@bitkyc08/opencodex
Version:
Universal provider proxy for OpenAI Codex & Claude Code — use any LLM with Codex CLI/App/SDK and Claude Code
290 lines (276 loc) • 15 kB
text/typescript
import { modelRecordValue } from "../../reasoning-effort";
import { modelInList } from "../../types";
import type { OcxConfig, OcxProviderConfig } from "../../types";
import { PROVIDER_REGISTRY } from "../../providers/registry";
import { fastPolicyForModel, serviceTierSupportFromPolicy } from "../../providers/service-tier";
import { resolveProviderAuthTransport } from "../../providers/fastwire";
import { localFingerprint } from "../../lab/digest";
import type { LabBehaviorSource, LabBehaviorValues } from "../../lab/live/types";
export function upstreamProtocolForAdapter(adapter: string): string {
switch (adapter) {
case "openai-responses":
return "openai-responses";
case "openai-chat":
case "command-code":
case "cursor":
case "devin":
case "azure":
case "azure-openai":
case "kiro":
case "mimo-free":
return "openai-chat";
case "anthropic":
return "anthropic-messages";
case "google":
return "google-generate";
default:
return "openai-responses";
}
}
export function surfaceForProtocols(inboundProtocol: string, upstreamProtocol: string): string {
if (inboundProtocol === "anthropic-messages") return "anthropic-messages-http";
if (upstreamProtocol === "anthropic-messages") return "responses-http";
if (upstreamProtocol === "openai-chat") return "responses-sse";
return "responses-http";
}
function behaviorRow(source: LabBehaviorSource, value: unknown) {
return { source, value };
}
/**
* Membership for the provider's `no*Models`-style lists.
*
* Delegates to modelInList so the report matches the wire: every runtime gate these
* rows describe (openai-chat's sampling/reasoning/tool-choice gates, reasoning-effort's
* noReasoningModels) matches through modelInList, which also accepts a bare entry for a
* tagged id. ollama-cloud serves `gpt-oss:120b` and lists the bare `gpt-oss`, so an
* exact-only check here reported "temperature is sent" on a request that omits it.
*/
function includesModel(list: string[] | undefined, modelId: string): boolean {
return modelInList(list, modelId);
}
/**
* Per-model override lookup for the nine family-aware report rows.
*
* Delegates to modelRecordValue so the report reads these maps the way the
* runtime does -- own properties only, then the pre-colon family, then a
* case-folded key. A bare index disagreed on all three: it missed the
* `gpt-oss` entry ollama-cloud's `gpt-oss:120b` actually resolves, missed a
* differently-cased key, and walked the prototype chain, so a routed model id
* of `constructor`/`toString` yielded an Object.prototype function. That last
* one made buildBehaviorFingerprintV1 throw ("unsupported value type
* function"); the caller catches it (`src/routing/compatibility/subject.ts:125`)
* and returns no route, so the subject is silently dropped -- and the linker
* contract says implementations do not throw.
*
* Not every override map belongs here. `modelPreferHostedTools` and
* `modelOpenRouterRouting` are exact-own at runtime and go through
* `exactOwnValue` below; widening those to the family would be this same bug
* with the sign flipped.
*/
function modelValue<T>(map: Record<string, T> | undefined, modelId: string): T | undefined {
return modelRecordValue(map, modelId);
}
/**
* Exact, own-property lookup for the two maps the runtime resolves that way.
*
* `modelPreferHostedTools` and `modelOpenRouterRouting` are deliberately exact: the
* adapter reads the first through `hasOwnProperty`
* (`src/adapters/openai-responses.ts:1001`) and the second through `Object.hasOwn`
* (`src/providers/openrouter-routing.ts:89`), and the type documents the first as
* "Exact-model hosted tools" (`src/types.ts:1584`). Sending them through
* `modelRecordValue` would make the report say a `gpt-oss` entry applies to
* `gpt-oss:120b` when the adapter will never apply it -- the same divergence this
* file exists to remove, pointed the other way. A bare index is not the answer
* either: it walks the prototype chain, which is the bug `modelValue` just fixed.
* Neither existing primitive is right for these two, so this is the third one.
*/
function exactOwnValue<T>(map: Record<string, T> | undefined, modelId: string): T | undefined {
return map !== undefined && Object.hasOwn(map, modelId) ? map[modelId] : undefined;
}
const CREDENTIAL_HEADER = /(authorization|api[-_]?key|token|secret|credential|cookie)/i;
function nonCredentialHeaderDigest(
headers: Record<string, string> | undefined,
installationSalt: Uint8Array | string,
): string {
const rows = Object.entries(headers ?? {})
.filter(([name]) => !CREDENTIAL_HEADER.test(name))
.map(([name, value]) => [name.toLowerCase().trim(), value] as const)
.sort(([a], [b]) => a < b ? -1 : a > b ? 1 : 0);
return localFingerprint("nonCredentialHeaders", rows, installationSalt);
}
function effectiveOpenRouterRouting(effective: OcxProviderConfig, modelId: string) {
return exactOwnValue(effective.modelOpenRouterRouting, modelId) ?? effective.openRouterRouting;
}
/**
* Production behavior resolver for route-subject fingerprinting.
*
* The caller supplies the already-resolved effective provider config. This is
* deliberately the same effective config used for routeContext identity so a
* model wire override cannot disagree with the behavior fingerprint.
*/
export function resolveProductionBehaviorValues(
config: OcxConfig,
providerName: string,
modelId: string,
effective: OcxProviderConfig,
installationSalt: Uint8Array | string,
): LabBehaviorValues | null {
const provider = config.providers[providerName];
if (!provider || provider.disabled === true) return null;
const adapter = effective.adapter ?? "openai-responses";
const upstreamProtocol = upstreamProtocolForAdapter(adapter);
const registryEntry = PROVIDER_REGISTRY.find(entry => entry.id === providerName);
const authMode = effective.authMode ?? registryEntry?.authKind ?? "key";
const reasoningEfforts = modelValue(effective.modelReasoningEfforts, modelId)
?? effective.reasoningEfforts
?? [];
const defaultReasoningEffort = modelValue(effective.modelDefaultReasoningEfforts, modelId) ?? null;
const reasoningEffortMap = modelValue(effective.modelReasoningEffortMap, modelId)
?? effective.reasoningEffortMap
?? {};
const openRouterRouting = effectiveOpenRouterRouting(effective, modelId);
const project = typeof effective.project === "string" && effective.project ? effective.project : null;
const location = typeof effective.location === "string" && effective.location ? effective.location : null;
const nativeLocalExec = effective.nativeLocalExec === "on" || effective.unsafeAllowNativeLocalExec === true;
const fastPolicy = fastPolicyForModel(effective, modelId, providerName, "responses", provider);
const values: LabBehaviorValues = {
"wire.adapter": behaviorRow("provider_config", adapter),
"wire.upstreamProtocol": behaviorRow("provider_config", upstreamProtocol),
"wire.responsesPath": behaviorRow("provider_config", effective.responsesPath ?? null),
"wire.chatCompletionsPath": behaviorRow("provider_config", effective.chatCompletionsPath ?? null),
"wire.commandCodeVersion": behaviorRow("provider_config", effective.commandCodeVersion ?? null),
"wire.modelSuffixMode": behaviorRow(
"provider_config",
effective.modelSuffixBracketStrip === true ? "bracket_strip" : "none",
),
"auth.mode": behaviorRow("provider_config", authMode),
"auth.transport": behaviorRow(
"provider_config",
resolveProviderAuthTransport(adapter, authMode, effective.apiKeyTransport),
),
"responses.stateful": behaviorRow("provider_config", effective.statelessResponses !== true),
"responses.serviceTier": behaviorRow(
"provider_config",
serviceTierSupportFromPolicy(fastPolicy) ?? null,
),
"responses.fastWireKind": behaviorRow(
"provider_config",
fastPolicy.fastWire?.kind ?? null,
),
"responses.fastWireValue": behaviorRow(
"provider_config",
fastPolicy.fastWire?.canonicalToWire.priority ?? null,
),
"responses.snapshotRepair": behaviorRow("provider_config", effective.responsesSnapshotRepair === true),
"responses.itemIdRepair": behaviorRow("provider_config", effective.responsesItemIdRepair ?? null),
"limits.contextWindow": behaviorRow(
"provider_config",
modelValue(effective.modelContextWindows, modelId) ?? effective.contextWindow ?? null,
),
"limits.maxInputTokens": behaviorRow(
"provider_config",
modelValue(effective.modelMaxInputTokens, modelId) ?? null,
),
"limits.maxOutputTokens": behaviorRow(
"provider_config",
modelValue(effective.modelMaxOutputTokens, modelId) ?? effective.defaultMaxOutputTokens ?? null,
),
"modalities.input": behaviorRow(
"provider_config",
modelValue(effective.modelInputModalities, modelId) ?? ["text"],
),
"sampling.omitTemperature": behaviorRow("provider_config", includesModel(effective.noTemperatureModels, modelId)),
"sampling.omitTopP": behaviorRow("provider_config", includesModel(effective.noTopPModels, modelId)),
"sampling.omitPenalties": behaviorRow("provider_config", includesModel(effective.noPenaltyModels, modelId)),
"reasoning.supported": behaviorRow(
"provider_config",
includesModel(effective.noReasoningModels, modelId)
? false
: reasoningEfforts.length > 0 || Boolean(registryEntry?.reasoningEfforts),
),
"reasoning.efforts": behaviorRow("provider_config", reasoningEfforts),
// Whether the configured ladder above is what the WIRE uses, or only what the catalog
// advertises. An adapter shipping its own per-model table (command-code) otherwise wins for
// models it knows, and it also repairs a refused rung by replaying without the effort. Two
// routes that disagree about this send different bytes and recover differently, so they are
// not the same compatibility subject and must not share evidence.
"reasoning.effortsAuthoritative": behaviorRow("provider_config", effective.modelReasoningEffortsAuthoritative === true),
"reasoning.defaultEffort": behaviorRow("provider_config", defaultReasoningEffort),
"reasoning.effortMap": behaviorRow("provider_config", reasoningEffortMap),
"reasoning.wireFormat": behaviorRow("provider_config", effective.reasoningWireFormat ?? null),
"reasoning.summaryMode": behaviorRow("provider_config", {
supported: modelValue(effective.modelSupportsReasoningSummaries, modelId) ?? null,
delivery: modelValue(effective.modelReasoningSummaryDelivery, modelId) ?? null,
}),
"reasoning.replayMode": behaviorRow("provider_config", {
preserveResponses: effective.preserveResponsesReasoningContent === true,
// Dropping replayed reasoning items changes the continuation body, so two routes that
// disagree about it are not the same compatibility subject and must not share evidence.
dropResponses: effective.dropResponsesReasoningItems === true,
preserveContent: includesModel(effective.preserveReasoningContentModels, modelId),
placeholder: includesModel(effective.requiresReasoningPlaceholderModels, modelId),
}),
"reasoning.splitMode": behaviorRow("provider_config", includesModel(effective.reasoningSplitModels, modelId)),
"reasoning.toggleMode": behaviorRow("provider_config", includesModel(effective.thinkingToggleModels, modelId)),
"reasoning.budgetMode": behaviorRow("provider_config", includesModel(effective.thinkingBudgetModels, modelId)),
"tools.choiceRestrictions": behaviorRow(
"provider_config",
includesModel(effective.autoToolChoiceOnlyModels, modelId) ? ["auto"] : [],
),
"tools.parallel": behaviorRow(
"provider_config",
effective.parallelToolCalls ?? (upstreamProtocol === "openai-chat"),
),
"tools.hostedPreference": behaviorRow("provider_config", {
tools: exactOwnValue(effective.modelPreferHostedTools, modelId) ?? [],
}),
"tools.builtinNameEscaping": behaviorRow("provider_config", effective.escapeBuiltinToolNames === true),
"cache.forwarding": behaviorRow("provider_config", effective.promptCacheKey === true),
"cache.retention": behaviorRow("global_config", config.cacheRetention ?? "short"),
"anthropic.eofPolicy": behaviorRow("provider_config", effective.anthropicEofTolerance === true ? "tolerant" : "strict"),
"openai-chat.eofPolicy": behaviorRow("provider_config", effective.openaiChatEofTolerance === true ? "tolerant" : "strict"),
"google.mode": behaviorRow("provider_config", effective.googleMode ?? null),
"google.projectFingerprint": behaviorRow(
"provider_config",
project ? localFingerprint("googleProject", project, installationSalt) : null,
),
"google.locationFingerprint": behaviorRow(
"provider_config",
location ? localFingerprint("googleLocation", location, installationSalt) : null,
),
"openrouter.order": behaviorRow("provider_config", openRouterRouting?.order ?? []),
"openrouter.only": behaviorRow("provider_config", openRouterRouting?.only ?? []),
"openrouter.allowFallbacks": behaviorRow("provider_config", openRouterRouting?.allowFallbacks ?? null),
"sidecars.vision": behaviorRow(
"global_config",
localFingerprint("visionSidecar", config.visionSidecar ?? null, installationSalt),
),
"sidecars.webSearch": behaviorRow(
"global_config",
localFingerprint("webSearchSidecar", config.webSearchSidecar ?? null, installationSalt),
),
"mcp.maxTools": behaviorRow("provider_config", effective.mcpMaxTools ?? null),
"mcp.maxSchemaBytes": behaviorRow("provider_config", effective.mcpMaxSchemaBytes ?? null),
"mcp.maxResultBytes": behaviorRow("provider_config", effective.mcpMaxResultBytes ?? null),
"mcp.nativeLocalExec": behaviorRow("provider_config", nativeLocalExec),
"runtime.bunVersion": behaviorRow("registry_runtime_default", Bun.version),
"runtime.platform": behaviorRow("registry_runtime_default", process.platform),
"runtime.arch": behaviorRow("registry_runtime_default", process.arch),
"runtime.streamMode": behaviorRow("global_config", config.streamMode ?? "auto"),
"runtime.fastMode": behaviorRow("global_config", config.fastMode === true),
"runtime.effortCap": behaviorRow("global_config", config.effortCap ?? null),
"headers.nonCredentialBehaviorDigest": behaviorRow(
"provider_config",
nonCredentialHeaderDigest(effective.headers, installationSalt),
),
};
return values;
}
export function providerInstanceKey(
providerName: string,
effective: OcxProviderConfig,
): string {
const baseUrl = typeof effective.baseUrl === "string" ? effective.baseUrl.trim() : "";
const adapter = typeof effective.adapter === "string" ? effective.adapter : "";
return `${providerName}\0${baseUrl}\0${adapter}`;
}