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.go
+++ b/internal/generatedtext/limits.go
@@ -0,0 +1,120 @@
+package generatedtext
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"unicode/utf8"
+)
+
+const (
+	// MaxGeneratedTextBytes is the largest provider response accepted for one
+	// generated-text report before JSON decoding begins.
+	MaxGeneratedTextBytes = 64 * 1024
+
+	maxGeneratedTextCharacters          = 20_000
+	maxGeneratedTextStringCharacters    = 4_000
+	maxHourlyDiscussionCharacters       = 12_000
+	maxDayStyleDiscussionParagraphs     = 12
+	maxDayStyleDiscussionParagraphChars = 4_000
+)
+
+// ValidateRawOutput rejects generated JSON before it is decoded or retained.
+func ValidateRawOutput(data []byte) error {
+	if len(data) > MaxGeneratedTextBytes {
+		return fmt.Errorf("generated text exceeds the %d-byte limit", MaxGeneratedTextBytes)
+	}
+	return nil
+}
+
+func validateGeneratedTextSize(data []byte, name string) error {
+	if err := ValidateRawOutput(data); err != nil {
+		return fmt.Errorf("%s %w", name, err)
+	}
+	return nil
+}
+
+func validateGeneratedTextField(name, field string, raw json.RawMessage, total *int) error {
+	switch field {
+	case "summary", "precipitation_timing":
+		return validateGeneratedTextString(name, field, raw, maxGeneratedTextStringCharacters, total)
+	case "forecast_discussion":
+		if name == "hourly" {
+			return validateGeneratedTextString(name, field, raw, maxHourlyDiscussionCharacters, total)
+		}
+		return validateDayStyleDiscussion(name, raw, total)
+	default:
+		return fmt.Errorf("decode %s generated text: unsupported field", name)
+	}
+}
+
+func validateGeneratedTextString(name, field string, raw json.RawMessage, limit int, total *int) error {
+	if bytes.Equal(bytes.TrimSpace(raw), []byte("null")) {
+		return fmt.Errorf("%s generated text %s must be a string", name, generatedTextFieldLabel(field))
+	}
+	var value string
+	if err := json.Unmarshal(raw, &value); err != nil {
+		return fmt.Errorf("%s generated text %s must be a string", name, generatedTextFieldLabel(field))
+	}
+	characters := utf8.RuneCountInString(value)
+	if characters > limit {
+		return fmt.Errorf("%s generated text %s exceeds the %d-character limit", name, generatedTextFieldLabel(field), limit)
+	}
+	return addGeneratedTextCharacters(name, characters, total)
+}
+
+func validateDayStyleDiscussion(name string, raw json.RawMessage, total *int) error {
+	decoder := json.NewDecoder(bytes.NewReader(raw))
+	token, err := decoder.Token()
+	if err != nil {
+		return fmt.Errorf("%s generated text forecast discussion must be an array", name)
+	}
+	if delimiter, ok := token.(json.Delim); !ok || delimiter != '[' {
+		return fmt.Errorf("%s generated text forecast discussion must be an array", name)
+	}
+
+	paragraphs := 0
+	for decoder.More() {
+		paragraphs++
+		if paragraphs > maxDayStyleDiscussionParagraphs {
+			return fmt.Errorf("%s generated text forecast discussion exceeds the %d-paragraph limit", name, maxDayStyleDiscussionParagraphs)
+		}
+		var rawParagraph json.RawMessage
+		if err := decoder.Decode(&rawParagraph); err != nil || bytes.Equal(bytes.TrimSpace(rawParagraph), []byte("null")) {
+			return fmt.Errorf("%s generated text forecast discussion paragraphs must be strings", name)
+		}
+		var paragraph string
+		if err := json.Unmarshal(rawParagraph, &paragraph); err != nil {
+			return fmt.Errorf("%s generated text forecast discussion paragraphs must be strings", name)
+		}
+		characters := utf8.RuneCountInString(paragraph)
+		if characters > maxDayStyleDiscussionParagraphChars {
+			return fmt.Errorf("%s generated text forecast discussion paragraphs exceed the %d-character limit", name, maxDayStyleDiscussionParagraphChars)
+		}
+		if err := addGeneratedTextCharacters(name, characters, total); err != nil {
+			return err
+		}
+	}
+	if _, err := decoder.Token(); err != nil {
+		return fmt.Errorf("%s generated text forecast discussion must be an array", name)
+	}
+	return nil
+}
+
+func addGeneratedTextCharacters(name string, characters int, total *int) error {
+	*total += characters
+	if *total > maxGeneratedTextCharacters {
+		return fmt.Errorf("%s generated text exceeds the %d-character limit", name, maxGeneratedTextCharacters)
+	}
+	return nil
+}
+
+func generatedTextFieldLabel(field string) string {
+	if field == "forecast_discussion" {
+		return "forecast discussion"
+	}
+	if field == "precipitation_timing" {
+		return "precipitation timing"
+	}
+	return field
+}