Bound generated text content and diagnostics

2026-08-13 01:47:47 +00:00
parent f8beed04cf
commit e520ffb13b
20 changed files with 400 additions and 45 deletions
--- a/internal/generatedtext/limits_test.go
+++ b/internal/generatedtext/limits_test.go
@@ -0,0 +1,112 @@
+package generatedtext
+
+import (
+	"fmt"
+	"strings"
+	"testing"
+)
+
+func TestValidateHourlyGeneratedTextEnforcesContentLimits(t *testing.T) {
+	input := func(summary, discussion, timing string) []byte {
+		return []byte(fmt.Sprintf(`{"summary":%q,"forecast_discussion":%q,"precipitation_timing":%q}`, summary, discussion, timing))
+	}
+	for _, test := range []struct {
+		name  string
+		input []byte
+		want  string
+	}{
+		{name: "summary at limit", input: input(strings.Repeat("s", maxGeneratedTextStringCharacters), "discussion", "")},
+		{name: "summary over limit", input: input(strings.Repeat("s", maxGeneratedTextStringCharacters+1), "discussion", ""), want: "summary exceeds"},
+		{name: "discussion at limit", input: input("summary", strings.Repeat("d", maxHourlyDiscussionCharacters), "")},
+		{name: "discussion over limit", input: input("summary", strings.Repeat("d", maxHourlyDiscussionCharacters+1), ""), want: "forecast discussion exceeds"},
+		{name: "total at limit", input: input(strings.Repeat("s", 4_000), strings.Repeat("d", 12_000), strings.Repeat("t", 4_000))},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			err := validateHourlyJSON(test.input)
+			if test.want == "" && err != nil {
+				t.Fatalf("validate() error = %v", err)
+			}
+			if test.want != "" && (err == nil || !strings.Contains(err.Error(), test.want)) {
+				t.Fatalf("validate() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestValidateDayStyleGeneratedTextEnforcesDiscussionLimits(t *testing.T) {
+	for _, test := range []struct {
+		name  string
+		input string
+		want  string
+	}{
+		{
+			name:  "paragraph count at limit",
+			input: fmt.Sprintf(`{"summary":"summary","forecast_discussion":[%s],"precipitation_timing":""}`, strings.TrimSuffix(strings.Repeat(`"p",`, maxDayStyleDiscussionParagraphs), ",")),
+		},
+		{
+			name:  "paragraph count over limit",
+			input: fmt.Sprintf(`{"summary":"summary","forecast_discussion":[%s],"precipitation_timing":""}`, strings.TrimSuffix(strings.Repeat(`"p",`, maxDayStyleDiscussionParagraphs+1), ",")),
+			want:  "12-paragraph limit",
+		},
+		{
+			name:  "paragraph at limit",
+			input: fmt.Sprintf(`{"summary":"summary","forecast_discussion":[%q],"precipitation_timing":""}`, strings.Repeat("p", maxDayStyleDiscussionParagraphChars)),
+		},
+		{
+			name:  "paragraph over limit",
+			input: fmt.Sprintf(`{"summary":"summary","forecast_discussion":[%q],"precipitation_timing":""}`, strings.Repeat("p", maxDayStyleDiscussionParagraphChars+1)),
+			want:  "paragraphs exceed the 4000-character limit",
+		},
+		{
+			name: "total at limit",
+			input: fmt.Sprintf(`{"summary":%q,"forecast_discussion":[%q,%q,%q,%q],"precipitation_timing":""}`,
+				strings.Repeat("s", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000)),
+		},
+		{
+			name: "total over limit",
+			input: fmt.Sprintf(`{"summary":%q,"forecast_discussion":[%q,%q,%q,%q,%q],"precipitation_timing":""}`,
+				strings.Repeat("s", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000), strings.Repeat("p", 4_000)),
+			want: "exceeds the 20000-character limit",
+		},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			err := validateDailyJSON([]byte(test.input))
+			if test.want == "" && err != nil {
+				t.Fatalf("validate() error = %v", err)
+			}
+			if test.want != "" && (err == nil || !strings.Contains(err.Error(), test.want)) {
+				t.Fatalf("validate() error = %v, want %q", err, test.want)
+			}
+		})
+	}
+}
+
+func TestValidateGeneratedTextRejectsOversizedRawOutputWithoutLeakingContent(t *testing.T) {
+	valid := []byte(`{"summary":"summary","forecast_discussion":"discussion","precipitation_timing":""}`)
+	atLimit := append([]byte(strings.Repeat(" ", MaxGeneratedTextBytes-len(valid))), valid...)
+	if err := validateHourlyJSON(atLimit); err != nil {
+		t.Fatalf("validate() at raw limit error = %v", err)
+	}
+
+	marker := "provider-controlled-marker"
+	unknownPrefix := `{"summary":"summary","forecast_discussion":"discussion","precipitation_timing":"","`
+	unknownSuffix := `":"value"}`
+	unknownKey := marker + strings.Repeat("x", MaxGeneratedTextBytes-len(unknownPrefix)-len(unknownSuffix)-len(marker))
+	unknownField := []byte(unknownPrefix + unknownKey + unknownSuffix)
+	err := validateHourlyJSON(unknownField)
+	if err == nil || !strings.Contains(err.Error(), "unsupported field") {
+		t.Fatalf("validate() error = %v, want unsupported field error", err)
+	}
+	if len(err.Error()) > 128 || strings.Contains(err.Error(), marker) {
+		t.Fatalf("validate() leaked unknown field content: %q", err)
+	}
+
+	input := append(atLimit, []byte(marker)...)
+	err = validateHourlyJSON(input)
+	if err == nil || !strings.Contains(err.Error(), "65536-byte limit") {
+		t.Fatalf("validate() error = %v, want raw size error", err)
+	}
+	if len(err.Error()) > 128 || strings.Contains(err.Error(), marker) {
+		t.Fatalf("validate() leaked oversized provider content: %q", err)
+	}
+}