint cap;
// min() with the model ceiling still applies if a variant's own limit is lower.
{ provider: "volcengine-ark", match: /kimi/i, maxOutputCap: 32768, clampToModelMaxOutput: true },
];
// Test a rule's match (regex or predicate) against the model id.
@@ -20,6 +29,12 @@ function matches(rule, model) {
return typeof rule.match === "function" ? rule.match(model) : rule.match.test(model);
}
function clampNumber(body, key, ceiling) {
if (typeof body[key] === "number" && Number.isFinite(body[key]) && body[key] > ceiling) {
body[key] = ceiling;
}
}
// Remove unsupported params from body in place; returns body.
export function stripUnsupportedParams(provider, model, body) {
if (!model || !body || typeof body !== "object") return body;
@@ -39,6 +54,22 @@ export function stripUnsupportedParams(provider, model, body) {
}
}
}
if (rule.clampToModelMaxOutput || Number.isFinite(rule.maxOutputCap)) {
const modelCeiling = getCapabilitiesForModel(provider, model).maxOutput;
const candidates = [];
if (rule.clampToModelMaxOutput && Number.isFinite(modelCeiling) && modelCeiling > 0) {
candidates.push(modelCeiling);
}
if (Number.isFinite(rule.maxOutputCap) && rule.maxOutputCap > 0) {
candidates.push(rule.maxOutputCap);
}
if (candidates.length > 0) {
const ceiling = Math.min(...candidates);
clampNumber(body, "max_tokens", ceiling);
clampNumber(body, "max_completion_tokens", ceiling);
clampNumber(body, "max_output_tokens", ceiling);
}
}
}
return body;
}