Initial commit

internal/classifier/parse.go

@@ -0,0 +1,123 @@
+package classifier
+
+import (
+	"encoding/json"
+	"fmt"
+	"regexp"
+	"strings"
+)
+
+// Tier1Result is the LLM response for Tier 1 (routing only) stories.
+type Tier1Result struct {
+	Channel     string   `json:"channel"`
+	Platforms   []string `json:"platforms"`
+	DuplicateOf *string  `json:"duplicate_of"`
+	Rationale   string   `json:"rationale"`
+}
+
+// Tier2Result is the LLM response for Tier 2 (gating + routing) stories.
+type Tier2Result struct {
+	Post        bool     `json:"post"`
+	Channel     string   `json:"channel"`
+	Platforms   []string `json:"platforms"`
+	DuplicateOf *string  `json:"duplicate_of"`
+	Rationale   string   `json:"rationale"`
+}
+
+var validChannels = map[string]bool{
+	"tech":     true,
+	"politics": true,
+	"gaming":   true,
+}
+
+var validPlatforms = map[string]bool{
+	"nintendo-switch": true,
+	"playstation":     true,
+	"xbox":            true,
+	"pc":              true,
+	"retro":           true,
+	"mobile":          true,
+	"multi":           true,
+}
+
+var thinkRe = regexp.MustCompile(`(?s)<think>.*?</think>`)
+var fenceRe = regexp.MustCompile("(?s)```(?:json)?\\s*(.+?)\\s*```")
+var trailingCommaRe = regexp.MustCompile(`,\s*([}\]])`)
+
+// ParseTier1Response parses an LLM response into a Tier1Result.
+func ParseTier1Response(raw string) (*Tier1Result, error) {
+	cleaned := repairJSON(raw)
+
+	var result Tier1Result
+	if err := json.Unmarshal([]byte(cleaned), &result); err != nil {
+		return nil, fmt.Errorf("parse tier1 response: %w (raw: %s)", err, truncate(raw, 200))
+	}
+
+	if !validChannels[result.Channel] {
+		return nil, fmt.Errorf("invalid channel: %q", result.Channel)
+	}
+
+	result.Platforms = filterPlatforms(result.Platforms)
+	return &result, nil
+}
+
+// ParseTier2Response parses an LLM response into a Tier2Result.
+func ParseTier2Response(raw string) (*Tier2Result, error) {
+	cleaned := repairJSON(raw)
+
+	var result Tier2Result
+	if err := json.Unmarshal([]byte(cleaned), &result); err != nil {
+		return nil, fmt.Errorf("parse tier2 response: %w (raw: %s)", err, truncate(raw, 200))
+	}
+
+	// Always validate channel — gating may override post=false to true later
+	if !validChannels[result.Channel] {
+		if result.Post {
+			return nil, fmt.Errorf("invalid channel: %q", result.Channel)
+		}
+		// Default to "politics" for discarded stories with garbage channels
+		result.Channel = "politics"
+	}
+
+	result.Platforms = filterPlatforms(result.Platforms)
+	return &result, nil
+}
+
+// repairJSON attempts to fix common LLM output issues.
+func repairJSON(raw string) string {
+	s := strings.TrimSpace(raw)
+
+	// Strip <think>...</think> blocks (some models include reasoning)
+	s = thinkRe.ReplaceAllString(s, "")
+
+	// Strip markdown code fences
+	if matches := fenceRe.FindStringSubmatch(s); len(matches) > 1 {
+		s = matches[1]
+	}
+
+	s = strings.TrimSpace(s)
+
+	// Remove trailing commas before closing braces/brackets (outside quoted strings)
+	s = trailingCommaRe.ReplaceAllString(s, "$1")
+
+	return s
+}
+
+func filterPlatforms(platforms []string) []string {
+	var valid []string
+	for _, p := range platforms {
+		p = strings.TrimSpace(strings.ToLower(p))
+		if validPlatforms[p] {
+			valid = append(valid, p)
+		}
+	}
+	return valid
+}
+
+func truncate(s string, n int) string {
+	runes := []rune(s)
+	if len(runes) <= n {
+		return s
+	}
+	return string(runes[:n]) + "..."
+}