detect license ID from full text when incidentally provided as a value (#3876)

---------
Signed-off-by: Christopher Phillips <32073428+spiffcs@users.noreply.github.com>
Signed-off-by: Alex Goodman <wagoodman@users.noreply.github.com>
Co-authored-by: Alex Goodman <wagoodman@users.noreply.github.com>
This commit is contained in:
Christopher Angelo Phillips
2025-05-13 16:37:18 -04:00
committed by GitHub
co-authored by Alex Goodman
parent b4d717fb30
commit f77d503892
117 changed files with 2082 additions and 1327 deletions
+12 -22
View File
@@ -25,10 +25,6 @@ import (
// parseWheelOrEgg takes the primary metadata file reference and returns the python package it represents. Contained
// fields are governed by the PyPA core metadata specification (https://packaging.python.org/en/latest/specifications/core-metadata/).
func parseWheelOrEgg(ctx context.Context, resolver file.Resolver, _ *generic.Environment, reader file.LocationReadCloser) ([]pkg.Package, []artifact.Relationship, error) {
licenseScanner, err := licenses.ContextLicenseScanner(ctx)
if err != nil {
return nil, nil, err
}
pd, sources, err := assembleEggOrWheelMetadata(resolver, reader.Location)
if err != nil {
return nil, nil, err
@@ -46,7 +42,7 @@ func parseWheelOrEgg(ctx context.Context, resolver file.Resolver, _ *generic.Env
pkgs := []pkg.Package{
newPackageForPackage(
*pd,
findLicenses(ctx, licenseScanner, resolver, *pd),
findLicenses(ctx, resolver, *pd),
sources...,
),
}
@@ -253,7 +249,7 @@ func assembleEggOrWheelMetadata(resolver file.Resolver, metadataLocation file.Lo
return &pd, sources, nil
}
func findLicenses(ctx context.Context, scanner licenses.Scanner, resolver file.Resolver, m parsedData) pkg.LicenseSet {
func findLicenses(ctx context.Context, resolver file.Resolver, m parsedData) pkg.LicenseSet {
var licenseSet pkg.LicenseSet
licenseLocations := file.NewLocationSet()
@@ -268,9 +264,9 @@ func findLicenses(ctx context.Context, scanner licenses.Scanner, resolver file.R
switch {
case m.LicenseExpression != "" || m.Licenses != "":
licenseSet = getLicenseSetFromValues(licenseLocations.ToSlice(), m.LicenseExpression, m.Licenses)
licenseSet = getLicenseSetFromValues(ctx, licenseLocations.ToSlice(), m.LicenseExpression, m.Licenses)
case !licenseLocations.Empty():
licenseSet = getLicenseSetFromFiles(ctx, scanner, resolver, licenseLocations.ToSlice()...)
licenseSet = getLicenseSetFromFiles(ctx, resolver, licenseLocations.ToSlice()...)
default:
// search for known license paths from RECORDS file
@@ -302,14 +298,14 @@ func findLicenses(ctx context.Context, scanner licenses.Scanner, resolver file.R
locationSet.Add(locs...)
}
licenseSet = getLicenseSetFromFiles(ctx, scanner, resolver, locationSet.ToSlice()...)
licenseSet = getLicenseSetFromFiles(ctx, resolver, locationSet.ToSlice()...)
}
return licenseSet
}
func getLicenseSetFromValues(locations []file.Location, licenseValues ...string) pkg.LicenseSet {
func getLicenseSetFromValues(ctx context.Context, locations []file.Location, licenseValues ...string) pkg.LicenseSet {
if len(locations) == 0 {
return pkg.NewLicenseSet(pkg.NewLicensesFromValues(licenseValues...)...)
return pkg.NewLicenseSet(pkg.NewLicensesFromValuesWithContext(ctx, licenseValues...)...)
}
licenseSet := pkg.NewLicenseSet()
@@ -318,31 +314,25 @@ func getLicenseSetFromValues(locations []file.Location, licenseValues ...string)
continue
}
licenseSet.Add(pkg.NewLicenseFromLocations(value, locations...))
licenseSet.Add(pkg.NewLicenseFromLocationsWithContext(ctx, value, locations...))
}
return licenseSet
}
func getLicenseSetFromFiles(ctx context.Context, scanner licenses.Scanner, resolver file.Resolver, locations ...file.Location) pkg.LicenseSet {
func getLicenseSetFromFiles(ctx context.Context, resolver file.Resolver, locations ...file.Location) pkg.LicenseSet {
licenseSet := pkg.NewLicenseSet()
for _, loc := range locations {
licenseSet.Add(getLicenseSetFromFile(ctx, scanner, resolver, loc)...)
licenseSet.Add(getLicenseSetFromFile(ctx, resolver, loc)...)
}
return licenseSet
}
func getLicenseSetFromFile(ctx context.Context, scanner licenses.Scanner, resolver file.Resolver, location file.Location) []pkg.License {
func getLicenseSetFromFile(ctx context.Context, resolver file.Resolver, location file.Location) []pkg.License {
metadataContents, err := resolver.FileContentsByLocation(location)
if err != nil {
log.WithFields("error", err, "path", location.Path()).Trace("unable to read file contents")
return nil
}
defer internal.CloseAndLogError(metadataContents, location.Path())
parsed, err := scanner.PkgSearch(ctx, file.NewLocationReadCloser(location, metadataContents))
if err != nil {
log.WithFields("error", err, "path", location.Path()).Trace("unable to parse a license from the file")
return nil
}
return parsed
return pkg.NewLicensesFromReadCloserWithContext(ctx, file.NewLocationReadCloser(location, metadataContents))
}