json_filter/tool/gabs.go

// Package json_tool ...
//
// Description : json_tool ...
//
// Author : go_developer@163.com<张德满>
//
// Date : 2022/01/22 9:19 PM
package json_tool

import (
	"fmt"
	"io"
	"log"
	"os"
	"strings"

	"github.com/pkg/errors"
	"github.com/tidwall/gjson"
	"github.com/tidwall/sjson"
)

const (
	// virtualRoot 虚拟根节点
	virtualRoot = "__VIRTUAL_ROOT__"
)

// FilterOption 过滤选项
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 17:59 2023/9/1
type FilterOption struct {
	DebugModel  bool      // 调试模式
	LogInstance io.Writer // 日志实例
}

// FilterDataRule 参数过滤规则
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 2022/1/22 9:44 PM
type FilterDataRule struct {
	SourceKey    string      // 原始数据路径
	MapKey       string      // 提取后映射到的数据路径
	DefaultValue interface{} // 原始数据路径不存在时的默认值
	WithDefault  bool        // 是否使用默认值
}

// NewDataFilter 获取数据过滤方法实例
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 2022/1/22 9:50 PM
func NewDataFilter(source string, filterRule []*FilterDataRule, filterOption *FilterOption) *DataFilter {
	if nil == filterOption {
		filterOption = &FilterOption{}
	}
	if filterOption.DebugModel && nil == filterOption.LogInstance {
		filterOption.LogInstance = os.Stdout
		log.SetOutput(filterOption.LogInstance)
	}
	df := &DataFilter{
		source:        source,
		filterRule:    make([]*FilterDataRule, 0),
		rewriteResult: "{}",
		filterOption:  filterOption,
	}
	// 去除末尾的 .[]
	for _, item := range filterRule {
		item.MapKey = strings.TrimRight(item.MapKey, ".[]")
		item.SourceKey = strings.TrimRight(item.SourceKey, ".[]")
		mapIsArr := df.isArrPath(item.MapKey)
		if !mapIsArr {
			df.filterRule = append(df.filterRule, item)
			continue
		}
		if len(df.getArrPathList(item.SourceKey)) < len(df.getArrPathList(item.MapKey)) {
			panic("map result deep more than source data deep")
		}
		formatRes := make([]*FilterDataRule, 0)
		r, _ := df.unfoldSameDeepArr(item.SourceKey, item.MapKey, "", "")
		for _, itemUnfoldResult := range r {
			itemMapArr := strings.Split(itemUnfoldResult.MapKey, ".")
			if len(itemMapArr) != len(strings.Split(item.MapKey, ".")) {
				continue
			}
			formatRes = append(formatRes, itemUnfoldResult)
		}
		df.filterRule = append(df.filterRule, formatRes...)
	}
	return df
}

// DataFilter 数据过滤
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 2022/1/22 9:20 PM
type DataFilter struct {
	source        string
	filterRule    []*FilterDataRule
	rewriteResult string        // json数据重写结果
	filterOption  *FilterOption // 过滤选项
}

// Filter 数据过滤
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 2022/1/22 9:36 PM
func (df *DataFilter) Filter() (string, error) {
	df.logPrint(logLevelDebug, "source_data => ", df.source)
	var (
		err error
	)

	for _, itemRule := range df.filterRule {
		sourceIsArr := df.isArrPath(itemRule.SourceKey)
		mapIsArr := df.isArrPath(itemRule.MapKey)
		if !sourceIsArr && !mapIsArr {
			// 输入输出均不是数组, 最简单的场景
			if err = df.setKV(itemRule); nil != err {
				return "", err
			}
			continue
		}
		sourcePathArr := df.getArrPathList(itemRule.SourceKey)
		mapPathArr := df.getArrPathList(itemRule.MapKey)
		if len(mapPathArr) > len(sourcePathArr) {
			df.logPrint(logLevelFatal, "映射的层级深度大于数据源深度", "source_path => "+itemRule.SourceKey, "map_path => "+itemRule.MapKey)
			return "", fmt.Errorf("映射的层级深度大于数据源深度, source_path => %v map_path => %v", itemRule.SourceKey, itemRule.MapKey)
		}

		// 映射至非数组字段
		if !mapIsArr {
			if err = df.setValue(itemRule.MapKey, df.getDataAsSlice(df.source, df.getArrPathList(itemRule.SourceKey))); nil != err {
				df.logPrint(logLevelFatal, "映射非数组, 数据源为数组, 设置失败", "source_path => "+itemRule.SourceKey, "map_path => "+itemRule.MapKey, " err => "+err.Error())
				return "", fmt.Errorf("映射的层级深度大于数据源深度, source_path => %v map_path => %v", itemRule.SourceKey, itemRule.MapKey)
			}
			continue
		}

		if len(mapPathArr) == len(sourcePathArr) {
			// 数组深度一致
			continue
		}
	}
	df.logPrint(logLevelDebug, "过滤结果", df.rewriteResult)
	return df.rewriteResult, nil
}

// isArrPath 是否为数组路径
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 16:00 2023/9/1
func (df *DataFilter) isArrPath(path string) bool {
	return strings.Contains(path, "[]")
}

// setSameDeepArr 展开同深度数组的
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 12:19 2023/9/2
func (df *DataFilter) unfoldSameDeepArr(sourcePath string, mapPath string, sourceRootPath string, mapRootPath string) ([]*FilterDataRule, error) {
	df.logPrint(logLevelDebug, " 同深数组展开", sourceRootPath, mapRootPath)
	result := make([]*FilterDataRule, 0)
	if len(sourcePath) == 0 {
		return result, nil
	}
	sourcePathArr := df.getArrPathList(sourcePath)
	mapPathArr := df.getArrPathList(mapPath)
	if 1 == len(mapPathArr) {
		sourceKey := sourceRootPath + "." + sourcePathArr[0]
		if len(sourcePathArr[0:]) > 0 {
			sourceKey = sourceRootPath + "." + strings.Join(sourcePathArr[0:], ".[].")
		}
		result = append(result, &FilterDataRule{
			SourceKey:    sourceKey,
			MapKey:       mapRootPath + "." + mapPathArr[0],
			DefaultValue: nil,
			WithDefault:  false,
		})
		return result, nil
	}
	// 数组 展开
	for idx := 0; idx < len(mapPathArr); idx++ {
		if len(sourceRootPath) > 0 {
			sourceRootPath = fmt.Sprintf("%v.%v", sourceRootPath, sourcePathArr[idx])
		} else {
			sourceRootPath = sourcePathArr[idx]
		}
		if len(mapRootPath) > 0 {
			mapRootPath = fmt.Sprintf("%v.%v", mapRootPath, mapPathArr[idx])
		} else {
			mapRootPath = mapPathArr[idx]
		}
		valList := gjson.Get(df.source, sourceRootPath)
		if valList.Value() == nil {
			continue
		}
		for i := 0; i < len(valList.Array()); i++ {
			result = append(result, &FilterDataRule{
				SourceKey:    sourceRootPath,
				MapKey:       mapRootPath,
				DefaultValue: nil,
				WithDefault:  false,
			})
			r, e := df.unfoldSameDeepArr(
				strings.Join(sourcePathArr[idx+1:], "[]"),
				strings.Join(mapPathArr[idx+1:], "[]"),
				fmt.Sprintf("%v.%v", sourceRootPath, i),
				fmt.Sprintf("%v.%v", mapRootPath, i),
			)
			if nil != e {
				return nil, e
			}
			result = append(result, r...)
		}
	}
	return result, nil
}

// a.[].b.[].c.[].d
// g
// getDataAsSlice 抽取制定深度，生成list
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 12:17 2023/9/2
func (df *DataFilter) getDataAsSlice(sourceData string, pathList []string) []interface{} {
	//fmt.Println(sourceData, pathList)
	result := make([]interface{}, 0)
	if len(pathList) == 0 {
		return result
	}
	if len(pathList) == 1 {
		return []interface{}{gjson.Get(sourceData, pathList[0]).Value()}
	}
	for idx, itemPath := range pathList {
		if len(pathList)-1 == idx {
			val := gjson.Get(sourceData, itemPath).Value()
			if nil == val {
				return result
			}
			result = append(result, val)
			return result
		}

		currentPathVal := gjson.Get(sourceData, itemPath).Array()
		for _, sonItem := range currentPathVal {
			result = append(result, df.getDataAsSlice(sonItem.String(), pathList[idx:])...)
		}
	}
	return result
}

// setValue 设置值
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 22:04 2023/9/1
func (df *DataFilter) setValue(path string, value interface{}) error {
	var (
		err error
	)

	df.rewriteResult, err = sjson.Set(df.rewriteResult, path, value)
	return err
}

// setKV 设置相关值
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 16:03 2023/9/1
func (df *DataFilter) setKV(rule *FilterDataRule) error {
	var (
		err error
	)
	sourceValue := gjson.Get(df.source, rule.SourceKey)
	if sourceValue.Exists() {
		// 原始数据存在对应路径
		df.rewriteResult, err = sjson.Set(df.rewriteResult, rule.MapKey, sourceValue.Value())
		return err
	}
	if !rule.WithDefault {
		// 路径不存在, 且禁用默认值
		return errors.New(rule.SourceKey + " : source path not found, and default value is forbidden")
	}
	// 使用默认值填充
	df.rewriteResult, err = sjson.Set(df.rewriteResult, rule.MapKey, rule.DefaultValue)
	return err
}

// getArrPathList 获取路径列表
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 21:32 2023/9/1
func (df *DataFilter) getArrPathList(inputPath string) []string {
	pathArr := strings.Split(inputPath, "[]")
	arr := make([]string, 0)
	for _, item := range pathArr {
		arr = append(arr, strings.Trim(item, "."))
	}
	return arr
}

// logPrint 打印日志
//
// Author : go_developer@163.com<白茶清欢>
//
// Date : 18:00 2023/9/1
func (df *DataFilter) logPrint(level string, msg string, logAttr ...interface{}) {
	if !df.filterOption.DebugModel {
		// 未开启调试模式
		return
	}
	logData := append([]interface{}{level, msg}, logAttr...)
	log.Println(logData...)
}

// 日志等级定义
const (
	logLevelPanic = "PANIC"
	logLevelFatal = "FATAL"
	logLevelWarn  = "WARN"
	logLevelInfo  = "INFO"
	logLevelDebug = "DEBUG"
)