iailab-module-model/iailab-module-model-biz/src/main/java/com/iailab/module/model/mdk/sample/PredictSampleDataConstructor.java

@@ -1,18 +1,19 @@
package com.iailab.module.model.mdk.sample;

import com.iailab.framework.common.util.object.ConvertUtils;
import com.iailab.module.data.api.point.DataPointApi;
import com.iailab.module.data.api.point.dto.ApiPointDTO;
import com.iailab.module.data.api.point.dto.ApiPointValueDTO;
import com.iailab.module.data.api.point.dto.ApiPointValueQueryDTO;
import com.iailab.module.model.mcs.pre.entity.MmItemOutputEntity;
import com.iailab.module.model.mcs.pre.service.MmItemOutputService;
import com.iailab.module.model.mcs.pre.service.MmItemResultService;
import com.iailab.module.model.mdk.factory.ItemEntityFactory;
import com.iailab.module.model.mcs.pre.service.MmItemTypeService;
import com.iailab.module.model.mdk.common.enums.ModelParamType;
import com.iailab.module.model.mdk.sample.dto.ColumnItem;
import com.iailab.module.model.mdk.sample.dto.ColumnItemPort;
import com.iailab.module.model.mdk.sample.dto.SampleData;
import com.iailab.module.model.mdk.sample.dto.SampleInfo;
import com.iailab.module.model.mdk.vo.DataValueVO;
import com.iailab.module.model.mdk.vo.MmItemOutputVO;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
@@ -20,6 +21,7 @@

import java.math.BigDecimal;
import java.util.*;
import java.util.stream.Collectors;

/**
 * 预测样本数据构造
@@ -36,7 +38,10 @@
    private MmItemResultService mmItemResultService;

    @Autowired
    private ItemEntityFactory itemEntityFactory;
    private MmItemTypeService mmItemTypeService;

    @Autowired
    private MmItemOutputService mmItemOutputService;

    /**
     * alter by zfc 2020.11.24 修改数据样本构造方案：sampleInfo中数据已按爪子进行分类，但爪内数据为无序的，
@@ -49,7 +54,6 @@
    public List<SampleData> prepareSampleData(SampleInfo sampleInfo) {
        List<SampleData> sampleDataList = new ArrayList<>();
        //对每个爪分别进行计算
        int deviationIndex = 0;
        for (ColumnItemPort entry : sampleInfo.getColumnInfo()) {
            //先依据爪内数据项的modelParamOrder进行排序——重写comparator匿名函数
            Collections.sort(entry.getColumnItemList(), new Comparator<ColumnItem>() {
@@ -67,31 +71,14 @@
                }
            }

            //找出对应的调整值
            BigDecimal[] deviationItem = null;
            if (sampleInfo.getDeviation() != null && sampleInfo.getDeviation().length > 0) {
                deviationItem = sampleInfo.getDeviation()[deviationIndex];
            }
            deviationIndex++;

            //对每一项依次进行数据查询,然后将查询出的值赋给matrix对应的位置
            for (int i = 0; i < entry.getColumnItemList().size(); i++) {
                try {
                    List<DataValueVO> dataEntityList = getData(entry.getColumnItemList().get(i));
                    //设置调整值
                    if (deviationItem != null && deviationItem.length > 0) {
                        logger.info("设置调整值, i = " + i);
                        if (deviationItem[i] != null && deviationItem[i].compareTo(BigDecimal.ZERO) != 0) {
                            for (int dataKey = 1; dataKey < dataEntityList.size(); dataKey++) {
                                DataValueVO item = dataEntityList.get(dataKey);
                                item.setDataValue(item.getDataValue() + deviationItem[i].doubleValue());
                            }
                        }
                    }
                    //补全数据
                    ColumnItem columnItem = entry.getColumnItemList().get(i);
//                    dataEntityList = super.completionData(matrix.length, dataEntityList, columnItem.startTime, columnItem.getEndTime(), columnItem.granularity);
                    dataEntityList = super.completionData(matrix.length, dataEntityList, columnItem.startTime, columnItem.endTime, columnItem.paramId,columnItem.getParamType());
                    dataEntityList = super.completionData(matrix.length, dataEntityList, columnItem.startTime, columnItem.endTime,
                            columnItem.paramId, columnItem.getParamType());

                    /** 如果数据取不满，把缺失的数据点放在后面 */
                    if (dataEntityList != null && dataEntityList.size() != 0) {
@@ -121,26 +108,29 @@
    private List<DataValueVO> getData(ColumnItem columnItem) throws Exception {
        List<DataValueVO> dataList = new ArrayList<>();
        String paramType = columnItem.getParamType();
        switch (paramType) {
            case "DATAPOINT":
        switch (ModelParamType.getEumByCode(paramType)) {
            case DATAPOINT:
                ApiPointDTO point = dataPointApi.getInfoById(columnItem.getParamId());
                ApiPointValueQueryDTO queryDto = new ApiPointValueQueryDTO();
                queryDto.setPointNo(point.getPointNo());
                queryDto.setStart(columnItem.getStartTime());
                queryDto.setEnd(columnItem.getEndTime());
                List<ApiPointValueDTO> pointValueList = dataPointApi.queryPointHistoryValue(queryDto);
                dataList = ConvertUtils.sourceToTarget(pointValueList, DataValueVO.class);
                dataList = pointValueList.stream().map(t -> {
                    DataValueVO vo = new DataValueVO();
                    vo.setDataTime(t.getT());
                    vo.setDataValue(t.getV());
                    return vo;
                }).collect(Collectors.toList());
                break;
            case "PREDICTITEM":
                MmItemOutputVO outPut = itemEntityFactory.getItemOutPutById(columnItem.getId());
            case PREDICTITEM:
                MmItemOutputEntity outPut = mmItemOutputService.getOutPutById(columnItem.getId());
                dataList = mmItemResultService.getPredictValue(outPut.getId(),
                        columnItem.getStartTime(), columnItem.getEndTime());
                if (dataList == null) {
                    throw new Exception("没有预测值");
                }
                break;


            default:
                break;
        }