163 semantic diff end to end

cmd/analyzer/main.go

@@ -10,6 +10,13 @@ import (
 	"github.com/np-guard/vpc-network-config-analyzer/pkg/vpcmodel"
 )
 
+const (
+	ParsingErr       = "error parsing arguments:"
+	OutGenerationErr = "output generation error:"
+	InGenerationErr  = "error generating cloud config from input vpc resources file:"
+	ErrorFormat      = "%s %w"
+)
+
 func getOutputFormat(inArgs *InArgs) vpcmodel.OutFormat {
 	switch *inArgs.OutputFormat {
 	case TEXTFormat:
@@ -36,12 +43,14 @@ func analysisTypeToUseCase(inArgs *InArgs) vpcmodel.OutputUseCase {
 		return vpcmodel.SingleSubnet
 	case allSubnets:
 		return vpcmodel.AllSubnets
+	case allSubnetsDiff:
+		return vpcmodel.AllSubnetsDiff
 	}
 	return vpcmodel.AllEndpoints
 }
 
 func analysisPerVPCConfig(c *vpcmodel.VPCConfig, inArgs *InArgs, outFile string) (*vpcmodel.VPCAnalysisOutput, error) {
-	og, err := vpcmodel.NewOutputGenerator(c,
+	og, err := vpcmodel.NewOutputGenerator(c, nil,
 		*inArgs.Grouping,
 		analysisTypeToUseCase(inArgs),
 		*inArgs.OutputFormat == ARCHDRAWIOFormat)
@@ -57,12 +66,31 @@ func analysisPerVPCConfig(c *vpcmodel.VPCConfig, inArgs *InArgs, outFile string)
 	outFormat := getOutputFormat(inArgs)
 	output, err := og.Generate(outFormat, genOutFile)
 	if err != nil {
-		return nil, fmt.Errorf("output generation error: %w", err)
+		return nil, fmt.Errorf(ErrorFormat, OutGenerationErr, err)
 	}
 
 	return output, nil
 }
 
+func analysisDiffVPCConfig(c1, c2 *vpcmodel.VPCConfig, inArgs *InArgs, outFile string) (*vpcmodel.VPCAnalysisOutput, error) {
+	og, err := vpcmodel.NewOutputGenerator(c1, c2,
+		*inArgs.Grouping,
+		analysisTypeToUseCase(inArgs),
+		false)
+	if err != nil {
+		return nil, err
+	}
+
+	var analysisOut *vpcmodel.VPCAnalysisOutput
+	outFormat := getOutputFormat(inArgs)
+	analysisOut, err = og.Generate(outFormat, outFile)
+	if err != nil {
+		return nil, fmt.Errorf(ErrorFormat, OutGenerationErr, err)
+	}
+
+	return analysisOut, nil
+}
+
 // The actual main function
 // Takes command-line flags and returns an error rather than exiting, so it can be more easily used in testing
 func _main(cmdlineArgs []string) error {
@@ -71,7 +99,7 @@ func _main(cmdlineArgs []string) error {
 		return nil
 	}
 	if err != nil {
-		return fmt.Errorf("error parsing arguments: %w", err)
+		return fmt.Errorf(ErrorFormat, ParsingErr, err)
 	}
 
 	rc, err := ibmvpc.ParseResourcesFromFile(*inArgs.InputConfigFile)
@@ -81,30 +109,61 @@ func _main(cmdlineArgs []string) error {
 
 	vpcConfigs, err := ibmvpc.VPCConfigsFromResources(rc, *inArgs.VPC, *inArgs.Debug)
 	if err != nil {
-		return fmt.Errorf("error generating cloud config from input vpc resources file: %w", err)
+		return fmt.Errorf(ErrorFormat, InGenerationErr, err)
 	}
+
 	outFile := ""
 	if inArgs.OutputFile != nil {
 		outFile = *inArgs.OutputFile
 	}
 
-	outputPerVPC := make([]*vpcmodel.VPCAnalysisOutput, len(vpcConfigs))
-	i := 0
-	for _, vpcConfig := range vpcConfigs {
-		vpcAnalysisOutput, err2 := analysisPerVPCConfig(vpcConfig, inArgs, outFile)
+	diffAnalysis := *inArgs.AnalysisType == allEndpointsDiff || *inArgs.AnalysisType == allSubnetsDiff
+	if !diffAnalysis {
+		outputPerVPC := make([]*vpcmodel.VPCAnalysisOutput, len(vpcConfigs))
+		i := 0
+		for _, vpcConfig := range vpcConfigs {
+			vpcAnalysisOutput, err2 := analysisPerVPCConfig(vpcConfig, inArgs, outFile)
+			if err2 != nil {
+				return err2
+			}
+			outputPerVPC[i] = vpcAnalysisOutput
+			i++
+		}
+
+		var out string
+		out, err = vpcmodel.AggregateVPCsOutput(outputPerVPC, getOutputFormat(inArgs), outFile)
+		if err != nil {
+			return err
+		}
+		fmt.Println(out)
+	} else {
+		// Diff analysis
+		// ToDo SM: for diff analysis assume 2 configs only, the 2nd given through vpc-config-second
+		var rc2ndForDiff *ibmvpc.ResourcesContainer
+		rc2ndForDiff, err = ibmvpc.ParseResourcesFromFile(*inArgs.InputSecondConfigFile)
+		if err != nil {
+			return fmt.Errorf(ErrorFormat, ParsingErr, err)
+		}
+		vpc2ndConfigs, err := ibmvpc.VPCConfigsFromResources(rc2ndForDiff, *inArgs.VPC, *inArgs.Debug)
+		if err != nil {
+			return fmt.Errorf(ErrorFormat, InGenerationErr, err)
+		}
+		// For diff analysis each vpcConfigs have a single element
+		c1 := getFirstCfg(vpcConfigs)
+		c2 := getFirstCfg(vpc2ndConfigs)
+		analysisOutput, err2 := analysisDiffVPCConfig(c1, c2, inArgs, outFile)
 		if err2 != nil {
 			return err2
 		}
-		outputPerVPC[i] = vpcAnalysisOutput
-		i++
+		fmt.Println(analysisOutput.Output)
 	}
+	return nil
+}
 
-	out, err := vpcmodel.AggregateVPCsOutput(outputPerVPC, getOutputFormat(inArgs), outFile)
-	if err != nil {
-		return err
+func getFirstCfg(vpcConfigs map[string]*vpcmodel.VPCConfig) *vpcmodel.VPCConfig {
+	for _, vpcConfig := range vpcConfigs {
+		return vpcConfig
 	}
-	fmt.Println(out)
-
 	return nil
 }
 

cmd/analyzer/parse_args.go

@@ -8,13 +8,14 @@ import (
 
 // InArgs contains the input arguments for the analyzer
 type InArgs struct {
-	InputConfigFile *string
-	OutputFile      *string
-	OutputFormat    *string
-	AnalysisType    *string
-	Grouping        *bool
-	VPC             *string
-	Debug           *bool
+	InputConfigFile       *string
+	InputSecondConfigFile *string
+	OutputFile            *string
+	OutputFormat          *string
+	AnalysisType          *string
+	Grouping              *bool
+	VPC                   *string
+	Debug                 *bool
 }
 
 const (
@@ -27,9 +28,11 @@ const (
 	DEBUGFormat      = "debug"
 
 	// connectivity analysis types supported
-	allEndpoints = "all_endpoints" // vsi to vsi connectivity analysis
-	allSubnets   = "all_subnets"   // subnet to subnet connectivity analysis
-	singleSubnet = "single_subnet" // single subnet connectivity analysis
+	allEndpoints     = "all_endpoints"      // vsi to vsi connectivity analysis
+	allSubnets       = "all_subnets"        // subnet to subnet connectivity analysis
+	singleSubnet     = "single_subnet"      // single subnet connectivity analysis
+	allEndpointsDiff = "diff_all_endpoints" // semantic diff of allEndpoints analysis between two configurations
+	allSubnetsDiff   = "diff_all_subnets"   // semantic diff of allSubnets analysis between two configurations
 )
 
 var supportedOutputFormats = map[string]bool{
@@ -41,9 +44,11 @@ var supportedOutputFormats = map[string]bool{
 	DEBUGFormat:      true,
 }
 var supportedAnalysisTypes = map[string]bool{
-	allEndpoints: true,
-	allSubnets:   true,
-	singleSubnet: true,
+	allEndpoints:     true,
+	allSubnets:       true,
+	singleSubnet:     true,
+	allSubnetsDiff:   true,
+	allEndpointsDiff: false,
 }
 
 func getSupportedValuesString(supportedValues map[string]bool) string {
@@ -60,6 +65,7 @@ func ParseInArgs(cmdlineArgs []string) (*InArgs, error) {
 	args := InArgs{}
 	flagset := flag.NewFlagSet("vpc-network-config-analyzer", flag.ContinueOnError)
 	args.InputConfigFile = flagset.String("vpc-config", "", "file path to input config")
+	args.InputSecondConfigFile = flagset.String("vpc-config-second", "", "file path to second input config for semantic diff")
 	args.OutputFile = flagset.String("output-file", "", "file path to store results")
 	args.OutputFormat = flagset.String("format", TEXTFormat, "output format; must be one of "+getSupportedValuesString(supportedOutputFormats))
 	args.AnalysisType = flagset.String("analysis-type", allEndpoints,
@@ -72,32 +78,55 @@ func ParseInArgs(cmdlineArgs []string) (*InArgs, error) {
 	if err != nil {
 		return nil, err
 	}
+	err = errorInErgs(&args, flagset)
+	if err != nil {
+		return nil, err
+	}
+	err = notSupportedYetArgs(&args)
+	if err != nil {
+		return nil, err
+	}
 
+	return &args, nil
+}
+
+func errorInErgs(args *InArgs, flagset *flag.FlagSet) error {
 	if args.InputConfigFile == nil || *args.InputConfigFile == "" {
 		flagset.PrintDefaults()
-		return nil, fmt.Errorf("missing parameter: vpc-config")
+		return fmt.Errorf("missing parameter: vpc-config")
 	}
-
-	if !supportedOutputFormats[*args.OutputFormat] {
+	if !supportedAnalysisTypes[*args.AnalysisType] {
 		flagset.PrintDefaults()
-		return nil, fmt.Errorf("wrong output format %s; must be one of %s", *args.OutputFormat, getSupportedValuesString(supportedOutputFormats))
+		return fmt.Errorf("wrong analysis type %s; must be one of: %s", *args.AnalysisType, getSupportedValuesString(supportedAnalysisTypes))
 	}
-
-	if !supportedAnalysisTypes[*args.AnalysisType] {
+	if !supportedOutputFormats[*args.OutputFormat] {
 		flagset.PrintDefaults()
-		return nil, fmt.Errorf("wrong analysis type %s; must be one of: %s", *args.AnalysisType, getSupportedValuesString(supportedAnalysisTypes))
+		return fmt.Errorf("wrong output format %s; must be one of %s", *args.OutputFormat, getSupportedValuesString(supportedOutputFormats))
+	}
+	diffAnalysis := *args.AnalysisType == allEndpointsDiff || *args.AnalysisType == allSubnetsDiff
+	fileForDiffSpecified := args.InputSecondConfigFile != nil && *args.InputSecondConfigFile != ""
+	if fileForDiffSpecified && !diffAnalysis {
+		return fmt.Errorf("wrong analysis type %s for 2nd file (%v) specified for diff",
+			*args.AnalysisType, *args.InputSecondConfigFile)
+	}
+	if !fileForDiffSpecified && diffAnalysis {
+		return fmt.Errorf("missing parameter vpc-config-second for diff analysis %s", *args.AnalysisType)
 	}
+	return nil
+}
 
+func notSupportedYetArgs(args *InArgs) error {
 	if *args.AnalysisType != allEndpoints && *args.OutputFormat != TEXTFormat && *args.OutputFormat != JSONFormat {
-		return nil, fmt.Errorf("currently only txt/json output format supported with %s analysis type", *args.AnalysisType)
+		return fmt.Errorf("currently only txt/json output format supported with %s analysis type", *args.AnalysisType)
 	}
-
 	if *args.AnalysisType == singleSubnet && *args.Grouping {
-		return nil, fmt.Errorf("currently singleSubnet analysis type does not support grouping")
+		return fmt.Errorf("currently singleSubnet analysis type does not support grouping")
 	}
-
 	if *args.OutputFormat == JSONFormat && *args.Grouping {
-		return nil, fmt.Errorf("json output format is not supported with grouping")
+		return fmt.Errorf("json output format is not supported with grouping")
 	}
-	return &args, nil
+	if *args.AnalysisType == allSubnetsDiff && *args.OutputFormat != TEXTFormat {
+		return fmt.Errorf("currently only txt output format supported with diff_all_subnets")
+	}
+	return nil
 }

pkg/ibmvpc/analysis_output_test.go

@@ -33,6 +33,7 @@ const (
 type vpcGeneralTest struct {
 	name           string                            // test name
 	inputConfig    string                            // name (relative path) of input config file (json)
+	inputConfig2nd string                            // 2nd input file for diff
 	expectedOutput map[vpcmodel.OutputUseCase]string // expected output file path
 	actualOutput   map[vpcmodel.OutputUseCase]string // actual output file path
 	useCases       []vpcmodel.OutputUseCase          // the list of output use cases to test
@@ -50,6 +51,7 @@ const (
 	suffixOutFileDebugSubnet       = "_analysisPerSubnetSeparately"
 	suffixOutFileSubnetsLevel      = "subnetsBased_withPGW"
 	suffixOutFileSubnetsLevelNoPGW = "subnetsBased_withoutPGW"
+	suffixOutFileDiffSubnets       = "subnetsDiff"
 	txtOutSuffix                   = ".txt"
 	debugOutSuffix                 = "_debug.txt"
 	mdOutSuffix                    = ".md"
@@ -93,6 +95,8 @@ func getTestFileName(testName string,
 		res = baseName + suffixOutFileSubnetsLevel
 	case vpcmodel.AllSubnetsNoPGW:
 		res = baseName + suffixOutFileSubnetsLevelNoPGW
+	case vpcmodel.AllSubnetsDiff:
+		res = baseName + suffixOutFileDiffSubnets
 	}
 	switch format {
 	case vpcmodel.Text:
@@ -120,9 +124,9 @@ func getTestFileName(testName string,
 // files names (actual and expected), per use case
 func (tt *vpcGeneralTest) initTest() {
 	tt.inputConfig = inputFilePrefix + tt.name + ".json"
+	tt.inputConfig2nd = inputFilePrefix + tt.name + "_2nd.json"
 	tt.expectedOutput = map[vpcmodel.OutputUseCase]string{}
 	tt.actualOutput = map[vpcmodel.OutputUseCase]string{}
-
 	// init field of expected errs
 	if tt.errPerUseCase == nil {
 		tt.errPerUseCase = map[vpcmodel.OutputUseCase]error{}
@@ -326,6 +330,11 @@ var tests = []*vpcGeneralTest{
 		useCases: []vpcmodel.OutputUseCase{vpcmodel.AllSubnets},
 		format:   vpcmodel.Text,
 	},
+	{
+		name:     "acl_testing5",
+		useCases: []vpcmodel.OutputUseCase{vpcmodel.AllSubnetsDiff},
+		format:   vpcmodel.Text,
+	},
 }
 
 var formatsAvoidComparison = map[vpcmodel.OutFormat]bool{vpcmodel.ARCHDRAWIO: true, vpcmodel.DRAWIO: true}
@@ -387,11 +396,23 @@ func (tt *vpcGeneralTest) runTest(t *testing.T) {
 	tt.initTest()
 
 	// get vpcConfigs obj from parsing + analyzing input config file
-	vpcConfigs := getVPCConfigs(t, tt)
+	vpcConfigs := getVPCConfigs(t, tt, true)
+	var vpcConfigs2nd map[string]*vpcmodel.VPCConfig
+	diffUseCase := false
+	for _, useCase := range tt.useCases {
+		if useCase == vpcmodel.AllSubnetsDiff {
+			diffUseCase = true
+		}
+	}
+	if diffUseCase {
+		vpcConfigs2nd = getVPCConfigs(t, tt, false)
+	} else { // inputConfig2nd should be ignored if not diffUseCase
+		tt.inputConfig2nd = ""
+	}
 
 	// generate actual output for all use cases specified for this test
 	for _, uc := range tt.useCases {
-		err := runTestPerUseCase(t, tt, vpcConfigs, uc, tt.mode)
+		err := runTestPerUseCase(t, tt, vpcConfigs, vpcConfigs2nd, uc, tt.mode)
 		require.Equal(t, tt.errPerUseCase[uc], err, "comparing actual err to expected err")
 	}
 	for uc, outFile := range tt.actualOutput {
@@ -400,8 +421,14 @@ func (tt *vpcGeneralTest) runTest(t *testing.T) {
 }
 
 // getVPCConfigs returns  map[string]*vpcmodel.VPCConfig obj for the input test (config json file)
-func getVPCConfigs(t *testing.T, tt *vpcGeneralTest) map[string]*vpcmodel.VPCConfig {
-	inputConfigFile := filepath.Join(getTestsDir(), tt.inputConfig)
+func getVPCConfigs(t *testing.T, tt *vpcGeneralTest, firstCfg bool) map[string]*vpcmodel.VPCConfig {
+	var inputConfig string
+	if firstCfg {
+		inputConfig = tt.inputConfig
+	} else {
+		inputConfig = tt.inputConfig2nd
+	}
+	inputConfigFile := filepath.Join(getTestsDir(), inputConfig)
 	inputConfigContent, err := os.ReadFile(inputConfigFile)
 	if err != nil {
 		t.Fatalf("err: %s", err)
@@ -461,18 +488,23 @@ func initTestFileNames(tt *vpcGeneralTest,
 // runTestPerUseCase runs the connectivity analysis for the required use case and compares/generates the output
 func runTestPerUseCase(t *testing.T,
 	tt *vpcGeneralTest,
-	c map[string]*vpcmodel.VPCConfig,
+	c1, c2 map[string]*vpcmodel.VPCConfig,
 	uc vpcmodel.OutputUseCase,
 	mode testMode) error {
-	numConfigs := len(c)
+	numConfigs := len(c1)
 	allVPCsOutput := make([]*vpcmodel.VPCAnalysisOutput, numConfigs)
 	i := 0
-	for _, vpcConfig := range c {
+	var vpcConfig2nd *vpcmodel.VPCConfig
+	// note that for diff analysis mode a single vpcConfig in c1 is provided; c2 is assumed to have a single cfg
+	for _, vpcConfig := range c2 {
+		vpcConfig2nd = vpcConfig
+	}
+	for _, vpcConfig := range c1 {
 		if err := initTestFileNames(tt, uc, numConfigs, vpcConfig.VPC.Name(), false); err != nil {
 			return err
 		}
 
-		og, err := vpcmodel.NewOutputGenerator(vpcConfig, tt.grouping, uc, tt.format == vpcmodel.ARCHDRAWIO)
+		og, err := vpcmodel.NewOutputGenerator(vpcConfig, vpcConfig2nd, tt.grouping, uc, tt.format == vpcmodel.ARCHDRAWIO)
 		if err != nil {
 			return err
 		}

pkg/ibmvpc/examples/acl_testing5subnetsBased_withPGW.txt

@@ -1,4 +1,4 @@
-Analysis for VPC test-vpc-ky
+Analysis for VPC test-vpc-ky1
 combined connections between subnets:
 sub1-1-ky => Public Internet 8.8.8.8/32 : protocol: UDP dst-ports: 53
 sub1-1-ky => sub1-2-ky : protocol: TCP