dsh-lcx-codex 0.5.1-pre.1 → 0.5.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +11 -11
- package/THIRD_PARTY_NOTICES.md +2 -2
- package/lib/auxiliary-usage.js +45 -6
- package/lib/client.js +12 -12
- package/lib/dsh-compat.js +17 -0
- package/lib/dsh-responses.js +66 -88
- package/lib/index.js +121 -50
- package/lib/model-catalog.js +48 -0
- package/lib/native-checkpoint.js +18 -51
- package/lib/pi-responses-runtime.js +4 -0
- package/lib/responses-request.js +4 -1
- package/lib/responses-stream.js +162 -12
- package/lib/route.js +75 -33
- package/lib/search-usage.js +12 -5
- package/lib/transport.js +8 -0
- package/package.json +30 -30
package/lib/responses-stream.js
CHANGED
|
@@ -73,12 +73,115 @@ function errorFacts(error) {
|
|
|
73
73
|
text: texts.join(" | "),
|
|
74
74
|
};
|
|
75
75
|
}
|
|
76
|
+
/** @param {unknown} value */
|
|
77
|
+
function safeMachineField(value) {
|
|
78
|
+
if (typeof value !== "string" || value.length === 0 || value.length > 96)
|
|
79
|
+
return undefined;
|
|
80
|
+
return /^[A-Za-z0-9_.:\[\]-]+$/u.test(value) ? value : undefined;
|
|
81
|
+
}
|
|
82
|
+
/**
|
|
83
|
+
* Record the structured facts of a failure the provider reports in-stream. Pi collapses every such
|
|
84
|
+
* event into one `Error("<code>: <message>")` and `response.completed` with status `failed` into a
|
|
85
|
+
* bare "stream ended in failure", so the machine fields are captured here, before that happens.
|
|
86
|
+
*/
|
|
87
|
+
function observeUpstreamFailure(event, meta) {
|
|
88
|
+
const type = String(event.type ?? "");
|
|
89
|
+
let source;
|
|
90
|
+
let details;
|
|
91
|
+
let hasError = false;
|
|
92
|
+
let incomplete = false;
|
|
93
|
+
if (type === "error") {
|
|
94
|
+
// OpenAI sends flat `{code, message}`; some gateways nest `{error: {code, type}}`. `type` is only
|
|
95
|
+
// ever the error type when nested, because the flat `type` is the event name itself.
|
|
96
|
+
const nested = isObject(event.error) ? event.error : undefined;
|
|
97
|
+
source = { code: nested?.code ?? event.code, type: nested?.type };
|
|
98
|
+
hasError = true;
|
|
99
|
+
}
|
|
100
|
+
else if (isObject(event.response)) {
|
|
101
|
+
const response = event.response;
|
|
102
|
+
const status = response.status;
|
|
103
|
+
const failedTerminal = type === "response.failed" ||
|
|
104
|
+
(type === "response.completed" && status === "failed");
|
|
105
|
+
const incompleteTerminal = type === "response.incomplete" ||
|
|
106
|
+
(type === "response.completed" && status === "incomplete");
|
|
107
|
+
if (!failedTerminal && !incompleteTerminal)
|
|
108
|
+
return;
|
|
109
|
+
incomplete = incompleteTerminal;
|
|
110
|
+
details = isObject(response.incomplete_details)
|
|
111
|
+
? response.incomplete_details.reason
|
|
112
|
+
: undefined;
|
|
113
|
+
// max_output_tokens is a length stop that Pi maps to a normal finish, not a failure.
|
|
114
|
+
if (incompleteTerminal && details === "max_output_tokens")
|
|
115
|
+
return;
|
|
116
|
+
source = isObject(response.error) ? response.error : undefined;
|
|
117
|
+
hasError = source !== undefined;
|
|
118
|
+
}
|
|
119
|
+
else
|
|
120
|
+
return;
|
|
121
|
+
const code = safeMachineField(source?.code);
|
|
122
|
+
const errorType = safeMachineField(source?.type);
|
|
123
|
+
const incompleteReason = safeMachineField(details);
|
|
124
|
+
meta.upstreamFailure = {
|
|
125
|
+
event: type,
|
|
126
|
+
...(code === undefined ? {} : { code }),
|
|
127
|
+
...(errorType === undefined ? {} : { type: errorType }),
|
|
128
|
+
...(incompleteReason === undefined ? {} : { incompleteReason }),
|
|
129
|
+
hasError,
|
|
130
|
+
incomplete,
|
|
131
|
+
};
|
|
132
|
+
}
|
|
133
|
+
/** Closed mapping from provider machine identifiers to the DSH failure taxonomy; unknown identifiers return undefined. */
|
|
134
|
+
function classifyProviderIdentifier(value) {
|
|
135
|
+
if (value === undefined)
|
|
136
|
+
return undefined;
|
|
137
|
+
const id = value.toLowerCase();
|
|
138
|
+
if (id === "insufficient_quota" || id === "insufficient_user_quota")
|
|
139
|
+
return "INSUFFICIENT_QUOTA";
|
|
140
|
+
if (/^(?:invalid_api_key|authentication_error|permission_error)$/u.test(id))
|
|
141
|
+
return "AUTH";
|
|
142
|
+
if (/^rate_limit|^too_many_requests$/u.test(id))
|
|
143
|
+
return "RATE_LIMIT";
|
|
144
|
+
if (id === "context_length_exceeded")
|
|
145
|
+
return "CONTEXT_WINDOW_EXCEEDED";
|
|
146
|
+
if (id === "content_filter" ||
|
|
147
|
+
id === "content_policy_violation" ||
|
|
148
|
+
id.startsWith("invalid_"))
|
|
149
|
+
return "INVALID_REQUEST";
|
|
150
|
+
if (/^(?:server_error|internal_error|internal_server_error|api_error|service_unavailable|bad_response_status_code)$/u.test(id) ||
|
|
151
|
+
id.includes("overloaded"))
|
|
152
|
+
return "SERVER";
|
|
153
|
+
return undefined;
|
|
154
|
+
}
|
|
155
|
+
/**
|
|
156
|
+
* Failure class for an in-stream provider failure, from its structured fields only.
|
|
157
|
+
* `definite` classes come from a recognised identifier and outrank everything else. A failed response
|
|
158
|
+
* with no usable `error.code` (no error object, `{}`, a message only, an empty code, or only an unknown
|
|
159
|
+
* type) is still a provider-side failure and defaults to SERVER, but not definitely: wording in the
|
|
160
|
+
* message that clearly names another class (rate limit, context window, timeout) keeps that class.
|
|
161
|
+
* An unrecognised non-empty code is left to the text heuristics and the generic fallback.
|
|
162
|
+
*/
|
|
163
|
+
function classifyUpstreamFailure(upstream) {
|
|
164
|
+
if (upstream === undefined)
|
|
165
|
+
return undefined;
|
|
166
|
+
const fromError = classifyProviderIdentifier(upstream.code) ??
|
|
167
|
+
classifyProviderIdentifier(upstream.type);
|
|
168
|
+
if (fromError !== undefined)
|
|
169
|
+
return { code: fromError, definite: true };
|
|
170
|
+
const fromReason = classifyProviderIdentifier(upstream.incompleteReason);
|
|
171
|
+
if (fromReason !== undefined)
|
|
172
|
+
return { code: fromReason, definite: true };
|
|
173
|
+
if (upstream.incomplete)
|
|
174
|
+
return undefined;
|
|
175
|
+
if (upstream.code === undefined)
|
|
176
|
+
return { code: "SERVER", definite: false };
|
|
177
|
+
return undefined;
|
|
178
|
+
}
|
|
76
179
|
/**
|
|
77
180
|
* Provider bodies/messages are deliberately not surfaced. Only a stable class and safe facts leave the wire boundary.
|
|
78
181
|
* @param {unknown} error
|
|
79
182
|
* @param {AbortSignal} [signal]
|
|
80
183
|
*/
|
|
81
|
-
export function managedFailure(error, signal) {
|
|
184
|
+
export function managedFailure(error, signal, upstream) {
|
|
82
185
|
const facts = errorFacts(error);
|
|
83
186
|
const sourceCodes = new Set(facts.codes);
|
|
84
187
|
const text = facts.text;
|
|
@@ -87,8 +190,11 @@ export function managedFailure(error, signal) {
|
|
|
87
190
|
(signalReason?.name === "TimeoutError" ||
|
|
88
191
|
signalReason?.code === "TimeoutError" ||
|
|
89
192
|
/time(?:d)?\s*out|timeout/iu.test(String(signalReason?.message ?? "")));
|
|
193
|
+
const upstreamClass = signal?.aborted ? undefined : classifyUpstreamFailure(upstream);
|
|
90
194
|
let code = "RESPONSES_ERROR";
|
|
91
|
-
if (
|
|
195
|
+
if (upstreamClass?.definite === true)
|
|
196
|
+
code = upstreamClass.code;
|
|
197
|
+
else if (signalTimedOut ||
|
|
92
198
|
facts.status === 408 ||
|
|
93
199
|
sourceCodes.has("TimeoutError") ||
|
|
94
200
|
/time(?:d)?\s*out|timeout/iu.test(text))
|
|
@@ -104,8 +210,9 @@ export function managedFailure(error, signal) {
|
|
|
104
210
|
else if (sourceCodes.has("LCX_GROK_NATIVE_PROTOCOL_ERROR"))
|
|
105
211
|
code = "GROK_NATIVE_PROTOCOL_ERROR";
|
|
106
212
|
else if (sourceCodes.has("LCX_RESPONSES_ROUTE_UNAVAILABLE") ||
|
|
107
|
-
sourceCodes.has("
|
|
108
|
-
sourceCodes.has("
|
|
213
|
+
sourceCodes.has("LCX_RESPONSES_API_UNRESOLVED") ||
|
|
214
|
+
sourceCodes.has("LCX_RESPONSES_MODEL_NOT_CONFIGURED") ||
|
|
215
|
+
sourceCodes.has("LCX_RESPONSES_MODEL_UNAVAILABLE"))
|
|
109
216
|
code = "NO_ADAPTER";
|
|
110
217
|
else if (sourceCodes.has("LCX_INSUFFICIENT_QUOTA"))
|
|
111
218
|
code = "INSUFFICIENT_QUOTA";
|
|
@@ -129,6 +236,7 @@ export function managedFailure(error, signal) {
|
|
|
129
236
|
facts.status === 422 ||
|
|
130
237
|
sourceCodes.has("LCX_RESPONSES_INVALID_INPUT") ||
|
|
131
238
|
sourceCodes.has("LCX_CHECKPOINT_UNSUPPORTED") ||
|
|
239
|
+
sourceCodes.has("LCX_CHECKPOINT_IMAGE_BUDGET") ||
|
|
132
240
|
/invalid.?request|payload too large|length limit exceeded/iu.test(text))
|
|
133
241
|
code = "INVALID_REQUEST";
|
|
134
242
|
else if (sourceCodes.has("LCX_RESPONSE_TOO_LARGE"))
|
|
@@ -139,19 +247,26 @@ export function managedFailure(error, signal) {
|
|
|
139
247
|
else if (error instanceof TypeError ||
|
|
140
248
|
/\bnetwork|connection|socket|fetch|ECONN|EAI_AGAIN|terminated|premature close\b/iu.test(text))
|
|
141
249
|
code = "TRANSPORT";
|
|
250
|
+
// Nothing in the text said otherwise: a failure the provider reported without a usable code is its own.
|
|
251
|
+
if (code === "RESPONSES_ERROR" && upstreamClass !== undefined)
|
|
252
|
+
code = upstreamClass.code;
|
|
142
253
|
const messages = /** @type {Record<string, string>} */ {
|
|
143
254
|
ABORTED: "Responses request was aborted",
|
|
144
255
|
AUTH: "Responses request was rejected by authentication",
|
|
145
256
|
INSUFFICIENT_QUOTA: "服务商额度不足或未满足请求预留额度,请检查中转额度(并非 API 密钥无效) / Provider quota is insufficient for this request",
|
|
146
257
|
UNSUPPORTED_OPTION: "LCX Responses does not support this request option",
|
|
147
|
-
NO_ADAPTER: sourceCodes.has("
|
|
148
|
-
? "LCX
|
|
149
|
-
: "
|
|
258
|
+
NO_ADAPTER: sourceCodes.has("LCX_RESPONSES_MODEL_NOT_CONFIGURED")
|
|
259
|
+
? "LCX is enabled but the selected model is not configured for this provider (the provider profile lists other models); add it to the profile's models list or select a listed model"
|
|
260
|
+
: sourceCodes.has("LCX_RESPONSES_API_UNRESOLVED")
|
|
261
|
+
? "LCX could not determine the wire protocol of the selected GPT route (the provider profile sets no api and the model is unknown to the catalog); add api: openai-responses to the provider profile or select a model DSH recognizes"
|
|
262
|
+
: "LCX could not resolve the selected Responses route",
|
|
150
263
|
RATE_LIMIT: "Responses provider rate limit was reached",
|
|
151
264
|
SERVER: "Responses provider returned a server failure",
|
|
152
265
|
INVALID_REQUEST: sourceCodes.has("LCX_CHECKPOINT_UNSUPPORTED")
|
|
153
266
|
? "Unsupported or invalid LCX checkpoint; start a new session"
|
|
154
|
-
: "
|
|
267
|
+
: sourceCodes.has("LCX_CHECKPOINT_IMAGE_BUDGET")
|
|
268
|
+
? "Images kept in the compaction checkpoint exceed the request image bound; start a new session or raise maxRequestImageBytes"
|
|
269
|
+
: "Responses provider rejected the request",
|
|
155
270
|
CONTEXT_WINDOW_EXCEEDED: "Responses request exceeded the model context window",
|
|
156
271
|
UNSUPPORTED_CONTENT: "LCX cannot serialize a DSH content block for this Responses route",
|
|
157
272
|
GROK_NATIVE_PROTOCOL_ERROR: "Grok returned an unrecognized client tool call while native search was enabled",
|
|
@@ -722,6 +837,7 @@ async function* normalizedResponseEvents(source, meta = {}, options = {}) {
|
|
|
722
837
|
if (!isObject(raw))
|
|
723
838
|
continue;
|
|
724
839
|
const event = /** @type {UnknownRecord} */ raw;
|
|
840
|
+
observeUpstreamFailure(event, meta);
|
|
725
841
|
if (options.serverToolTypes && ['response.completed', 'response.failed', 'response.incomplete'].includes(String(event.type))) {
|
|
726
842
|
meta.contextDetails = sanitizeContextDetails(isObject(event.response) && isObject(event.response.usage)
|
|
727
843
|
? event.response.usage.context_details : undefined);
|
|
@@ -1335,7 +1451,33 @@ async function* toDshChunks(events, signal, wireMeta, nativeReplayRoute) {
|
|
|
1335
1451
|
* @param {ReadonlySet<string>} [options.serverToolTypes]
|
|
1336
1452
|
* @param {(usage: ServerToolUsage) => void} [options.onServerToolUsage]
|
|
1337
1453
|
*/
|
|
1338
|
-
export async function* streamResponsesRequest({ baseURL, provider, model, piModel, body, grammarToolInputProperties, headers, signal, timeoutMs, applyDefaultTimeout = true, streamIdleTimeoutMs, maxAttempts = 1, maxResponseBytes, serverToolTypes, isServerToolItem, declaredToolNames, nativeReplayRoute, onServerToolUsage, }) {
|
|
1454
|
+
export async function* streamResponsesRequest({ baseURL, provider, model, piModel, body, grammarToolInputProperties, headers, signal, timeoutMs, applyDefaultTimeout = true, streamIdleTimeoutMs, maxAttempts = 1, maxResponseBytes, serverToolTypes, isServerToolItem, declaredToolNames, nativeReplayRoute, onServerToolUsage, onFailure, }) {
|
|
1455
|
+
let contentEmitted = false;
|
|
1456
|
+
let failureReported = false;
|
|
1457
|
+
/** One redacted report per request; user aborts are not failures. */
|
|
1458
|
+
const reportFailure = (chunk, managed, upstream) => {
|
|
1459
|
+
if (failureReported || onFailure === undefined)
|
|
1460
|
+
return;
|
|
1461
|
+
if (chunk.type !== "finish" || chunk.reason.kind !== "error")
|
|
1462
|
+
return;
|
|
1463
|
+
failureReported = true;
|
|
1464
|
+
const failure = chunk.reason.failure;
|
|
1465
|
+
try {
|
|
1466
|
+
onFailure({
|
|
1467
|
+
code: String(failure.code),
|
|
1468
|
+
...(typeof failure.status === "number" ? { status: failure.status } : {}),
|
|
1469
|
+
...(managed?.requestId === undefined ? {} : { requestId: managed.requestId }),
|
|
1470
|
+
...(upstream?.code === undefined ? {} : { providerCode: upstream.code }),
|
|
1471
|
+
...(upstream?.type === undefined ? {} : { providerType: upstream.type }),
|
|
1472
|
+
...(upstream?.incompleteReason === undefined ? {} : { incompleteReason: upstream.incompleteReason }),
|
|
1473
|
+
...(upstream === undefined ? {} : { event: upstream.event }),
|
|
1474
|
+
contentEmitted,
|
|
1475
|
+
});
|
|
1476
|
+
}
|
|
1477
|
+
catch {
|
|
1478
|
+
// Diagnostics must never change the stream outcome.
|
|
1479
|
+
}
|
|
1480
|
+
};
|
|
1339
1481
|
const effectiveTimeoutMs = timeoutMs ?? (applyDefaultTimeout ? 300_000 : undefined);
|
|
1340
1482
|
const deadline = requestDeadline(signal, effectiveTimeoutMs);
|
|
1341
1483
|
const watchdog = streamIdleWatchdog(deadline.signal, streamIdleTimeoutMs);
|
|
@@ -1383,7 +1525,7 @@ export async function* streamResponsesRequest({ baseURL, provider, model, piMode
|
|
|
1383
1525
|
output.stopReason = signal?.aborted ? "aborted" : "error";
|
|
1384
1526
|
output.errorMessage =
|
|
1385
1527
|
error instanceof Error ? error.message : "Responses stream failed";
|
|
1386
|
-
output.__lcxFailure = managedFailure(error, signal);
|
|
1528
|
+
output.__lcxFailure = managedFailure(error, signal, wireMeta.upstreamFailure);
|
|
1387
1529
|
piEvents.push({
|
|
1388
1530
|
type: "error",
|
|
1389
1531
|
reason: /** @type {'aborted' | 'error'} */ output.stopReason,
|
|
@@ -1405,7 +1547,13 @@ export async function* streamResponsesRequest({ baseURL, provider, model, piMode
|
|
|
1405
1547
|
break;
|
|
1406
1548
|
}
|
|
1407
1549
|
watchdog.disarm();
|
|
1408
|
-
|
|
1550
|
+
const value = result.value;
|
|
1551
|
+
if (value.type === "text-delta" ||
|
|
1552
|
+
value.type === "reasoning-delta" ||
|
|
1553
|
+
value.type === "tool-call-delta")
|
|
1554
|
+
contentEmitted = true;
|
|
1555
|
+
reportFailure(value, output.__lcxFailure, wireMeta.upstreamFailure);
|
|
1556
|
+
yield value;
|
|
1409
1557
|
watchdog.arm();
|
|
1410
1558
|
}
|
|
1411
1559
|
}
|
|
@@ -1417,7 +1565,9 @@ export async function* streamResponsesRequest({ baseURL, provider, model, piMode
|
|
|
1417
1565
|
await parser;
|
|
1418
1566
|
}
|
|
1419
1567
|
catch (error) {
|
|
1420
|
-
|
|
1568
|
+
const chunk = managedFailureChunk(error, signal);
|
|
1569
|
+
reportFailure(chunk, managedFailure(error, signal), undefined);
|
|
1570
|
+
yield chunk;
|
|
1421
1571
|
}
|
|
1422
1572
|
finally {
|
|
1423
1573
|
watchdog.dispose();
|
package/lib/route.js
CHANGED
|
@@ -3,8 +3,7 @@ import { credentialRef } from "@deepseek-ai/dsh-credentials";
|
|
|
3
3
|
import { SessionId } from "@deepseek-ai/dsh-session";
|
|
4
4
|
import "@deepseek-ai/dsh-settings";
|
|
5
5
|
import { attributionHeaders, resolveRetryPolicy } from "@deepseek-ai/dsh-llm";
|
|
6
|
-
import {
|
|
7
|
-
import { wasPi086ResponsesModel } from "./pi086-responses-allowlist.js";
|
|
6
|
+
import { resolveCatalogModel } from "./model-catalog.js";
|
|
8
7
|
/** @param {unknown} value */
|
|
9
8
|
function isRecord(value) {
|
|
10
9
|
return value !== null && typeof value === "object" && !Array.isArray(value);
|
|
@@ -125,17 +124,6 @@ function copyResponsesCompat(target, source) {
|
|
|
125
124
|
Object.assign(target, { [field]: value });
|
|
126
125
|
}
|
|
127
126
|
}
|
|
128
|
-
/** @param {unknown} provider @param {unknown} modelId */
|
|
129
|
-
function builtinResponsesModel(provider, modelId) {
|
|
130
|
-
const providerName = String(provider ?? "");
|
|
131
|
-
// A newer plugin-Pi catalog must not grant LCX a route that Pi 0.86 lacked.
|
|
132
|
-
if (!wasPi086ResponsesModel(providerName, String(modelId ?? "")))
|
|
133
|
-
return undefined;
|
|
134
|
-
const builtinProvider = getBuiltinProviders().find((candidate) => candidate === providerName);
|
|
135
|
-
if (builtinProvider === undefined)
|
|
136
|
-
return undefined;
|
|
137
|
-
return getBuiltinModels(builtinProvider).find((model) => model.id === String(modelId ?? "") && model.api === "openai-responses");
|
|
138
|
-
}
|
|
139
127
|
function modelDescriptorDefaults(model) {
|
|
140
128
|
if (!model)
|
|
141
129
|
return undefined;
|
|
@@ -144,6 +132,18 @@ function modelDescriptorDefaults(model) {
|
|
|
144
132
|
thinkingLevelMap: model.thinkingLevelMap,
|
|
145
133
|
};
|
|
146
134
|
}
|
|
135
|
+
/** The profile lists this model in `models`: the one way DSH serves a model its catalog does not describe. */
|
|
136
|
+
function profileDeclaresModel(profile, modelId) {
|
|
137
|
+
return Array.isArray(profile?.models) && profile.models.some((entry) => String(entry?.id ?? "") === String(modelId));
|
|
138
|
+
}
|
|
139
|
+
/**
|
|
140
|
+
* A nonempty `models` list replaces the served catalog: DSH serves exactly the listed models, and LCX answers
|
|
141
|
+
* `llm/stream` before DSH's adapter does, so a claim here would send a request DSH itself refuses.
|
|
142
|
+
*/
|
|
143
|
+
function profileExcludesModel(profile, modelId) {
|
|
144
|
+
return Array.isArray(profile?.models) && profile.models.length > 0 &&
|
|
145
|
+
!profile.models.some((entry) => String(entry?.id ?? "") === String(modelId));
|
|
146
|
+
}
|
|
147
147
|
function configuredModelProfile(profile, modelId) {
|
|
148
148
|
const configuredModels = Array.isArray(profile?.models) ? profile.models : [];
|
|
149
149
|
return configuredModels.length > 0
|
|
@@ -169,34 +169,48 @@ const MODEL_THINKING_LEVELS = [
|
|
|
169
169
|
function profileReasoning(value) {
|
|
170
170
|
return MODEL_THINKING_LEVELS.find((level) => level === value);
|
|
171
171
|
}
|
|
172
|
-
|
|
173
|
-
|
|
174
|
-
|
|
172
|
+
/**
|
|
173
|
+
* The reasoning projection of a model whose profile entry declares `reasoningEfforts` (the mapping
|
|
174
|
+
* from DSH levels to wire spellings, or `false` for a non-reasoning model); undefined when the
|
|
175
|
+
* profile says nothing, so the caller falls back to the catalog or to what DSH resolved.
|
|
176
|
+
*/
|
|
177
|
+
function declaredReasoningControls(profile, modelId) {
|
|
178
|
+
const efforts = configuredModelProfile(profile, modelId)?.reasoningEfforts;
|
|
175
179
|
if (efforts === undefined)
|
|
176
|
-
return
|
|
177
|
-
reasoning: base?.reasoning ?? false,
|
|
178
|
-
...(base?.thinkingLevelMap === undefined
|
|
179
|
-
? {}
|
|
180
|
-
: { thinkingLevelMap: base.thinkingLevelMap }),
|
|
181
|
-
};
|
|
180
|
+
return undefined;
|
|
182
181
|
if (efforts === false)
|
|
183
182
|
return { reasoning: false };
|
|
184
183
|
if (!isRecord(efforts))
|
|
185
184
|
return { reasoning: false };
|
|
186
185
|
const thinkingLevelMap = {};
|
|
186
|
+
let offOmitsReasoning = false;
|
|
187
187
|
for (const level of MODEL_THINKING_LEVELS) {
|
|
188
188
|
const value = efforts[level];
|
|
189
189
|
if (value === undefined)
|
|
190
190
|
thinkingLevelMap[level] = null;
|
|
191
191
|
else if (value === null) {
|
|
192
|
-
// DSH permits only off:null: it means supported-off by omitting the wire option.
|
|
193
|
-
|
|
194
|
-
|
|
192
|
+
// DSH permits only off:null: it means supported-off by omitting the wire option. Pi 1.1.0 would
|
|
193
|
+
// read an absent `off` as "send none", so the omission is made explicit here.
|
|
194
|
+
thinkingLevelMap[level] = null;
|
|
195
|
+
if (level === "off")
|
|
196
|
+
offOmitsReasoning = true;
|
|
195
197
|
}
|
|
196
198
|
else if (typeof value === "string")
|
|
197
199
|
thinkingLevelMap[level] = value;
|
|
198
200
|
}
|
|
199
|
-
return {
|
|
201
|
+
return {
|
|
202
|
+
reasoning: true,
|
|
203
|
+
thinkingLevelMap,
|
|
204
|
+
...(offOmitsReasoning ? { offOmitsReasoning: true } : {}),
|
|
205
|
+
};
|
|
206
|
+
}
|
|
207
|
+
function configuredModelControls(profile, modelId, base) {
|
|
208
|
+
return (declaredReasoningControls(profile, modelId) ?? {
|
|
209
|
+
reasoning: base?.reasoning ?? false,
|
|
210
|
+
...(base?.thinkingLevelMap === undefined
|
|
211
|
+
? {}
|
|
212
|
+
: { thinkingLevelMap: base.thinkingLevelMap }),
|
|
213
|
+
});
|
|
200
214
|
}
|
|
201
215
|
function isLcxCapabilityRoute(provider, model) {
|
|
202
216
|
return provider === "lcx" &&
|
|
@@ -213,11 +227,15 @@ export function resolveResponsesRouteConfig(ctx, options, policy) {
|
|
|
213
227
|
if (!isRecord(profile))
|
|
214
228
|
return undefined;
|
|
215
229
|
const configured = profile;
|
|
216
|
-
|
|
217
|
-
|
|
230
|
+
if (profileExcludesModel(configured, model))
|
|
231
|
+
return undefined;
|
|
232
|
+
// Protocol and endpoint: the profile, then what DSH resolves for the model (see `resolveCatalogModel`).
|
|
233
|
+
const catalog = resolveCatalogModel(provider, model, profileDeclaresModel(configured, model));
|
|
234
|
+
const builtin = catalog.descriptor;
|
|
235
|
+
const api = configured.api ?? catalog.api;
|
|
218
236
|
if (api !== "openai-responses")
|
|
219
237
|
return undefined;
|
|
220
|
-
const baseURL = configured.baseURL ??
|
|
238
|
+
const baseURL = configured.baseURL ?? catalog.baseUrl;
|
|
221
239
|
const apiKeyEnv = configured.apiKeyEnv;
|
|
222
240
|
if (typeof baseURL !== "string" || !normalizeBaseURL(baseURL) || typeof apiKeyEnv !== "string" || !apiKeyEnv.trim())
|
|
223
241
|
return undefined;
|
|
@@ -225,7 +243,7 @@ export function resolveResponsesRouteConfig(ctx, options, policy) {
|
|
|
225
243
|
copyResponsesCompat(responsesCompat, builtin?.compat);
|
|
226
244
|
copyResponsesCompat(responsesCompat, lcxCapabilityRoute ? policy.responsesCompat : undefined);
|
|
227
245
|
copyResponsesCompat(responsesCompat, configuredResponsesCompat(configured, model));
|
|
228
|
-
const promptCacheModel = lcxCapabilityRoute && policy.supportsExplicitPromptCacheMode === true ?
|
|
246
|
+
const promptCacheModel = lcxCapabilityRoute && policy.supportsExplicitPromptCacheMode === true ? resolveCatalogModel("openai", model).descriptor : undefined;
|
|
229
247
|
const promptCacheRaw = promptCacheModel;
|
|
230
248
|
const promptCacheCompat = isRecord(promptCacheRaw)
|
|
231
249
|
? promptCacheRaw.compat
|
|
@@ -241,6 +259,7 @@ export function resolveResponsesRouteConfig(ctx, options, policy) {
|
|
|
241
259
|
const fallbackRetention = lcxCapabilityRoute ? policy.cacheRetention : undefined;
|
|
242
260
|
const requestedRetention = configuredRetention === "none" || configuredRetention === "short" || configuredRetention === "long" ? configuredRetention : fallbackRetention === "none" || fallbackRetention === "short" || fallbackRetention === "long" ? fallbackRetention : undefined;
|
|
243
261
|
const cacheRetention = requestedRetention === "long" && !supportsLongCacheRetention ? "short" : requestedRetention ?? (supportsLongCacheRetention ? "long" : "short");
|
|
262
|
+
const declaredControls = declaredReasoningControls(configured, model);
|
|
244
263
|
return {
|
|
245
264
|
provider,
|
|
246
265
|
model,
|
|
@@ -251,6 +270,9 @@ export function resolveResponsesRouteConfig(ctx, options, policy) {
|
|
|
251
270
|
cacheRetention,
|
|
252
271
|
supportsLongCacheRetention,
|
|
253
272
|
responsesCompat,
|
|
273
|
+
// Only a declared `reasoningEfforts` is projected here; an undeclared model follows the catalog
|
|
274
|
+
// and what DSH resolved for it (see `gptModelOverride`).
|
|
275
|
+
...(declaredControls === undefined ? {} : { modelControls: declaredControls }),
|
|
254
276
|
// These are explicitly LCX-owned transport and request-capability policies.
|
|
255
277
|
timeoutMs: isPositiveInteger(configured.timeoutMs) ? configured.timeoutMs : policy.timeoutMs,
|
|
256
278
|
maxAttempts: retryAttempts(configured.retryPolicy, policy.maxAttempts),
|
|
@@ -259,6 +281,23 @@ export function resolveResponsesRouteConfig(ctx, options, policy) {
|
|
|
259
281
|
requestImageMaxBytes: isPositiveInteger(configured.requestImageMaxBytes) ? configured.requestImageMaxBytes : policy.requestImageMaxBytes,
|
|
260
282
|
};
|
|
261
283
|
}
|
|
284
|
+
/**
|
|
285
|
+
* Name the reason `resolveResponsesRouteConfig` returned nothing, so the failure can tell the user what to
|
|
286
|
+
* change: `api-unresolved` means neither the profile nor the model catalog names a protocol for this model.
|
|
287
|
+
*/
|
|
288
|
+
export function diagnoseResponsesRoute(ctx, options) {
|
|
289
|
+
const provider = String(options?.provider ?? "");
|
|
290
|
+
const model = String(options?.model ?? "");
|
|
291
|
+
const profile = provider.trim() ? configuredProviderProfile(ctx, provider) : undefined;
|
|
292
|
+
if (!isRecord(profile))
|
|
293
|
+
return "no-profile";
|
|
294
|
+
if (profileExcludesModel(profile, model))
|
|
295
|
+
return "model-not-configured";
|
|
296
|
+
const api = profile.api ?? resolveCatalogModel(provider, model, profileDeclaresModel(profile, model)).api;
|
|
297
|
+
if (api === undefined)
|
|
298
|
+
return "api-unresolved";
|
|
299
|
+
return api === "openai-responses" ? "incomplete" : "api-not-responses";
|
|
300
|
+
}
|
|
262
301
|
/** Resolve a Grok-prefixed model through its selected DSH Responses profile. */
|
|
263
302
|
export function resolveGrokResponsesRouteConfig(ctx, options, policy) {
|
|
264
303
|
const provider = String(options?.provider ?? "");
|
|
@@ -268,11 +307,14 @@ export function resolveGrokResponsesRouteConfig(ctx, options, policy) {
|
|
|
268
307
|
const configured = configuredProviderProfile(ctx, provider);
|
|
269
308
|
if (!isRecord(configured))
|
|
270
309
|
return undefined;
|
|
271
|
-
|
|
272
|
-
|
|
310
|
+
if (profileExcludesModel(configured, model))
|
|
311
|
+
return undefined;
|
|
312
|
+
const catalog = resolveCatalogModel(provider, model, profileDeclaresModel(configured, model));
|
|
313
|
+
const selectedBuiltin = catalog.descriptor;
|
|
314
|
+
const api = configured.api ?? catalog.api;
|
|
273
315
|
if (api !== "openai-responses")
|
|
274
316
|
return undefined;
|
|
275
|
-
const baseURL = configured.baseURL ??
|
|
317
|
+
const baseURL = configured.baseURL ?? catalog.baseUrl;
|
|
276
318
|
const apiKeyEnv = configured.apiKeyEnv;
|
|
277
319
|
if (typeof baseURL !== "string" ||
|
|
278
320
|
!normalizeBaseURL(baseURL) ||
|
package/lib/search-usage.js
CHANGED
|
@@ -10,15 +10,22 @@ export const searchUsageProjection = {
|
|
|
10
10
|
key: 'lcxSearchUsage', stateVersion: 3, stateSchema: schema.extend({ pending: z.record(z.string(), z.string()), headerKey: z.string().optional(), sampleHeaderKey: z.string().optional() }),
|
|
11
11
|
init: () => ({ auxiliary: zeroBuckets(), aggregateContext: false, pending: {} }),
|
|
12
12
|
apply(state, event) {
|
|
13
|
-
|
|
14
|
-
|
|
13
|
+
// Unit state is checkpointed by the host's projection cache, which refuses any state that is not
|
|
14
|
+
// lossless plain JSON. An optional field is therefore omitted when absent, never set to `undefined`.
|
|
15
|
+
if (event.type === 'request/header') {
|
|
16
|
+
const { headerKey: _previous, ...rest } = state;
|
|
17
|
+
const key = headerKey(event.data.header);
|
|
18
|
+
return key === undefined ? rest : { ...rest, headerKey: key };
|
|
19
|
+
}
|
|
15
20
|
// DSH measure anchors assistant/message only. Attempts must not transplant a
|
|
16
21
|
// newer context sample onto an older successful measurement anchor.
|
|
17
22
|
if (event.type === 'assistant/message' || event.type === 'assistant/attempt') {
|
|
18
23
|
const contextDetails = event.type === 'assistant/message' ? providerContextOf(event) : undefined;
|
|
19
|
-
|
|
20
|
-
|
|
21
|
-
|
|
24
|
+
// The newest reply replaces the whole provider-context sample: drop the old keys, then add the new ones.
|
|
25
|
+
const { contextDetails: _details, contextProvenance: _provenance, sampleHeaderKey: _sample, ...rest } = state;
|
|
26
|
+
return { ...rest, aggregateContext: aggregateContextOf(event) ?? state.aggregateContext,
|
|
27
|
+
...(contextDetails ? { contextDetails, contextProvenance: 'provider-context-details',
|
|
28
|
+
...(state.headerKey === undefined ? {} : { sampleHeaderKey: state.headerKey }) } : {}) };
|
|
22
29
|
}
|
|
23
30
|
if (event.type === 'tool/call' && isSearchTool(event.data.name))
|
|
24
31
|
return { ...state, pending: { ...state.pending, [event.data.callId]: event.data.name } };
|
package/lib/transport.js
CHANGED
|
@@ -279,10 +279,17 @@ export async function consumeSse(response, onEvent, options = {}) {
|
|
|
279
279
|
}
|
|
280
280
|
onEvent(event);
|
|
281
281
|
};
|
|
282
|
+
// An abort must interrupt a read that never settles, not only be noticed between reads. Cancelling
|
|
283
|
+
// the reader settles the pending read as done, so the abort is re-checked right after every read.
|
|
284
|
+
const onAbort = () => {
|
|
285
|
+
reader.cancel(options.signal?.reason).catch(() => { });
|
|
286
|
+
};
|
|
287
|
+
options.signal?.addEventListener("abort", onAbort, { once: true });
|
|
282
288
|
try {
|
|
283
289
|
while (true) {
|
|
284
290
|
abortIfNeeded(options.signal);
|
|
285
291
|
const { done, value } = await reader.read();
|
|
292
|
+
abortIfNeeded(options.signal);
|
|
286
293
|
if (done)
|
|
287
294
|
break;
|
|
288
295
|
bytes += value.byteLength;
|
|
@@ -305,6 +312,7 @@ export async function consumeSse(response, onEvent, options = {}) {
|
|
|
305
312
|
dispatch();
|
|
306
313
|
}
|
|
307
314
|
finally {
|
|
315
|
+
options.signal?.removeEventListener("abort", onAbort);
|
|
308
316
|
try {
|
|
309
317
|
await reader.cancel();
|
|
310
318
|
}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "dsh-lcx-codex",
|
|
3
|
-
"version": "0.5.
|
|
3
|
+
"version": "0.5.2",
|
|
4
4
|
"description": "Model-scoped DSH plugin: GPT Responses lifecycle, Native V2 compaction and Hosted/Alpha Search; Grok native Web/X Search",
|
|
5
5
|
"keywords": [
|
|
6
6
|
"deepseek-harness",
|
|
@@ -62,9 +62,9 @@
|
|
|
62
62
|
}
|
|
63
63
|
},
|
|
64
64
|
"devDependencies": {
|
|
65
|
-
"@deepseek-ai/dsh": "0.2.
|
|
65
|
+
"@deepseek-ai/dsh": "0.2.0-rc.2",
|
|
66
66
|
"@earendil-works/pi-ai": "0.87.1",
|
|
67
|
-
"@deepseek-ai/schemastery": "
|
|
67
|
+
"@deepseek-ai/schemastery": "^3.18.2",
|
|
68
68
|
"@types/node": "^22.20.0",
|
|
69
69
|
"@types/react": "~18.3.1",
|
|
70
70
|
"esbuild": "^0.28.2",
|
|
@@ -72,31 +72,31 @@
|
|
|
72
72
|
"typescript": "7.0.2"
|
|
73
73
|
},
|
|
74
74
|
"peerDependencies": {
|
|
75
|
-
"@deepseek-ai/cordis": "
|
|
76
|
-
"@deepseek-ai/cordis-plugin-loader": "
|
|
77
|
-
"@deepseek-ai/dsh-agent": ">=0.2.
|
|
78
|
-
"@deepseek-ai/dsh-attachment": ">=0.2.
|
|
79
|
-
"@deepseek-ai/dsh-client-locale": ">=0.2.
|
|
80
|
-
"@deepseek-ai/dsh-client-modules": ">=0.2.
|
|
81
|
-
"@deepseek-ai/dsh-client-store": ">=0.2.
|
|
82
|
-
"@deepseek-ai/dsh-client-ui-chat": ">=0.2.
|
|
83
|
-
"@deepseek-ai/dsh-client-ui-conversation": ">=0.2.
|
|
84
|
-
"@deepseek-ai/dsh-client-ui-plugin-manager": ">=0.2.
|
|
85
|
-
"@deepseek-ai/dsh-client-ui-settings": ">=0.2.
|
|
86
|
-
"@deepseek-ai/dsh-compaction": ">=0.2.
|
|
87
|
-
"@deepseek-ai/dsh-credentials": ">=0.2.
|
|
88
|
-
"@deepseek-ai/dsh-fs": ">=0.2.
|
|
89
|
-
"@deepseek-ai/dsh-llm": ">=0.2.
|
|
90
|
-
"@deepseek-ai/dsh-session": ">=0.2.
|
|
91
|
-
"@deepseek-ai/dsh-session-query": ">=0.2.
|
|
92
|
-
"@deepseek-ai/dsh-settings": ">=0.2.
|
|
93
|
-
"@deepseek-ai/dsh-tool-web": ">=0.2.
|
|
94
|
-
"@deepseek-ai/dsh-tools": ">=0.2.
|
|
95
|
-
"@deepseek-ai/dsh-web": ">=0.2.
|
|
96
|
-
"@deepseek-ai/schemastery": "
|
|
75
|
+
"@deepseek-ai/cordis": "^4.0.2",
|
|
76
|
+
"@deepseek-ai/cordis-plugin-loader": "~1.0.5",
|
|
77
|
+
"@deepseek-ai/dsh-agent": ">=0.2.0-rc.2 <0.2.1",
|
|
78
|
+
"@deepseek-ai/dsh-attachment": ">=0.2.0-rc.2 <0.2.1",
|
|
79
|
+
"@deepseek-ai/dsh-client-locale": ">=0.2.0-rc.2 <0.2.1",
|
|
80
|
+
"@deepseek-ai/dsh-client-modules": ">=0.2.0-rc.2 <0.2.1",
|
|
81
|
+
"@deepseek-ai/dsh-client-store": ">=0.2.0-rc.2 <0.2.1",
|
|
82
|
+
"@deepseek-ai/dsh-client-ui-chat": ">=0.2.0-rc.2 <0.2.1",
|
|
83
|
+
"@deepseek-ai/dsh-client-ui-conversation": ">=0.2.0-rc.2 <0.2.1",
|
|
84
|
+
"@deepseek-ai/dsh-client-ui-plugin-manager": ">=0.2.0-rc.2 <0.2.1",
|
|
85
|
+
"@deepseek-ai/dsh-client-ui-settings": ">=0.2.0-rc.2 <0.2.1",
|
|
86
|
+
"@deepseek-ai/dsh-compaction": ">=0.2.0-rc.2 <0.2.1",
|
|
87
|
+
"@deepseek-ai/dsh-credentials": ">=0.2.0-rc.2 <0.2.1",
|
|
88
|
+
"@deepseek-ai/dsh-fs": ">=0.2.0-rc.2 <0.2.1",
|
|
89
|
+
"@deepseek-ai/dsh-llm": ">=0.2.0-rc.2 <0.2.1",
|
|
90
|
+
"@deepseek-ai/dsh-session": ">=0.2.0-rc.2 <0.2.1",
|
|
91
|
+
"@deepseek-ai/dsh-session-query": ">=0.2.0-rc.2 <0.2.1",
|
|
92
|
+
"@deepseek-ai/dsh-settings": ">=0.2.0-rc.2 <0.2.1",
|
|
93
|
+
"@deepseek-ai/dsh-tool-web": ">=0.2.0-rc.2 <0.2.1",
|
|
94
|
+
"@deepseek-ai/dsh-tools": ">=0.2.0-rc.2 <0.2.1",
|
|
95
|
+
"@deepseek-ai/dsh-web": ">=0.2.0-rc.2 <0.2.1",
|
|
96
|
+
"@deepseek-ai/schemastery": "^3.18.2",
|
|
97
97
|
"react": "^18.2.0",
|
|
98
|
-
"@deepseek-ai/dsh-token-meter": ">=0.2.
|
|
99
|
-
"@deepseek-ai/dsh-session-projection": ">=0.2.
|
|
98
|
+
"@deepseek-ai/dsh-token-meter": ">=0.2.0-rc.2 <0.2.1",
|
|
99
|
+
"@deepseek-ai/dsh-session-projection": ">=0.2.0-rc.2 <0.2.1",
|
|
100
100
|
"zod": "^4.5.4"
|
|
101
101
|
},
|
|
102
102
|
"peerDependenciesMeta": {
|
|
@@ -179,9 +179,9 @@
|
|
|
179
179
|
},
|
|
180
180
|
"license": "MIT",
|
|
181
181
|
"lcxCompatibility": {
|
|
182
|
-
"dshInstallRange": ">=0.2.
|
|
183
|
-
"verifiedDsh": "0.2.
|
|
184
|
-
"verifiedHostPi": "
|
|
182
|
+
"dshInstallRange": ">=0.2.0-rc.2 <0.2.1",
|
|
183
|
+
"verifiedDsh": "0.2.0-rc.2",
|
|
184
|
+
"verifiedHostPi": "0.87.1",
|
|
185
185
|
"verifiedPluginPi": "0.87.1"
|
|
186
186
|
}
|
|
187
187
|
}
|