i/next-ai-draw-io
1
0
Fork 0
mirror of https://github.com/DayuanJiang/next-ai-draw-io.git synced 2026-10-12 20:49:53 +08:00
Code Issues Packages Projects Releases Wiki Activity

refactor(server): drop stale prompt text, duplicated routes and dead options

Browse Source 
- lib/system-prompts.ts: delete the extended prompt kept only to reach
  the 4096-token cache threshold of two old Claude models; describe the
  current canvas-first UI in the App Features section
- app/api/chat/route.ts: remove the server-side example cache (the
  browser answers examples itself), the ENABLE_HISTORY_XML_REPLACE
  option and its helper, the unread messageMetadata fields; the
  DEBUG_LLM_PAYLOAD blocks become one describeMessages summary
- app/api/log-feedback and log-save share scoreLatestTrace in
  lib/langfuse.ts
- app/api/admin/test-model and validate-model call a shared
  testModelConnection() instead of one route faking a request to the
  other and passing trust through the body
- lib/dynamo-quota-manager.ts reads the failing item from the
  conditional update instead of a second GetItem
- lib/ai-providers.ts: geminiThinkingOptions helper for the google and
  vertexai cases, drop the unused fourth resolveBaseURL parameter
- edge-functions: always stream, shrink messageSchema, drop model
  aliases nothing sends
- delete app/api/chat/xml_guide.md, parseOutputTokenLimit, unused type
  exports and export keywords on file-internal symbols
- env.example: drop variables nothing reads, document the quota,
  validation and analytics variables, mention the apiKeyEnv array form
  (also in docs/*/ai-providers.md)
This commit is contained in:
dayuan.jiang
2026-10-12 08:15:15 +09:00
parent 610ba6872d
commit 8ca2d03e7f
36 changed files with 694 additions and 1505 deletions
Download Patch File Download Diff File Expand all files Collapse all files
app/api/chat/route.ts
+26 -132
View File
@@ -1,8 +1,6 @@
import {
APICallError,
convertToModelMessages,
createUIMessageStream,
createUIMessageStreamResponse,
InvalidToolInputError,
stepCountIs,
streamText,
@@ -21,12 +19,10 @@ import {
usesServerCredentials,
usesServerEndpoint,
} from "@/lib/ai-providers"
import { findCachedResponse } from "@/lib/cached-responses"
import {
dropInvalidToolCalls,
fixToolInputJson,
formatSelectionContext,
replaceHistoricalToolInputs,
validateFileParts,
} from "@/lib/chat-helpers"
import { withDeprecatedParamsFallback } from "@/lib/deprecated-params"
@@ -56,7 +52,6 @@ import { getSystemPrompt } from "@/lib/system-prompts"
import { normalizeBaseUrl } from "@/lib/types/model-config"
import { getUserIdFromRequest } from "@/lib/user-id"
import { foldCells } from "@/packages/mcp-server/src/compact-cells.ts"
import { hasCells } from "@/packages/mcp-server/src/pages.ts"
import {
getShapeLibrary,
SHAPE_LIBRARY_LIST,
@@ -70,36 +65,6 @@ import {
// the tool call, so take whatever the host allows. Vercel's own default is 300s,
// which is also where Node's response-body timeout on the upstream stream lands.
// Helper function to create cached stream response
function createCachedStreamResponse(xml: string): Response {
const toolCallId = `cached-${Date.now()}`
const stream = createUIMessageStream({
execute: async ({ writer }) => {
writer.write({ type: "start" })
writer.write({
type: "tool-input-start",
toolCallId,
toolName: "display_diagram",
})
writer.write({
type: "tool-input-delta",
toolCallId,
inputTextDelta: xml,
})
writer.write({
type: "tool-input-available",
toolCallId,
toolName: "display_diagram",
input: { xml },
})
writer.write({ type: "finish" })
},
})
return createUIMessageStreamResponse({ stream })
}
// Responses streamed from the model, whose trace streamText's callbacks end
const modelStreamResponses = new WeakSet<Response>()
// A Request's signal follows the client's disconnect only while the Request
@@ -111,6 +76,21 @@ const requestOfResponse = new WeakMap<Response, Request>()
// Inner handler function
const DEBUG_LLM_PAYLOAD = process.env.DEBUG_LLM_PAYLOAD === "true"
// Message structure for DEBUG_LLM_PAYLOAD logs: roles, part types, tool names
// and input keys, without the diagram XML or image data
const describeMessages = (msgs: any[]) =>
msgs.map((m) => ({
role: m.role,
parts: (m.parts ?? m.content ?? []).map(
(p: any) =>
p.type +
(p.toolName ? `:${p.toolName}` : "") +
(p.input && typeof p.input === "object"
? `(${Object.keys(p.input).join(",")})`
: ""),
),
}))
async function handleChatRequest(req: Request): Promise<Response> {
const crossSite = rejectCrossSite(req)
if (crossSite) return crossSite
@@ -156,23 +136,6 @@ async function handleChatRequest(req: Request): Promise<Response> {
}
// === FILE VALIDATION END ===
// === CACHE CHECK START ===
const isFirstMessage = messages.length === 1
const isEmptyDiagram = !xml || !hasCells(xml)
if (isFirstMessage && isEmptyDiagram) {
const lastMessage = messages[0]
const textPart = lastMessage.parts?.find((p: any) => p.type === "text")
const filePart = lastMessage.parts?.find((p: any) => p.type === "file")
const cached = findCachedResponse(textPart?.text || "", !!filePart)
if (cached) {
return createCachedStreamResponse(cached.xml)
}
}
// === CACHE CHECK END ===
// Read client AI provider overrides from headers
const provider = req.headers.get("x-ai-provider")
let baseUrl = req.headers.get("x-ai-base-url")
@@ -348,7 +311,7 @@ async function handleChatRequest(req: Request): Promise<Response> {
`[Prompt Caching] ${shouldCache ? "ENABLED" : "DISABLED"} for model: ${modelId}`,
)
// Get the appropriate system prompt based on model (extended for Opus/Haiku 4.5)
// System prompt, with the model name filled in and the style section chosen
const systemMessage = getSystemPrompt(modelId, minimalStyle)
const finalSystemMessage = customSystemMessage
? `${systemMessage}\n\n## Custom Instructions\n${customSystemMessage}`
@@ -380,47 +343,15 @@ ${userInputText}
// DEBUG_LLM_PAYLOAD=true logs the incoming message structure
if (DEBUG_LLM_PAYLOAD) {
console.log("[route.ts] Incoming messages count:", messages.length)
messages.forEach((msg: any, idx: number) => {
console.log(
`[route.ts] Message ${idx} role:`,
msg.role,
"parts count:",
msg.parts?.length,
)
if (msg.parts) {
msg.parts.forEach((part: any, partIdx: number) => {
if (
part.type === "tool-invocation" ||
part.type === "tool-result"
) {
console.log(`[route.ts] Part ${partIdx}:`, {
type: part.type,
toolName: part.toolName,
hasInput: !!part.input,
inputType: typeof part.input,
inputKeys:
part.input && typeof part.input === "object"
? Object.keys(part.input)
: null,
})
}
})
}
})
console.log(
"[route.ts] incoming",
JSON.stringify(describeMessages(messages)),
)
}
// Replace historical tool call XML with placeholders to reduce tokens
// Disabled by default - some models (e.g. minimax) copy placeholders instead of generating XML
const enableHistoryReplace =
process.env.ENABLE_HISTORY_XML_REPLACE === "true"
const placeholderMessages = enableHistoryReplace
? replaceHistoricalToolInputs(modelMessages)
: modelMessages
// Filter out messages with empty content arrays (Bedrock API rejects these)
// This is a safety measure - ideally convertToModelMessages should handle all cases
let enhancedMessages = placeholderMessages.filter(
let enhancedMessages = modelMessages.filter(
(msg: any) =>
msg.content && Array.isArray(msg.content) && msg.content.length > 0,
)
@@ -432,36 +363,10 @@ ${userInputText}
// DEBUG_LLM_PAYLOAD=true logs what is sent to the model
if (DEBUG_LLM_PAYLOAD) {
console.log("[route.ts] Model messages count:", enhancedMessages.length)
enhancedMessages.forEach((msg: any, idx: number) => {
console.log(
`[route.ts] ModelMsg ${idx} role:`,
msg.role,
"content count:",
msg.content?.length,
)
if (msg.content) {
msg.content.forEach((part: any, partIdx: number) => {
if (
part.type === "tool-call" ||
part.type === "tool-result"
) {
console.log(`[route.ts] Content ${partIdx}:`, {
type: part.type,
toolName: part.toolName,
hasInput: !!part.input,
inputType: typeof part.input,
inputValue:
part.input === undefined
? "undefined"
: part.input === null
? "null"
: "object",
})
}
})
}
})
console.log(
"[route.ts] to model",
JSON.stringify(describeMessages(enhancedMessages)),
)
}
// Update the last message with user input only (XML moved to separate cached system message)
@@ -790,17 +695,6 @@ Call this tool to get shape names and usage syntax for a specific library.`,
// Keyless endpoints keep theirs: the desktop app's Ollama is the
// user's own, and EdgeOne's text is our function's explanation.
onError: (error) => streamErrorText(error, onServerCredentials),
messageMetadata: ({ part }) => {
if (part.type === "finish") {
const usage = (part as any).totalUsage
// AI SDK 6 provides totalTokens directly
return {
totalTokens: usage?.totalTokens ?? 0,
finishReason: (part as any).finishReason,
}
}
return undefined
},
})
modelStreamResponses.add(response)
requestOfResponse.set(response, req)
@@ -840,7 +734,7 @@ async function safeHandler(req: Request): Promise<Response> {
} catch (error) {
response = handleError(error)
}
// Early returns, cache hits and errors never reach streamText's callbacks,
// Early returns and errors never reach streamText's callbacks,
// so their Langfuse trace has to be ended here
if (!modelStreamResponses.has(response)) endTrace()
return response