public ListDownStreamReportCollectVO searchDownApiSumBIReportSearchVo biReportSearchVo ListDownStreamReportCollectVO voList = biReportMappersearchApiReportListbiReportSearchVosetProductIdListL

这段代码可以进行优化。以下是一些可能的优化方法：

1. 使用方法引用：可以使用方法引用来简化代码。例如，将Function<Function<DownstreamAPIReportVO, Long>, Long>简化为ToLongFunction<DownstreamAPIReportVO>，将Function<Function<DownstreamAPIReportVO, BigDecimal>, BigDecimal>简化为Function<DownstreamAPIReportVO, BigDecimal>。

2. 使用静态导入：通过使用静态导入，可以将代码更清晰地表达出来，减少代码量。

3. 减少重复代码：可以将重复的代码提取到单独的方法中，以避免重复编写相同的代码。

4. 使用并行流：如果数据量很大，可以考虑使用并行流来加快处理速度。

下面是经过优化的代码示例：

public List<DownStreamReportCollectVO> searchDownApiSum(BIReportSearchVo biReportSearchVo) {
    List<DownStreamReportCollectVO> voList = biReportMapper.searchApiReportList(biReportSearchVo.setProductIdList(Lists.emptyList()))
            .stream()
            .collect(Collectors.groupingBy(DownstreamAPIReportVO::getProductName))
            .entrySet()
            .stream()
            .map(item -> {
                Supplier<Stream<DownstreamAPIReportVO>> sup = item.getValue()::stream;
                // Long 求和
                ToLongFunction<DownstreamAPIReportVO> longFun = t -> sup.get().mapToLong(tlf).filter(Objects::nonNull).sum();
                // BigDecimal 求和
                Function<DownstreamAPIReportVO, BigDecimal> bigFun = fn -> sup.get().map(fn).filter(Objects::nonNull).reduce(BigDecimal.ZERO, BigDecimal::add);

                Long threeItemCheckCount = longFun.applyAsLong(DownstreamAPIReportVO::getThreeItemCheckCount);
                Long authSuccessCount = longFun.applyAsLong(DownstreamAPIReportVO::getAuthSuccessCount);
                BigDecimal authSuccessAmount = bigFun.apply(DownstreamAPIReportVO::getAuthSuccessAmount);
                BigDecimal importIncome = bigFun.apply(DownstreamAPIReportVO::getImportTrafficIncome);
                return new DownStreamReportCollectVO()
                        .setProductName(item.getKey())
                        .setApplyCount(threeItemCheckCount)
                        .setAuthingCount(longFun.applyAsLong(DownstreamAPIReportVO::getAuthingCount))
                        .setAuthSuccessCount(longFun.applyAsLong(DownstreamAPIReportVO::getThreeItemCheckPassCount))
                        .setAuthSuccessCount(authSuccessCount)
                        .setAuthSuccessAmount(authSuccessAmount)
                        .setAuthSuccessAmountAvg(authSuccessCount > 0 ? authSuccessAmount.divide(new BigDecimal(authSuccessCount), 2, RoundingMode.HALF_UP) : BigDecimal.ZERO)
                        .setClickGetCashAccountCount(longFun.applyAsLong(DownstreamAPIReportVO::getClickGetCashAccountCount))
                        .setFirstLoanAccountCount(longFun.applyAsLong(DownstreamAPIReportVO::getFirstLoanAccountCount))
                        .setFirstLoanAmount(bigFun.apply(DownstreamAPIReportVO::getFirstLoanAmount))
                        .setAuthSuccessRate(BigDecimal.valueOf(Arith.div(authSuccessCount, threeItemCheckCount, 4)).multiply(new BigDecimal(100)))
                        .setImportTrafficIncome(importIncome)
                        .setSingleIncome(threeItemCheckCount > 0 ? importIncome.divide(new BigDecimal(threeItemCheckCount), 2, RoundingMode.HALF_UP) : BigDecimal.ZERO);
            }).collect(Collectors.toList());

    Supplier<Stream<DownStreamReportCollectVO>> sup = voList::stream;
    // Long 求和
    ToLongFunction<DownStreamReportCollectVO> longFun = t -> sup.get().mapToLong(tlf).filter(Objects::nonNull).sum();
    // BigDecimal 求和
    Function<DownStreamReportCollectVO, BigDecimal> bigFun = fn -> sup.get().map(fn).filter(Objects::nonNull).reduce(BigDecimal.ZERO, BigDecimal::add);
    Long applyCount = longFun.applyAsLong(DownStreamReportCollectVO::getApplyCount);
    Long authSuccessCount = longFun.applyAsLong(DownStreamReportCollectVO::getAuthSuccessCount);
    BigDecimal authSuccessAmount = bigFun.apply(DownStreamReportCollectVO::getAuthSuccessAmount);
    BigDecimal importIncome = bigFun.apply(DownStreamReportCollectVO::getImportTrafficIncome);
    return Stream.concat(Stream.of(new DownStreamReportCollectVO()
            .setApplyCount(applyCount)
            .setAuthingCount(longFun.applyAsLong(DownStreamReportCollectVO::getAuthingCount))
            .setAuthSuccessCount(authSuccessCount)
            .setAuthSuccessAmount(authSuccessAmount)
            .setAuthSuccessAmountAvg(authSuccessCount > 0 ? authSuccessAmount.divide(new BigDecimal(authSuccessCount), 2, RoundingMode.HALF_UP) : BigDecimal.ZERO)
            .setClickGetCashAccountCount(longFun.applyAsLong(DownStreamReportCollectVO::getClickGetCashAccountCount))
            .setFirstLoanAccountCount(longFun.applyAsLong(DownStreamReportCollectVO::getFirstLoanAccountCount))
            .setFirstLoanAmount(bigFun.apply(DownStreamReportCollectVO::getFirstLoanAmount))
            .setAuthSuccessRate(BigDecimal.valueOf(Arith.div(authSuccessCount, applyCount, 4)).multiply(new BigDecimal(100)))
            .setImportTrafficIncome(importIncome)
            .setSingleIncome(applyCount > 0 ? importIncome.divide(new BigDecimal(applyCount), 2, RoundingMode.HALF_UP) : BigDecimal.ZERO)
    ), voList.stream()).collect(Collectors.toList());
}

请注意，由于我无法了解您的完整代码和相关环境，因此无法保证上述优化方法一定能提高性能。建议在进行任何优化之前，先进行性能测试和评估，以确保优化的有效性