json2excel/cmd/main/main.go

package main

import (
	"encoding/json"
	"flag"
	"fmt"
	"log"
	"os"
	"reflect"
	"strings"
	"time"
	"unicode/utf8"

	"github.com/iancoleman/orderedmap"
	"github.com/xuri/excelize/v2"
)

// Initial concept from https://stackoverflow.com/q/68621039

type Config struct {
	keyOrder       map[int]string
	parentOverride string
}

var config Config

func main() {
	//jsonFile := "test.json"
	parentNode := "input"
	configNode := "config"
	//worksheetName := "Sheet2"
	//outputFilename := "test.xlsx"

	// Command line arguments
	var inputJson string
	var worksheetName string
	var outputFilename string
	var boldTopRow bool
	var freezeTopRow bool
	var autoFilter bool
	var autoWidth bool
	var overwriteFile bool

	// Process command line arguments
	flag.StringVar(&inputJson, "inputJson", "./input.json", "Full path to input json data file")
	flag.StringVar(&outputFilename, "outputFilename", "./output.xlsx", "Filename for excel worksheet output")
	flag.StringVar(&worksheetName, "worksheetName", "Sheet1", "Label to set on worksheet")
	flag.BoolVar(&boldTopRow, "bold-toprow", true, "Sets the top row of the worksheet to bold")
	flag.BoolVar(&freezeTopRow, "freeze-toprow", true, "Freezes the first row of the Excel worksheet")
	flag.BoolVar(&autoFilter, "autofilter", true, "Sets the auto filter on the first row")
	flag.BoolVar(&autoWidth, "autowidth", true, "Automatically set the column width to fit contents")
	flag.BoolVar(&overwriteFile, "overwrite", false, "Set to true to overwrite existing file rather than modifying in place")
	flag.Parse()

	var xlsx *excelize.File
	var s []byte

	//var sheetIndex int
	var err error

	var cell string
	var row, column int

	// Truncate worksheetName to the maximum 31 characters
	worksheetName = TruncateString(worksheetName, 31)

	// Read the json input file
	if fileExists(inputJson) {
		s, err = os.ReadFile(inputJson)
		if err != nil {
			panic(err)
		}
	} else {
		fmt.Printf("Input JSON file '%s' does not exist.\n", inputJson)
		os.Exit(1)
	}

	// Unmarshal the json into an orderedmap to preserve the ordering of json structure
	o := orderedmap.New()
	err = json.Unmarshal([]byte(s), &o)
	if err != nil {
		error := fmt.Sprintf("JSON Unmarshal error %s\n", err)
		panic(error)
	}

	// Assume that our content is within the first top-level key
	topLevel := o.Keys()
	fmt.Printf("Found %d top-level keys in json data\n", len(topLevel))
	for i, key := range topLevel {
		fmt.Printf("[%d] : %s\n", i, key)
	}

	// Check for config embedded in json
	if strings.EqualFold(topLevel[0], configNode) && len(topLevel) > 1 {
		fmt.Printf("Found configNode as toplevel json key, setting parentNode as '%s'\n", topLevel[1])
		parentNode = topLevel[1]

		config.keyOrder = make(map[int]string)

		// Get a reference to the top level node we specified earlier
		configInterface, ok := o.Get(configNode)
		if !ok {
			fmt.Printf("Missing key for multitype array when reading embedded config")
		}

		// Get an interface that we can work with to access the sub elements
		// This doesn't seem necessary for some reason - maybe because there's only one level of depth to the configNode
		//configSlice := configInterface
		//fmt.Printf("%v\n", configSlice)

		// Get the keys for the first element so we know what config options have been specified
		configMap := configInterface.(orderedmap.OrderedMap)
		configKeys := configMap.Keys()

		// Parse each key into our config struct
		for _, key := range configKeys {
			if strings.EqualFold(key, "key-order") {
				fmt.Printf("Found config element for key-order\n")
				e, _ := configMap.Get(key)
				for i, e := range strings.Split(e.(string), ",") {
					config.keyOrder[i] = e
				}
				fmt.Printf("Column order is now : '%v'\n", config.keyOrder)
			} else if strings.EqualFold(key, "overwrite-file") {
				fmt.Printf("Found config element for overwriting output file\n")
				e, _ := configMap.Get(key)
				overwriteFile = e.(bool)
			} else if strings.EqualFold(key, "parent-node") {
				fmt.Printf("Found config element for forcing parent key for spreadsheet data\n")
				e, _ := configMap.Get(key)
				config.parentOverride = e.(string)
			}
		}
	} else if strings.EqualFold(topLevel[0], configNode) {
		error := "Only found config in first level of json keys"
		panic(error)
	} else {
		fmt.Printf("Detected toplevel json key as: '%s'\n", topLevel[0])
		parentNode = topLevel[0]
	}

	if config.parentOverride != "" {
		fmt.Printf("Overriding parent node to '%s'\n", config.parentOverride)
		parentNode = config.parentOverride
	}

	// Get a reference to the top level node we specified earlier
	vislice, ok := o.Get(parentNode)
	if !ok {
		fmt.Printf("Missing key for multitype array")
	}

	// Get an interface that we can work with to access the sub elements
	vslice := vislice.([]interface{})

	// Check that the first element is what we expected
	if _, ok := vslice[0].(orderedmap.OrderedMap); !ok {
		error := fmt.Sprintf("Type of first vslice element is not an ordered map. It appears to be '%v'\n", reflect.TypeOf(vslice[0]))
		panic(error)
	}
	// Get the keys for the first element so we know what the column names will be
	columnMap := vslice[0].(orderedmap.OrderedMap)
	//fmt.Printf("First vslice element is an ordered map\n")
	columnNames := columnMap.Keys()

	// Check if xlsx file exists already, and if it does then open and append data
	if fileExists(outputFilename) {
		if overwriteFile {
			// Delete existing file
			fmt.Printf("Overwrite flag is set, removing existing file '%s'\n", outputFilename)
			err = os.Remove(outputFilename)
			if err != nil {
				panic(err)
			}
			// Create new file
			xlsx = createWorkbook(worksheetName, outputFilename)

		} else {
			xlsx = modifyWorkbook(worksheetName, outputFilename)
		}

	} else {
		xlsx = createWorkbook(worksheetName, outputFilename)
	}

	// Run code to add column names to the first row of the workbook
	createHeadingRow(xlsx, worksheetName, columnNames)

	// Set the style for the header values
	// Just handling bold for now but we can do other styles too as per https://xuri.me/excelize/en/style.html#NewStyle
	fmt.Printf("Bolding top row : %v\n", boldTopRow)
	headerStyle, err2 := xlsx.NewStyle(&excelize.Style{
		Font: &excelize.Font{
			Bold: true,
		},
	})
	if err2 != nil {
		fmt.Printf("Error generating header style : '%s'\n", err2)
	}

	// Set the style
	if boldTopRow {
		err = xlsx.SetRowStyle(worksheetName, 1, 1, headerStyle)
		if err != nil {
			fmt.Printf("Error setting header style : '%s'\n", err)
		}
	}

	// Freeze top row if requested, see https://xuri.me/excelize/en/utils.html#SetPanes
	if freezeTopRow {
		err = xlsx.SetPanes(worksheetName, &excelize.Panes{
			Freeze:      true,
			Split:       false,
			XSplit:      0,
			YSplit:      1,
			TopLeftCell: "A2",
			ActivePane:  "bottomLeft",
			Panes: []excelize.PaneOptions{
				{SQRef: "A2", ActiveCell: "A2", Pane: "bottomLeft"},
			},
		})
		if err != nil {
			fmt.Printf("Error freezing top row : '%s'\n", err)
		}
	}

	// Handle autofilter
	if autoFilter {
		cell, _ = excelize.CoordinatesToCellName(len(columnNames), 1)
		filterRange := "A1:" + cell
		fmt.Printf("Setting autofilter to range '%s'\n", filterRange)
		// As per docs any filters applied need to be manually processed by us (eg hiding rows with blanks)
		err = xlsx.AutoFilter(worksheetName, filterRange, nil)
		if err != nil {
			fmt.Printf("Error setting autofilter : '%s'\n", err)
		}
	}

	// Now process the remaining data in our json input
	// Set starting row for data
	row = 2

	// Iterate the whole slice to get the data and add to the worksheet
	fmt.Printf("Adding %d rows of data to spreadsheet.\n", len(vslice))
	for i, v := range vslice {
		// Print the contents each slice
		//fmt.Printf("'%d' : '%v'\n", i, v)

		//fmt.Printf("Slice element %d\n", i)
		_ = i

		// Each iteration should start back at column 1
		column = 1

		// Get the key-value pairs contained in this slice
		vmap := v.(orderedmap.OrderedMap)
		k := vmap.Keys()

		if len(config.keyOrder) > 0 {
			// If we have a specified order for the values then use that

			for j := 0; j < len(config.keyOrder); j++ {
				cell, _ = excelize.CoordinatesToCellName(column, row)

				e, _ := vmap.Get(config.keyOrder[j])
				//fmt.Printf("Setting cell %s to value %v\n", cell, e)
				xlsx.SetCellValue(worksheetName, cell, e)

				// Move to the next column
				column++
			}
		} else {
			// Otherwise use the order the json was in
			for j := range k {
				cell, _ = excelize.CoordinatesToCellName(column, row)

				e, _ := vmap.Get(k[j])
				//fmt.Printf("Setting cell %s to value %v\n", cell, e)
				xlsx.SetCellValue(worksheetName, cell, e)

				// Move to the next column
				column++
			}
		}

		//fmt.Printf("k: %v\n", k)

		// Move to next row
		row++
	}

	// Perform any post processing now that the data exists
	if autoWidth {
		err = SetColAutoWidth(xlsx, worksheetName)
		if err != nil {
			fmt.Printf("Error setting auto width : '%s'\n", err)
		}
	}

	// Close off the file
	if err := xlsx.SaveAs(outputFilename); err != nil {
		log.Fatal(err)
		return
	}

	fmt.Printf("Excel file '%s' generated sucessfully.\n", outputFilename)

	/*
		//Test https://gitlab.com/c0b/go-ordered-json
		var om *ordered.OrderedMap = ordered.NewOrderedMap()
		err = json.Unmarshal([]byte(file), om)
		if err != nil {
			fmt.Printf("error: '%s'\n", err)
		}

		iter := om.EntriesIter()
		for {
			pair, ok := iter()
			if !ok {
				break
			}
			fmt.Printf("%-12s: %v\n", pair.Key, pair.Value)
		}
	*/
}

func fileExists(filename string) bool {
	info, err := os.Stat(filename)
	if os.IsNotExist(err) {
		return false
	}
	return !info.IsDir()
}

// From https://dev.to/takakd/go-safe-truncate-string-9h0#comment-1hp14
func TruncateString(str string, length int) string {
	if length <= 0 {
		return ""
	}

	if utf8.RuneCountInString(str) < length {
		return str
	}

	return string([]rune(str)[:length])
}

func createWorkbook(worksheetName string, outputFilename string) *excelize.File {
	fmt.Printf("Creating output spreadsheet '%s'\n", outputFilename)
	var err error
	// Create the file
	xlsx := excelize.NewFile()

	// Rename the default Sheet1 to this worksheet name
	if worksheetName != "Sheet1" {
		fmt.Printf("Renaming default worksheet to '%s'\n", worksheetName)
		err = xlsx.SetSheetName("Sheet1", worksheetName)
		if err != nil {
			fmt.Printf("Error setting sheet name to '%s': %s\n", worksheetName, err)
		}
	}

	// Set the document properties
	err = xlsx.SetDocProps(&excelize.DocProperties{
		Creator: "json2excel",
		Created: time.Now().Format(time.RFC3339),
	})
	if err != nil {
		fmt.Printf("Error setting document properties: %s\n", err)
	}
	return xlsx
}

func modifyWorkbook(worksheetName string, outputFilename string) *excelize.File {
	fmt.Printf("Output spreadsheet '%s' already exists.\n", outputFilename)
	xlsx, err := excelize.OpenFile(outputFilename)
	if err != nil {
		fmt.Println(err)
		os.Exit(1)
	}

	// Since we have an existing workbook, check if the sheet we want to write to already exists
	sheetFound := false
	for index, name := range xlsx.GetSheetMap() {
		if name == worksheetName {
			fmt.Printf("Found worksheet '%s' at index '%d'\n", worksheetName, index)
			sheetFound = true
		}
	}
	if !sheetFound {
		// Create the sheet
		fmt.Printf("Creating worksheet '%s'\n", worksheetName)
		sheetIndex, err := xlsx.NewSheet(worksheetName)
		if err != nil {
			fmt.Printf("Error creating worksheet '%s' : %s\n", worksheetName, err)
		}
		// Set active worksheet
		fmt.Printf("Setting active sheet to index %d", sheetIndex)
		xlsx.SetActiveSheet(sheetIndex)
	}
	return xlsx
}

func createHeadingRow(xlsx *excelize.File, worksheetName string, columnNames []string) {
	var cell string
	row := 1
	column := 1

	if len(config.keyOrder) > 0 {
		fmt.Printf("Creating excel workbook with heading order as per config key-order\n")
		// Check that the number of specified columns matches input data
		if len(config.keyOrder) != len(columnNames) {
			error := fmt.Sprintf("Column order specified in json key-order but mismatch found in json data. %d specifed columns does not match %d found columns.", len(config.keyOrder), len(columnNames))
			panic(error)
		}

		// Iterate the map and add the columns as per that order
		for i := 0; i < len(config.keyOrder); i++ {
			cell, _ = excelize.CoordinatesToCellName(column, row)
			fmt.Printf("Setting cell %s to value %s at index %d\n", cell, config.keyOrder[i], i)
			xlsx.SetCellValue(worksheetName, cell, config.keyOrder[i])
			column++
		}
	} else {
		fmt.Printf("Creating excel workbook with following headings : '%v'\n", columnNames)
		// Add the header row
		for i := 0; i < len(columnNames); i++ {
			cell, _ = excelize.CoordinatesToCellName(column, row)
			fmt.Printf("Setting cell %s to value %s\n", cell, columnNames[i])
			xlsx.SetCellValue(worksheetName, cell, columnNames[i])
			//xlsx.SetCellStyle(worksheetName, cell, cell, headerStyle)
			column++
		}
	}

}

// Taken from https://github.com/qax-os/excelize/issues/92#issuecomment-821578446
func SetColAutoWidth(xlsx *excelize.File, sheetName string) error {
	// Autofit all columns according to their text content
	cols, err := xlsx.GetCols(sheetName)
	if err != nil {
		return err
	}
	for idx, col := range cols {
		largestWidth := 0
		for _, rowCell := range col {
			cellWidth := utf8.RuneCountInString(rowCell) + 2 // + 2 for margin
			if cellWidth > largestWidth {
				largestWidth = cellWidth
			}
		}
		fmt.Printf("SetColAutoWidth calculated largest width for column index '%d' is '%d'\n", idx, largestWidth)
		name, err := excelize.ColumnNumberToName(idx + 1)
		if err != nil {
			return err
		}
		xlsx.SetColWidth(sheetName, name, name, float64(largestWidth))
	}
	// No errors at this point
	return nil
}

func TestUnmarshalJSON() {
	s := `{
  "number": 4,
  "string": "x",
  "z": 1,
  "a": "should not break with unclosed { character in value",
  "b": 3,
  "slice": [
    "1",
    1
  ],
  "orderedmap": {
    "e": 1,
    "a { nested key with brace": "with a }}}} }} {{{ brace value",
	"after": {
		"link": "test {{{ with even deeper nested braces }"
	}
  },
  "test\"ing": 9,
  "after": 1,
  "multitype_array": [
    "test",
	1,
	{ "map": "obj", "it" : 5, ":colon in key": "colon: in value" },
	[{"inner": "map"}]
  ],
  "should not break with { character in key": 1
}`
	o := orderedmap.New()
	err := json.Unmarshal([]byte(s), &o)
	if err != nil {
		fmt.Printf("JSON Unmarshal error", err)
	}
	// Check the root keys
	expectedKeys := []string{
		"number",
		"string",
		"z",
		"a",
		"b",
		"slice",
		"orderedmap",
		"test\"ing",
		"after",
		"multitype_array",
		"should not break with { character in key",
	}
	k := o.Keys()
	for i := range k {
		if k[i] != expectedKeys[i] {
			fmt.Printf("Unmarshal root key order", i, k[i], "!=", expectedKeys[i])
		}
	}
	// Check nested maps are converted to orderedmaps
	// nested 1 level deep
	expectedKeys = []string{
		"e",
		"a { nested key with brace",
		"after",
	}
	vi, ok := o.Get("orderedmap")
	if !ok {
		fmt.Printf("Missing key for nested map 1 deep")
	}
	v := vi.(orderedmap.OrderedMap)
	k = v.Keys()
	for i := range k {
		if k[i] != expectedKeys[i] {
			fmt.Printf("Key order for nested map 1 deep ", i, k[i], "!=", expectedKeys[i])
		}
	}
	// nested 2 levels deep
	expectedKeys = []string{
		"link",
	}
	vi, ok = v.Get("after")
	if !ok {
		fmt.Printf("Missing key for nested map 2 deep")
	}
	v = vi.(orderedmap.OrderedMap)
	k = v.Keys()
	for i := range k {
		if k[i] != expectedKeys[i] {
			fmt.Printf("Key order for nested map 2 deep", i, k[i], "!=", expectedKeys[i])
		}
	}
	// multitype array
	expectedKeys = []string{
		"map",
		"it",
		":colon in key",
	}
	vislice, ok := o.Get("multitype_array")
	if !ok {
		fmt.Printf("Missing key for multitype array")
	}
	vslice := vislice.([]interface{})
	vmap := vslice[2].(orderedmap.OrderedMap)
	k = vmap.Keys()
	for i := range k {
		if k[i] != expectedKeys[i] {
			fmt.Printf("Key order for nested map 2 deep", i, k[i], "!=", expectedKeys[i])
		}
	}
	// nested map 3 deep
	vislice, _ = o.Get("multitype_array")
	vslice = vislice.([]interface{})
	expectedKeys = []string{"inner"}
	vinnerslice := vslice[3].([]interface{})
	vinnermap := vinnerslice[0].(orderedmap.OrderedMap)
	k = vinnermap.Keys()
	for i := range k {
		if k[i] != expectedKeys[i] {
			fmt.Printf("Key order for nested map 3 deep", i, k[i], "!=", expectedKeys[i])
		}
	}
}