Files
OmniRoute/src/lib/images/imageRouteModel.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

166 lines
6.5 KiB
TypeScript

/**
* Shared model resolution for the image routes (#3214 / #3215).
*
* `/v1/images/generations` and `/v1/images/edits` must resolve a requested model the
* same way, and as close as practical to how chat routing resolves models:
*
* 1. Bare combo / alias name with no slash (`image`) — resolved to the combo's single
* image target, then that target is itself prefix-resolved. Bare combos intentionally
* override built-in image aliases with the same name.
* 2. Built-in image model id / alias (`cgpt-web/...`, `gpt-image-1`, …) — untouched.
* 3. Custom provider *prefix* form (`myImg/gpt-image-2`) — rewritten to the internal
* `<nodeId>/<model>` id (#3205 did this inline in the generations route only).
*
* Anything that does not match falls through unchanged, so existing built-in and
* already-internal ids keep working.
*/
import { parseImageModel } from "@omniroute/open-sse/config/imageRegistry.ts";
import { resolveComboTargets } from "@omniroute/open-sse/services/combo.ts";
import { getComboByName, getCombos } from "@/lib/db/combos";
import { getProviderNodes } from "@/lib/db/providers";
/**
* Rewrite a `prefix/model` custom image model to its internal `<nodeId>/<model>` form.
* Returns the original string when no openai-compatible node prefix matches (so built-in
* and already-internal ids pass through). Mirrors `src/sse/services/model.ts` (match on
* `node.prefix` OR `node.id`).
*/
export async function resolveImageModelPrefix(modelStr: string): Promise<string> {
if (typeof modelStr !== "string") return modelStr;
const slash = modelStr.indexOf("/");
if (slash <= 0) return modelStr;
const prefixPart = modelStr.slice(0, slash);
const rest = modelStr.slice(slash + 1);
if (!rest) return modelStr;
try {
const nodes = await getProviderNodes({ type: "openai-compatible" });
// node.id (internal UUID) is already a valid internal id; only rewrite when a
// user-defined prefix differs from the node id.
const matched = nodes.find((node: { prefix?: unknown }) => node.prefix === prefixPart);
if (matched && typeof matched.id === "string" && matched.id && matched.id !== prefixPart) {
return `${matched.id}/${rest}`;
}
} catch {
// DB unavailable (pre-migration / tests) — leave the model untouched.
}
return modelStr;
}
/**
* Resolve a bare combo/alias name (e.g. `image`) to its first image model target's
* model string, or null when the name is not a combo / has no usable target.
*/
export async function resolveSingleImageComboTarget(name: string): Promise<string | null> {
if (typeof name !== "string" || !name.trim()) return null;
try {
const combo = await getComboByName(name);
if (!combo) return null;
const allCombos = await getCombos();
const targets = resolveComboTargets(combo as never, allCombos as never);
const first = targets.find(
(t: { modelStr?: unknown }) =>
typeof t?.modelStr === "string" && (t.modelStr as string).trim()
);
return (first?.modelStr as string) ?? null;
} catch {
return null;
}
}
/**
* Full image-route model resolver. See module header for the resolution order.
*/
export async function resolveImageRouteModel(modelStr: string): Promise<string> {
if (typeof modelStr !== "string" || !modelStr.trim()) return modelStr;
// 1. Bare combo/alias name (no slash): resolve to its single image target, then
// prefix-resolve that target (it may itself be a `prefix/model` custom id).
// This intentionally precedes built-in aliases so user combos can shadow names
// like `gpt-image-2`; explicit `provider/model` ids still bypass this branch.
if (!modelStr.includes("/")) {
const target = await resolveSingleImageComboTarget(modelStr);
if (target && target !== modelStr) return resolveImageModelPrefix(target);
}
// 2. Built-in image model (alias or provider/model) — leave untouched.
if (parseImageModel(modelStr).provider) return modelStr;
if (!modelStr.includes("/")) return modelStr;
// 3. Custom provider prefix form — rewrite to internal `<nodeId>/<model>`.
return resolveImageModelPrefix(modelStr);
}
interface ParsedImageEditInput {
prompt: string;
model: string | null;
size: string | null;
responseFormat: string | null;
imageBytes: Buffer | null;
imageMime: string | null;
}
/** Parse a `data:<mime>;base64,<data>` URL into raw bytes + mime, or null when invalid. */
export function parseDataUrl(value: unknown): { bytes: Buffer; mime: string } | null {
if (typeof value !== "string") return null;
const match = /^data:([^;,]+)?(;base64)?,(.*)$/s.exec(value.trim());
if (!match) return null;
const mime = match[1] || "image/png";
const isBase64 = Boolean(match[2]);
const payload = match[3] ?? "";
try {
const bytes = isBase64
? Buffer.from(payload, "base64")
: Buffer.from(decodeURIComponent(payload), "utf8");
if (bytes.length === 0) return null;
return { bytes, mime };
} catch {
return null;
}
}
/**
* Extract an OpenAI-compatible image-edit payload from a JSON body. Some clients send
* edit input as JSON with data-URL images instead of multipart/form-data; accept the
* common shapes (`image: "data:..."`, `images: [{ image_url: "data:..." }]` or
* `images: ["data:..."]`) and surface the same fields the multipart reader produces.
*/
export function extractImageEditInputFromJson(body: unknown): ParsedImageEditInput {
const obj = (body && typeof body === "object" ? body : {}) as Record<string, unknown>;
const str = (v: unknown): string | null => (typeof v === "string" && v.trim() ? v.trim() : null);
const prompt = typeof obj.prompt === "string" ? obj.prompt.trim() : "";
const model = str(obj.model);
const size = str(obj.size);
const responseFormat = str(obj.response_format);
const candidates: unknown[] = [];
if (obj.image !== undefined) candidates.push(obj.image);
const images = obj.images;
if (Array.isArray(images)) {
for (const entry of images) {
if (typeof entry === "string") candidates.push(entry);
else if (entry && typeof entry === "object") {
const e = entry as Record<string, unknown>;
candidates.push(e.image_url ?? e.url ?? e.b64_json);
}
}
}
let imageBytes: Buffer | null = null;
let imageMime: string | null = null;
for (const candidate of candidates) {
const parsed = parseDataUrl(candidate);
if (parsed) {
imageBytes = parsed.bytes;
imageMime = parsed.mime;
break;
}
}
return { prompt, model, size, responseFormat, imageBytes, imageMime };
}