Skip to content

Commit 5e4c32e

Browse files
mechanicmuthumechanicmuthu
committed
flex with heavy console logs
1 parent b0c1eed commit 5e4c32e

File tree

3 files changed

+133
-13
lines changed

3 files changed

+133
-13
lines changed

‎src/api/providers/openai-native.ts

Lines changed: 108 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -57,7 +57,11 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
5757
this.options.enableGpt5ReasoningSummary = true
5858
}
5959
const apiKey = this.options.openAiNativeApiKey ?? "not-provided"
60-
this.client = new OpenAI({ baseURL: this.options.openAiNativeBaseUrl, apiKey })
60+
this.client = new OpenAI({
61+
baseURL: this.options.openAiNativeBaseUrl,
62+
apiKey,
63+
timeout: 15 * 1000 * 60, // 15 minutes default timeout
64+
})
6165
}
6266

6367
private normalizeGpt5Usage(usage: any, model: OpenAiNativeModel): ApiStreamUsageChunk | undefined {
@@ -136,7 +140,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
136140
const isOriginalO1 = model.id === "o1"
137141
const { reasoning } = this.getModel()
138142

139-
const response = await this.client.chat.completions.create({
143+
const params: any = {
140144
model: model.id,
141145
messages: [
142146
{
@@ -148,9 +152,30 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
148152
stream: true,
149153
stream_options: { include_usage: true },
150154
...(reasoning && reasoning),
151-
})
155+
}
156+
157+
// Add service_tier parameter if configured and not "auto"
158+
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
159+
params.service_tier = this.options.serviceTier
160+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
161+
console.log("[DEBUG] Full request params:", JSON.stringify(params, null, 2))
162+
} else {
163+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
164+
}
165+
166+
const response = await this.client.chat.completions.create(params, { timeout: 15 * 1000 * 60 })
167+
console.log("[DEBUG] OpenAI Chat Completions Response (O1Family):", response)
152168

153-
yield* this.handleStreamResponse(response, model)
169+
if (typeof (response as any)[Symbol.asyncIterator] !== "function") {
170+
throw new Error(
171+
"OpenAI SDK did not return an AsyncIterable for streaming response. Please check SDK version and usage.",
172+
)
173+
}
174+
175+
yield* this.handleStreamResponse(
176+
response as unknown as AsyncIterable<OpenAI.Chat.Completions.ChatCompletionChunk>,
177+
model,
178+
)
154179
}
155180

156181
private async *handleReasonerMessage(
@@ -161,7 +186,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
161186
): ApiStream {
162187
const { reasoning } = this.getModel()
163188

164-
const stream = await this.client.chat.completions.create({
189+
const params: any = {
165190
model: family,
166191
messages: [
167192
{
@@ -173,9 +198,29 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
173198
stream: true,
174199
stream_options: { include_usage: true },
175200
...(reasoning && reasoning),
176-
})
201+
}
202+
203+
// Add service_tier parameter if configured and not "auto"
204+
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
205+
params.service_tier = this.options.serviceTier
206+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
207+
console.log("[DEBUG] Full request params:", JSON.stringify(params, null, 2))
208+
} else {
209+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
210+
}
211+
212+
const stream = await this.client.chat.completions.create(params, { timeout: 15 * 1000 * 60 })
177213

178-
yield* this.handleStreamResponse(stream, model)
214+
if (typeof (stream as any)[Symbol.asyncIterator] !== "function") {
215+
throw new Error(
216+
"OpenAI SDK did not return an AsyncIterable for streaming response. Please check SDK version and usage.",
217+
)
218+
}
219+
220+
yield* this.handleStreamResponse(
221+
stream as unknown as AsyncIterable<OpenAI.Chat.Completions.ChatCompletionChunk>,
222+
model,
223+
)
179224
}
180225

181226
private async *handleDefaultModelMessage(
@@ -200,7 +245,16 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
200245
params.verbosity = verbosity
201246
}
202247

203-
const stream = await this.client.chat.completions.create(params)
248+
// Add service_tier parameter if configured and not "auto"
249+
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
250+
params.service_tier = this.options.serviceTier
251+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
252+
console.log("[DEBUG] Full request params:", JSON.stringify(params, null, 2))
253+
} else {
254+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
255+
}
256+
257+
const stream = await this.client.chat.completions.create(params, { timeout: 15 * 1000 * 60 })
204258

205259
if (typeof (stream as any)[Symbol.asyncIterator] !== "function") {
206260
throw new Error(
@@ -277,6 +331,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
277331
temperature?: number
278332
max_output_tokens?: number
279333
previous_response_id?: string
334+
service_tier?: string
280335
}
281336

282337
const requestBody: Gpt5RequestBody = {
@@ -297,9 +352,20 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
297352
...(requestPreviousResponseId && { previous_response_id: requestPreviousResponseId }),
298353
}
299354

355+
// Add service_tier parameter if configured and not "auto"
356+
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
357+
requestBody.service_tier = this.options.serviceTier
358+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
359+
console.log("[DEBUG] Full request body:", JSON.stringify(requestBody, null, 2))
360+
} else {
361+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
362+
}
363+
300364
try {
301365
// Use the official SDK
302-
const stream = (await (this.client as any).responses.create(requestBody)) as AsyncIterable<any>
366+
const stream = (await (this.client as any).responses.create(requestBody, {
367+
timeout: 15 * 1000 * 60,
368+
})) as AsyncIterable<any>
303369

304370
if (typeof (stream as any)[Symbol.asyncIterator] !== "function") {
305371
throw new Error(
@@ -308,11 +374,13 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
308374
}
309375

310376
for await (const event of stream) {
377+
console.log("[DEBUG] GPT-5 Responses API Stream Event:", event) // Log each event
311378
for await (const outChunk of this.processGpt5Event(event, model)) {
312379
yield outChunk
313380
}
314381
}
315382
} catch (sdkErr: any) {
383+
console.error("[DEBUG] OpenAI Responses API SDK Error:", sdkErr) // Log SDK errors
316384
// Check if this is a 400 error about previous_response_id not found
317385
const errorMessage = sdkErr?.message || sdkErr?.error?.message || ""
318386
const is400Error = sdkErr?.status === 400 || sdkErr?.response?.status === 400
@@ -419,6 +487,15 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
419487
const baseUrl = this.options.openAiNativeBaseUrl || "https://api.openai.com"
420488
const url = `${baseUrl}/v1/responses`
421489

490+
// Log the exact request being sent
491+
console.log("[DEBUG] GPT-5 Responses API Request URL:", url)
492+
console.log("[DEBUG] GPT-5 Responses API Request Headers:", {
493+
"Content-Type": "application/json",
494+
Authorization: `Bearer ${apiKey.substring(0, 8)}...`, // Log only first 8 chars of API key for security
495+
Accept: "text/event-stream",
496+
})
497+
console.log("[DEBUG] GPT-5 Responses API Request Body:", JSON.stringify(requestBody, null, 2))
498+
422499
try {
423500
const response = await fetch(url, {
424501
method: "POST",
@@ -430,8 +507,18 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
430507
body: JSON.stringify(requestBody),
431508
})
432509

510+
// Log the response status
511+
console.log("[DEBUG] GPT-5 Responses API Response Status:", response.status)
512+
// Convert headers to a plain object for logging
513+
const headersObj: Record<string, string> = {}
514+
response.headers.forEach((value, key) => {
515+
headersObj[key] = value
516+
})
517+
console.log("[DEBUG] GPT-5 Responses API Response Headers:", headersObj)
518+
433519
if (!response.ok) {
434520
const errorText = await response.text()
521+
console.log("[DEBUG] GPT-5 Responses API Error Response Body:", errorText)
435522

436523
let errorMessage = `GPT-5 API request failed (${response.status})`
437524
let errorDetails = ""
@@ -471,6 +558,11 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
471558
this.resolveResponseId(undefined)
472559

473560
// Retry the request without the previous_response_id
561+
console.log(
562+
"[DEBUG] GPT-5 Responses API Retry Request Body:",
563+
JSON.stringify(retryRequestBody, null, 2),
564+
)
565+
474566
const retryResponse = await fetch(url, {
475567
method: "POST",
476568
headers: {
@@ -481,7 +573,11 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
481573
body: JSON.stringify(retryRequestBody),
482574
})
483575

576+
console.log("[DEBUG] GPT-5 Responses API Retry Response Status:", retryResponse.status)
577+
484578
if (!retryResponse.ok) {
579+
const retryErrorText = await retryResponse.text()
580+
console.log("[DEBUG] GPT-5 Responses API Retry Error Response Body:", retryErrorText)
485581
// If retry also fails, throw the original error
486582
throw new Error(`GPT-5 API retry failed (${retryResponse.status})`)
487583
}
@@ -537,6 +633,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
537633
// Handle streaming response
538634
yield* this.handleGpt5StreamResponse(response.body, model)
539635
} catch (error) {
636+
console.error("[DEBUG] GPT-5 Responses API Fetch Error:", error) // Log fetch errors
540637
if (error instanceof Error) {
541638
// Re-throw with the original error message if it's already formatted
542639
if (error.message.includes("GPT-5")) {
@@ -1040,6 +1137,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
10401137
* Used by both the official SDK streaming path and (optionally) by the SSE fallback.
10411138
*/
10421139
private async *processGpt5Event(event: any, model: OpenAiNativeModel): ApiStream {
1140+
console.log("[DEBUG] processGpt5Event: Processing event type:", event?.type)
10431141
// Persist response id for conversation continuity when available
10441142
if (event?.response?.id) {
10451143
this.resolveResponseId(event.response.id)
@@ -1148,6 +1246,7 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
11481246
model: OpenAiNativeModel,
11491247
): ApiStream {
11501248
for await (const chunk of stream) {
1249+
console.log("[DEBUG] handleStreamResponse: OpenAI Chat Completions Stream Chunk:", chunk) // Log each chunk here
11511250
const delta = chunk.choices[0]?.delta
11521251

11531252
if (delta?.content) {

‎src/api/providers/openai.ts

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -160,6 +160,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
160160

161161
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
162162
;(requestOptions as any).service_tier = this.options.serviceTier
163+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
164+
console.log("[DEBUG] Full request options:", JSON.stringify(requestOptions, null, 2))
165+
} else {
166+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
163167
}
164168

165169
// Add max_tokens if needed
@@ -226,6 +230,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
226230

227231
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
228232
;(requestOptions as any).service_tier = this.options.serviceTier
233+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
234+
console.log("[DEBUG] Full request options:", JSON.stringify(requestOptions, null, 2))
235+
} else {
236+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
229237
}
230238

231239
// Add max_tokens if needed
@@ -275,6 +283,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
275283

276284
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
277285
;(requestOptions as any).service_tier = this.options.serviceTier
286+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
287+
console.log("[DEBUG] Full request options:", JSON.stringify(requestOptions, null, 2))
288+
} else {
289+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
278290
}
279291

280292
// Add max_tokens if needed
@@ -323,6 +335,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
323335

324336
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
325337
;(requestOptions as any).service_tier = this.options.serviceTier
338+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
339+
console.log("[DEBUG] Full request options:", JSON.stringify(requestOptions, null, 2))
340+
} else {
341+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
326342
}
327343

328344
// O3 family models do not support the deprecated max_tokens parameter
@@ -352,6 +368,10 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
352368

353369
if (this.options.serviceTier && this.options.serviceTier !== "auto") {
354370
;(requestOptions as any).service_tier = this.options.serviceTier
371+
console.log("[DEBUG] Setting service_tier parameter:", this.options.serviceTier)
372+
console.log("[DEBUG] Full request options:", JSON.stringify(requestOptions, null, 2))
373+
} else {
374+
console.log("[DEBUG] Service tier not set or is 'auto'. Current value:", this.options.serviceTier)
355375
}
356376

357377
// O3 family models do not support the deprecated max_tokens parameter

‎webview-ui/src/components/settings/ServiceTier.tsx

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -18,10 +18,11 @@ export const ServiceTier = ({ apiConfiguration, setApiConfigurationField, modelI
1818
const { t } = useAppTranslation()
1919
const effectiveModelId = modelId || apiConfiguration.openAiModelId || ""
2020

21-
const isSupported = useMemo(
22-
() => !!effectiveModelId && FLEX_COMPATIBLE_MODELS.some((m) => effectiveModelId.includes(m)),
23-
[effectiveModelId],
24-
)
21+
const isSupported = useMemo(() => {
22+
const supported = !!effectiveModelId && FLEX_COMPATIBLE_MODELS.some((m) => effectiveModelId.includes(m))
23+
console.log("[DEBUG] Service tier supported check:", { effectiveModelId, supported, FLEX_COMPATIBLE_MODELS })
24+
return supported
25+
}, [effectiveModelId])
2526

2627
// Initialize to auto when supported and unset; clear when unsupported
2728
useEffect(() => {

0 commit comments

Comments
 (0)