file-plugins: enabling messagepack format

conduit/pipeline/pipeline.go

@@ -6,15 +6,13 @@
 	"fmt"
 	"net/http"
 	"os"
-	"path"
 	"runtime/pprof"
 	"sync"
 	"time"
 
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	log "github.com/sirupsen/logrus"
-	yaml "gopkg.in/yaml.v3"
 
 	sdk "github.com/algorand/go-algorand-sdk/v2/types"
 
@@ -103,29 +101,23 @@
 	}
 }
 
-// makeConfig creates a plugin config from a name and config pair.
+// configWithLogger creates a plugin config from a name and config pair.
 // It also creates a logger for the plugin and configures it using the pipeline's log settings.
-func (p *pipelineImpl) makeConfig(cfg data.NameConfigPair, pluginType plugins.PluginType) (*log.Logger, plugins.PluginConfig, error) {
-	configs, err := yaml.Marshal(cfg.Config)
+func (p *pipelineImpl) configWithLogger(cfg data.NameConfigPair, pluginType plugins.PluginType) (*log.Logger, plugins.PluginConfig, error) {
+	var dataDir string
+	if p.cfg.ConduitArgs != nil {
+		dataDir = p.cfg.ConduitArgs.ConduitDataDir
+	}
+	config, err := pluginType.GetConfig(cfg, dataDir)
 	if err != nil {
-		return nil, plugins.PluginConfig{}, fmt.Errorf("makeConfig(): could not serialize config: %w", err)
+		return nil, plugins.PluginConfig{}, fmt.Errorf("configWithLogger(): unable to create plugin config: %w", err)
 	}
 
 	lgr := log.New()
 	lgr.SetOutput(p.logger.Out)
 	lgr.SetLevel(p.logger.Level)
 	lgr.SetFormatter(makePluginLogFormatter(string(pluginType), cfg.Name))
 
-	var config plugins.PluginConfig
-	config.Config = string(configs)
-	if p.cfg != nil && p.cfg.ConduitArgs != nil {
-		config.DataDir = path.Join(p.cfg.ConduitArgs.ConduitDataDir, fmt.Sprintf("%s_%s", pluginType, cfg.Name))
-		err = os.MkdirAll(config.DataDir, os.ModePerm)
-		if err != nil {
-			return nil, plugins.PluginConfig{}, fmt.Errorf("makeConfig: unable to create plugin data directory: %w", err)
-		}
-	}
-
 	return lgr, config, nil
 }
 
@@ -171,10 +163,10 @@
 	var pluginOverride uint64
 	var pluginOverrideName string // cache this in case of error.
 	for _, part := range parts {
-		_, config, err := p.makeConfig(part.cfg, part.t)
+		_, config, err := p.configWithLogger(part.cfg, part.t)
 		if err != nil {
 			return 0, err
 		}
 		rnd, err := part.RoundRequest(config)
 		if err != nil {
 			return 0, err
@@ -204,15 +196,15 @@
 	telemetryConfig := telemetry.MakeTelemetryConfig(p.cfg.Telemetry.URI, p.cfg.Telemetry.Index, p.cfg.Telemetry.UserName, p.cfg.Telemetry.Password)
 	telemetryClient, err := telemetry.MakeOpenSearchClient(telemetryConfig)
 	if err != nil {
 		return nil, fmt.Errorf("failed to initialize telemetry: %w", err)
 	}
 	p.logger.Infof("Telemetry initialized with URI: %s", telemetryConfig.URI)
 
 	// If GUID is not in metadata, save it. Otherwise, use the GUID from metadata.
 	if p.pipelineMetadata.TelemetryID == "" {
 		p.pipelineMetadata.TelemetryID = telemetryClient.TelemetryConfig.GUID
 	} else {
 		telemetryClient.TelemetryConfig.GUID = p.pipelineMetadata.TelemetryID
 	}
 
 	return telemetryClient, nil
@@ -232,12 +224,12 @@
 		var err error
 		profFile, err := os.Create(p.cfg.CPUProfile)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): unable to create profile: %w", err)
 		}
 		p.profFile = profFile
 		err = pprof.StartCPUProfile(profFile)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): unable to start pprof: %w", err)
 		}
 	}
 
@@ -286,7 +278,7 @@
 		var telemetryErr error
 		telemetryClient, telemetryErr = p.initializeTelemetry()
 		if telemetryErr != nil {
 			p.logger.Warnf("Telemetry initialization failed, continuing without telemetry: %s", telemetryErr)
 		} else {
 			// Try sending a startup event. If it fails, log a warning and continue
 			event := telemetryClient.MakeTelemetryStartupEvent()
@@ -302,18 +294,18 @@
 
 	// Initialize Importer
 	{
-		importerLogger, pluginConfig, err := p.makeConfig(p.cfg.Importer, plugins.Importer)
+		importerLogger, pluginConfig, err := p.configWithLogger(p.cfg.Importer, plugins.Importer)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): could not make %s config: %w", p.cfg.Importer.Name, err)
 		}
 		err = p.importer.Init(p.ctx, *p.initProvider, pluginConfig, importerLogger)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): could not initialize importer (%s): %w", p.cfg.Importer.Name, err)
 		}
 		genesis, err := p.importer.GetGenesis()
 		if err != nil {
 			return fmt.Errorf("Pipeline.GetGenesis(): could not obtain Genesis from the importer (%s): %w", p.cfg.Importer.Name, err)
 		}
 		(*p.initProvider).SetGenesis(genesis)
 
 		// write pipeline metadata
@@ -326,8 +318,8 @@
 		p.pipelineMetadata.Network = genesis.Network
 		err = p.pipelineMetadata.encodeToFile(p.cfg.ConduitArgs.ConduitDataDir)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init() failed to write metadata to file: %w", err)
 		}
 
 		p.logger.Infof("Initialized Importer: %s", p.cfg.Importer.Name)
 	}
@@ -335,27 +327,27 @@
 	// Initialize Processors
 	for idx, processor := range p.processors {
 		ncPair := p.cfg.Processors[idx]
-		logger, config, err := p.makeConfig(ncPair, plugins.Processor)
+		logger, config, err := p.configWithLogger(ncPair, plugins.Processor)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): could not initialize processor (%s): %w", ncPair, err)
 		}
 		err = processor.Init(p.ctx, *p.initProvider, config, logger)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): could not initialize processor (%s): %w", ncPair.Name, err)
 		}
 		p.logger.Infof("Initialized Processor: %s", ncPair.Name)
 	}
 
 	// Initialize Exporter
 	{
-		logger, config, err := p.makeConfig(p.cfg.Exporter, plugins.Exporter)
+		logger, config, err := p.configWithLogger(p.cfg.Exporter, plugins.Exporter)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): could not initialize processor (%s): %w", p.cfg.Exporter.Name, err)
 		}
 		err = p.exporter.Init(p.ctx, *p.initProvider, config, logger)
 		if err != nil {
 			return fmt.Errorf("Pipeline.Init(): could not initialize Exporter (%s): %w", p.cfg.Exporter.Name, err)
 		}
 		p.logger.Infof("Initialized Exporter: %s", p.cfg.Exporter.Name)
 	}
 
@@ -388,20 +380,20 @@
 		}
 	}
 
 	if err := p.importer.Close(); err != nil {
 		// Log and continue on closing the rest of the pipeline
 		p.logger.Errorf("Pipeline.Stop(): Importer (%s) error on close: %v", p.importer.Metadata().Name, err)
 	}
 
 	for _, processor := range p.processors {
 		if err := processor.Close(); err != nil {
 			// Log and continue on closing the rest of the pipeline
 			p.logger.Errorf("Pipeline.Stop(): Processor (%s) error on close: %v", processor.Metadata().Name, err)
 		}
 	}
 
 	if err := p.exporter.Close(); err != nil {
 		p.logger.Errorf("Pipeline.Stop(): Exporter (%s) error on close: %v", p.exporter.Metadata().Name, err)
 	}
 }
 
@@ -574,7 +566,7 @@
 		return nil, fmt.Errorf("MakePipeline(): could not build importer '%s': %w", importerName, err)
 	}
 
 	pipeline.importer = importerConstructor.New()
 	logger.Infof("Found Importer: %s", importerName)
 
 	// ---
@@ -582,12 +574,12 @@
 	for _, processorConfig := range cfg.Processors {
 		processorName := processorConfig.Name
 
 		processorConstructor, err := processors.ProcessorConstructorByName(processorName)
 		if err != nil {
 			return nil, fmt.Errorf("MakePipeline(): could not build processor '%s': %w", processorName, err)
 		}
 
 		pipeline.processors = append(pipeline.processors, processorConstructor.New())
 		logger.Infof("Found Processor: %s", processorName)
 	}
 
@@ -595,12 +587,12 @@
 
 	exporterName := cfg.Exporter.Name
 
 	exporterConstructor, err := exporters.ExporterConstructorByName(exporterName)
 	if err != nil {
 		return nil, fmt.Errorf("MakePipeline(): could not build exporter '%s': %w", exporterName, err)
 	}
 
 	pipeline.exporter = exporterConstructor.New()
 	logger.Infof("Found Exporter: %s", exporterName)
 
 	return pipeline, nil
@@ -612,22 +604,22 @@
 	logger.Infof("Creating PID file at: %s", pidFilePath)
 	fout, err := os.Create(pidFilePath)
 	if err != nil {
 		err = fmt.Errorf("%s: could not create pid file, %v", pidFilePath, err)
 		logger.Error(err)
 		return err
 	}
 
 	if _, err = fmt.Fprintf(fout, "%d", os.Getpid()); err != nil {
 		err = fmt.Errorf("%s: could not write pid file, %v", pidFilePath, err)
 		logger.Error(err)
 		return err
 	}
 
 	err = fout.Close()
 	if err != nil {
 		err = fmt.Errorf("%s: could not close pid file, %v", pidFilePath, err)
 		logger.Error(err)
 		return err
 	}
 	return err
 }

conduit/plugins/exporters/filewriter/README.md

@@ -3,6 +3,7 @@
 Write block data to files. This plugin works with the file rerader plugin to create a simple file-based pipeine.
 
 ## Configuration
+
 ```yml @sample.yaml
 name: file_writer
 config:
@@ -13,9 +14,11 @@ config:
 
     # FilenamePattern is the format used to write block files. It uses go
     # string formatting and should accept one number for the round.
-    # If the file has a '.gz' extension, blocks will be gzipped.
-    # Default: "%[1]d_block.json"
-    filename-pattern: "%[1]d_block.json"
+    # To specify JSON encoding, add a '.json' extension to the filename.
+    # To specify MessagePack encoding, add a '.msgp' extension to the filename.
+    # If the file has a '.gz' extension, blocks will be gzipped regardless of encoding.
+    # Default: "%[1]d_block.msgp.gz"
+    filename-pattern: "%[1]d_block.msgp.gz"
 
     # DropCertificate is used to remove the vote certificate from the block data before writing files.
     drop-certificate: true

conduit/plugins/exporters/filewriter/file_exporter.go

@@ -18,13 +18,16 @@
 const (
 	// PluginName to use when configuring.
 	PluginName = "file_writer"
+
 	// FilePattern is used to name the output files.
-	FilePattern = "%[1]d_block.json"
+	FilePattern = "%[1]d_block.msgp.gz"
 )
 
 type fileExporter struct {
 	round  uint64
 	cfg    Config
+	gzip   bool
+	format EncodingFormat
 	logger *logrus.Logger
 }
 
@@ -51,20 +54,38 @@
 	if exp.cfg.FilenamePattern == "" {
 		exp.cfg.FilenamePattern = FilePattern
 	}
+	exp.format, exp.gzip, err = ParseFilenamePattern(exp.cfg.FilenamePattern)
+	if err != nil {
+		return fmt.Errorf("Init() error: %w", err)
+	}
+
 	// default to the data directory if no override provided.
 	if exp.cfg.BlocksDir == "" {
 		exp.cfg.BlocksDir = cfg.DataDir
 	}
 	// create block directory
 	err = os.Mkdir(exp.cfg.BlocksDir, 0755)
-	if err != nil && errors.Is(err, os.ErrExist) {
-		// Ignore mkdir if the dir exists
-		err = nil
-	} else if err != nil {
+	if err != nil && !errors.Is(err, os.ErrExist) {
+		// Ignore mkdir err if the dir exists (case errors.Is(err, os.ErrExist))
 		return fmt.Errorf("Init() error: %w", err)
 	}
+
 	exp.round = uint64(initProvider.NextDBRound())
-	return err
+
+	// export the genesis as well in the same format
+	genesis := initProvider.GetGenesis()
+	genesisFile, err := GenesisFilename(exp.format, exp.gzip)
+	if err != nil {
+		return fmt.Errorf("Init() error: %w", err)
+	}
+
+	genesisPath := path.Join(exp.cfg.BlocksDir, genesisFile)
+	err = EncodeToFile(genesisPath, genesis, exp.format, exp.gzip)
+	if err != nil {
+		return fmt.Errorf("Init() error sending to genesisPath=%s: %w", genesisPath, err)
+	}
+
+	return nil
 }
 
 func (exp *fileExporter) Close() error {
@@ -87,10 +108,12 @@
 		}
 
 		blockFile := path.Join(exp.cfg.BlocksDir, fmt.Sprintf(exp.cfg.FilenamePattern, exportData.Round()))
-		err := EncodeJSONToFile(blockFile, exportData, true)
+
+		err := EncodeToFile(blockFile, &exportData, exp.format, exp.gzip)
 		if err != nil {
 			return fmt.Errorf("Receive(): failed to write file %s: %w", blockFile, err)
 		}
+
 		exp.logger.Infof("Wrote block %d to %s", exportData.Round(), blockFile)
 	}
 

conduit/plugins/exporters/filewriter/file_exporter_test.go

@@ -9,7 +9,6 @@ import (
 
 	"github.com/sirupsen/logrus"
 	"github.com/sirupsen/logrus/hooks/test"
-	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	"gopkg.in/yaml.v3"
 
@@ -21,29 +20,45 @@ import (
 	sdk "github.com/algorand/go-algorand-sdk/v2/types"
 )
 
+const (
+	defaultEncodingFormat = MessagepackFormat
+	defaultIsGzip         = true
+)
+
 var logger *logrus.Logger
 var fileCons = exporters.ExporterConstructorFunc(func() exporters.Exporter {
 	return &fileExporter{}
 })
-var configTemplate = "block-dir: %s/blocks\n"
+var configTemplatePrefix = "block-dir: %s/blocks\n"
 var round = sdk.Round(2)
 
 func init() {
 	logger, _ = test.NewNullLogger()
 }
 
-func getConfig(t *testing.T) (config, tempdir string) {
+func getConfigWithoutPattern(t *testing.T) (config, tempdir string) {
 	tempdir = t.TempDir()
-	config = fmt.Sprintf(configTemplate, tempdir)
+	config = fmt.Sprintf(configTemplatePrefix, tempdir)
+	return
+}
+
+func getConfigWithPattern(t *testing.T, pattern string) (config, tempdir string) {
+	config, tempdir = getConfigWithoutPattern(t)
+	config = fmt.Sprintf("%sfilename-pattern: '%s'\n", config, pattern)
 	return
 }
 
+func TestDefaults(t *testing.T) {
+	require.Equal(t, defaultEncodingFormat, MessagepackFormat)
+	require.Equal(t, defaultIsGzip, true)
+}
+
 func TestExporterMetadata(t *testing.T) {
 	fileExp := fileCons.New()
 	meta := fileExp.Metadata()
-	assert.Equal(t, metadata.Name, meta.Name)
-	assert.Equal(t, metadata.Description, meta.Description)
-	assert.Equal(t, metadata.Deprecated, meta.Deprecated)
+	require.Equal(t, metadata.Name, meta.Name)
+	require.Equal(t, metadata.Description, meta.Description)
+	require.Equal(t, metadata.Deprecated, meta.Deprecated)
 }
 
 func TestExporterInitDefaults(t *testing.T) {
@@ -87,18 +102,18 @@ func TestExporterInitDefaults(t *testing.T) {
 }
 
 func TestExporterInit(t *testing.T) {
-	config, _ := getConfig(t)
+	config, _ := getConfigWithPattern(t, "%[1]d_block.json")
 	fileExp := fileCons.New()
 	defer fileExp.Close()
 
 	// creates a new output file
 	err := fileExp.Init(context.Background(), conduit.MakePipelineInitProvider(&round, nil, nil), plugins.MakePluginConfig(config), logger)
-	assert.NoError(t, err)
+	require.NoError(t, err)
 	fileExp.Close()
 
 	// can open existing file
 	err = fileExp.Init(context.Background(), conduit.MakePipelineInitProvider(&round, nil, nil), plugins.MakePluginConfig(config), logger)
-	assert.NoError(t, err)
+	require.NoError(t, err)
 	fileExp.Close()
 }
 
@@ -155,55 +170,67 @@ func sendData(t *testing.T, fileExp exporters.Exporter, config string, numRounds
 }
 
 func TestExporterReceive(t *testing.T) {
-	config, tempdir := getConfig(t)
-	fileExp := fileCons.New()
-	numRounds := 5
-	sendData(t, fileExp, config, numRounds)
-
-	// block data is valid
-	for i := 0; i < 5; i++ {
-		filename := fmt.Sprintf(FilePattern, i)
-		path := fmt.Sprintf("%s/blocks/%s", tempdir, filename)
-		require.FileExists(t, path)
-
-		blockBytes, err := os.ReadFile(path)
-		require.NoError(t, err)
-		assert.NotContains(t, string(blockBytes), " 0: ")
+	patterns := []string{
+		"%[1]d_block.json",
+		"%[1]d_block.json.gz",
+		"%[1]d_block.msgp",
+		"%[1]d_block.msgp.gz",
+	}
+	for _, pattern := range patterns {
+		pattern := pattern
+		t.Run(pattern, func(t *testing.T) {
+			t.Parallel()
 
-		var blockData data.BlockData
-		err = DecodeJSONFromFile(path, &blockData, true)
-		require.Equal(t, sdk.Round(i), blockData.BlockHeader.Round)
-		require.NoError(t, err)
-		require.NotNil(t, blockData.Certificate)
+			format, isGzip, err := ParseFilenamePattern(pattern)
+			require.NoError(t, err)
+			config, tempdir := getConfigWithPattern(t, pattern)
+			fileExp := fileCons.New()
+			numRounds := 5
+			sendData(t, fileExp, config, numRounds)
+
+			// block data is valid
+			for i := 0; i < 5; i++ {
+				filename := fmt.Sprintf(pattern, i)
+				path := fmt.Sprintf("%s/blocks/%s", tempdir, filename)
+				require.FileExists(t, path)
+
+				blockBytes, err := os.ReadFile(path)
+				require.NoError(t, err)
+				require.NotContains(t, string(blockBytes), " 0: ")
+
+				var blockData data.BlockData
+				err = DecodeFromFile(path, &blockData, format, isGzip)
+				require.NoError(t, err)
+				require.Equal(t, sdk.Round(i), blockData.BlockHeader.Round)
+				require.NotNil(t, blockData.Certificate)
+			}
+		})
 	}
 }
 
 func TestExporterClose(t *testing.T) {
-	config, _ := getConfig(t)
+	config, _ := getConfigWithoutPattern(t)
 	fileExp := fileCons.New()
 	rnd := sdk.Round(0)
 	fileExp.Init(context.Background(), conduit.MakePipelineInitProvider(&rnd, nil, nil), plugins.MakePluginConfig(config), logger)
 	require.NoError(t, fileExp.Close())
 }
 
-func TestPatternOverride(t *testing.T) {
-	config, tempdir := getConfig(t)
+func TestPatternDefault(t *testing.T) {
+	config, tempdir := getConfigWithoutPattern(t)
 	fileExp := fileCons.New()
 
-	patternOverride := "PREFIX_%[1]d_block.json"
-	config = fmt.Sprintf("%sfilename-pattern: '%s'\n", config, patternOverride)
-
 	numRounds := 5
 	sendData(t, fileExp, config, numRounds)
 
 	// block data is valid
 	for i := 0; i < 5; i++ {
-		filename := fmt.Sprintf(patternOverride, i)
+		filename := fmt.Sprintf(FilePattern, i)
 		path := fmt.Sprintf("%s/blocks/%s", tempdir, filename)
-		assert.FileExists(t, path)
+		require.FileExists(t, path)
 
 		var blockData data.BlockData
-		err := DecodeJSONFromFile(path, &blockData, true)
+		err := DecodeFromFile(path, &blockData, defaultEncodingFormat, defaultIsGzip)
 		require.Equal(t, sdk.Round(i), blockData.BlockHeader.Round)
 		require.NoError(t, err)
 		require.NotNil(t, blockData.Certificate)
@@ -227,10 +254,10 @@ func TestDropCertificate(t *testing.T) {
 	for i := 0; i < numRounds; i++ {
 		filename := fmt.Sprintf(FilePattern, i)
 		path := fmt.Sprintf("%s/%s", tempdir, filename)
-		assert.FileExists(t, path)
+		require.FileExists(t, path)
 		var blockData data.BlockData
-		err := DecodeJSONFromFile(path, &blockData, true)
-		assert.NoError(t, err)
-		assert.Nil(t, blockData.Certificate)
+		err := DecodeFromFile(path, &blockData, defaultEncodingFormat, defaultIsGzip)
+		require.NoError(t, err)
+		require.Nil(t, blockData.Certificate)
 	}
 }

conduit/plugins/exporters/filewriter/sample.yaml

@@ -7,9 +7,11 @@ config:
 
     # FilenamePattern is the format used to write block files. It uses go
     # string formatting and should accept one number for the round.
-    # If the file has a '.gz' extension, blocks will be gzipped.
-    # Default: "%[1]d_block.json"
-    filename-pattern: "%[1]d_block.json"
+    # To specify JSON encoding, add a '.json' extension to the filename.
+    # To specify MessagePack encoding, add a '.msgp' extension to the filename.
+    # If the file has a '.gz' extension, blocks will be gzipped regardless of encoding.
+    # Default: "%[1]d_block.msgp.gz"
+    filename-pattern: "%[1]d_block.msgp.gz"
 
     # DropCertificate is used to remove the vote certificate from the block data before writing files.
     drop-certificate: true