feat: use publiccode-parser-go v4

crawler/crawler.go

@@ -18,7 +18,7 @@ import (
 	"github.com/italia/publiccode-crawler/v4/git"
 	"github.com/italia/publiccode-crawler/v4/metrics"
 	"github.com/italia/publiccode-crawler/v4/scanner"
-	publiccode "github.com/italia/publiccode-parser-go/v3"
+	publiccode "github.com/italia/publiccode-parser-go/v4"
 	log "github.com/sirupsen/logrus"
 	"github.com/spf13/viper"
 	"golang.org/x/exp/slices"
@@ -327,15 +327,6 @@ func (c *Crawler) ProcessRepo(repository common.Repository) { //nolint:maintidx
 		),
 	)
 
-	var parser *publiccode.Parser
-	parser, err = publiccode.NewParser(repository.FileRawURL)
-	if err != nil {
-		logEntries = append(logEntries, fmt.Sprintf("[%s] BAD publiccode.yml: %s\n", repository.Name, err.Error()))
-		metrics.GetCounter("repository_bad_publiccodeyml", c.Index).Inc()
-
-		return
-	}
-
 	//nolint:godox
 	// FIXME: this is hardcoded for now, because it requires changes to publiccode-parser-go.
 	domain := publiccode.Domain{
@@ -344,9 +335,22 @@ func (c *Crawler) ProcessRepo(repository common.Repository) { //nolint:maintidx
 		BasicAuth:   []string{os.Getenv("GITHUB_TOKEN")},
 	}
 
+	var parser *publiccode.Parser
+	parser, err = publiccode.NewParser(publiccode.ParserConfig{Domain: domain})
+	if err != nil {
+		logEntries = append(
+			logEntries,
+			fmt.Sprintf("[%s] can't create a Parser: %s\n", repository.Name, err.Error()),
+		)
+
+		return
+	}
+
+	var parsed publiccode.PublicCode
+	parsed, err = parser.Parse(repository.FileRawURL)
+
 	valid := true
 
-	err = parser.ParseInDomain(resp.Body, domain.Host, domain.UseTokenFor, domain.BasicAuth)
 	if err != nil {
 		var validationResults publiccode.ValidationResults
 		if errors.As(err, &validationResults) {
@@ -363,7 +367,7 @@ func (c *Crawler) ProcessRepo(repository common.Repository) { //nolint:maintidx
 
 	publisherID := viper.GetString("MAIN_PUBLISHER_ID")
 	if valid && repository.Publisher.ID != publisherID {
-		err = validateFile(repository.Publisher, *parser, repository.FileRawURL)
+		err = validateFile(repository.Publisher, parsed.(publiccode.PublicCodeV0), repository.FileRawURL)
 		if err != nil {
 			valid = false
 		}
@@ -391,7 +395,7 @@ func (c *Crawler) ProcessRepo(repository common.Repository) { //nolint:maintidx
 		aliases = append(aliases, repository.URL.String())
 	}
 
-	publiccodeYml, err := parser.ToYAML()
+	publiccodeYml, err := parsed.ToYAML()
 	if err != nil {
 		logEntries = append(logEntries, fmt.Sprintf("[%s] parsing error: %s", repository.Name, err.Error()))
 
@@ -433,7 +437,7 @@ func (c *Crawler) ProcessRepo(repository common.Repository) { //nolint:maintidx
 
 	if !viper.GetBool("SKIP_VITALITY") && !c.DryRun {
 		// Clone repository.
-		err = git.CloneRepository(repository.URL.Host, repository.Name, parser.PublicCode.URL.String(), c.Index)
+		err = git.CloneRepository(repository.URL.Host, repository.Name, parsed.Url().String(), c.Index)
 		if err != nil {
 			logEntries = append(logEntries, fmt.Sprintf("[%s] error while cloning: %v\n", repository.Name, err))
 		}
@@ -460,11 +464,11 @@ func (c *Crawler) ProcessRepo(repository common.Repository) { //nolint:maintidx
 // validateFile performs additional validations that are not strictly mandated
 // by the publiccode.yml Standard.
 // Using `one` command this check will be skipped.
-func validateFile(publisher common.Publisher, parser publiccode.Parser, fileRawURL string) error {
+func validateFile(publisher common.Publisher, parsed publiccode.PublicCodeV0, fileRawURL string) error {
 	u, _ := url.Parse(fileRawURL)
 	repo1 := vcsurl.GetRepo(u)
 
-	repo2 := vcsurl.GetRepo((*url.URL)(parser.PublicCode.URL))
+	repo2 := vcsurl.GetRepo((*url.URL)(parsed.Url()))
 
 	if repo1 != nil && repo2 != nil {
 		// Let's ignore the schema when checking for equality.
@@ -477,7 +481,7 @@ func validateFile(publisher common.Publisher, parser publiccode.Parser, fileRawU
 			return fmt.Errorf(
 				"declared url (%s) and actual publiccode.yml location URL (%s) "+
 					"are not in the same repo: '%s' vs '%s'",
-				parser.PublicCode.URL, fileRawURL, repo2, repo1,
+				parsed.Url(), fileRawURL, repo2, repo1,
 			)
 		}
 	}
@@ -494,11 +498,11 @@ func validateFile(publisher common.Publisher, parser publiccode.Parser, fileRawU
 
 	if !idIsUUID && !strings.EqualFold(
 		strings.TrimSpace(publisher.ID),
-		strings.TrimSpace(parser.PublicCode.It.Riuso.CodiceIPA),
+		strings.TrimSpace(parsed.It.Riuso.CodiceIPA),
 	) {
 		return fmt.Errorf(
 			"codiceIPA is '%s', but '%s' was expected for '%s' in %s",
-			parser.PublicCode.It.Riuso.CodiceIPA,
+			parsed.It.Riuso.CodiceIPA,
 			publisher.ID,
 			publisher.Name,
 			fileRawURL,

go.mod

@@ -6,7 +6,7 @@ require (
 	github.com/google/go-github/v43 v43.0.0
 	github.com/hashicorp/go-retryablehttp v0.7.1
 	github.com/italia/httpclient-lib-go v0.0.2
-	github.com/italia/publiccode-parser-go/v3 v3.1.3
+	github.com/italia/publiccode-parser-go/v4 v4.0.0
 	github.com/ktrysmt/go-bitbucket v0.9.63
 	github.com/prometheus/client_golang v1.11.1
 	github.com/prometheus/client_model v0.3.0

go.sum

@@ -914,8 +914,8 @@ github.com/inconshreveable/mousetrap v1.0.0 h1:Z8tu5sraLXCXIcARxBp/8cbvlwVa7Z1NH
 github.com/inconshreveable/mousetrap v1.0.0/go.mod h1:PxqpIevigyE2G7u3NXJIT2ANytuPF1OarO4DADm73n8=
 github.com/italia/httpclient-lib-go v0.0.2 h1:4bJLywTVd7qHPdKxJXvvhlXp436JTC4KA6dLhIl5a6c=
 github.com/italia/httpclient-lib-go v0.0.2/go.mod h1:b0/D3ULsBw8X+zEl7j/kSZmiMlUdj+agppneOvSq6eA=
-github.com/italia/publiccode-parser-go/v3 v3.1.3 h1:o4x0K6dJgBUh1Tac99AJEyM6uNV+e1GEJ7F5r+uMNMI=
-github.com/italia/publiccode-parser-go/v3 v3.1.3/go.mod h1:TvC+rGxBbIE+riQyey4GFtyC7GgKKTGmtsHINIp4aQ8=
+github.com/italia/publiccode-parser-go/v4 v4.0.0 h1:1/q4VH7WtnkaJyq3Cn0HU0KUz9X6uRPzmxNariKuI3w=
+github.com/italia/publiccode-parser-go/v4 v4.0.0/go.mod h1:qmxP/BgLwCeNMjfcXbRz8WCMPn85Pekcy+oGTUmEF4U=
 github.com/jbenet/go-context v0.0.0-20150711004518-d14ea06fba99 h1:BQSFePA1RWJOlocH6Fxy8MmwDt+yVQYULKfN0RoTN8A=
 github.com/jbenet/go-context v0.0.0-20150711004518-d14ea06fba99/go.mod h1:1lJo3i6rXxKeerYnT8Nvf0QmHCRC1n8sfWVwXF2Frvo=
 github.com/jonboulle/clockwork v0.1.0/go.mod h1:Ii8DK3G1RaLaWxj9trq07+26W01tbo22gdxWY5EU2bo=