knox/internal/ingest/log.go

package ingest

import (
	"bufio"
	"fmt"
	"os"
	"path/filepath"
	"regexp"
	"strings"
	"time"
)

var _ Ingester = (*LogIngester)(nil)

type LogIngester struct {
	Version string
}

func NewLogIngester() *LogIngester {
	return &LogIngester{Version: "log-parser/v1"}
}

func (l *LogIngester) SourceID() string { return "opencode-log" }

var (
	sessionIDRE = regexp.MustCompile(`session[=_ ]?([a-zA-Z0-9_-]+)`)
	projectIDRE = regexp.MustCompile(`project[=_ ]?([a-zA-Z0-9_.-]+)`)
	modelRE     = regexp.MustCompile(`model[=_ ]?([a-zA-Z0-9_.-]+)`)
	agentRE     = regexp.MustCompile(`agent[=_ ]?([a-zA-Z0-9_.-]+)`)
	errorRE     = regexp.MustCompile(`(?i)(error|fail|exception|panic|timeout)`)
)

func (l *LogIngester) Ingest(path string) (*IngestResult, error) {
	f, err := os.Open(path)
	if err != nil {
		return nil, fmt.Errorf("open log: %w", err)
	}
	defer f.Close()

	var (
		sessionIDs []string
		models     []string
		agents     []string
		errors     []string
		totalLines int
		firstError int
	)

	scanner := bufio.NewScanner(f)
	for scanner.Scan() {
		line := scanner.Text()
		totalLines++

		if m := sessionIDRE.FindStringSubmatch(line); len(m) > 1 {
			sessionIDs = append(sessionIDs, m[1])
		}
		if m := modelRE.FindStringSubmatch(line); len(m) > 1 {
			models = append(models, m[1])
		}
		if m := agentRE.FindStringSubmatch(line); len(m) > 1 {
			agents = append(agents, m[1])
		}
		if errorRE.MatchString(line) {
			if len(errors) == 0 {
				firstError = totalLines
			}
			errors = append(errors, truncate(line, 120))
		}
	}

	base := filepath.Base(path)
	fp := Fingerprint([]byte("log:" + base))
	title := fmt.Sprintf("Log %s (%d lines)", base, totalLines)
	createdAt := ""
	if fi, err := os.Stat(path); err == nil {
		createdAt = fi.ModTime().UTC().Format(time.RFC3339)
	}

	var summaryParts []string
	if len(sessionIDs) > 0 {
		summaryParts = append(summaryParts, fmt.Sprintf("sessions: %s", uniqueJoin(sessionIDs, 5)))
	}
	if len(models) > 0 {
		summaryParts = append(summaryParts, fmt.Sprintf("models: %s", uniqueJoin(models, 3)))
	}
	if len(errors) > 0 {
		summaryParts = append(summaryParts, fmt.Sprintf("%d errors (line %d)", len(errors), firstError))
	}
	summary := strings.Join(summaryParts, " | ")
	if summary == "" {
		summary = fmt.Sprintf("%d lines", totalLines)
	}

	confidence := 0.6
	if len(errors) > 0 {
		confidence = 0.8
	}

	return &IngestResult{
		Fingerprint:     fp,
		SourceID:        l.SourceID(),
		SourcePath:      base,
		ContentType:     ".log",
		Title:           title,
		Summary:         summary,
		CreatedAt:       createdAt,
		Confidence:      confidence,
		IngesterVersion: l.Version,
		LineEnd:         totalLines,
		Provenance: map[string]any{
			"session_ids": uniqueSlice(sessionIDs),
			"models":      uniqueSlice(models),
			"agents":      uniqueSlice(agents),
			"error_count": len(errors),
			"file_size":   fmt.Sprintf("%d lines", totalLines),
		},
	}, nil
}

func uniqueJoin(items []string, max int) string {
	uniq := uniqueSlice(items)
	if len(uniq) > max {
		uniq = uniq[:max]
	}
	return strings.Join(uniq, ", ")
}

func uniqueSlice(items []string) []string {
	seen := make(map[string]bool)
	var uniq []string
	for _, s := range items {
		if !seen[s] {
			seen[s] = true
			uniq = append(uniq, s)
		}
	}
	return uniq
}