.gitignore

@@ -1,7 +1,11 @@
 .idea
 .vscode
 .DS_Store
-*.json
+*data.json
+*mapping.json
+*setting.json
+*output.json
+*test.json
 *.txt
 dist
 xtest

go.mod

@@ -5,14 +5,14 @@ go 1.18
 require (
 	github.com/elastic/go-elasticsearch/v6 v6.8.10
 	github.com/elastic/go-elasticsearch/v7 v7.17.10
-	github.com/fatih/color v1.17.0
-	github.com/go-resty/resty/v2 v2.16.5
 	github.com/jedib0t/go-pretty/v6 v6.6.4
+	github.com/loveuer/nf v0.2.12
 	github.com/samber/lo v1.39.0
 	github.com/spf13/cobra v1.8.1
 )
 
 require (
+	github.com/fatih/color v1.17.0 // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
 	github.com/mattn/go-colorable v0.1.13 // indirect
 	github.com/mattn/go-isatty v0.0.20 // indirect
@@ -20,6 +20,5 @@ require (
 	github.com/rivo/uniseg v0.2.0 // indirect
 	github.com/spf13/pflag v1.0.5 // indirect
 	golang.org/x/exp v0.0.0-20220303212507-bbda1eaf7a17 // indirect
-	golang.org/x/net v0.33.0 // indirect
+	golang.org/x/sys v0.20.0 // indirect
-	golang.org/x/sys v0.28.0 // indirect
 )

go.sum

@@ -6,12 +6,12 @@ github.com/elastic/go-elasticsearch/v7 v7.17.10 h1:TCQ8i4PmIJuBunvBS6bwT2ybzVFxx
 github.com/elastic/go-elasticsearch/v7 v7.17.10/go.mod h1:OJ4wdbtDNk5g503kvlHLyErCgQwwzmDtaFC4XyOxXA4=
 github.com/fatih/color v1.17.0 h1:GlRw1BRJxkpqUCBKzKOw098ed57fEsKeNjpTe3cSjK4=
 github.com/fatih/color v1.17.0/go.mod h1:YZ7TlrGPkiz6ku9fK3TLD/pl3CpsiFyu8N92HLgmosI=
-github.com/go-resty/resty/v2 v2.16.5 h1:hBKqmWrr7uRc3euHVqmh1HTHcKn99Smr7o5spptdhTM=
-github.com/go-resty/resty/v2 v2.16.5/go.mod h1:hkJtXbA2iKHzJheXYvQ8snQES5ZLGKMwQ07xAwp/fiA=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
 github.com/jedib0t/go-pretty/v6 v6.6.4 h1:B51RjA+Sytv0C0Je7PHGDXZBF2JpS5dZEWWRueBLP6U=
 github.com/jedib0t/go-pretty/v6 v6.6.4/go.mod h1:zbn98qrYlh95FIhwwsbIip0LYpwSG8SUOScs+v9/t0E=
+github.com/loveuer/nf v0.2.12 h1:1Og+ORHsOWKFmy9kKJhjvXDkdbaurH82HjIxuGA3nNM=
+github.com/loveuer/nf v0.2.12/go.mod h1:M6reF17/kJBis30H4DxR5hrtgo/oJL4AV4cBe4HzJLw=
 github.com/mattn/go-colorable v0.1.13 h1:fFA4WZxdEF4tXPZVKMLwD8oUnCTTo08duU7wxecdEvA=
 github.com/mattn/go-colorable v0.1.13/go.mod h1:7S9/ev0klgBDR4GtXTXX8a3vIGJpMovkB8vQcUbaXHg=
 github.com/mattn/go-isatty v0.0.16/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
@@ -32,13 +32,10 @@ github.com/spf13/pflag v1.0.5/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An
 github.com/stretchr/testify v1.8.4 h1:CcVxjf3Q8PM0mHUKJCdn+eZZtm5yQwehR5yeSVQQcUk=
 golang.org/x/exp v0.0.0-20220303212507-bbda1eaf7a17 h1:3MTrJm4PyNL9NBqvYDSj3DHl46qQakyfqfWo4jgfaEM=
 golang.org/x/exp v0.0.0-20220303212507-bbda1eaf7a17/go.mod h1:lgLbSvA5ygNOMpwM/9anMpWVlVJ7Z+cHWq/eFuinpGE=
-golang.org/x/net v0.33.0 h1:74SYHlV8BIgHIFC/LrYkOGIwL19eTYXQ5wc6TBuO36I=
-golang.org/x/net v0.33.0/go.mod h1:HXLR5J+9DxmrqMwG9qjGCxZ+zKXxBru04zlTvWlWuN4=
 golang.org/x/sys v0.0.0-20220811171246-fbc7d0a398ab/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.28.0 h1:Fksou7UEQUWlKvIdsqzJmUmCX3cZuD2+P3XyyzwMhlA=
+golang.org/x/sys v0.20.0 h1:Od9JTbYCk261bKm4M/mw7AklTlFYIa0bIp9BgSm1S8Y=
-golang.org/x/sys v0.28.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/sys v0.20.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
-golang.org/x/time v0.6.0 h1:eTDhh4ZXt5Qf0augr54TN6suAUudPcawVZeIAPU7D4U=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

internal/cmd/cmd.go

@@ -2,59 +2,76 @@ package cmd
 
 import (
 	"context"
+	"fmt"
+	"os"
+
+	"github.com/loveuer/nf/nft/log"
+
 	"github.com/loveuer/esgo2dump/internal/opt"
+	"github.com/loveuer/esgo2dump/internal/tool"
 	"github.com/spf13/cobra"
 )
 
-const (
+var (
-	example = `
+	rootCommand = &cobra.Command{
-esgo2dump -i https://<user>:<password>@<es_node1_host>:<es_node1_port>,<es_node2_host>:<es_node2_port>/some_index?ping=false&sniff=false -o ./data.json
+		Use:           "esgo2dump",
+		Short:         "esgo2dump is alternative to elasticdump",
+		SilenceUsage:  true,
+		SilenceErrors: true,
+		RunE:          run,
+		PersistentPreRun: func(cmd *cobra.Command, args []string) {
+			if opt.Cfg.Debug {
+				log.SetLogLevel(log.LogLevelDebug)
+			}
 
+			if opt.Cfg.Args.Version {
+				fmt.Printf("esgo2dump version: %s\n", opt.Version)
+				os.Exit(0)
+			}
+
+			if opt.Cfg.Debug {
+				tool.TablePrinter(opt.Cfg)
+			}
+		},
+		Example: `
 esgo2dump --input=http://127.0.0.1:9200/some_index --output=./data.json
 
 esgo2dump --input=http://127.0.0.1:9200/some_index --output=http://192.168.1.1:9200/some_index --limit=5000
 
+esgo2dump --input=http://127.0.0.1:9200/some_index --i-version 6 --output=./data.json
+
+esgo2dump --output=http://127.0.0.1:9200/some_index --o-version 6 --input=./data.json
+
 esgo2dump --input=https://username:password@127.0.0.1:9200/some_index --output=./data.json
 
 esgo2dump --input=http://127.0.0.1:9200/some_index --source='id;name;age;address' --output=./data.json
 
 esgo2dump --input=http://127.0.0.1:9200/some_index --output=./data.json --query='{"match": {"name": "some_name"}}'
 
-esgo2dump --input=http://127.0.0.1:9200/some_index --output=./data.json --query_file=my_queries.json`
+esgo2dump --input=http://127.0.0.1:9200/some_index --output=./data.json --query_file=my_queries.json`,
-)
-
-var rootCommand = &cobra.Command{
-	Use:           "esgo2dump",
-	Short:         "esgo2dump is alternative to elasticdump",
-	Example:       example,
-	SilenceUsage:  true,
-	SilenceErrors: true,
-	PreRunE:       preRun,
-	RunE:          run,
 	}
 
-func initRoot(cmds ...*cobra.Command) *cobra.Command {
+	es_iversion, es_oversion string
+)
+
+func init() {
 	rootCommand.PersistentFlags().BoolVar(&opt.Cfg.Debug, "debug", false, "")
 	rootCommand.PersistentFlags().BoolVar(&opt.Cfg.Dev, "dev", false, "")
-	rootCommand.PersistentFlags().BoolVar(&opt.Cfg.DisablePing, "disable-ping", false, "")
 	rootCommand.PersistentFlags().BoolVarP(&opt.Cfg.Args.Version, "version", "v", false, "print esgo2dump version")
 
-	// rootCommand.Flags().IntVar(&opt.Cfg.Args.Timeout, "timeout", 30, "max timeout seconds per operation with limit")
+	rootCommand.Flags().IntVar(&opt.Cfg.Args.Timeout, "timeout", 30, "max timeout seconds per operation with limit")
 	rootCommand.Flags().StringVarP(&opt.Cfg.Args.Input, "input", "i", "", "*required: input file or es url (example :data.json / http://127.0.0.1:9200/my_index)")
 	rootCommand.Flags().StringVarP(&opt.Cfg.Args.Output, "output", "o", "output.json", "")
+	rootCommand.Flags().StringVar(&es_iversion, "i-version", "7", "input(es) version")
+	rootCommand.Flags().StringVar(&es_oversion, "o-version", "7", "output(es) version")
 	rootCommand.Flags().StringVarP(&opt.Cfg.Args.Type, "type", "t", "data", "data/mapping/setting")
-	rootCommand.Flags().StringVar(&opt.Cfg.Args.Field, "field", "", "query include field, use ',' to separate")
+	rootCommand.Flags().StringVar(&opt.Cfg.Args.Source, "source", "", "query source, use ';' to separate")
-	rootCommand.Flags().StringVar(&opt.Cfg.Args.Sort, "sort", "", "sort, <field>:<direction> format, for example: time:desc or name:asc, user ',' to separate")
+	rootCommand.Flags().StringVar(&opt.Cfg.Args.Sort, "sort", "", "sort, <field>:<direction> format, for example: time:desc or name:asc")
 	rootCommand.Flags().StringVar(&opt.Cfg.Args.Query, "query", "", `query dsl, example: {"bool":{"must":[{"term":{"name":{"value":"some_name"}}}],"must_not":[{"range":{"age":{"gte":18,"lt":60}}}]}}`)
 	rootCommand.Flags().StringVar(&opt.Cfg.Args.QueryFile, "query_file", "", `query json file (will execute line by line)`)
 	rootCommand.Flags().IntVar(&opt.Cfg.Args.Limit, "limit", 100, "")
-	rootCommand.Flags().IntVar(&opt.Cfg.Args.Max, "max", 0, "max dump records")
-
-	rootCommand.AddCommand(cmds...)
-
-	return rootCommand
 }
 
-func Run(ctx context.Context) error {
+func Start(ctx context.Context) error {
 	return rootCommand.ExecuteContext(ctx)
 }

internal/cmd/init.go

@@ -1,9 +0,0 @@
-package cmd
-
-import "time"
-
-func init() {
-	time.Local = time.FixedZone("CST", 8*3600)
-
-	initRoot()
-}

internal/cmd/root.run.go

@@ -1,74 +0,0 @@
-package cmd
-
-import (
-	"fmt"
-	"github.com/loveuer/esgo2dump/internal/core"
-	"github.com/loveuer/esgo2dump/internal/opt"
-	"github.com/loveuer/esgo2dump/internal/tool"
-	"github.com/loveuer/esgo2dump/pkg/log"
-	"github.com/loveuer/esgo2dump/pkg/model"
-	"github.com/spf13/cobra"
-	"os"
-)
-
-func preRun(cmd *cobra.Command, args []string) error {
-	if opt.Cfg.Debug {
-		log.SetLogLevel(log.LogLevelDebug)
-	}
-
-	if opt.Cfg.Args.Version {
-		fmt.Printf("esgo2dump version: %s\n", opt.Version)
-		os.Exit(0)
-	}
-
-	if opt.Cfg.Debug {
-		tool.TablePrinter(opt.Cfg)
-	}
-
-	// check args
-	if opt.Cfg.Args.Input == "" {
-		return cmd.Help()
-	}
-
-	if opt.Cfg.Args.Limit == 0 || opt.Cfg.Args.Limit > 10000 {
-		return fmt.Errorf("invalid limit(1 - 10000)")
-	}
-
-	if opt.Cfg.Args.Query != "" && opt.Cfg.Args.QueryFile != "" {
-		return fmt.Errorf("cannot specify both query and query_file at the same time")
-	}
-
-	switch opt.Cfg.Args.Type {
-	case "data", "mapping", "setting":
-	default:
-		return fmt.Errorf("unknown type=%s", opt.Cfg.Args.Type)
-	}
-
-	return nil
-}
-
-func run(cmd *cobra.Command, args []string) error {
-	var (
-		err    error
-		input  model.IO[map[string]any]
-		output model.IO[map[string]any]
-	)
-
-	if input, err = core.NewIO(cmd.Context(), opt.Cfg.Args.Input, model.Input); err != nil {
-		return err
-	}
-
-	if output, err = core.NewIO(cmd.Context(), opt.Cfg.Args.Output, model.Output); err != nil {
-		return err
-	}
-
-	switch opt.Cfg.Args.Type {
-	case "data":
-		return core.RunData(cmd, input, output)
-	case "mapping":
-		return core.RunMapping(cmd, input, output)
-	case "setting":
-		return core.RunSetting(cmd, input, output)
-	}
-	return fmt.Errorf("unknown args: type = %s", opt.Cfg.Args.Type)
-}

internal/cmd/run.go

@@ -0,0 +1,292 @@
+package cmd
+
+import (
+	"bufio"
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"net/url"
+	"os"
+	"strings"
+	"sync"
+
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/nf/nft/log"
+
+	"github.com/loveuer/esgo2dump/internal/interfaces"
+	"github.com/loveuer/esgo2dump/internal/opt"
+	"github.com/loveuer/esgo2dump/internal/xes"
+	"github.com/loveuer/esgo2dump/internal/xfile"
+	"github.com/samber/lo"
+	"github.com/spf13/cobra"
+)
+
+func check(cmd *cobra.Command) error {
+	if opt.Cfg.Args.Input == "" {
+		return cmd.Help()
+		// return fmt.Errorf("must specify input(example: data.json/http://127.0.0.1:9200/my_index)")
+	}
+
+	if opt.Cfg.Args.Limit == 0 || opt.Cfg.Args.Limit > 10000 {
+		return fmt.Errorf("invalid limit(1 - 10000)")
+	}
+
+	if opt.Cfg.Args.Query != "" && opt.Cfg.Args.QueryFile != "" {
+		return fmt.Errorf("cannot specify both query and query_file at the same time")
+	}
+
+	switch opt.Cfg.Args.Type {
+	case "data", "mapping", "setting":
+	default:
+		return fmt.Errorf("unknown type=%s", opt.Cfg.Args.Type)
+	}
+
+	return nil
+}
+
+func run(cmd *cobra.Command, args []string) error {
+	var (
+		err error
+		ioi interfaces.DumpIO
+		ioo interfaces.DumpIO
+	)
+
+	if err = check(cmd); err != nil {
+		return err
+	}
+
+	if ioi, err = newIO(opt.Cfg.Args.Input, interfaces.IOInput, es_iversion); err != nil {
+		return err
+	}
+
+	if ioo, err = newIO(opt.Cfg.Args.Output, interfaces.IOOutput, es_oversion); err != nil {
+		return err
+	}
+
+	defer func() {
+		_ = ioi.Close()
+		_ = ioo.Close()
+	}()
+
+	if (opt.Cfg.Args.Query != "" || opt.Cfg.Args.QueryFile != "") && ioi.IsFile() {
+		return fmt.Errorf("with file input, query or query_file can't be supported")
+	}
+
+	if (opt.Cfg.Args.Source != "") && ioi.IsFile() {
+		return fmt.Errorf("with file input, source can't be supported")
+	}
+
+	switch opt.Cfg.Args.Type {
+	case "data":
+		if err = executeData(cmd.Context(), ioi, ioo); err != nil {
+			return err
+		}
+
+		log.Info("Dump: write data succeed!!!")
+
+		return nil
+	case "mapping":
+		var mapping map[string]any
+		if mapping, err = ioi.ReadMapping(cmd.Context()); err != nil {
+			return err
+		}
+
+		if err = ioo.WriteMapping(cmd.Context(), mapping); err != nil {
+			return err
+		}
+
+		log.Info("Dump: write mapping succeed!!!")
+
+		return nil
+	case "setting":
+		var setting map[string]any
+		if setting, err = ioi.ReadSetting(cmd.Context()); err != nil {
+			return err
+		}
+
+		if err = ioo.WriteSetting(cmd.Context(), setting); err != nil {
+			return err
+		}
+
+		log.Info("Dump: write setting succeed!!!")
+
+		return nil
+	default:
+		return fmt.Errorf("unknown type=%s", opt.Cfg.Args.Type)
+	}
+}
+
+func executeData(ctx context.Context, input, output interfaces.DumpIO) error {
+	var (
+		err     error
+		queries = make([]map[string]any, 0)
+		sources = make([]string, 0)
+	)
+
+	if opt.Cfg.Args.Source != "" {
+		sources = lo.Map(strings.Split(opt.Cfg.Args.Source, ";"), func(item string, idx int) string {
+			return strings.TrimSpace(item)
+		})
+	}
+
+	if opt.Cfg.Args.Query != "" {
+		query := make(map[string]any)
+		if err = json.Unmarshal([]byte(opt.Cfg.Args.Query), &query); err != nil {
+			return fmt.Errorf("invalid query err=%v", err)
+		}
+
+		queries = append(queries, query)
+	}
+
+	if opt.Cfg.Args.QueryFile != "" {
+		var qf *os.File
+
+		if qf, err = os.Open(opt.Cfg.Args.QueryFile); err != nil {
+			return fmt.Errorf("open query_file err=%v", err)
+		}
+
+		defer func() {
+			_ = qf.Close()
+		}()
+
+		scanner := bufio.NewScanner(qf)
+		scanner.Buffer(make([]byte, 1*1024*1024), 5*1024*1024)
+		lineCount := 1
+		for scanner.Scan() {
+			line := scanner.Text()
+			oq := make(map[string]any)
+			if err = json.Unmarshal([]byte(line), &oq); err != nil {
+				return fmt.Errorf("query file line=%d invalid err=%v", lineCount, err)
+			}
+
+			queries = append(queries, oq)
+
+			if len(queries) > 10000 {
+				return fmt.Errorf("query_file support max lines=%d", 10000)
+			}
+
+			lineCount++
+		}
+
+	}
+
+	if len(queries) == 0 {
+		queries = append(queries, nil)
+	}
+
+	var (
+		ok   bool
+		docs []*model.ESSource
+		dch  <-chan []*model.ESSource
+		ech  <-chan error
+
+		e2ch = make(chan error)
+		wch  = make(chan []*model.ESSource)
+		wg   = sync.WaitGroup{}
+	)
+
+	go func() {
+		wg.Add(1)
+		if err = output.WriteData(ctx, wch); err != nil {
+			e2ch <- err
+		}
+
+		wg.Done()
+	}()
+
+	log.Info("Query: got queries=%d", len(queries))
+
+Loop:
+	for queryIdx, query := range queries {
+		bs, _ := json.Marshal(query)
+
+		log.Debug("Query[%d]: %s", queryIdx, string(bs))
+
+		dch, ech = input.ReadData(ctx, opt.Cfg.Args.Limit, query, sources, []string{opt.Cfg.Args.Sort})
+
+		for {
+			select {
+			case <-ctx.Done():
+				return ctx.Err()
+			case err, ok = <-ech:
+				if err != nil {
+					return err
+				}
+
+				continue Loop
+			case err, _ = <-e2ch:
+				return err
+			case docs, ok = <-dch:
+				if !ok || len(docs) == 0 {
+					continue Loop
+				}
+
+				wch <- docs
+			}
+		}
+	}
+
+	close(wch)
+
+	wg.Wait()
+
+	return nil
+}
+
+func newIO(source string, ioType interfaces.IO, esv string) (interfaces.DumpIO, error) {
+	var (
+		err  error
+		iurl *url.URL
+		file *os.File
+		qm   = make(map[string]any)
+	)
+
+	log.Debug("action=%s, type=%s, source=%s, es_version=%s", "new_io", ioType.Code(), source, esv)
+
+	if iurl, err = url.Parse(source); err != nil {
+		log.Debug("action=%s, type=%s, source=%s, err=%s", "new_io url parse err", ioType.Code(), source, err.Error())
+		goto ClientByFile
+	}
+
+	if !(iurl.Scheme == "http" || iurl.Scheme == "https") {
+		log.Debug("action=%s, type=%s, source=%s, scheme=%s", "new_io url scheme error", ioType.Code(), source, iurl.Scheme)
+		goto ClientByFile
+	}
+
+	if iurl.Host == "" {
+		log.Debug("action=%s, type=%s, source=%s", "new_io url host empty", ioType.Code(), source)
+		goto ClientByFile
+	}
+
+	if ioType == interfaces.IOInput && opt.Cfg.Args.Query != "" {
+		if err = json.Unmarshal([]byte(opt.Cfg.Args.Query), &qm); err != nil {
+			log.Debug("action=%s, type=%s, source=%s, query=%s", "new_io query string invalid", ioType.Code(), source, opt.Cfg.Args.Query)
+			return nil, fmt.Errorf("invalid query err=%v", err)
+		}
+	}
+
+	switch esv {
+	case "7":
+		return xes.NewClient(iurl, ioType)
+	case "6":
+		return xes.NewClientV6(iurl, ioType)
+	case "8":
+		return nil, errors.New("es version 8 coming soon")
+	default:
+		return nil, fmt.Errorf("unknown es version=%s", esv)
+	}
+
+ClientByFile:
+	if ioType == interfaces.IOOutput {
+		if _, err = os.Stat(source); !os.IsNotExist(err) {
+			return nil, fmt.Errorf("output_file=%s already exist", source)
+		}
+	}
+
+	if file, err = os.OpenFile(source, os.O_CREATE|os.O_RDWR, 0o644); err != nil {
+		return nil, err
+	}
+
+	return xfile.NewClient(file, ioType)
+}

internal/core/io.go

@@ -1,81 +0,0 @@
-package core
-
-import (
-	"context"
-	"encoding/json"
-	"fmt"
-	elastic7 "github.com/elastic/go-elasticsearch/v7"
-	"github.com/go-resty/resty/v2"
-	"github.com/loveuer/esgo2dump/internal/opt"
-	"github.com/loveuer/esgo2dump/internal/tool"
-	"github.com/loveuer/esgo2dump/internal/xfile"
-	"github.com/loveuer/esgo2dump/pkg/log"
-	"github.com/loveuer/esgo2dump/pkg/model"
-	"github.com/loveuer/esgo2dump/xes/es7"
-	"net/url"
-	"strings"
-)
-
-func NewIO(ctx context.Context, uri string, ioType model.IOType) (model.IO[map[string]any], error) {
-	type Version struct {
-		Name    string
-		Version struct {
-			Number string `json:"number"`
-		} `json:"version"`
-	}
-
-	var (
-		err    error
-		target *url.URL
-		rr     *resty.Response
-		v      Version
-	)
-
-	if target, err = url.Parse(uri); err != nil {
-		log.Debug("parse uri failed, type = %s, uri = %s, err = %s", ioType, uri, err.Error())
-		return xfile.NewClient(uri, ioType)
-	}
-
-	if err = tool.ValidScheme(target.Scheme); err != nil {
-		log.Debug("uri scheme check failed, type = %s, uri = %s", ioType, uri)
-		return xfile.NewClient(uri, ioType)
-	}
-
-	// elastic uri
-	index := strings.TrimPrefix(target.Path, "/")
-	if index == "" {
-		return nil, fmt.Errorf("uri invalid without index(path)")
-	}
-
-	log.Debug("%s uri es index = %s", ioType, index)
-
-	versionURL := fmt.Sprintf("%s://%s", target.Scheme, strings.Split(target.Host, ",")[0])
-	log.Debug("%s version url = %s", ioType, versionURL)
-	if rr, err = opt.HttpClient.R().Get(versionURL); err != nil {
-		log.Debug("get uri es version failed, type = %s, uri = %s, version_url = %s, err = %s", ioType, uri, versionURL, err.Error())
-	}
-
-	if err = json.Unmarshal(rr.Body(), &v); err != nil {
-		log.Debug("decode uri es version failed, type = %s, uri = %s, version_url = %s, err = %s", ioType, uri, versionURL, err.Error())
-		return nil, err
-	}
-
-	log.Debug("%s uri es version = %s", ioType, v.Version.Number)
-
-	mainVersion := strings.Split(v.Version.Number, ".")[0]
-	switch mainVersion {
-	case "8":
-	case "7":
-		var client *elastic7.Client
-		if client, err = es7.NewClient(ctx, uri); err != nil {
-			return nil, err
-		}
-
-		return es7.NewStreamer(ctx, client, index)
-	case "6":
-	default:
-		return nil, fmt.Errorf("es version not supported yet: %s", mainVersion)
-	}
-
-	return nil, nil
-}

internal/core/run.data.go

@@ -1,125 +0,0 @@
-package core
-
-import (
-	"bufio"
-	"encoding/json"
-	"fmt"
-	"github.com/loveuer/esgo2dump/internal/opt"
-	"github.com/loveuer/esgo2dump/internal/tool"
-	"github.com/loveuer/esgo2dump/pkg/log"
-	"github.com/loveuer/esgo2dump/pkg/model"
-	"github.com/samber/lo"
-	"github.com/spf13/cobra"
-	"os"
-	"strings"
-	"sync"
-)
-
-func RunData(cmd *cobra.Command, input, output model.IO[map[string]any]) error {
-	var (
-		err error
-		// query chan
-		qc = make(chan map[string]any)
-		// error chan
-		ec = make(chan error)
-		// done chan
-		wc    = &sync.WaitGroup{}
-		total = 0
-	)
-
-	wc.Add(1)
-
-	go func() {
-		var (
-			wroteCount = 0
-			items      []map[string]any
-		)
-
-		defer wc.Done()
-
-		for query := range qc {
-			for {
-				limit := tool.CalculateLimit(opt.Cfg.Args.Limit, total, opt.Cfg.Args.Max)
-				log.Debug("one-step dump: arg.limit = %d, total = %d, arg.max = %d, calculate.limit = %d", opt.Cfg.Args.Limit, total, opt.Cfg.Args.Max, limit)
-				if limit == 0 {
-					break
-				}
-
-				if items, err = input.ReadData(
-					cmd.Context(),
-					limit,
-					query,
-					lo.Filter(strings.Split(opt.Cfg.Args.Field, ","), func(x string, _ int) bool { return x != "" }),
-					lo.Filter(strings.Split(opt.Cfg.Args.Sort, ","), func(x string, _ int) bool { return x != "" }),
-				); err != nil {
-					ec <- err
-					return
-				}
-
-				if len(items) == 0 {
-					break
-				}
-
-				if wroteCount, err = output.WriteData(cmd.Context(), items); err != nil {
-					ec <- err
-					return
-				}
-
-				total += wroteCount
-
-				if wroteCount != len(items) {
-					ec <- fmt.Errorf("got items %d, but wrote %d", len(items), wroteCount)
-					return
-				}
-
-				log.Info("Dump: dump data success = %d total = %d", wroteCount, total)
-			}
-		}
-	}()
-
-	switch {
-	case opt.Cfg.Args.QueryFile != "":
-		var (
-			// query file
-			qf         *os.File
-			queryCount = 0
-		)
-		if qf, err = os.Open(opt.Cfg.Args.QueryFile); err != nil {
-			return err
-		}
-
-		scanner := bufio.NewScanner(qf)
-		for scanner.Scan() {
-			queryCount++
-			qm := make(map[string]any)
-			if err = json.Unmarshal(scanner.Bytes(), &qm); err != nil {
-				return err
-			}
-
-			qc <- qm
-
-			log.Debug("Dump: queries[%06d]", queryCount)
-		}
-	case opt.Cfg.Args.Query != "":
-		var (
-			qm = make(map[string]any)
-		)
-
-		if err = json.Unmarshal([]byte(opt.Cfg.Args.Query), &qm); err != nil {
-			return err
-		}
-
-		qc <- qm
-	default:
-		qc <- nil
-	}
-
-	// close query chan to stop trans_io_goroutine
-	close(qc)
-
-	wc.Wait()
-
-	log.Info("Dump: dump all data success, total = %d", total)
-
-	return nil
-}

internal/core/run.mapping.go

@@ -1,19 +0,0 @@
-package core
-
-import (
-	"github.com/loveuer/esgo2dump/pkg/model"
-	"github.com/spf13/cobra"
-)
-
-func RunMapping(cmd *cobra.Command, input model.IO[map[string]any], output model.IO[map[string]any]) error {
-	mapping, err := input.ReadMapping(cmd.Context())
-	if err != nil {
-		return err
-	}
-
-	if err = output.WriteMapping(cmd.Context(), mapping); err != nil {
-		return err
-	}
-
-	return nil
-}

internal/core/run.setting.go

@@ -1,19 +0,0 @@
-package core
-
-import (
-	"github.com/loveuer/esgo2dump/pkg/model"
-	"github.com/spf13/cobra"
-)
-
-func RunSetting(cmd *cobra.Command, input model.IO[map[string]any], output model.IO[map[string]any]) error {
-	setting, err := input.ReadSetting(cmd.Context())
-	if err != nil {
-		return err
-	}
-
-	if err = output.WriteSetting(cmd.Context(), setting); err != nil {
-		return err
-	}
-
-	return nil
-}

internal/interfaces/dumpio.go

@@ -0,0 +1,23 @@
+package interfaces
+
+import (
+	"context"
+
+	"github.com/loveuer/esgo2dump/model"
+)
+
+type DumpIO interface {
+	ReadData(ctx context.Context, size int, query map[string]any, includeFields []string, sort []string) (<-chan []*model.ESSource, <-chan error)
+	WriteData(ctx context.Context, docsCh <-chan []*model.ESSource) error
+
+	ReadMapping(context.Context) (map[string]any, error)
+	WriteMapping(context.Context, map[string]any) error
+
+	ReadSetting(ctx context.Context) (map[string]any, error)
+	WriteSetting(context.Context, map[string]any) error
+
+	Close() error
+
+	IOType() IO
+	IsFile() bool
+}

internal/interfaces/enum.go

@@ -0,0 +1,27 @@
+package interfaces
+
+type IO int64
+
+const (
+	IOInput IO = iota
+	IOOutput
+)
+
+func (io IO) Code() string {
+	switch io {
+	case IOInput:
+		return "input"
+	case IOOutput:
+		return "output"
+	default:
+		return "unknown"
+	}
+}
+
+type DataType int64
+
+const (
+	DataTypeData DataType = iota
+	DataTypeMapping
+	DataTypeSetting
+)

internal/opt/opt.go

@@ -8,7 +8,7 @@ type args struct {
 	Max       int
 	Type      string
 	Timeout   int
-	Field     string
+	Source    string
 	Sort      string
 	Query     string
 	QueryFile string
@@ -17,7 +17,6 @@ type args struct {
 type config struct {
 	Debug bool `json:"-"`
 	Dev   bool `json:"-"`
-	DisablePing bool `json:"-"`
 	Args  args `json:"-"`
 }
 

internal/opt/var.go

@@ -1,11 +1,5 @@
 package opt
 
-import (
-	"crypto/tls"
-
-	"github.com/go-resty/resty/v2"
-)
-
 const (
 	ScrollDurationSeconds = 10 * 60
 	DefaultSize           = 100
@@ -17,6 +11,4 @@ var (
 
 	BuffSize    = 5 * 1024 * 1024   // 5M
 	MaxBuffSize = 100 * 1024 * 1024 // 100M, default elastic_search doc max size
-
-	HttpClient = resty.New().SetTLSClientConfig(&tls.Config{InsecureSkipVerify: true})
 )

internal/tool/min.go

@@ -1,5 +1,7 @@
 package tool
 
+import "github.com/loveuer/esgo2dump/internal/opt"
+
 func Min[T ~string | ~int | ~int64 | ~uint64 | ~float64 | ~float32 | ~int32 | ~uint32 | ~int16 | ~uint16 | ~int8 | ~uint8](a, b T) T {
 	if a <= b {
 		return a
@@ -8,14 +10,23 @@ func Min[T ~string | ~int | ~int64 | ~uint64 | ~float64 | ~float32 | ~int32 | ~u
 	return b
 }
 
-func CalculateLimit(limit, total, max int) int {
+func CalcSize(size, max, total int) int {
+	fs := size
+	if fs == 0 {
+		fs = opt.DefaultSize
+	}
+
 	if max == 0 {
-		return limit
+		return fs
-	}
-
-	if max-total > 0 {
-		return Min(max-total, limit)
 	}
 
+	if max > 0 && total >= max {
 		return 0
 	}
+
+	if max-total > fs {
+		return max - total
+	}
+
+	return fs
+}

internal/tool/table.go

@@ -9,7 +9,7 @@ import (
 	"strings"
 
 	"github.com/jedib0t/go-pretty/v6/table"
-	"github.com/loveuer/esgo2dump/pkg/log"
+	"github.com/loveuer/nf/nft/log"
 )
 
 func TablePrinter(data any, writers ...io.Writer) {

internal/tool/validate.go

@@ -1,15 +0,0 @@
-package tool
-
-import (
-	"fmt"
-	"strings"
-)
-
-func ValidScheme(scheme string) error {
-	switch strings.ToLower(scheme) {
-	case "http", "https":
-		return nil
-	default:
-		return fmt.Errorf("invalid scheme: %s", scheme)
-	}
-}

internal/xes/xes6.go

@@ -0,0 +1,234 @@
+package xes
+
+import (
+	"bytes"
+	"context"
+	"crypto/tls"
+	"encoding/json"
+	"fmt"
+	"net"
+	"net/http"
+	"net/url"
+	"strings"
+	"time"
+
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/esgo2dump/xes/es6"
+	"github.com/loveuer/nf/nft/log"
+
+	elastic "github.com/elastic/go-elasticsearch/v6"
+	"github.com/elastic/go-elasticsearch/v6/esapi"
+	"github.com/loveuer/esgo2dump/internal/interfaces"
+	"github.com/loveuer/esgo2dump/internal/opt"
+	"github.com/loveuer/esgo2dump/internal/tool"
+)
+
+func NewClientV6(url *url.URL, iot interfaces.IO) (interfaces.DumpIO, error) {
+	var (
+		address     = fmt.Sprintf("%s://%s", url.Scheme, url.Host)
+		urlIndex    = strings.TrimPrefix(url.Path, "/")
+		urlUsername string
+		urlPassword string
+		errCh       = make(chan error)
+		cliCh       = make(chan *elastic.Client)
+	)
+
+	if url.User != nil {
+		urlUsername = url.User.Username()
+		if p, ok := url.User.Password(); ok {
+			urlPassword = p
+		}
+	}
+
+	log.Debug("action=%s, endpoint=%s, index=%s, username=%s, password=%s", "new es client v6", address, urlIndex, urlUsername, urlPassword)
+
+	if urlIndex == "" {
+		return nil, fmt.Errorf("please specify index name: (like => http://127.0.0.1:9200/my_index)")
+	}
+
+	ncFunc := func(endpoints []string, username, password, index string) {
+		var (
+			err      error
+			cli      *elastic.Client
+			infoResp *esapi.Response
+		)
+
+		if cli, err = elastic.NewClient(
+			elastic.Config{
+				Addresses:     endpoints,
+				Username:      username,
+				Password:      password,
+				CACert:        nil,
+				RetryOnStatus: []int{429},
+				MaxRetries:    3,
+				RetryBackoff:  nil,
+				Transport: &http.Transport{
+					TLSClientConfig: &tls.Config{InsecureSkipVerify: true},
+					DialContext:     (&net.Dialer{Timeout: 10 * time.Second}).DialContext,
+				},
+			},
+		); err != nil {
+			log.Debug("action=%s, endpoints=%v, err=%s", "new es client v6 error", endpoints, err.Error())
+			errCh <- err
+			return
+		}
+
+		if infoResp, err = cli.Info(); err != nil {
+			log.Debug("action=%s, endpoints=%v, err=%s", "new es client v6 info error", endpoints, err.Error())
+			errCh <- err
+			return
+		}
+
+		if infoResp.StatusCode != 200 {
+			err = fmt.Errorf("info xes status=%d", infoResp.StatusCode)
+			log.Debug("action=%s, endpoints=%v, err=%s", "es client v6 ping status error", endpoints, err.Error())
+			errCh <- err
+			return
+		}
+
+		cliCh <- cli
+	}
+
+	go ncFunc([]string{address}, urlUsername, urlPassword, urlIndex)
+
+	select {
+	case <-tool.Timeout(10).Done():
+		return nil, fmt.Errorf("dial es=%s err=%v", address, context.DeadlineExceeded)
+	case c := <-cliCh:
+		return &clientv6{client: c, index: urlIndex, iot: iot}, nil
+	case e := <-errCh:
+		return nil, e
+	}
+}
+
+type clientv6 struct {
+	client *elastic.Client
+	iot    interfaces.IO
+	index  string
+}
+
+func (c *clientv6) Info(msg string, data ...any) {
+	log.Info(msg, data...)
+}
+
+func (c *clientv6) WriteData(ctx context.Context, docsCh <-chan []*model.ESSource) error {
+	return es6.WriteData(ctx, c.client, c.index, docsCh, c)
+}
+
+func (c *clientv6) checkResponse(r *esapi.Response) error {
+	if r.StatusCode == 200 {
+		return nil
+	}
+
+	return fmt.Errorf("status=%d msg=%s", r.StatusCode, r.String())
+}
+
+func (c *clientv6) IOType() interfaces.IO {
+	return c.iot
+}
+
+func (c *clientv6) IsFile() bool {
+	return false
+}
+
+func (c *clientv6) Close() error {
+	return nil
+}
+
+func (c *clientv6) ReadData(ctx context.Context, size int, query map[string]any, source []string, sort []string) (<-chan []*model.ESSource, <-chan error) {
+	dch, ech := es6.ReadData(ctx, c.client, c.index, size, 0, query, source, sort)
+
+	return dch, ech
+}
+
+func (c *clientv6) ReadMapping(ctx context.Context) (map[string]any, error) {
+	r, err := c.client.Indices.GetMapping(
+		c.client.Indices.GetMapping.WithIndex(c.index),
+	)
+	if err != nil {
+		return nil, err
+	}
+
+	if r.StatusCode != 200 {
+		return nil, fmt.Errorf("status=%d, msg=%s", r.StatusCode, r.String())
+	}
+
+	m := make(map[string]any)
+	decoder := json.NewDecoder(r.Body)
+	if err = decoder.Decode(&m); err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}
+
+func (c *clientv6) WriteMapping(ctx context.Context, m map[string]any) error {
+	var (
+		err    error
+		bs     []byte
+		result *esapi.Response
+	)
+
+	for idxKey := range m {
+		if bs, err = json.Marshal(m[idxKey]); err != nil {
+			return err
+		}
+
+		if result, err = c.client.Indices.Create(
+			c.index,
+			c.client.Indices.Create.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+			c.client.Indices.Create.WithBody(bytes.NewReader(bs)),
+		); err != nil {
+			return err
+		}
+
+		if err = c.checkResponse(result); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (c *clientv6) ReadSetting(ctx context.Context) (map[string]any, error) {
+	r, err := c.client.Indices.GetSettings(
+		c.client.Indices.GetSettings.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+		c.client.Indices.GetSettings.WithIndex(c.index),
+	)
+	if err != nil {
+		return nil, err
+	}
+
+	if r.StatusCode != 200 {
+		return nil, fmt.Errorf("status=%d, msg=%s", r.StatusCode, r.String())
+	}
+
+	m := make(map[string]any)
+	decoder := json.NewDecoder(r.Body)
+	if err = decoder.Decode(&m); err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}
+
+func (c *clientv6) WriteSetting(ctx context.Context, m map[string]any) error {
+	var (
+		err    error
+		bs     []byte
+		result *esapi.Response
+	)
+
+	if bs, err = json.Marshal(m); err != nil {
+		return err
+	}
+
+	if result, err = c.client.Indices.PutSettings(
+		bytes.NewReader(bs),
+		c.client.Indices.PutSettings.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+	); err != nil {
+		return err
+	}
+
+	return c.checkResponse(result)
+}

internal/xes/xes7.go

@@ -0,0 +1,169 @@
+package xes
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	"net/url"
+	"strings"
+
+	elastic "github.com/elastic/go-elasticsearch/v7"
+	"github.com/elastic/go-elasticsearch/v7/esapi"
+	"github.com/loveuer/esgo2dump/internal/interfaces"
+	"github.com/loveuer/esgo2dump/internal/opt"
+	"github.com/loveuer/esgo2dump/internal/tool"
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/esgo2dump/xes/es7"
+	"github.com/loveuer/nf/nft/log"
+)
+
+type client struct {
+	client *elastic.Client
+	iot    interfaces.IO
+	index  string
+}
+
+func (c *client) Info(msg string, data ...any) {
+	log.Info(msg, data...)
+}
+
+func (c *client) WriteData(ctx context.Context, docsCh <-chan []*model.ESSource) error {
+	return es7.WriteData(ctx, c.client, c.index, docsCh, c)
+}
+
+func NewClient(url *url.URL, iot interfaces.IO) (interfaces.DumpIO, error) {
+	var (
+		urlIndex = strings.TrimPrefix(url.Path, "/")
+		cli      *elastic.Client
+		err      error
+	)
+
+	if urlIndex == "" {
+		return nil, fmt.Errorf("please specify index name: (like => http://127.0.0.1:9200/my_index)")
+	}
+
+	if cli, err = es7.NewClient(context.TODO(), url); err != nil {
+		return nil, err
+	}
+
+	return &client{client: cli, iot: iot, index: urlIndex}, nil
+}
+
+func (c *client) checkResponse(r *esapi.Response) error {
+	if r.StatusCode == 200 {
+		return nil
+	}
+
+	return fmt.Errorf("status=%d msg=%s", r.StatusCode, r.String())
+}
+
+func (c *client) IOType() interfaces.IO {
+	return c.iot
+}
+
+func (c *client) IsFile() bool {
+	return false
+}
+
+func (c *client) Close() error {
+	return nil
+}
+
+func (c *client) ReadData(ctx context.Context, size int, query map[string]any, source []string, sort []string) (<-chan []*model.ESSource, <-chan error) {
+	dch, ech := es7.ReadData(ctx, c.client, c.index, size, 0, query, source, sort)
+
+	return dch, ech
+}
+
+func (c *client) ReadMapping(ctx context.Context) (map[string]any, error) {
+	r, err := c.client.Indices.GetMapping(
+		c.client.Indices.GetMapping.WithIndex(c.index),
+	)
+	if err != nil {
+		return nil, err
+	}
+
+	if r.StatusCode != 200 {
+		return nil, fmt.Errorf("status=%d, msg=%s", r.StatusCode, r.String())
+	}
+
+	m := make(map[string]any)
+	decoder := json.NewDecoder(r.Body)
+	if err = decoder.Decode(&m); err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}
+
+func (c *client) WriteMapping(ctx context.Context, m map[string]any) error {
+	var (
+		err    error
+		bs     []byte
+		result *esapi.Response
+	)
+
+	for idxKey := range m {
+		if bs, err = json.Marshal(m[idxKey]); err != nil {
+			return err
+		}
+
+		if result, err = c.client.Indices.Create(
+			c.index,
+			c.client.Indices.Create.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+			c.client.Indices.Create.WithBody(bytes.NewReader(bs)),
+		); err != nil {
+			return err
+		}
+
+		if err = c.checkResponse(result); err != nil {
+			return err
+		}
+	}
+
+	return nil
+}
+
+func (c *client) ReadSetting(ctx context.Context) (map[string]any, error) {
+	r, err := c.client.Indices.GetSettings(
+		c.client.Indices.GetSettings.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+		c.client.Indices.GetSettings.WithIndex(c.index),
+	)
+	if err != nil {
+		return nil, err
+	}
+
+	if r.StatusCode != 200 {
+		return nil, fmt.Errorf("status=%d, msg=%s", r.StatusCode, r.String())
+	}
+
+	m := make(map[string]any)
+	decoder := json.NewDecoder(r.Body)
+	if err = decoder.Decode(&m); err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}
+
+func (c *client) WriteSetting(ctx context.Context, m map[string]any) error {
+	var (
+		err    error
+		bs     []byte
+		result *esapi.Response
+	)
+
+	if bs, err = json.Marshal(m); err != nil {
+		return err
+	}
+
+	if result, err = c.client.Indices.PutSettings(
+		bytes.NewReader(bs),
+		c.client.Indices.PutSettings.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+	); err != nil {
+		return err
+	}
+
+	return c.checkResponse(result)
+}

internal/xes/xes7_test.go

@@ -0,0 +1,107 @@
+package xes
+
+import (
+	"bufio"
+	"fmt"
+	"os"
+	"testing"
+
+	elastic "github.com/elastic/go-elasticsearch/v7"
+	"github.com/loveuer/esgo2dump/internal/tool"
+)
+
+func TestGetESMapping(t *testing.T) {
+	endpoint := "http://127.0.0.1:9200"
+	index := "some_index"
+
+	cli, err := elastic.NewClient(elastic.Config{
+		Addresses: []string{endpoint},
+	})
+	if err != nil {
+		t.Error(1, err)
+		return
+	}
+
+	resp, err := cli.Info(cli.Info.WithContext(tool.Timeout(5)))
+	if err != nil {
+		t.Error(2, err)
+		return
+	}
+
+	t.Log("info:", resp.String())
+
+	r, err := cli.Indices.GetMapping(
+		cli.Indices.GetMapping.WithIndex(index),
+	)
+	if err != nil {
+		t.Error(3, err)
+		return
+	}
+
+	t.Log("get source:", r.String())
+}
+
+func TestScanWithInterrupt(t *testing.T) {
+	filename := "test_scan.txt"
+	f, err := os.OpenFile(filename, os.O_RDWR|os.O_CREATE, 0o644)
+	if err != nil {
+		t.Error(1, err)
+		return
+	}
+	defer func() {
+		os.Remove(filename)
+	}()
+	f.WriteString(`line 01
+line 02
+line 03
+line 04
+line 05
+line 06
+line 07
+line 08
+line 09
+line 10
+line 11
+line 12
+line 13
+line 14
+line 15`)
+	f.Close()
+
+	of, err := os.Open(filename)
+	if err != nil {
+		t.Error(2, err)
+		return
+	}
+
+	scanner := bufio.NewScanner(of)
+
+	count := 0
+	for scanner.Scan() {
+		text := scanner.Text()
+		fmt.Printf("[line: %2d] = %s\n", count, text)
+		count++
+
+		if count > 5 {
+			break
+		}
+	}
+
+	count = 0
+	for scanner.Scan() {
+		text := scanner.Text()
+		fmt.Printf("[line: %2d] = %s\n", count, text)
+		count++
+
+		if count > 5 {
+			break
+		}
+	}
+
+	count = 0
+	for scanner.Scan() {
+		text := scanner.Text()
+		fmt.Printf("[line: %2d] = %s\n", count, text)
+		count++
+	}
+}

internal/xfile/xfile.go

@@ -4,74 +4,43 @@ import (
 	"bufio"
 	"context"
 	"encoding/json"
-	"fmt"
 	"io"
 	"os"
 
 	"github.com/loveuer/esgo2dump/internal/opt"
-	"github.com/loveuer/esgo2dump/pkg/log"
+	"github.com/loveuer/esgo2dump/model"
-	"github.com/loveuer/esgo2dump/pkg/model"
+	"github.com/loveuer/nf/nft/log"
+
+	"github.com/loveuer/esgo2dump/internal/interfaces"
 )
 
 type client struct {
-	info    os.FileInfo
 	f       *os.File
+	iot     interfaces.IO
 	scanner *bufio.Scanner
 }
 
-func (c *client) ReadData(ctx context.Context, limit int, query map[string]any, fields []string, sort []string) ([]map[string]any, error) {
+func (c *client) WriteData(ctx context.Context, docsCh <-chan []*model.ESSource) error {
-	if len(query) != 0 {
-		return nil, fmt.Errorf("file with query is unsupported")
-	}
-
-	if len(sort) != 0 {
-		return nil, fmt.Errorf("file with sort is unsupported")
-	}
-
-	list := make([]map[string]any, 0, limit)
-
-	for c.scanner.Scan() {
-		line := c.scanner.Bytes()
-		item := make(map[string]any)
-
-		if err := json.Unmarshal(line, &item); err != nil {
-			return nil, err
-		}
-
-		if len(fields) > 0 {
-			// todo: pick fields
-		}
-
-		list = append(list, item)
-
-		if len(list) >= limit {
-			return list, nil
-		}
-	}
-
-	return list, nil
-}
-
-func (c *client) WriteData(ctx context.Context, items []map[string]any) (int, error) {
 	total := 0
-	for _, item := range items {
+	for line := range docsCh {
-		bs, err := json.Marshal(item)
+		for _, doc := range line {
+			bs, err := json.Marshal(doc)
 			if err != nil {
-			return total, err
+				return err
 			}
 
-		if _, err = c.f.Write(bs); err != nil {
+			if _, err = c.f.Write(append(bs, '\n')); err != nil {
-			return total, err
+				return err
-		}
-
-		total++
-
-		if _, err = c.f.WriteString("\n"); err != nil {
-			return total, err
 			}
 		}
 
-	return total, nil
+		count := len(line)
+		total += count
+
+		log.Info("Dump: succeed=%d total=%d docs succeed!!!", count, total)
+	}
+
+	return nil
 }
 
 func (c *client) ReadMapping(ctx context.Context) (map[string]any, error) {
@@ -93,17 +62,6 @@ func (c *client) ReadMapping(ctx context.Context) (map[string]any, error) {
 	return m, nil
 }
 
-func (c *client) WriteMapping(ctx context.Context, mapping map[string]any) error {
-	bs, err := json.Marshal(mapping)
-	if err != nil {
-		return err
-	}
-
-	_, err = c.f.Write(bs)
-
-	return err
-}
-
 func (c *client) ReadSetting(ctx context.Context) (map[string]any, error) {
 	var (
 		err error
@@ -123,8 +81,8 @@ func (c *client) ReadSetting(ctx context.Context) (map[string]any, error) {
 	return m, nil
 }
 
-func (c *client) WriteSetting(ctx context.Context, setting map[string]any) error {
+func (c *client) WriteMapping(ctx context.Context, m map[string]any) error {
-	bs, err := json.Marshal(setting)
+	bs, err := json.Marshal(m)
 	if err != nil {
 		return err
 	}
@@ -134,45 +92,95 @@ func (c *client) WriteSetting(ctx context.Context, setting map[string]any) error
 	return err
 }
 
-func NewClient(path string, t model.IOType) (model.IO[map[string]any], error) {
+func (c *client) WriteSetting(ctx context.Context, m map[string]any) error {
+	bs, err := json.Marshal(m)
+	if err != nil {
+		return err
+	}
+
+	_, err = c.f.Write(bs)
+
+	return err
+}
+
+func (c *client) IOType() interfaces.IO {
+	return c.iot
+}
+
+func (c *client) IsFile() bool {
+	return true
+}
+
+func (c *client) ReadData(ctx context.Context, size int, _ map[string]any, _ []string, _ []string) (<-chan []*model.ESSource, <-chan error) {
 	var (
-		info os.FileInfo
 		err   error
-		f    *os.File
+		count int = 0
+		list      = make([]*model.ESSource, 0, size)
+		dch       = make(chan []*model.ESSource)
+		ech       = make(chan error)
+		ready     = make(chan bool)
 	)
 
-	switch t {
+	go func(ctx context.Context) {
-	case model.Input:
+		defer func() {
-		if info, err = os.Stat(path); err != nil {
+			close(dch)
-			return nil, err
+			close(ech)
-		}
+		}()
 
-		log.Debug("input file: %s, size: %d", path, info.Size())
+		ready <- true
 
-		if f, err = os.Open(path); err != nil {
+		for c.scanner.Scan() {
-			return nil, err
+			select {
-		}
+			case <-ctx.Done():
-	case model.Output:
+				return
-		if info, err = os.Stat(path); err == nil {
-			return nil, fmt.Errorf("file already exists: %s", path)
-		}
-
-		if !os.IsNotExist(err) {
-			return nil, err
-		}
-
-		if f, err = os.OpenFile(path, os.O_RDWR|os.O_CREATE|os.O_TRUNC|os.O_APPEND, 0o644); err != nil {
-			return nil, err
-		}
 			default:
-		return nil, fmt.Errorf("unknown type: %s", t)
+				item := new(model.ESSource)
+				line := c.scanner.Bytes()
+
+				if err = json.Unmarshal(line, item); err != nil {
+					ech <- err
+					return
 				}
 
-	c := &client{f: f, info: info}
+				list = append(list, item)
+				count++
+
+				if count >= size {
+					dch <- list
+					list = list[:0]
+					count = 0
+				}
+			}
+		}
+
+		if len(list) > 0 {
+			dch <- list
+			list = list[:0]
+			count = 0
+		}
+
+		if err = c.scanner.Err(); err != nil {
+			ech <- err
+		}
+	}(ctx)
+
+	<-ready
+
+	return dch, ech
+}
+
+func (c *client) Close() error {
+	return c.f.Close()
+}
+
+func NewClient(file *os.File, ioType interfaces.IO) (interfaces.DumpIO, error) {
+	c := &client{f: file, iot: ioType}
+
+	if ioType == interfaces.IOInput {
+		c.scanner = bufio.NewScanner(c.f)
 		buf := make([]byte, opt.BuffSize)
-	scanner := bufio.NewScanner(c.f)
+		c.scanner.Buffer(buf, opt.MaxBuffSize)
-	scanner.Buffer(buf, opt.MaxBuffSize)
+	}
-	c.scanner = scanner
 
 	return c, nil
 }

main.go

@@ -5,7 +5,7 @@ import (
 	"os/signal"
 	"syscall"
 
-	"github.com/loveuer/esgo2dump/pkg/log"
+	"github.com/loveuer/nf/nft/log"
 
 	"github.com/loveuer/esgo2dump/internal/cmd"
 )
@@ -14,12 +14,7 @@ func main() {
 	ctx, cancel := signal.NotifyContext(context.Background(), syscall.SIGHUP, syscall.SIGINT, syscall.SIGTERM, syscall.SIGQUIT)
 	defer cancel()
 
-	go func() {
+	if err := cmd.Start(ctx); err != nil {
-		<-ctx.Done()
-		log.Fatal(ctx.Err().Error())
-	}()
-
-	if err := cmd.Run(ctx); err != nil {
 		log.Error(err.Error())
 		return
 	}

model/es.go

@@ -1,13 +1,13 @@
 package model
 
-type ESSource[T any] struct {
+type ESSource struct {
 	DocId   string         `json:"_id"`
 	Index   string         `json:"_index"`
-	Content T      `json:"_source"`
+	Content map[string]any `json:"_source"`
 	Sort    []any          `json:"sort"`
 }
 
-type ESResponseV6[T any] struct {
+type ESResponseV6 struct {
 	ScrollId string `json:"_scroll_id"`
 	Took     int    `json:"took"`
 	TimedOut bool   `json:"timed_out"`
@@ -20,11 +20,11 @@ type ESResponseV6[T any] struct {
 	Hits struct {
 		Total    int         `json:"total"`
 		MaxScore float64     `json:"max_score"`
-		Hits     []*ESSource[T] `json:"hits"`
+		Hits     []*ESSource `json:"hits"`
 	} `json:"hits"`
 }
 
-type ESResponseV7[T any] struct {
+type ESResponseV7 struct {
 	ScrollId string `json:"_scroll_id"`
 	Took     int    `json:"took"`
 	TimedOut bool   `json:"timed_out"`
@@ -40,6 +40,6 @@ type ESResponseV7[T any] struct {
 			Relation string `json:"relation"`
 		} `json:"total"`
 		MaxScore float64     `json:"max_score"`
-		Hits     []*ESSource[T] `json:"hits"`
+		Hits     []*ESSource `json:"hits"`
 	} `json:"hits"`
 }

pkg/log/default.go

@@ -1,67 +0,0 @@
-package log
-
-import (
-	"fmt"
-	"os"
-	"sync"
-)
-
-var (
-	nilLogger    = func(prefix, timestamp, msg string, data ...any) {}
-	normalLogger = func(prefix, timestamp, msg string, data ...any) {
-		fmt.Printf(prefix+"| "+timestamp+" | "+msg+"\n", data...)
-	}
-
-	panicLogger = func(prefix, timestamp, msg string, data ...any) {
-		panic(fmt.Sprintf(prefix+"| "+timestamp+" | "+msg+"\n", data...))
-	}
-
-	fatalLogger = func(prefix, timestamp, msg string, data ...any) {
-		fmt.Printf(prefix+"| "+timestamp+" | "+msg+"\n", data...)
-		os.Exit(1)
-	}
-
-	DefaultLogger = &logger{
-		Mutex:      sync.Mutex{},
-		timeFormat: "2006-01-02T15:04:05",
-		writer:     os.Stdout,
-		level:      LogLevelInfo,
-		debug:      nilLogger,
-		info:       normalLogger,
-		warn:       normalLogger,
-		error:      normalLogger,
-		panic:      panicLogger,
-		fatal:      fatalLogger,
-	}
-)
-
-func SetTimeFormat(format string) {
-	DefaultLogger.SetTimeFormat(format)
-}
-
-func SetLogLevel(level LogLevel) {
-	DefaultLogger.SetLogLevel(level)
-}
-
-func Debug(msg string, data ...any) {
-	DefaultLogger.Debug(msg, data...)
-}
-func Info(msg string, data ...any) {
-	DefaultLogger.Info(msg, data...)
-}
-
-func Warn(msg string, data ...any) {
-	DefaultLogger.Warn(msg, data...)
-}
-
-func Error(msg string, data ...any) {
-	DefaultLogger.Error(msg, data...)
-}
-
-func Panic(msg string, data ...any) {
-	DefaultLogger.Panic(msg, data...)
-}
-
-func Fatal(msg string, data ...any) {
-	DefaultLogger.Fatal(msg, data...)
-}

pkg/log/log.go

@@ -1,115 +0,0 @@
-package log
-
-import (
-	"github.com/fatih/color"
-	"io"
-	"sync"
-	"time"
-)
-
-type LogLevel uint32
-
-const (
-	LogLevelDebug = iota
-	LogLevelInfo
-	LogLevelWarn
-	LogLevelError
-	LogLevelPanic
-	LogLevelFatal
-)
-
-type logger struct {
-	sync.Mutex
-	timeFormat string
-	writer     io.Writer
-	level      LogLevel
-	debug      func(prefix, timestamp, msg string, data ...any)
-	info       func(prefix, timestamp, msg string, data ...any)
-	warn       func(prefix, timestamp, msg string, data ...any)
-	error      func(prefix, timestamp, msg string, data ...any)
-	panic      func(prefix, timestamp, msg string, data ...any)
-	fatal      func(prefix, timestamp, msg string, data ...any)
-}
-
-var (
-	red    = color.New(color.FgRed)
-	hired  = color.New(color.FgHiRed)
-	green  = color.New(color.FgGreen)
-	yellow = color.New(color.FgYellow)
-	white  = color.New(color.FgWhite)
-)
-
-func (l *logger) SetTimeFormat(format string) {
-	l.Lock()
-	defer l.Unlock()
-	l.timeFormat = format
-}
-
-func (l *logger) SetLogLevel(level LogLevel) {
-	l.Lock()
-	defer l.Unlock()
-
-	if level > LogLevelDebug {
-		l.debug = nilLogger
-	} else {
-		l.debug = normalLogger
-	}
-
-	if level > LogLevelInfo {
-		l.info = nilLogger
-	} else {
-		l.info = normalLogger
-	}
-
-	if level > LogLevelWarn {
-		l.warn = nilLogger
-	} else {
-		l.warn = normalLogger
-	}
-
-	if level > LogLevelError {
-		l.error = nilLogger
-	} else {
-		l.error = normalLogger
-	}
-
-	if level > LogLevelPanic {
-		l.panic = nilLogger
-	} else {
-		l.panic = panicLogger
-	}
-
-	if level > LogLevelFatal {
-		l.fatal = nilLogger
-	} else {
-		l.fatal = fatalLogger
-	}
-}
-
-func (l *logger) Debug(msg string, data ...any) {
-	l.debug(white.Sprint("Debug "), time.Now().Format(l.timeFormat), msg, data...)
-}
-
-func (l *logger) Info(msg string, data ...any) {
-	l.info(green.Sprint("Info  "), time.Now().Format(l.timeFormat), msg, data...)
-}
-
-func (l *logger) Warn(msg string, data ...any) {
-	l.warn(yellow.Sprint("Warn  "), time.Now().Format(l.timeFormat), msg, data...)
-}
-
-func (l *logger) Error(msg string, data ...any) {
-	l.error(red.Sprint("Error "), time.Now().Format(l.timeFormat), msg, data...)
-}
-
-func (l *logger) Panic(msg string, data ...any) {
-	l.panic(hired.Sprint("Panic "), time.Now().Format(l.timeFormat), msg, data...)
-}
-
-func (l *logger) Fatal(msg string, data ...any) {
-	l.fatal(hired.Sprint("Fatal "), time.Now().Format(l.timeFormat), msg, data...)
-}
-
-type WroteLogger interface {
-	Info(msg string, data ...any)
-}

pkg/log/new.go

@@ -1,21 +0,0 @@
-package log
-
-import (
-	"os"
-	"sync"
-)
-
-func New() *logger {
-	return &logger{
-		Mutex:      sync.Mutex{},
-		timeFormat: "2006-01-02T15:04:05",
-		writer:     os.Stdout,
-		level:      LogLevelInfo,
-		debug:      nilLogger,
-		info:       normalLogger,
-		warn:       normalLogger,
-		error:      normalLogger,
-		panic:      panicLogger,
-		fatal:      fatalLogger,
-	}
-}

pkg/model/io.go

@@ -1,19 +0,0 @@
-package model
-
-import "context"
-
-type IOType string
-
-const (
-	Input  IOType = "input"
-	Output IOType = "output"
-)
-
-type IO[T any] interface {
-	ReadData(ctx context.Context, limit int, query map[string]any, fields []string, sort []string) ([]T, error)
-	WriteData(ctx context.Context, items []T) (int, error)
-	ReadMapping(ctx context.Context) (map[string]any, error)
-	WriteMapping(ctx context.Context, mapping map[string]any) error
-	ReadSetting(ctx context.Context) (map[string]any, error)
-	WriteSetting(ctx context.Context, setting map[string]any) error
-}

xes/es6/read.go

@@ -0,0 +1,146 @@
+package es6
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	"time"
+
+	elastic "github.com/elastic/go-elasticsearch/v6"
+	"github.com/elastic/go-elasticsearch/v6/esapi"
+	"github.com/loveuer/esgo2dump/internal/tool"
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/nf/nft/log"
+	"github.com/samber/lo"
+)
+
+func ReadData(ctx context.Context, client *elastic.Client, index string, size, max int, query map[string]any, source []string, sort []string) (<-chan []*model.ESSource, <-chan error) {
+	var (
+		dataCh = make(chan []*model.ESSource)
+		errCh  = make(chan error)
+	)
+
+	go func() {
+		var (
+			err      error
+			resp     *esapi.Response
+			result   = new(model.ESResponseV6)
+			scrollId string
+			total    int
+		)
+
+		defer func() {
+			close(dataCh)
+			close(errCh)
+
+			if scrollId != "" {
+				bs, _ := json.Marshal(map[string]string{
+					"scroll_id": scrollId,
+				})
+
+				var rr *esapi.Response
+
+				if rr, err = client.ClearScroll(
+					client.ClearScroll.WithContext(tool.Timeout(3)),
+					client.ClearScroll.WithBody(bytes.NewReader(bs)),
+				); err != nil {
+					log.Warn("clear scroll id=%s err=%v", scrollId, err)
+					return
+				}
+
+				if rr.StatusCode != 200 {
+					log.Warn("clear scroll id=%s status=%d msg=%s", scrollId, rr.StatusCode, rr.String())
+				}
+			}
+		}()
+
+		if client == nil {
+			errCh <- fmt.Errorf("client is nil")
+		}
+
+		qs := []func(*esapi.SearchRequest){
+			client.Search.WithContext(tool.TimeoutCtx(ctx, 20)),
+			client.Search.WithIndex(index),
+			client.Search.WithSize(int(size)),
+			client.Search.WithFrom(0),
+			client.Search.WithScroll(time.Duration(120) * time.Second),
+		}
+
+		if len(source) > 0 {
+			qs = append(qs, client.Search.WithSourceIncludes(source...))
+		}
+
+		if len(sort) > 0 {
+			sorts := lo.Filter(sort, func(item string, index int) bool {
+				return item != ""
+			})
+
+			if len(sorts) > 0 {
+				qs = append(qs, client.Search.WithSort(sorts...))
+			}
+		}
+
+		if query != nil && len(query) > 0 {
+			queryBs, _ := json.Marshal(map[string]any{"query": query})
+			qs = append(qs, client.Search.WithBody(bytes.NewReader(queryBs)))
+		}
+
+		if resp, err = client.Search(qs...); err != nil {
+			errCh <- err
+			return
+		}
+
+		if resp.StatusCode != 200 {
+			errCh <- fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
+			return
+		}
+
+		decoder := json.NewDecoder(resp.Body)
+		if err = decoder.Decode(result); err != nil {
+			errCh <- err
+			return
+		}
+
+		scrollId = result.ScrollId
+
+		dataCh <- result.Hits.Hits
+		total += len(result.Hits.Hits)
+
+		if len(result.Hits.Hits) < size || (max > 0 && total >= max) {
+			return
+		}
+
+		for {
+			if resp, err = client.Scroll(
+				client.Scroll.WithScrollID(scrollId),
+				client.Scroll.WithScroll(time.Duration(120)*time.Second),
+			); err != nil {
+				errCh <- err
+				return
+			}
+
+			result = new(model.ESResponseV6)
+
+			decoder = json.NewDecoder(resp.Body)
+			if err = decoder.Decode(result); err != nil {
+				errCh <- err
+				return
+			}
+
+			if resp.StatusCode != 200 {
+				errCh <- fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
+				return
+			}
+
+			dataCh <- result.Hits.Hits
+			total += len(result.Hits.Hits)
+
+			if len(result.Hits.Hits) < size || (max > 0 && total >= max) {
+				break
+			}
+		}
+	}()
+
+	return dataCh, errCh
+}

xes/es6/write.go

@@ -0,0 +1,85 @@
+package es6
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+
+	elastic "github.com/elastic/go-elasticsearch/v6"
+	"github.com/elastic/go-elasticsearch/v6/esutil"
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/nf/nft/log"
+)
+
+func WriteData(ctx context.Context, client *elastic.Client, index string, docsCh <-chan []*model.ESSource, logs ...log.WroteLogger) error {
+	var (
+		err     error
+		indexer esutil.BulkIndexer
+		total   = 0
+	)
+
+	for {
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		case docs, ok := <-docsCh:
+			if !ok {
+				return nil
+			}
+
+			if len(docs) == 0 {
+				continue
+			}
+
+			count := 0
+
+			if indexer, err = esutil.NewBulkIndexer(esutil.BulkIndexerConfig{
+				Client:     client,
+				Index:      index,
+				ErrorTrace: true,
+				OnError: func(ctx context.Context, err error) {
+				},
+			}); err != nil {
+				return err
+			}
+
+			for _, doc := range docs {
+				var bs []byte
+
+				if bs, err = json.Marshal(doc.Content); err != nil {
+					return err
+				}
+
+				if err = indexer.Add(context.Background(), esutil.BulkIndexerItem{
+					Action:       "index",
+					Index:        index,
+					DocumentID:   doc.DocId,
+					DocumentType: "_doc",
+					Body:         bytes.NewReader(bs),
+					OnFailure: func(ctx context.Context, item esutil.BulkIndexerItem, item2 esutil.BulkIndexerResponseItem, bulkErr error) {
+					},
+				}); err != nil {
+					return err
+				}
+
+				count++
+			}
+
+			total += count
+
+			if err = indexer.Close(ctx); err != nil {
+				return err
+			}
+
+			stats := indexer.Stats()
+			if stats.NumFailed > 0 {
+				return fmt.Errorf("write to es failed_count=%d bulk_count=%d", stats.NumFailed, count)
+			}
+
+			if len(logs) > 0 && logs[0] != nil {
+				logs[0].Info("Dump: succeed=%d total=%d docs succeed!!!", count, total)
+			}
+		}
+	}
+}

xes/es7/client.go

@@ -16,39 +16,37 @@ import (
 	"github.com/samber/lo"
 )
 
-// NewClient
+func NewClient(ctx context.Context, url *url.URL) (*elastic.Client, error) {
-// new esv7 client
-// uri example:
-//   - http://127.0.0.1:9200
-//   - https://<username>:<password>@node1.dev:9200,node2.dev:19200,node3.dev:29200
-func NewClient(ctx context.Context, uri string) (*elastic.Client, error) {
 	var (
 		err         error
-		username string
+		urlUsername string
-		password string
+		urlPassword string
 		client      *elastic.Client
-		ins      *url.URL
+		errCh       = make(chan error)
-	)
+		cliCh       = make(chan *elastic.Client)
-
+		endpoints   = lo.Map(
-	if ins, err = url.Parse(uri); err != nil {
+			strings.Split(url.Host, ","),
-		return nil, err
-	}
-
-	endpoints := lo.Map(
-		strings.Split(ins.Host, ","),
 			func(item string, index int) string {
-			return fmt.Sprintf("%s://%s", ins.Scheme, item)
+				return fmt.Sprintf("%s://%s", url.Scheme, item)
 			},
 		)
+	)
 
-	if ins.User != nil {
+	if url.User != nil {
-		username = ins.User.Username()
+		urlUsername = url.User.Username()
-		password, _ = ins.User.Password()
+		if p, ok := url.User.Password(); ok {
+			urlPassword = p
+		}
 	}
 
-	query := ins.Query()
+	ncFunc := func(endpoints []string, username, password string) {
+		var (
+			err      error
+			cli      *elastic.Client
+			infoResp *esapi.Response
+		)
 
-	if client, err = elastic.NewClient(
+		if cli, err = elastic.NewClient(
 			elastic.Config{
 				Addresses:     endpoints,
 				Username:      username,
@@ -61,23 +59,35 @@ func NewClient(ctx context.Context, uri string) (*elastic.Client, error) {
 					TLSClientConfig: &tls.Config{InsecureSkipVerify: true},
 					DialContext:     (&net.Dialer{Timeout: 10 * time.Second}).DialContext,
 				},
-			DiscoverNodesOnStart: lo.If(query.Get("sniff") == "true", true).Else(false),
 			},
 		); err != nil {
-		return nil, err
+			errCh <- err
+			return
 		}
 
-	if query.Get("ping") != "false" {
+		if infoResp, err = cli.Info(); err != nil {
-		var res *esapi.Response
+			errCh <- err
-		if res, err = client.Ping(client.Ping.WithContext(tool.TimeoutCtx(ctx, 5))); err != nil {
+			return
-			return nil, err
 		}
 
-		if res.StatusCode != 200 {
+		if infoResp.StatusCode != 200 {
-			err = fmt.Errorf("ping es server response: %s", res.String())
+			err = fmt.Errorf("info es7 status=%d", infoResp.StatusCode)
-			return nil, err
+			errCh <- err
-		}
+			return
 		}
 
+		cliCh <- cli
+	}
+
+	go ncFunc(endpoints, urlUsername, urlPassword)
+	timeout := tool.TimeoutCtx(ctx, 10)
+
+	select {
+	case <-timeout.Done():
+		return nil, fmt.Errorf("dial es=%v err=%v", endpoints, context.DeadlineExceeded)
+	case client = <-cliCh:
 		return client, nil
+	case err = <-errCh:
+		return nil, err
+	}
 }

xes/es7/client_test.go

@@ -1,6 +1,7 @@
 package es7
 
 import (
+	"net/url"
 	"testing"
 
 	"github.com/loveuer/esgo2dump/internal/tool"
@@ -8,8 +9,9 @@ import (
 
 func TestNewClient(t *testing.T) {
 	uri := "http://es1.dev:9200,es2.dev:9200"
+	ins, _ := url.Parse(uri)
 
-	c, err := NewClient(tool.Timeout(5), uri)
+	c, err := NewClient(tool.Timeout(5), ins)
 	if err != nil {
 		t.Fatal(err.Error())
 	}

xes/es7/read.go

@@ -5,272 +5,260 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
-	"github.com/elastic/go-elasticsearch/v7/esutil"
-	"github.com/loveuer/esgo2dump/internal/opt"
-	"github.com/loveuer/esgo2dump/pkg/log"
 	"time"
 
 	elastic "github.com/elastic/go-elasticsearch/v7"
 	"github.com/elastic/go-elasticsearch/v7/esapi"
 	"github.com/loveuer/esgo2dump/internal/tool"
-	"github.com/loveuer/esgo2dump/pkg/model"
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/nf/nft/log"
 	"github.com/samber/lo"
 )
 
-type streamer struct {
+// ReadData
-	ctx    context.Context
+// @param[source]: a list of include fields to extract and return from the _source field.
-	client *elastic.Client
+// @param[sort]:   a list of <field>:<direction> pairs.
-	index  string
+func ReadData(ctx context.Context, client *elastic.Client, index string, size, max int, query map[string]any, source []string, sort []string) (<-chan []*model.ESSource, <-chan error) {
-	scroll string
-}
-
-// ReadData implements model.IO.
-func (s *streamer) ReadData(ctx context.Context, limit int, query map[string]any, fields []string, sort []string) ([]map[string]any, error) {
 	var (
-		err    error
+		dataCh = make(chan []*model.ESSource)
-		qs     []func(*esapi.SearchRequest)
+		errCh  = make(chan error)
-		resp   *esapi.Response
-		result = new(model.ESResponseV7[map[string]any])
 	)
 
-	if limit == 0 {
+	go func() {
-		return nil, nil
-	}
-
-	if s.scroll != "" {
-		if resp, err = s.client.Scroll(
-			s.client.Scroll.WithContext(tool.TimeoutCtx(s.ctx)),
-			s.client.Scroll.WithScrollID(s.scroll),
-			s.client.Scroll.WithScroll(35*time.Second),
-		); err != nil {
-			return nil, err
-		}
-
-		goto HandleResp
-	}
-
-	qs = []func(*esapi.SearchRequest){
-		s.client.Search.WithContext(tool.TimeoutCtx(s.ctx)),
-		s.client.Search.WithIndex(s.index),
-		s.client.Search.WithSize(limit),
-		s.client.Search.WithScroll(35 * time.Second),
-	}
-
-	if len(fields) > 0 {
-		qs = append(qs, s.client.Search.WithSourceIncludes(fields...))
-	}
-
-	if len(sort) > 0 {
-		qs = append(qs, s.client.Search.WithSort(sort...))
-	}
-
-	if len(query) > 0 {
-		queryBs, err := json.Marshal(map[string]any{"query": query})
-		if err != nil {
-			return nil, err
-		}
-
-		qs = append(qs, s.client.Search.WithBody(bytes.NewReader(queryBs)))
-	}
-
-	if resp, err = s.client.Search(qs...); err != nil {
-		return nil, err
-	}
-
-HandleResp:
-
-	if resp.StatusCode != 200 {
-		return nil, fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
-	}
-
-	if err = json.NewDecoder(resp.Body).Decode(result); err != nil {
-		return nil, err
-	}
-
-	s.scroll = result.ScrollId
-
-	return lo.Slice(
-		lo.Map(
-			result.Hits.Hits,
-			func(item *model.ESSource[map[string]any], _ int) map[string]any {
-				return item.Content
-			},
-		),
-		0,
-		limit,
-	), nil
-}
-
-// WriteData implements model.IO.
-func (s *streamer) WriteData(ctx context.Context, items []map[string]any) (int, error) {
 		var (
 			err      error
-		indexer esutil.BulkIndexer
+			resp     *esapi.Response
+			result   = new(model.ESResponseV7)
+			scrollId string
 			total    int
 		)
 
-	if len(items) == 0 {
+		defer func() {
-		return 0, nil
+			close(dataCh)
+			close(errCh)
+
+			if scrollId != "" {
+				bs, _ := json.Marshal(map[string]string{
+					"scroll_id": scrollId,
+				})
+
+				var rr *esapi.Response
+
+				if rr, err = client.ClearScroll(
+					client.ClearScroll.WithContext(tool.Timeout(3)),
+					client.ClearScroll.WithBody(bytes.NewReader(bs)),
+				); err != nil {
+					log.Warn("clear scroll id=%s err=%v", scrollId, err)
+					return
 				}
 
-	count := 0
+				if rr.StatusCode != 200 {
+					log.Warn("clear scroll id=%s status=%d msg=%s", scrollId, rr.StatusCode, rr.String())
+				}
+			}
+		}()
 
-	if indexer, err = esutil.NewBulkIndexer(esutil.BulkIndexerConfig{
+		if client == nil {
-		NumWorkers:    0,
+			errCh <- fmt.Errorf("client is nil")
-		FlushBytes:    0,
-		FlushInterval: 0,
-		Client:        s.client,
-		Decoder:       nil,
-		OnError: func(ctx context.Context, err error) {
-			log.Error("es7.writer: on error log, err = %s", err.Error())
-		},
-		Index:               s.index,
-		ErrorTrace:          true,
-		FilterPath:          []string{},
-		Header:              map[string][]string{},
-		Human:               false,
-		Pipeline:            "",
-		Pretty:              false,
-		Refresh:             "",
-		Routing:             "",
-		Source:              []string{},
-		SourceExcludes:      []string{},
-		SourceIncludes:      []string{},
-		Timeout:             0,
-		WaitForActiveShards: "",
-	}); err != nil {
-		return 0, err
 		}
 
-	for _, item := range items {
+		qs := []func(*esapi.SearchRequest){
-		var bs []byte
+			client.Search.WithContext(tool.TimeoutCtx(ctx, 20)),
-
+			client.Search.WithIndex(index),
-		if bs, err = json.Marshal(item); err != nil {
+			client.Search.WithSize(size),
-			return 0, err
+			client.Search.WithFrom(0),
+			client.Search.WithScroll(time.Duration(120) * time.Second),
 		}
 
-		if err = indexer.Add(context.Background(), esutil.BulkIndexerItem{
+		if len(source) > 0 {
-			Action: "index",
+			qs = append(qs, client.Search.WithSourceIncludes(source...))
-			Index:  s.index,
-			Body:   bytes.NewReader(bs),
-			OnFailure: func(ctx context.Context, item esutil.BulkIndexerItem, item2 esutil.BulkIndexerResponseItem, bulkErr error) {
-				log.Error("es7.writer: on failure err log, err = %s", bulkErr.Error())
-			},
-		}); err != nil {
-			return 0, err
 		}
 
-		count++
+		if len(sort) > 0 {
+			sorts := lo.Filter(sort, func(item string, index int) bool {
+				return item != ""
+			})
+
+			if len(sorts) > 0 {
+				qs = append(qs, client.Search.WithSort(sorts...))
+			}
 		}
 
-	total += count
+		if query != nil && len(query) > 0 {
-
+			queryBs, _ := json.Marshal(map[string]any{"query": query})
-	if err = indexer.Close(ctx); err != nil {
+			qs = append(qs, client.Search.WithBody(bytes.NewReader(queryBs)))
-		return 0, err
 		}
 
-	stats := indexer.Stats()
+		if resp, err = client.Search(qs...); err != nil {
-
+			errCh <- err
-	return len(items) - int(stats.NumFailed), nil
+			return
 		}
 
-func (s *streamer) ReadMapping(ctx context.Context) (map[string]any, error) {
+		if resp.StatusCode != 200 {
-	r, err := s.client.Indices.GetMapping(
+			errCh <- fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
-		s.client.Indices.GetMapping.WithIndex(s.index),
+			return
+		}
+
+		decoder := json.NewDecoder(resp.Body)
+		if err = decoder.Decode(result); err != nil {
+			errCh <- err
+			return
+		}
+
+		scrollId = result.ScrollId
+
+		dataCh <- result.Hits.Hits
+		total += len(result.Hits.Hits)
+
+		if len(result.Hits.Hits) < size || (max > 0 && total >= max) {
+			return
+		}
+
+		for {
+			if resp, err = client.Scroll(
+				client.Scroll.WithScrollID(scrollId),
+				client.Scroll.WithScroll(time.Duration(120)*time.Second),
+			); err != nil {
+				errCh <- err
+				return
+			}
+
+			result = new(model.ESResponseV7)
+
+			decoder = json.NewDecoder(resp.Body)
+			if err = decoder.Decode(result); err != nil {
+				errCh <- err
+				return
+			}
+
+			if resp.StatusCode != 200 {
+				errCh <- fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
+				return
+			}
+
+			dataCh <- result.Hits.Hits
+			total += len(result.Hits.Hits)
+
+			if len(result.Hits.Hits) < size || (max > 0 && total >= max) {
+				break
+			}
+		}
+	}()
+
+	return dataCh, errCh
+}
+
+// ReadDataV2 es7 read data
+// Deprecated: bug, when can't sort by _id
+/*
+	- @param[source]: a list of include fields to extract and return from the _source field.
+    - @param[sort]:   a list of <field>:<direction> pairs.
+*/
+func ReadDataV2(
+	ctx context.Context,
+	client *elastic.Client,
+	index string,
+	size, max int,
+	query map[string]any,
+	source []string,
+	sort []string,
+) (<-chan []*model.ESSource, <-chan error) {
+	var (
+		dataCh = make(chan []*model.ESSource)
+		errCh  = make(chan error)
 	)
-	if err != nil {
-		return nil, err
-	}
 
-	if r.StatusCode != 200 {
+	log.Debug("es7.ReadDataV2: arg.index = %s, arg.size = %d, arg.max = %d", index, size, max)
-		return nil, fmt.Errorf("status=%d, msg=%s", r.StatusCode, r.String())
-	}
 
-	m := make(map[string]any)
+	go func() {
-	decoder := json.NewDecoder(r.Body)
-	if err = decoder.Decode(&m); err != nil {
-		return nil, err
-	}
-
-	return m, nil
-}
-
-func (s *streamer) WriteMapping(ctx context.Context, mapping map[string]any) error {
 		var (
 			err         error
 			bs          []byte
-		result *esapi.Response
+			resp        *esapi.Response
+			searchAfter     = make([]any, 0)
+			total       int = 0
+			body            = make(map[string]any)
+			qs          []func(request *esapi.SearchRequest)
 		)
 
-	for idxKey := range mapping {
+		if sort == nil {
-		if bs, err = json.Marshal(mapping[idxKey]); err != nil {
+			sort = []string{}
-			return err
 		}
 
-		if result, err = s.client.Indices.Create(
+		if len(query) > 0 {
-			s.index,
+			body["query"] = query
-			s.client.Indices.Create.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
-			s.client.Indices.Create.WithBody(bytes.NewReader(bs)),
-		); err != nil {
-			return err
 		}
 
-		if result.StatusCode != 200 {
+		sort = append(sort, "_id:ASC")
-			return fmt.Errorf("status=%d, msg=%s", result.StatusCode, result.String())
+
-		}
+		sorts := lo.Filter(sort, func(item string, index int) bool {
+			return item != ""
+		})
+
+		defer func() {
+			close(dataCh)
+			close(errCh)
+		}()
+
+		for {
+			finaSize := tool.CalcSize(size, max, total)
+			qs = []func(*esapi.SearchRequest){
+				client.Search.WithContext(tool.TimeoutCtx(ctx, 30)),
+				client.Search.WithIndex(index),
+				client.Search.WithSize(finaSize),
+				client.Search.WithSort(sorts...),
 			}
 
-	return nil
+			if len(source) > 0 {
+				qs = append(qs, client.Search.WithSourceIncludes(source...))
 			}
 
-func (s *streamer) ReadSetting(ctx context.Context) (map[string]any, error) {
+			delete(body, "search_after")
-	r, err := s.client.Indices.GetSettings(
+			if len(searchAfter) > 0 {
-		s.client.Indices.GetSettings.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+				body["search_after"] = searchAfter
-		s.client.Indices.GetSettings.WithIndex(s.index),
-	)
-	if err != nil {
-		return nil, err
 			}
 
-	if r.StatusCode != 200 {
+			if bs, err = json.Marshal(body); err != nil {
-		return nil, fmt.Errorf("status=%d, msg=%s", r.StatusCode, r.String())
+				errCh <- err
+				return
 			}
 
-	m := make(map[string]any)
+			log.Debug("es7.ReadDataV2: search request size = %d, body = %s", finaSize, string(bs))
-	decoder := json.NewDecoder(r.Body)
+
-	if err = decoder.Decode(&m); err != nil {
+			qs = append(qs, client.Search.WithBody(bytes.NewReader(bs)))
-		return nil, err
+			if resp, err = client.Search(qs...); err != nil {
+				errCh <- err
+				return
 			}
 
-	return m, nil
+			if resp.StatusCode != 200 {
+				errCh <- fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
+				return
 			}
 
-func (s *streamer) WriteSetting(ctx context.Context, setting map[string]any) error {
+			result := new(model.ESResponseV7)
-	var (
+			decoder := json.NewDecoder(resp.Body)
-		err    error
+			if err = decoder.Decode(result); err != nil {
-		bs     []byte
+				errCh <- err
-		result *esapi.Response
+				return
-	)
-
-	if bs, err = json.Marshal(setting); err != nil {
-		return err
 			}
 
-	if result, err = s.client.Indices.PutSettings(
+			if resp.StatusCode != 200 {
-		bytes.NewReader(bs),
+				errCh <- fmt.Errorf("resp status=%d, resp=%s", resp.StatusCode, resp.String())
-		s.client.Indices.PutSettings.WithContext(tool.TimeoutCtx(ctx, opt.Timeout)),
+				return
-	); err != nil {
-		return err
 			}
 
-	if result.StatusCode != 200 {
+			dataCh <- result.Hits.Hits
-		return fmt.Errorf("status=%d, msg=%s", result.StatusCode, result.String())
+			log.Debug("es7.ReadDataV2: search response hits = %d", len(result.Hits.Hits))
+			total += len(result.Hits.Hits)
+
+			if len(result.Hits.Hits) < size || (max > 0 && total >= max) {
+				break
 			}
 
-	return nil
+			searchAfter = result.Hits.Hits[len(result.Hits.Hits)-1].Sort
 		}
+	}()
 
-func NewStreamer(ctx context.Context, client *elastic.Client, index string) (model.IO[map[string]any], error) {
+	return dataCh, errCh
-	s := &streamer{ctx: ctx, client: client, index: index}
-	return s, nil
 }

xes/es7/write.go

@@ -8,46 +8,38 @@ import (
 
 	elastic "github.com/elastic/go-elasticsearch/v7"
 	"github.com/elastic/go-elasticsearch/v7/esutil"
-	"github.com/loveuer/esgo2dump/pkg/log"
+	"github.com/loveuer/esgo2dump/model"
-	"github.com/loveuer/esgo2dump/pkg/model"
+	"github.com/loveuer/nf/nft/log"
 )
 
-func WriteData[T any](ctx context.Context, client *elastic.Client, index string, docs ...*model.ESSource[T]) error {
+func WriteData(ctx context.Context, client *elastic.Client, index string, docsCh <-chan []*model.ESSource, logs ...log.WroteLogger) error {
 	var (
 		err     error
 		indexer esutil.BulkIndexer
 		total   int
 	)
 
-	if len(docs) == 0 {
+	for {
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		case docs, ok := <-docsCh:
+			if !ok {
 				return nil
 			}
 
+			if len(docs) == 0 {
+				continue
+			}
+
 			count := 0
 
 			if indexer, err = esutil.NewBulkIndexer(esutil.BulkIndexerConfig{
-		NumWorkers:    0,
-		FlushBytes:    0,
-		FlushInterval: 0,
 				Client:     client,
-		Decoder:       nil,
-		OnError: func(ctx context.Context, err error) {
-			log.Error("es7.writer: on error log, err = %s", err.Error())
-		},
 				Index:      index,
 				ErrorTrace: true,
-		FilterPath:          []string{},
+				OnError: func(ctx context.Context, err error) {
-		Header:              map[string][]string{},
+				},
-		Human:               false,
-		Pipeline:            "",
-		Pretty:              false,
-		Refresh:             "",
-		Routing:             "",
-		Source:              []string{},
-		SourceExcludes:      []string{},
-		SourceIncludes:      []string{},
-		Timeout:             0,
-		WaitForActiveShards: "",
 			}); err != nil {
 				return err
 			}
@@ -65,7 +57,6 @@ func WriteData[T any](ctx context.Context, client *elastic.Client, index string,
 					DocumentID: doc.DocId,
 					Body:       bytes.NewReader(bs),
 					OnFailure: func(ctx context.Context, item esutil.BulkIndexerItem, item2 esutil.BulkIndexerResponseItem, bulkErr error) {
-				log.Error("es7.writer: on failure err log, err = %s", bulkErr.Error())
 					},
 				}); err != nil {
 					return err
@@ -85,5 +76,9 @@ func WriteData[T any](ctx context.Context, client *elastic.Client, index string,
 				return fmt.Errorf("write to es failed_count=%d bulk_count=%d", stats.NumFailed, count)
 			}
 
-	return nil
+			if len(logs) > 0 && logs[0] != nil {
+				logs[0].Info("Dump: succeed=%d total=%d docs succeed!!!", count, total)
+			}
+		}
+	}
 }