package com.yami.trading.huobi.data.job;
|
|
|
import com.yami.trading.bean.data.domain.Realtime;
|
import com.yami.trading.bean.item.domain.Item;
|
import com.yami.trading.common.util.Arith;
|
import com.yami.trading.huobi.data.AdjustmentValueCache;
|
import com.yami.trading.huobi.data.DataCache;
|
import com.yami.trading.huobi.data.internal.DataDBService;
|
import com.yami.trading.huobi.data.model.AdjustmentValue;
|
import com.yami.trading.huobi.hobi.HobiDataService;
|
import com.yami.trading.service.item.ItemService;
|
import com.yami.trading.service.syspara.SysparaService;
|
import org.slf4j.Logger;
|
import org.slf4j.LoggerFactory;
|
import org.springframework.beans.factory.annotation.Autowired;
|
|
import java.math.BigDecimal;
|
import java.math.RoundingMode;
|
import java.util.ArrayList;
|
import java.util.List;
|
import java.util.Random;
|
|
|
public abstract class AbstractGetDataJob implements Runnable {
|
public static volatile boolean first = true;
|
protected static Logger logger = LoggerFactory.getLogger(StockGetDataJob.class);
|
/**
|
* 数据接口调用间隔时长(毫秒)
|
*/
|
protected int interval;
|
@Autowired
|
protected SysparaService sysparaService;
|
@Autowired
|
protected DataDBService dataDBService;
|
@Autowired
|
protected HobiDataService hobiDataService;
|
@Autowired
|
protected ItemService itemService;
|
|
public void start() {
|
new Thread(this, getName()).start();
|
}
|
|
public abstract void run();
|
|
public abstract String getName();
|
|
// 在类中定义静态Random实例
|
private static final Random random = new Random();
|
|
public abstract void realtimeHandle(String symbols);
|
|
public void handleRealTimeList(List<Realtime> realtimeList) {
|
for (Realtime realtime : realtimeList) {
|
|
try {
|
String symbol = realtime.getSymbol();
|
Integer decimal = itemService.getDecimal(symbol); // 虚拟币通常8位小数，需保留足够精度
|
Item item = this.itemService.findBySymbol(symbol);
|
BigDecimal currentValue = AdjustmentValueCache.getCurrentValue().get(symbol);
|
AdjustmentValue delayValue = AdjustmentValueCache.getDelayValue().get(symbol);
|
|
|
if (delayValue != null) {
|
|
// K线场景专属参数（可根据周期动态调整）
|
//int kLineCycle = getKLineCycle(symbol); // 获取K线周期（如1min=1, 5min=5）
|
//double baseFluctuation = kLineCycle <= 1 ? 0.05 : 0.08; // 短周期（1min）±5%，长周期（5min）±8%
|
double baseFluctuation = 0.1;
|
int trendConsistencyRate = 70; // 70%概率保持与前一个值的波动方向一致（模拟趋势）
|
int maxAdjacentFluctuation = 5; // 相邻值波动不超过前一个值的5%（避免视觉跳变）
|
|
|
Integer frequency = AdjustmentValueCache.getFrequency().get(symbol);
|
List<BigDecimal> preAllocationList = AdjustmentValueCache.getPreAllocationList().get(symbol);
|
Integer currentIndex = AdjustmentValueCache.getCurrentAllocationIndex().get(symbol);
|
// 新增：记录前一个值的波动方向（用于趋势一致性）
|
List<Boolean> upDownTrend = AdjustmentValueCache.getUpDownTrend().getOrDefault(symbol, new ArrayList<>());
|
|
if (frequency == null) {
|
frequency = (int) Arith.div(Arith.mul(delayValue.getSecond(), 1000.0D), this.interval);
|
AdjustmentValueCache.getFrequency().put(symbol, frequency);
|
|
if (frequency > 1) {
|
preAllocationList = new ArrayList<>(frequency);
|
BigDecimal totalValue = delayValue.getValue();
|
// 虚拟币精度高，中间计算保留decimal+4位（避免精度丢失）
|
BigDecimal average = totalValue.divide(new BigDecimal(frequency), decimal + 4, RoundingMode.HALF_UP);
|
BigDecimal sum = BigDecimal.ZERO;
|
|
// 优化1：最后k个值分散偏差（k=频率的2%，最少5个，确保最后一根K线无异常）
|
int lastK = Math.max(5, frequency / 50);
|
int normalCount = frequency - lastK;
|
|
// 优化2：前n-k个值——模拟真实行情波动（截断正态分布+趋势一致性）
|
for (int i = 0; i < normalCount; i++) {
|
BigDecimal randomValue;
|
if (i == 0) {
|
// 第一个值：在基础波动范围内随机
|
randomValue = generateTruncatedGaussianValue(average, baseFluctuation, decimal + 4);
|
} else {
|
// 非第一个值：70%概率保持与前一个值的波动方向一致
|
boolean lastUp = upDownTrend.get(i - 1);
|
boolean keepTrend = Math.random() * 100 <= trendConsistencyRate;
|
|
if (keepTrend) {
|
// 保持趋势：上涨则继续涨（或小跌），下跌则继续跌（或小涨）
|
if (lastUp) {
|
// 前一个上涨：本次波动范围 [0, +baseFluctuation*1.2]（允许小回调）
|
randomValue = generateDirectionalValue(average, 0, baseFluctuation * 1.2, decimal + 4);
|
} else {
|
// 前一个下跌：本次波动范围 [-baseFluctuation*1.2, 0]（允许小反弹）
|
randomValue = generateDirectionalValue(average, -baseFluctuation * 1.2, 0, decimal + 4);
|
}
|
} else {
|
// 反转趋势：正常基础波动
|
randomValue = generateTruncatedGaussianValue(average, baseFluctuation, decimal + 4);
|
}
|
|
// 优化3：强约束相邻值波动——不超过前一个值的5%
|
BigDecimal lastValue = preAllocationList.get(i - 1);
|
BigDecimal maxAdjacent = lastValue.multiply(new BigDecimal(maxAdjacentFluctuation / 100.0))
|
.setScale(decimal + 4, RoundingMode.HALF_UP);
|
BigDecimal adjacentDiff = randomValue.subtract(lastValue).abs();
|
if (adjacentDiff.compareTo(maxAdjacent) > 0) {
|
randomValue = lastValue.add(
|
adjacentDiff.divide(randomValue.subtract(lastValue), decimal + 4, RoundingMode.HALF_UP)
|
.multiply(maxAdjacent)
|
);
|
}
|
}
|
|
preAllocationList.add(randomValue);
|
sum = sum.add(randomValue);
|
// 记录当前值的波动方向（与平均值对比）
|
upDownTrend.add(randomValue.compareTo(average) > 0);
|
}
|
|
// 优化4：最后k个值——极小波动+偏差分散（确保最后一根K线平滑）
|
BigDecimal remaining = totalValue.subtract(sum);
|
BigDecimal avgLastK = remaining.divide(new BigDecimal(lastK), decimal + 4, RoundingMode.HALF_UP);
|
for (int i = 0; i < lastK - 1; i++) {
|
// 最后k-1个值：波动范围缩小到基础波动的1/3（±1.7%~2.7%）
|
BigDecimal smallFluctValue = generateTruncatedGaussianValue(avgLastK, baseFluctuation / 3, decimal + 4);
|
preAllocationList.add(smallFluctValue);
|
sum = sum.add(smallFluctValue);
|
upDownTrend.add(smallFluctValue.compareTo(avgLastK) > 0);
|
}
|
// 最后1个值：仅承担剩余微小偏差（波动≤基础波动的1/5）
|
BigDecimal finalValue = totalValue.subtract(sum);
|
BigDecimal maxFinalFluct = avgLastK.multiply(new BigDecimal(baseFluctuation / 5))
|
.setScale(decimal + 4, RoundingMode.HALF_UP);
|
if (finalValue.abs().compareTo(maxFinalFluct) > 0) {
|
finalValue = maxFinalFluct.multiply(finalValue.signum() == 1 ? BigDecimal.ONE : BigDecimal.ONE.negate());
|
// 若超出范围，反向微调前一个值（确保总和正确）
|
BigDecimal prevLastValue = preAllocationList.get(preAllocationList.size() - 1);
|
preAllocationList.set(preAllocationList.size() - 1, prevLastValue.add(totalValue.subtract(sum).subtract(finalValue)));
|
}
|
preAllocationList.add(finalValue);
|
upDownTrend.add(finalValue.compareTo(avgLastK) > 0);
|
|
// 缓存新增：波动方向列表（用于趋势一致性）
|
AdjustmentValueCache.getPreAllocationList().put(symbol, preAllocationList);
|
AdjustmentValueCache.getCurrentAllocationIndex().put(symbol, 0);
|
AdjustmentValueCache.getUpDownTrend().put(symbol, upDownTrend);
|
currentIndex = 0;
|
}
|
}
|
|
// 后续分配逻辑（不变，仅需在清理缓存时移除upDownTrend）
|
if (frequency <= 1) {
|
// 单次分配逻辑
|
if (currentValue == null) {
|
AdjustmentValueCache.getCurrentValue().put(symbol, delayValue.getValue().setScale(decimal, RoundingMode.HALF_UP));
|
} else {
|
AdjustmentValueCache.getCurrentValue().put(symbol,
|
currentValue.add(delayValue.getValue()).setScale(decimal, RoundingMode.HALF_UP));
|
}
|
if (!item.getAdjustmentValue().equals(AdjustmentValueCache.getCurrentValue().get(symbol))) {
|
item.setAdjustmentValue(AdjustmentValueCache.getCurrentValue().get(symbol));
|
itemService.saveOrUpdate(item);
|
}
|
cleanUpKLineCache(symbol); // K线场景专属缓存清理
|
} else {
|
// 按预分配列表取值
|
if (preAllocationList != null && currentIndex != null && currentIndex < preAllocationList.size()) {
|
BigDecimal currentValue_frequency = preAllocationList.get(currentIndex)
|
.setScale(decimal, RoundingMode.HALF_UP);
|
|
// 更新当前值（K线价格）
|
if (currentValue == null) {
|
AdjustmentValueCache.getCurrentValue().put(symbol, currentValue_frequency);
|
} else {
|
AdjustmentValueCache.getCurrentValue().put(symbol,
|
currentValue.add(currentValue_frequency).setScale(decimal, RoundingMode.HALF_UP));
|
}
|
|
// 更新延迟值和索引
|
delayValue.setValue(delayValue.getValue().subtract(currentValue_frequency)
|
.setScale(decimal, RoundingMode.HALF_UP));
|
delayValue.setSecond(Arith.sub(delayValue.getSecond(), Arith.div(this.interval, 1000.0D)));
|
AdjustmentValueCache.getDelayValue().put(symbol, delayValue);
|
|
int nextIndex = currentIndex + 1;
|
AdjustmentValueCache.getCurrentAllocationIndex().put(symbol, nextIndex);
|
|
// 分配完成，清理缓存
|
if (nextIndex >= frequency) {
|
cleanUpKLineCache(symbol);
|
}
|
|
// 保存K线数据更新
|
if (!item.getAdjustmentValue().equals(AdjustmentValueCache.getCurrentValue().get(symbol))) {
|
item.setAdjustmentValue(AdjustmentValueCache.getCurrentValue().get(symbol));
|
itemService.saveOrUpdate(item);
|
}
|
}
|
}
|
}
|
|
currentValue = AdjustmentValueCache.getCurrentValue().get(realtime.getSymbol());
|
|
if (currentValue != null && currentValue.compareTo(BigDecimal.ZERO) != 0) {
|
realtime.setClose(realtime.getClose().add(currentValue).setScale(decimal, RoundingMode.HALF_UP));
|
BigDecimal ask = realtime.getAsk();
|
if(ask!=null){
|
realtime.setAsk(ask.add(currentValue).setScale(decimal, RoundingMode.HALF_UP));
|
}
|
BigDecimal bid = realtime.getBid();
|
if(bid!=null){
|
realtime.setBid(bid.add(currentValue).setScale(decimal, RoundingMode.HALF_UP));
|
}
|
// realtime.setVolume(Arith.add(realtime.getVolume(), Arith.mul(Arith.div(currentValue, realtime.getClose()), realtime.getVolume())));
|
// realtime.setAmount(Arith.add(realtime.getAmount(), Arith.mul(Arith.div(currentValue, realtime.getClose()), realtime.getAmount())));
|
}
|
|
// 缓存中最新一条Realtime数据
|
Realtime realtimeLast = DataCache.getRealtime(symbol);
|
// 临时处理：正常10秒超过25%也不合理,丢弃.只有虚拟货币才这样执行
|
boolean checkRate = getName().contains("虚拟货币");
|
double rate = 0;
|
if (!checkRate) {
|
saveData(realtime, symbol, item);
|
}else{
|
if (realtimeLast != null) {
|
rate = Math.abs(Arith.sub(realtime.getClose().doubleValue(), realtimeLast.getClose().doubleValue()));
|
}
|
if (null == realtimeLast || Arith.div(rate, realtimeLast.getClose().doubleValue()) < 0.25D) {
|
saveData(realtime, symbol, item);
|
} else {
|
logger.error("当前{}价格{},上一次价格为{}过25%也不合理,丢弃Realtime,不入库", realtime.getSymbol(),realtimeLast.getClose(), realtime.getClose());
|
}
|
}
|
|
|
} catch (Exception e) {
|
logger.error("数据采集失败", e);
|
|
}
|
}
|
}
|
|
private void saveData(Realtime realtime, String symbol, Item item) {
|
Double high = DataCache.getRealtimeHigh().get(symbol);
|
Double low = DataCache.getRealtimeLow().get(symbol);
|
if (realtime.getTs().toString().length() <= 10) {
|
realtime.setTs(Long.valueOf(realtime.getTs() + "000"));
|
}
|
realtime.setName(item.getName());
|
if (high == null || realtime.getHigh().doubleValue() > high) {
|
DataCache.getRealtimeHigh().put(symbol, realtime.getHigh().doubleValue());
|
}
|
if ((low == null || realtime.getLow().doubleValue() < low) && realtime.getLow().doubleValue() > 0) {
|
DataCache.getRealtimeLow().put(symbol, realtime.getLow().doubleValue());
|
}
|
this.dataDBService.saveAsyn(realtime);
|
}
|
|
|
|
// ------------------------------ 工具方法 ------------------------------
|
/**
|
* 生成截断正态分布值（限制在平均值±fluctuation范围内，模拟真实小波动）
|
*/
|
private BigDecimal generateTruncatedGaussianValue(BigDecimal average, double fluctuation, int scale) {
|
double factor = nextGaussian() * 0.25; // 标准差0.25，99.7%概率在±0.75内（更集中）
|
factor = Math.max(-1.0, Math.min(1.0, factor)); // 截断极端值
|
BigDecimal fluctuationValue = average.multiply(new BigDecimal(factor * fluctuation))
|
.setScale(scale, RoundingMode.HALF_UP);
|
return average.add(fluctuationValue);
|
}
|
|
/**
|
* 生成指定方向的波动值（如[0, +0.1]表示只涨不跌）
|
*/
|
private BigDecimal generateDirectionalValue(BigDecimal average, double minFactor, double maxFactor, int scale) {
|
double factor = minFactor + Math.random() * (maxFactor - minFactor);
|
BigDecimal fluctuationValue = average.multiply(new BigDecimal(factor))
|
.setScale(scale, RoundingMode.HALF_UP);
|
return average.add(fluctuationValue);
|
}
|
|
/**
|
* K线场景专属缓存清理（含波动方向列表）
|
*/
|
private void cleanUpKLineCache(String symbol) {
|
AdjustmentValueCache.getDelayValue().remove(symbol);
|
AdjustmentValueCache.getFrequency().remove(symbol);
|
AdjustmentValueCache.getPreAllocationList().remove(symbol);
|
AdjustmentValueCache.getCurrentAllocationIndex().remove(symbol);
|
AdjustmentValueCache.getUpDownTrend().remove(symbol); // 新增：清理波动方向缓存
|
}
|
|
/**
|
* 获取K线周期（根据symbol或配置判断，示例返回1=1min，5=5min等）
|
*/
|
private int getKLineCycle(String symbol) {
|
// 实际场景可从配置或symbol后缀获取（如BTC-USDT-1MIN → 1）
|
return 1;
|
}
|
|
/**
|
* 正态分布随机数生成（均值0，标准差1）
|
*/
|
private static double nextGaussian() {
|
double u1 = Math.random();
|
double u2 = Math.random();
|
return Math.sqrt(-2 * Math.log(u1)) * Math.cos(2 * Math.PI * u2);
|
}
|
}
|
