feat: truncate "summary" and fallback to description if advisory does not have a summary

go/internal/database/datastore/listed_vulnerability.go

@@ -18,6 +18,8 @@ import (
 	"regexp"
 	"slices"
 	"strings"
+	"unicode"
+	"unicode/utf8"
 
 	"github.com/ossf/osv-schema/bindings/go/osvschema"
 )
@@ -72,9 +74,50 @@ func removeVariants(ecosystem string) string {
 	return ""
 }
 
+// truncate ensures that the given string is shorter than 80 characters.
+//
+// If the string is longer than that limit, it's trimmed and suffixed with an ellipsis.
+// Ideally the string will be trimmed at the space that's closest to the limit to
+// preserve whole words; if a string has no spaces before the limit, it'll be forcefully truncated.
+func truncate(str string) string {
+	count := 0
+	truncateAt := -1
+
+	for i, c := range str {
+		if unicode.IsSpace(c) {
+			truncateAt = i
+		}
+
+		count++
+
+		if count >= 80 {
+			// ideally we want to keep words whole when truncating,
+			// but if we can't find a space just truncate at the limit
+			if truncateAt == -1 {
+				_, size := utf8.DecodeRuneInString(str[i:])
+				truncateAt = i + size
+			}
+
+			return str[:truncateAt] + "..."
+		}
+	}
+
+	return str
+}
+
+func describe(vuln *osvschema.Vulnerability) string {
+	summary := vuln.GetSummary()
+
+	if summary == "" {
+		summary = vuln.GetDetails()
+	}
+
+	return truncate(summary)
+}
+
 func NewListedVulnerabilityFromProto(vuln *osvschema.Vulnerability) *ListedVulnerability {
 	published := vuln.GetPublished().AsTime()
-	summary := vuln.GetSummary()
+	summary := describe(vuln)
 
 	allEcosystems := make(map[string]struct{})
 	allPackages := make(map[string]struct{})

go/internal/database/datastore/listed_vulnerability_test.go

@@ -1,8 +1,10 @@
 package datastore
 
 import (
+	"strings"
 	"testing"
 	"time"
+	"unicode/utf8"
 
 	"github.com/google/go-cmp/cmp"
 	"github.com/google/go-cmp/cmp/cmpopts"
@@ -106,3 +108,81 @@ func TestNewListedVulnerabilityFromProto(t *testing.T) {
 		t.Errorf("NewListedVulnerabilityFromProto mismatch (-want +got):\n%s", diff)
 	}
 }
+
+func TestNewListedVulnerabilityFromProto_Summary(t *testing.T) {
+	tests := []struct {
+		name    string
+		summary string
+		details string
+		want    string
+	}{
+		{want: ""},
+		{
+			summary: "hello world",
+			details: "hello sunshine",
+			want:    "hello world",
+		},
+		{
+			summary: "",
+			details: "hello sunshine",
+			want:    "hello sunshine",
+		},
+		{
+			summary: " ",
+			details: "hello sunshine",
+			want:    " ",
+		},
+		{
+			summary: "   ",
+			details: "hello sunshine",
+			want:    "   ",
+		},
+		{
+			summary: strings.Repeat("!", 100),
+			details: strings.Repeat("?", 100),
+			want:    strings.Repeat("!", 80) + "...",
+		},
+		{
+			summary: strings.Repeat("!", 100),
+			details: "hello sunshine",
+			want:    strings.Repeat("!", 80) + "...",
+		},
+		{
+			summary: "",
+			details: strings.Repeat("!", 100),
+			want:    strings.Repeat("!", 80) + "...",
+		},
+		{
+			summary: "",
+			details: strings.Repeat("\u754c", 100),
+			want:    strings.Repeat("\u754c", 80) + "...",
+		},
+		{
+			summary: "",
+			details: strings.Repeat("\u754c", 30) + " " + strings.Repeat("\u8a9e", 60),
+			want:    strings.Repeat("\u754c", 30) + "...",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Parallel()
+
+			vuln := &osvschema.Vulnerability{
+				Id:        "TEST-123",
+				Published: timestamppb.New(time.Date(2025, time.January, 1, 0, 0, 0, 0, time.UTC)),
+				Summary:   tt.summary,
+				Details:   tt.details,
+			}
+
+			got := NewListedVulnerabilityFromProto(vuln).Summary
+
+			if got != tt.want {
+				t.Fatalf("Summary = %q, want %q", got, tt.want)
+			}
+			if !utf8.ValidString(got) {
+				t.Fatalf("Summary returned invalid UTF-8: %q", got)
+			}
+		})
+	}
+}