refactory: read,write use channel

.github/workflows/build.yml

@@ -42,6 +42,17 @@ jobs:
       - name: build darwin arm64
         run: CGO_ENABLE=0 GOOS=darwin GOARCH=arm64 go build -ldflags='-s -w' -o dist/esgo2dump_${{ github.ref_name }}_darwin_arm64 .
 
+
+      - name: run upx
+        uses: crazy-max/ghaction-upx@v3
+        with:
+          version: latest
+          args: --best --ultra-brute
+          files: |
+            dist/esgo2dump_${{ github.ref_name }}_linux_amd64
+            dist/esgo2dump_${{ github.ref_name }}_linux_arm64
+            dist/esgo2dump_${{ github.ref_name }}_windows_amd64.exe
+
       - name: create releases
         id: create_releases
         uses: "marvinpinto/action-automatic-releases@latest"

.gitignore

@@ -8,3 +8,4 @@
 *test.json
 *.txt
 dist
+xtest

go.mod

@@ -5,6 +5,7 @@ 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 v1.16.0
 	github.com/samber/lo v1.39.0
 	github.com/sirupsen/logrus v1.9.3
 	github.com/spf13/cobra v1.8.0
@@ -12,6 +13,8 @@ require (
 
 require (
 	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
 	github.com/spf13/pflag v1.0.5 // indirect
 	github.com/stretchr/testify v1.8.4 // indirect
 	golang.org/x/exp v0.0.0-20220303212507-bbda1eaf7a17 // indirect

go.sum

@@ -6,8 +6,15 @@ github.com/elastic/go-elasticsearch/v6 v6.8.10 h1:2lN0gJ93gMBXvkhwih5xquldszpm8F
 github.com/elastic/go-elasticsearch/v6 v6.8.10/go.mod h1:UwaDJsD3rWLM5rKNFzv9hgox93HoX8utj1kxD9aFUcI=
 github.com/elastic/go-elasticsearch/v7 v7.17.10 h1:TCQ8i4PmIJuBunvBS6bwT2ybzVFxxUhhltAs3Gyu1yo=
 github.com/elastic/go-elasticsearch/v7 v7.17.10/go.mod h1:OJ4wdbtDNk5g503kvlHLyErCgQwwzmDtaFC4XyOxXA4=
+github.com/fatih/color v1.16.0 h1:zmkK9Ngbjj+K0yRhTVONQh1p/HknKYSlNT+vZCzyokM=
+github.com/fatih/color v1.16.0/go.mod h1:fL2Sau1YI5c0pdGEVCbKQbLXB6edEj1ZgiY4NijnWvE=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
+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=
+github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
+github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
@@ -26,6 +33,8 @@ github.com/stretchr/testify v1.8.4/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXl
 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/sys v0.0.0-20220715151400-c0bba94af5f8/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+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 v0.14.0 h1:Vz7Qs629MkJkGyHxUlRHizWJRG2j8fbQKjELVSNhy7Q=
 golang.org/x/sys v0.14.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=

internal/cmd/run.go

@@ -6,6 +6,8 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
 	"net/url"
 	"os"
 	"strings"
@@ -52,12 +54,12 @@ func run(cmd *cobra.Command, args []string) error {
 	if opt.Debug {
 		logrus.SetLevel(logrus.DebugLevel)
 		logrus.SetReportCaller(true)
-		logrus.SetFormatter(&logrus.TextFormatter{})
+		logrus.SetFormatter(&logrus.JSONFormatter{})
 	}
 
 	if f_version {
-		logrus.Infof("esgo2dump (Version: %s)", opt.Version)
+		fmt.Printf("esgo2dump (Version: %s)\n", opt.Version)
-		return nil
+		os.Exit(0)
 	}
 
 	if err = check(cmd); err != nil {
@@ -91,7 +93,7 @@ func run(cmd *cobra.Command, args []string) error {
 			return err
 		}
 
-		logrus.Info("Dump: write data succeed!!!")
+		log.Info("Dump: write data succeed!!!")
 
 		return nil
 	case "mapping":
@@ -104,7 +106,7 @@ func run(cmd *cobra.Command, args []string) error {
 			return err
 		}
 
-		logrus.Info("Dump: write mapping succeed!!!")
+		log.Info("Dump: write mapping succeed!!!")
 
 		return nil
 	case "setting":
@@ -117,7 +119,7 @@ func run(cmd *cobra.Command, args []string) error {
 			return err
 		}
 
-		logrus.Info("Dump: write setting succeed!!!")
+		log.Info("Dump: write setting succeed!!!")
 
 		return nil
 	default:
@@ -128,8 +130,6 @@ func run(cmd *cobra.Command, args []string) error {
 func executeData(ctx context.Context, input, output interfaces.DumpIO) error {
 	var (
 		err     error
-		ch      = make(chan []*interfaces.ESSource, 1)
-		errCh   = make(chan error)
 		queries = make([]map[string]any, 0)
 		sources = make([]string, 0)
 	)
@@ -187,80 +187,56 @@ func executeData(ctx context.Context, input, output interfaces.DumpIO) error {
 		queries = append(queries, nil)
 	}
 
-	go func(c context.Context) {
 	var (
-			lines []*interfaces.ESSource
+		ok   bool
+		docs []*model.ESSource
+		dch  <-chan []*model.ESSource
+		ech  <-chan error
+
+		e2ch = make(chan error)
+		wch  = make(chan []*model.ESSource)
 	)
 
+	go func() {
 		defer func() {
-			close(ch)
+			close(wch)
+			close(e2ch)
 		}()
 
-	Loop:
+		if err = output.WriteData(ctx, wch); err != nil {
+			e2ch <- err
+		}
+	}()
+
+	log.Info("Query: got queries=%d", len(queries))
+
+Loop:
 	for _, query := range queries {
-			for {
+		dch, ech = input.ReadData(ctx, f_limit, query, sources)
-				select {
-				case <-c.Done():
-					return
-				default:
-					if lines, err = input.ReadData(c, f_limit, query, sources); err != nil {
-						errCh <- err
-						return
-					}
-
-					logrus.Debugf("executeData: input read_data got lines=%d", len(lines))
-
-					if len(lines) == 0 {
-						input.ResetOffset()
-						if query != nil {
-							bs, _ := json.Marshal(query)
-							logrus.Infof("Dump: query_file query=%s read done!!!", string(bs))
-						}
-						continue Loop
-					}
-
-					ch <- lines
-				}
-			}
-		}
-	}(ctx)
-
-	var (
-		succeed int
-		total   int
-		docs    []*interfaces.ESSource
-		ok      bool
-	)
 
 		for {
 			select {
 			case <-ctx.Done():
-		case err = <-errCh:
+				return ctx.Err()
-			return err
+			case err, ok = <-ech:
-		case docs, ok = <-ch:
+				if err != nil {
-			if !ok {
 					return err
 				}
 
-			if len(docs) == 0 {
+				continue Loop
+			case err, _ = <-e2ch:
+				return err
+			case docs, ok = <-dch:
+				if !ok || len(docs) == 0 {
+					continue Loop
+				}
+
+				wch <- docs
+			}
+		}
+	}
+
 	return nil
-			}
-
-			if succeed, err = output.WriteData(ctx, docs); err != nil {
-				return err
-			}
-
-			logrus.Debugf("executeData: output write_data succeed lines=%d", succeed)
-
-			if succeed != len(docs) {
-				return fmt.Errorf("cmd.run: got lines=%d, only succeed=%d", len(docs), succeed)
-			}
-
-			total += succeed
-
-			logrus.Infof("Dump: succeed=%d total=%d docs succeed!!!", succeed, total)
-		}
-	}
 }
 
 func newIO(source string, ioType interfaces.IO, esv string) (interfaces.DumpIO, error) {
@@ -271,39 +247,61 @@ func newIO(source string, ioType interfaces.IO, esv string) (interfaces.DumpIO,
 		qm   = make(map[string]any)
 	)
 
-	logrus.Debugf("newIO.%s: source string=%s", ioType.Code(), source)
+	logrus.
+		WithField("action", "new_io").
+		WithField("type", ioType.Code()).
+		WithField("source", source).
+		WithField("es_version", esv).
+		Debug()
 
 	if iurl, err = url.Parse(source); err != nil {
-		logrus.Debugf("newIO.%s: url parse source err=%v", ioType.Code(), err)
+		logrus.
+			WithField("action", "new_io url parse error").
+			WithField("type", ioType.Code()).
+			WithField("source", source).
+			WithField("err", err).
+			Debug()
 		goto ClientByFile
 	}
 
 	if !(iurl.Scheme == "http" || iurl.Scheme == "https") {
-		logrus.Debugf("newIO.%s: url scheme=%s invalid", ioType.Code(), iurl.Scheme)
+		logrus.
+			WithField("action", "new_io url scheme error").
+			WithField("type", ioType.Code()).
+			WithField("source", source).
+			WithField("scheme", iurl.Scheme).
+			Debug()
 		goto ClientByFile
 	}
 
 	if iurl.Host == "" {
-		logrus.Debugf("newIO.%s: url host empty", ioType.Code())
+		logrus.
+			WithField("action", "new_io url host empty").
+			WithField("type", ioType.Code()).
+			WithField("source", source).
+			Debug()
 		goto ClientByFile
 	}
 
 	if ioType == interfaces.IOInput && f_query != "" {
 		if err = json.Unmarshal([]byte(f_query), &qm); err != nil {
-			logrus.Debugf("newIO.%s: query=%s invalid to map[string]any", ioType.Code(), f_query)
+			logrus.
+				WithField("action", "new_io query string invalid").
+				WithField("type", ioType.Code()).
+				WithField("source", source).
+				WithField("query", f_query).
+				Debug()
 			return nil, fmt.Errorf("invalid query err=%v", err)
 		}
 	}
 
-	logrus.Debugf("newIO.%s: source as url=%+v version=%s", ioType.Code(), *iurl, esv)
-
 	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 comming soon")
+		return nil, errors.New("es version 8 coming soon")
 	default:
 		return nil, fmt.Errorf("unknown es version=%s", esv)
 	}

internal/interfaces/dumpio.go

@@ -1,12 +1,13 @@
 package interfaces
 
-import "context"
+import (
+	"context"
+	"github.com/loveuer/esgo2dump/model"
+)
 
 type DumpIO interface {
-	ReadData(context.Context, int, map[string]any, []string) ([]*ESSource, error)
+	ReadData(ctx context.Context, size int, query map[string]any, includeFields []string) (<-chan []*model.ESSource, <-chan error)
-	WriteData(ctx context.Context, docs []*ESSource) (int, error)
+	WriteData(ctx context.Context, docsCh <-chan []*model.ESSource) error
-
-	ResetOffset()
 
 	ReadMapping(context.Context) (map[string]any, error)
 	WriteMapping(context.Context, map[string]any) error

internal/opt/version.go

@@ -1,3 +1,3 @@
 package opt
 
-const Version = "v0.1.2"
+const Version = "v0.2.1"

internal/xes/xes.go

@@ -1,351 +0,0 @@
-package xes
-
-import (
-	"bytes"
-	"context"
-	"crypto/tls"
-	"encoding/json"
-	"fmt"
-	"net"
-	"net/http"
-	"net/url"
-	"strings"
-	"time"
-
-	elastic "github.com/elastic/go-elasticsearch/v7"
-	"github.com/elastic/go-elasticsearch/v7/esapi"
-	"github.com/elastic/go-elasticsearch/v7/esutil"
-	"github.com/loveuer/esgo2dump/internal/interfaces"
-	"github.com/loveuer/esgo2dump/internal/opt"
-	"github.com/loveuer/esgo2dump/internal/util"
-	"github.com/sirupsen/logrus"
-)
-
-func NewClient(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
-		}
-	}
-
-	logrus.Debugf("xes.NewClient: endpoint=%s index=%s (username=%s password=%s)", 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 {
-			logrus.Debugf("xes.NewClient: elastic new client with endpont=%s err=%v", endpoints, err)
-			errCh <- err
-			return
-		}
-
-		if infoResp, err = cli.Info(); err != nil {
-			logrus.Debugf("xes.NewClient: ping err=%v", err)
-			errCh <- err
-			return
-		}
-
-		if infoResp.StatusCode != 200 {
-			err = fmt.Errorf("info xes status=%d", infoResp.StatusCode)
-			logrus.Debugf("xes.NewClient: status err=%v", err)
-			errCh <- err
-			return
-		}
-
-		cliCh <- cli
-	}
-
-	go ncFunc([]string{address}, urlUsername, urlPassword, urlIndex)
-
-	select {
-	case <-util.Timeout(10).Done():
-		return nil, fmt.Errorf("dial es=%s err=%v", address, context.DeadlineExceeded)
-	case c := <-cliCh:
-		return &client{c: c, index: urlIndex, iot: iot}, nil
-	case e := <-errCh:
-		return nil, e
-	}
-}
-
-type client struct {
-	c        *elastic.Client
-	iot      interfaces.IO
-	index    string
-	scrollId string
-}
-
-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) ResetOffset() {
-	defer func() {
-		c.scrollId = ""
-	}()
-
-	bs, _ := json.Marshal(map[string]string{
-		"scroll_id": c.scrollId,
-	})
-
-	rr, err := c.c.ClearScroll(
-		c.c.ClearScroll.WithContext(util.Timeout(3)),
-		c.c.ClearScroll.WithBody(bytes.NewReader(bs)),
-	)
-	if err != nil {
-		logrus.Warnf("ResetOffset: clear scroll id=%s err=%v", c.scrollId, err)
-		return
-	}
-
-	if rr.StatusCode != 200 {
-		logrus.Warnf("ResetOffset: clear scroll id=%s msg=%s", c.scrollId, rr.String())
-	}
-}
-func (c *client) WriteData(ctx context.Context, docs []*interfaces.ESSource) (int, error) {
-	var (
-		err     error
-		indexer esutil.BulkIndexer
-		count   int
-		be      error
-	)
-	if indexer, err = esutil.NewBulkIndexer(esutil.BulkIndexerConfig{
-		Client:  c.c,
-		Index:   c.index,
-		Refresh: "",
-	}); err != nil {
-		return 0, err
-	}
-
-	for _, doc := range docs {
-		var bs []byte
-
-		if bs, err = json.Marshal(doc.Content); err != nil {
-			return 0, err
-		}
-
-		if err = indexer.Add(context.Background(), esutil.BulkIndexerItem{
-			Action:     "index",
-			Index:      c.index,
-			DocumentID: doc.DocId,
-			Body:       bytes.NewReader(bs),
-			OnFailure: func(ctx context.Context, item esutil.BulkIndexerItem, item2 esutil.BulkIndexerResponseItem, bulkErr error) {
-				be = bulkErr
-			},
-		}); err != nil {
-			return 0, err
-		}
-		count++
-	}
-
-	if err = indexer.Close(util.TimeoutCtx(ctx, opt.Timeout)); err != nil {
-		return 0, err
-	}
-
-	if be != nil {
-		return 0, be
-	}
-
-	stats := indexer.Stats()
-	if stats.NumFailed > 0 {
-		return count, fmt.Errorf("write to xes failed_count=%d bulk_count=%d", stats.NumFailed, count)
-	}
-
-	return count, nil
-}
-
-func (c *client) ReadData(ctx context.Context, i int, query map[string]any, source []string) ([]*interfaces.ESSource, error) {
-	var (
-		err    error
-		resp   *esapi.Response
-		result = new(interfaces.ESResponse)
-	)
-
-	if c.scrollId == "" {
-		qs := []func(*esapi.SearchRequest){
-			c.c.Search.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-			c.c.Search.WithIndex(c.index),
-			c.c.Search.WithSize(i),
-			c.c.Search.WithFrom(0),
-			c.c.Search.WithScroll(time.Duration(opt.Timeout*2) * time.Second),
-		}
-
-		if len(source) > 0 {
-			qs = append(qs, c.c.Search.WithSourceIncludes(source...))
-		}
-
-		if query != nil && len(query) > 0 {
-			queryBs, _ := json.Marshal(map[string]any{"query": query})
-			qs = append(qs, c.c.Search.WithBody(bytes.NewReader(queryBs)))
-		}
-
-		if resp, err = c.c.Search(qs...); err != nil {
-			return nil, err
-		}
-
-		if resp.StatusCode != 200 {
-			return nil, fmt.Errorf(resp.String())
-		}
-
-		decoder := json.NewDecoder(resp.Body)
-		if err = decoder.Decode(result); err != nil {
-			return nil, err
-		}
-
-		c.scrollId = result.ScrollId
-
-		return result.Hits.Hits, nil
-	}
-
-	if resp, err = c.c.Scroll(
-		c.c.Scroll.WithScrollID(c.scrollId),
-		c.c.Scroll.WithScroll(time.Duration(opt.Timeout*2)*time.Second),
-	); err != nil {
-		return result.Hits.Hits, nil
-	}
-
-	decoder := json.NewDecoder(resp.Body)
-	if err = decoder.Decode(result); err != nil {
-		return nil, err
-	}
-
-	return result.Hits.Hits, nil
-}
-
-func (c *client) ReadMapping(ctx context.Context) (map[string]any, error) {
-	r, err := c.c.Indices.GetMapping(
-		c.c.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.c.Indices.Create(
-			c.index,
-			c.c.Indices.Create.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-			c.c.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.c.Indices.GetSettings(
-		c.c.Indices.GetSettings.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-		c.c.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.c.Indices.PutSettings(
-		bytes.NewReader(bs),
-		c.c.Indices.PutSettings.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-	); err != nil {
-		return err
-	}
-
-	return c.checkResponse(result)
-}

internal/xes/xes6.go

@@ -6,7 +6,9 @@ import (
 	"crypto/tls"
 	"encoding/json"
 	"fmt"
-	"io"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/esgo2dump/xes/es6"
 	"net"
 	"net/http"
 	"net/url"
@@ -15,7 +17,6 @@ import (
 
 	elastic "github.com/elastic/go-elasticsearch/v6"
 	"github.com/elastic/go-elasticsearch/v6/esapi"
-	"github.com/elastic/go-elasticsearch/v6/esutil"
 	"github.com/loveuer/esgo2dump/internal/interfaces"
 	"github.com/loveuer/esgo2dump/internal/opt"
 	"github.com/loveuer/esgo2dump/internal/util"
@@ -40,7 +41,13 @@ func NewClientV6(url *url.URL, iot interfaces.IO) (interfaces.DumpIO, error) {
 		}
 	}
 
-	logrus.Debugf("xes.NewClient: endpoint=%s index=%s (username=%s password=%s)", address, urlIndex, urlUsername, urlPassword)
+	logrus.
+		WithField("action", "new es client v6").
+		WithField("endpoint", address).
+		WithField("index", urlIndex).
+		WithField("username", urlUsername).
+		WithField("password", urlPassword).
+		Debug()
 
 	if urlIndex == "" {
 		return nil, fmt.Errorf("please specify index name: (like => http://127.0.0.1:9200/my_index)")
@@ -68,20 +75,30 @@ func NewClientV6(url *url.URL, iot interfaces.IO) (interfaces.DumpIO, error) {
 				},
 			},
 		); err != nil {
-			logrus.Debugf("xes.NewClient: elastic new client with endpont=%s err=%v", endpoints, err)
+			logrus.
+				WithField("action", "new es client v6 error").
+				WithField("endpoints", endpoints).
+				WithField("err", err).
+				Debug()
 			errCh <- err
 			return
 		}
 
 		if infoResp, err = cli.Info(); err != nil {
-			logrus.Debugf("xes.NewClient: ping err=%v", err)
+			logrus.
+				WithField("action", "es client v6 ping error").
+				WithField("err", err).
+				Debug()
 			errCh <- err
 			return
 		}
 
 		if infoResp.StatusCode != 200 {
 			err = fmt.Errorf("info xes status=%d", infoResp.StatusCode)
-			logrus.Debugf("xes.NewClient: status err=%v", err)
+			logrus.
+				WithField("action", "es client v6 ping status error").
+				WithField("status", infoResp.StatusCode).
+				Debug()
 			errCh <- err
 			return
 		}
@@ -95,17 +112,24 @@ func NewClientV6(url *url.URL, iot interfaces.IO) (interfaces.DumpIO, error) {
 	case <-util.Timeout(10).Done():
 		return nil, fmt.Errorf("dial es=%s err=%v", address, context.DeadlineExceeded)
 	case c := <-cliCh:
-		return &clientv6{c: c, index: urlIndex, iot: iot}, nil
+		return &clientv6{client: c, index: urlIndex, iot: iot}, nil
 	case e := <-errCh:
 		return nil, e
 	}
 }
 
 type clientv6 struct {
-	c        *elastic.Client
+	client *elastic.Client
 	iot    interfaces.IO
 	index  string
-	scrollId 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 {
@@ -128,153 +152,15 @@ func (c *clientv6) Close() error {
 	return nil
 }
 
-func (c *clientv6) ResetOffset() {
+func (c *clientv6) ReadData(ctx context.Context, size int, query map[string]any, source []string) (<-chan []*model.ESSource, <-chan error) {
-	defer func() {
+	dch, ech := es6.ReadData(ctx, c.client, c.index, size, 0, query, source)
-		c.scrollId = ""
-	}()
 
-	bs, _ := json.Marshal(map[string]string{
+	return dch, ech
-		"scroll_id": c.scrollId,
-	})
-
-	rr, err := c.c.ClearScroll(
-		c.c.ClearScroll.WithContext(util.Timeout(3)),
-		c.c.ClearScroll.WithBody(bytes.NewReader(bs)),
-	)
-	if err != nil {
-		logrus.Warnf("ResetOffset: clear scroll id=%s err=%v", c.scrollId, err)
-		return
-	}
-
-	if rr.StatusCode != 200 {
-		logrus.Warnf("ResetOffset: clear scroll id=%s msg=%s", c.scrollId, rr.String())
-	}
-}
-func (c *clientv6) WriteData(ctx context.Context, docs []*interfaces.ESSource) (int, error) {
-	var (
-		err     error
-		indexer esutil.BulkIndexer
-		count   int
-		be      error
-	)
-	if indexer, err = esutil.NewBulkIndexer(esutil.BulkIndexerConfig{
-		Client:       c.c,
-		Index:        c.index,
-		Refresh:      "",
-		DocumentType: "_doc",
-	}); err != nil {
-		return 0, err
-	}
-
-	for _, doc := range docs {
-		var bs []byte
-
-		if bs, err = json.Marshal(doc.Content); err != nil {
-			return 0, err
-		}
-
-		logrus.WithField("raw", string(bs)).Debug()
-
-		if err = indexer.Add(context.Background(), esutil.BulkIndexerItem{
-			Action:     "index",
-			Index:      c.index,
-			DocumentID: doc.DocId,
-			Body:       bytes.NewReader(bs),
-			OnFailure: func(ctx context.Context, item esutil.BulkIndexerItem, item2 esutil.BulkIndexerResponseItem, bulkErr error) {
-				be = bulkErr
-			},
-		}); err != nil {
-			return 0, err
-		}
-		count++
-	}
-
-	if err = indexer.Close(util.TimeoutCtx(ctx, opt.Timeout)); err != nil {
-		return 0, err
-	}
-
-	if be != nil {
-		return 0, be
-	}
-
-	stats := indexer.Stats()
-	if stats.NumFailed > 0 {
-		return count, fmt.Errorf("write to xes failed_count=%d bulk_count=%d", stats.NumFailed, count)
-	}
-
-	return count, nil
-}
-
-func (c *clientv6) ReadData(ctx context.Context, i int, query map[string]any, source []string) ([]*interfaces.ESSource, error) {
-	var (
-		err    error
-		resp   *esapi.Response
-		result = new(interfaces.ESResponseV6)
-		bs     []byte
-	)
-
-	if c.scrollId == "" {
-		qs := []func(*esapi.SearchRequest){
-			c.c.Search.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-			c.c.Search.WithIndex(c.index),
-			c.c.Search.WithSize(i),
-			c.c.Search.WithFrom(0),
-			c.c.Search.WithScroll(time.Duration(opt.Timeout*2) * time.Second),
-		}
-
-		if len(source) > 0 {
-			qs = append(qs, c.c.Search.WithSourceIncludes(source...))
-		}
-
-		if query != nil && len(query) > 0 {
-			queryBs, _ := json.Marshal(map[string]any{"query": query})
-			qs = append(qs, c.c.Search.WithBody(bytes.NewReader(queryBs)))
-		}
-
-		if resp, err = c.c.Search(qs...); err != nil {
-			return nil, err
-		}
-		defer resp.Body.Close()
-
-		if resp.StatusCode != 200 {
-			return nil, fmt.Errorf(resp.String())
-		}
-
-		if bs, err = io.ReadAll(resp.Body); err != nil {
-			return nil, err
-		}
-
-		if err = json.Unmarshal(bs, result); err != nil {
-			logrus.
-				WithField("err", err.Error()).
-				WithField("raw", string(bs)).
-				Debug()
-			return nil, err
-		}
-
-		c.scrollId = result.ScrollId
-
-		return result.Hits.Hits, nil
-	}
-
-	if resp, err = c.c.Scroll(
-		c.c.Scroll.WithScrollID(c.scrollId),
-		c.c.Scroll.WithScroll(time.Duration(opt.Timeout*2)*time.Second),
-	); err != nil {
-		return result.Hits.Hits, nil
-	}
-
-	decoder := json.NewDecoder(resp.Body)
-	if err = decoder.Decode(result); err != nil {
-		return nil, err
-	}
-
-	return result.Hits.Hits, nil
 }
 
 func (c *clientv6) ReadMapping(ctx context.Context) (map[string]any, error) {
-	r, err := c.c.Indices.GetMapping(
+	r, err := c.client.Indices.GetMapping(
-		c.c.Indices.GetMapping.WithIndex(c.index),
+		c.client.Indices.GetMapping.WithIndex(c.index),
 	)
 	if err != nil {
 		return nil, err
@@ -304,10 +190,10 @@ func (c *clientv6) WriteMapping(ctx context.Context, m map[string]any) error {
 			return err
 		}
 
-		if result, err = c.c.Indices.Create(
+		if result, err = c.client.Indices.Create(
 			c.index,
-			c.c.Indices.Create.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
+			c.client.Indices.Create.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-			c.c.Indices.Create.WithBody(bytes.NewReader(bs)),
+			c.client.Indices.Create.WithBody(bytes.NewReader(bs)),
 		); err != nil {
 			return err
 		}
@@ -321,9 +207,9 @@ func (c *clientv6) WriteMapping(ctx context.Context, m map[string]any) error {
 }
 
 func (c *clientv6) ReadSetting(ctx context.Context) (map[string]any, error) {
-	r, err := c.c.Indices.GetSettings(
+	r, err := c.client.Indices.GetSettings(
-		c.c.Indices.GetSettings.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
+		c.client.Indices.GetSettings.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
-		c.c.Indices.GetSettings.WithIndex(c.index),
+		c.client.Indices.GetSettings.WithIndex(c.index),
 	)
 	if err != nil {
 		return nil, err
@@ -353,9 +239,9 @@ func (c *clientv6) WriteSetting(ctx context.Context, m map[string]any) error {
 		return err
 	}
 
-	if result, err = c.c.Indices.PutSettings(
+	if result, err = c.client.Indices.PutSettings(
 		bytes.NewReader(bs),
-		c.c.Indices.PutSettings.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
+		c.client.Indices.PutSettings.WithContext(util.TimeoutCtx(ctx, opt.Timeout)),
 	); err != nil {
 		return err
 	}

internal/xes/xes7.go

@@ -0,0 +1,223 @@
+package xes
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	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/util"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
+	"github.com/loveuer/esgo2dump/xes/es7"
+	"net/url"
+	"strings"
+)
+
+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) WriteData(ctx context.Context, docs []*model.ESSource) (int, error) {
+//	var (
+//		err     error
+//		indexer esutil.BulkIndexer
+//		count   int
+//		be      error
+//	)
+//	if indexer, err = esutil.NewBulkIndexer(esutil.BulkIndexerConfig{
+//		Client:     c.client,
+//		Index:      c.index,
+//		ErrorTrace: true,
+//		OnError: func(ctx context.Context, err error) {
+//
+//		},
+//	}); err != nil {
+//		return 0, err
+//	}
+//
+//	for _, doc := range docs {
+//		var bs []byte
+//
+//		if bs, err = json.Marshal(doc.Content); err != nil {
+//			return 0, err
+//		}
+//
+//		if err = indexer.Add(context.Background(), esutil.BulkIndexerItem{
+//			Action:     "index",
+//			Index:      c.index,
+//			DocumentID: doc.DocId,
+//			Body:       bytes.NewReader(bs),
+//			OnFailure: func(ctx context.Context, item esutil.BulkIndexerItem, item2 esutil.BulkIndexerResponseItem, bulkErr error) {
+//				be = bulkErr
+//			},
+//		}); err != nil {
+//			return 0, err
+//		}
+//		count++
+//	}
+//
+//	if err = indexer.Close(util.TimeoutCtx(ctx, opt.Timeout)); err != nil {
+//		return 0, err
+//	}
+//
+//	if be != nil {
+//		return 0, be
+//	}
+//
+//	stats := indexer.Stats()
+//	if stats.NumFailed > 0 {
+//		return count, fmt.Errorf("write to xes failed_count=%d bulk_count=%d", stats.NumFailed, count)
+//	}
+//
+//	return count, nil
+//}
+
+func (c *client) ReadData(ctx context.Context, size int, query map[string]any, source []string) (<-chan []*model.ESSource, <-chan error) {
+	dch, ech := es7.ReadData(ctx, c.client, c.index, size, 0, query, source)
+
+	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(util.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(util.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(util.TimeoutCtx(ctx, opt.Timeout)),
+	); err != nil {
+		return err
+	}
+
+	return c.checkResponse(result)
+}

internal/xfile/xfile.go

@@ -4,11 +4,12 @@ import (
 	"bufio"
 	"context"
 	"encoding/json"
+	"github.com/loveuer/esgo2dump/internal/opt"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
 	"io"
 	"os"
 
-	"github.com/loveuer/esgo2dump/internal/opt"
-
 	"github.com/loveuer/esgo2dump/internal/interfaces"
 )
 
@@ -18,6 +19,29 @@ type client struct {
 	scanner *bufio.Scanner
 }
 
+func (c *client) WriteData(ctx context.Context, docsCh <-chan []*model.ESSource) error {
+	total := 0
+	for line := range docsCh {
+		for _, doc := range line {
+			bs, err := json.Marshal(doc)
+			if err != nil {
+				return err
+			}
+
+			if _, err = c.f.Write(append(bs, '\n')); err != nil {
+				return err
+			}
+		}
+
+		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) {
 	var (
 		err error
@@ -86,60 +110,62 @@ func (c *client) IsFile() bool {
 	return true
 }
 
-func (c *client) ResetOffset() {}
+func (c *client) ReadData(ctx context.Context, size int, _ map[string]any, _ []string) (<-chan []*model.ESSource, <-chan error) {
-
-func (c *client) WriteData(ctx context.Context, docs []*interfaces.ESSource) (int, error) {
-	var (
-		err   error
-		bs    []byte
-		count = 0
-	)
-
-	for _, doc := range docs {
-		if bs, err = json.Marshal(doc); err != nil {
-			return count, err
-		}
-
-		bs = append(bs, '\n')
-
-		if _, err = c.f.Write(bs); err != nil {
-			return count, err
-		}
-
-		count++
-	}
-
-	return count, nil
-}
-
-func (c *client) ReadData(ctx context.Context, i int, _ map[string]any, _ []string) ([]*interfaces.ESSource, error) {
 	var (
 		err   error
 		count = 0
-		list  = make([]*interfaces.ESSource, 0, i)
+		list  = make([]*model.ESSource, 0, size)
+		dch   = make(chan []*model.ESSource)
+		ech   = make(chan error)
+		ready = make(chan bool)
 	)
 
+	go func(ctx context.Context) {
+		defer func() {
+			close(dch)
+			close(ech)
+		}()
+
+		ready <- true
+
 		for c.scanner.Scan() {
-		line := c.scanner.Text()
+			select {
+			case <-ctx.Done():
+				return
+			default:
+				item := new(model.ESSource)
+				line := c.scanner.Bytes()
 
-		item := new(interfaces.ESSource)
+				if err = json.Unmarshal(line, item); err != nil {
-		if err = json.Unmarshal([]byte(line), item); err != nil {
+					ech <- err
-			return list, err
+					return
 				}
 
 				list = append(list, item)
-
 				count++
-		if count >= i {
+
-			break
+				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 {
-		return list, err
+			ech <- err
 		}
+	}(ctx)
 
-	return list, nil
+	<-ready
+
+	return dch, ech
 }
 
 func (c *client) Close() error {

log/log.go

@@ -0,0 +1,54 @@
+package log
+
+import (
+	"bytes"
+	"fmt"
+	"github.com/fatih/color"
+	"sync"
+	"time"
+)
+
+var (
+	red    = color.New(color.FgRed)
+	green  = color.New(color.FgGreen)
+	yellow = color.New(color.FgYellow)
+
+	locker = &sync.Mutex{}
+
+	timeFormat = "06-01-02T15:04:05"
+)
+
+func SetTimeFormat(format string) {
+	locker.Lock()
+	defer locker.Unlock()
+
+	timeFormat = format
+}
+
+func Info(msg string, data ...any) {
+	buf := &bytes.Buffer{}
+	_, _ = green.Fprint(buf, "Info  ")
+	_, _ = fmt.Fprintf(buf, "| %s | ", time.Now().Format(timeFormat))
+	_, _ = fmt.Fprintf(buf, msg, data...)
+	fmt.Println(buf.String())
+}
+
+func Warn(msg string, data ...any) {
+	buf := &bytes.Buffer{}
+	_, _ = yellow.Fprint(buf, "Warn  ")
+	_, _ = fmt.Fprintf(buf, "| %s | ", time.Now().Format(timeFormat))
+	_, _ = fmt.Fprintf(buf, msg, data...)
+	fmt.Println(buf.String())
+}
+
+func Error(msg string, data ...any) {
+	buf := &bytes.Buffer{}
+	_, _ = red.Fprint(buf, "Error ")
+	_, _ = fmt.Fprintf(buf, "| %s | ", time.Now().Format(timeFormat))
+	_, _ = fmt.Fprintf(buf, msg, data...)
+	fmt.Println(buf.String())
+}
+
+type WroteLogger interface {
+	Info(msg string, data ...any)
+}

main.go

@@ -2,12 +2,11 @@ package main
 
 import (
 	"context"
+	"github.com/loveuer/esgo2dump/log"
 	"os/signal"
 	"syscall"
 
 	"github.com/loveuer/esgo2dump/internal/cmd"
-
-	"github.com/sirupsen/logrus"
 )
 
 func main() {
@@ -16,9 +15,7 @@ func main() {
 	defer cancel()
 
 	if err := cmd.Start(ctx); err != nil {
-		logrus.Error(err)
+		log.Error(err.Error())
 		return
 	}
-
-	logrus.Debug("main: cmd start success!!!")
 }

model/es.go

@@ -1,4 +1,4 @@
-package interfaces
+package model
 
 type ESSource struct {
 	DocId   string         `json:"_id"`
@@ -25,26 +25,3 @@ type ESResponse struct {
 		Hits     []*ESSource `json:"hits"`
 	} `json:"hits"`
 }
-
-type ESMapping map[string]struct {
-	Mappings struct {
-		Properties map[string]any `json:"properties"`
-	} `json:"mappings"`
-}
-
-type ESResponseV6 struct {
-	ScrollId string `json:"_scroll_id"`
-	Took     int    `json:"took"`
-	TimedOut bool   `json:"timed_out"`
-	Shards   struct {
-		Total      int `json:"total"`
-		Successful int `json:"successful"`
-		Skipped    int `json:"skipped"`
-		Failed     int `json:"failed"`
-	} `json:"_shards"`
-	Hits struct {
-		Total    int         `json:"total"`
-		MaxScore float64     `json:"max_score"`
-		Hits     []*ESSource `json:"hits"`
-	} `json:"hits"`
-}

readme.md

@@ -3,7 +3,7 @@
 
 ---
 
-- 当前仅支持 elasticsearch 7
+- 支持 elasticsearch 7, elasticsearch 6
 
 ---
 
@@ -36,7 +36,7 @@ esgo2dump --input=http://127.0.0.1:9200/some_index --source='id;name;age;address
 
 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
 ```
 
 - example_queries.json
@@ -53,6 +53,7 @@ esgo2dump --input=http://127.0.0.1:9200/some_index --output=./data.json --query_
 - [x] es to file
 - [x] es to es
 - [x] auto create index with mapping
+- [ ] args: split_size (auto split json output file)
 - [ ] auto create index with mapping,setting
-- [ ] support es8
 - [x] support es6
+- [ ] support es8

xes/es6/client.go

@@ -0,0 +1,85 @@
+package es6
+
+import (
+	"context"
+	"crypto/tls"
+	"fmt"
+	elastic "github.com/elastic/go-elasticsearch/v6"
+	"github.com/elastic/go-elasticsearch/v6/esapi"
+	"github.com/loveuer/esgo2dump/internal/util"
+	"net"
+	"net/http"
+	"net/url"
+	"time"
+)
+
+func NewClient(ctx context.Context, url *url.URL) (*elastic.Client, error) {
+	var (
+		err         error
+		urlUsername string
+		urlPassword string
+		client      *elastic.Client
+		errCh       = make(chan error)
+		cliCh       = make(chan *elastic.Client)
+		address     = fmt.Sprintf("%s://%s", url.Scheme, url.Host)
+	)
+
+	if url.User != nil {
+		urlUsername = url.User.Username()
+		if p, ok := url.User.Password(); ok {
+			urlPassword = p
+		}
+	}
+
+	ncFunc := func(endpoints []string, username, password 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 {
+			errCh <- err
+			return
+		}
+
+		if infoResp, err = cli.Info(); err != nil {
+			errCh <- err
+			return
+		}
+
+		if infoResp.StatusCode != 200 {
+			err = fmt.Errorf("info es7 status=%d", infoResp.StatusCode)
+			errCh <- err
+			return
+		}
+
+		cliCh <- cli
+	}
+
+	go ncFunc([]string{address}, urlUsername, urlPassword)
+	timeout := util.TimeoutCtx(ctx, 10)
+
+	select {
+	case <-timeout.Done():
+		return nil, fmt.Errorf("dial es=%s err=%v", address, context.DeadlineExceeded)
+	case client = <-cliCh:
+		return client, nil
+	case err = <-errCh:
+		return nil, err
+	}
+}

xes/es6/read.go

@@ -0,0 +1,136 @@
+package es6
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	elastic "github.com/elastic/go-elasticsearch/v6"
+	"github.com/elastic/go-elasticsearch/v6/esapi"
+	"github.com/loveuer/esgo2dump/internal/util"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
+	"time"
+)
+
+func ReadData(ctx context.Context, client *elastic.Client, index string, size, max int, query map[string]any, source []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.ESResponse)
+			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(util.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(util.TimeoutCtx(ctx, 20)),
+			client.Search.WithIndex(index),
+			client.Search.WithSize(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 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.ESResponse)
+
+			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/log"
+	"github.com/loveuer/esgo2dump/model"
+)
+
+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

@@ -0,0 +1,85 @@
+package es7
+
+import (
+	"context"
+	"crypto/tls"
+	"fmt"
+	elastic "github.com/elastic/go-elasticsearch/v7"
+	"github.com/elastic/go-elasticsearch/v7/esapi"
+	"github.com/loveuer/esgo2dump/internal/util"
+	"net"
+	"net/http"
+	"net/url"
+	"time"
+)
+
+func NewClient(ctx context.Context, url *url.URL) (*elastic.Client, error) {
+	var (
+		err         error
+		urlUsername string
+		urlPassword string
+		client      *elastic.Client
+		errCh       = make(chan error)
+		cliCh       = make(chan *elastic.Client)
+		address     = fmt.Sprintf("%s://%s", url.Scheme, url.Host)
+	)
+
+	if url.User != nil {
+		urlUsername = url.User.Username()
+		if p, ok := url.User.Password(); ok {
+			urlPassword = p
+		}
+	}
+
+	ncFunc := func(endpoints []string, username, password 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 {
+			errCh <- err
+			return
+		}
+
+		if infoResp, err = cli.Info(); err != nil {
+			errCh <- err
+			return
+		}
+
+		if infoResp.StatusCode != 200 {
+			err = fmt.Errorf("info es7 status=%d", infoResp.StatusCode)
+			errCh <- err
+			return
+		}
+
+		cliCh <- cli
+	}
+
+	go ncFunc([]string{address}, urlUsername, urlPassword)
+	timeout := util.TimeoutCtx(ctx, 10)
+
+	select {
+	case <-timeout.Done():
+		return nil, fmt.Errorf("dial es=%s err=%v", address, context.DeadlineExceeded)
+	case client = <-cliCh:
+		return client, nil
+	case err = <-errCh:
+		return nil, err
+	}
+}

xes/es7/read.go

@@ -0,0 +1,136 @@
+package es7
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	elastic "github.com/elastic/go-elasticsearch/v7"
+	"github.com/elastic/go-elasticsearch/v7/esapi"
+	"github.com/loveuer/esgo2dump/internal/util"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
+	"time"
+)
+
+func ReadData(ctx context.Context, client *elastic.Client, index string, size, max int, query map[string]any, source []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.ESResponse)
+			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(util.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(util.TimeoutCtx(ctx, 20)),
+			client.Search.WithIndex(index),
+			client.Search.WithSize(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 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.ESResponse)
+
+			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/es7/write.go

@@ -0,0 +1,84 @@
+package es7
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	elastic "github.com/elastic/go-elasticsearch/v7"
+	"github.com/elastic/go-elasticsearch/v7/esutil"
+	"github.com/loveuer/esgo2dump/log"
+	"github.com/loveuer/esgo2dump/model"
+)
+
+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
+	)
+
+	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,
+					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)
+			}
+		}
+	}
+}