Function handleStream

src/handler.ts:1143–1814

(res: Response, cursorReq: CursorChatRequest, body: AnthropicRequest, log: RequestLogger, clientRequestedThinking: boolean = false)

// ==================== 流式处理 ====================

async function handleStream(res: Response, cursorReq: CursorChatRequest, body: AnthropicRequest, log: RequestLogger, clientRequestedThinking: boolean = false): Promise<void> {
    // 设置 SSE headers
    res.writeHead(200, {
        'Content-Type': 'text/event-stream',
        'Cache-Control': 'no-cache',
        'Connection': 'keep-alive',
        'X-Accel-Buffering': 'no',
    });

    const id = msgId();
    const model = body.model;
    const hasTools = (body.tools?.length ?? 0) > 0;

    // 发送 message_start
    writeSSE(res, 'message_start', {
        type: 'message_start',
        message: {
            id, type: 'message', role: 'assistant', content: [],
            model, stop_reason: null, stop_sequence: null,
            usage: { input_tokens: estimateInputTokens(body), output_tokens: 0 },
        },
    });

    // ★ 流式保活 — 注意：无工具的增量流式路径（handleDirectTextStream）有自己的 keepalive
    // 这里的 keepalive 仅用于工具模式下的缓冲/续写期间
    let keepaliveInterval: ReturnType<typeof setInterval> | undefined;

    let fullResponse = '';
    let sentText = '';
    let blockIndex = 0;
    let textBlockStarted = false;
    let thinkingBlockEmitted = false;
    let cursorUsage: { inputTokens?: number; outputTokens?: number; totalTokens?: number } | undefined;

    // 无工具模式：先缓冲全部响应再检测拒绝，如果是拒绝则重试
    let activeCursorReq = cursorReq;
    let retryCount = 0;

    const executeStream = async (detectRefusalEarly = false, onTextDelta?: (delta: string) => void): Promise<{ earlyAborted: boolean }> => {
        fullResponse = '';
        const apiStart = Date.now();
        let firstChunk = true;
        let earlyAborted = false;
        log.startPhase('send', '发送到 Cursor');

        // ★ 早期中止支持：检测到拒绝后立即中断流，不等完整响应
        const abortController = detectRefusalEarly ? new AbortController() : undefined;

        try {
            await sendCursorRequest(activeCursorReq, (event: CursorSSEEvent) => {
                if (event.type === 'finish') {
                    if (event.messageMetadata?.usage) cursorUsage = event.messageMetadata.usage;
                    return;
                }
                if (event.type !== 'text-delta' || !event.delta) return;
                if (firstChunk) { log.recordTTFT(); log.endPhase(); log.startPhase('response', '接收响应'); firstChunk = false; }
                fullResponse += event.delta;
                onTextDelta?.(event.delta);