memby/server/cmd/memby-credits/main.go

// Command memby-credits is the credits subsystem's bench, run as
//
//	memby-credits benchmark <emby-item-id>
//
// It exists because "is this cheap" is not a question a unit test can answer and not one an
// opinion should settle. Every claim the package makes — that a scan reads a couple of
// minutes rather than a file, that season history narrows the window, that analysis finishes
// in seconds on modest hardware — is a number, and this is what prints them against real
// media on the NAS.
//
// The line that matters most is the window's provenance. A run reporting
// "generic-tail-window" every time is a run in which demand-driven narrowing is doing
// nothing, and the whole design would need revisiting.
package main

import (
	"context"
	"fmt"
	"os"
	"os/signal"
	"runtime"
	"syscall"
	"time"

	"github.com/ponzischeme89/memby/server/internal/config"
	"github.com/ponzischeme89/memby/server/internal/credits"
	"github.com/ponzischeme89/memby/server/internal/emby"
	"github.com/ponzischeme89/memby/server/internal/store"
)

func main() {
	if len(os.Args) < 3 || os.Args[1] != "benchmark" {
		fmt.Fprintln(os.Stderr, "usage: memby-credits benchmark <emby-item-id>")
		os.Exit(2)
	}
	if err := run(os.Args[2]); err != nil {
		fmt.Fprintln(os.Stderr, "error:", err)
		os.Exit(1)
	}
}

func run(itemID string) error {
	ctx, stop := signal.NotifyContext(context.Background(), os.Interrupt, syscall.SIGTERM)
	defer stop()

	cfg, err := config.Load()
	if err != nil {
		return err
	}
	embyClient := emby.New(
		cfg.EmbyURL, cfg.EmbyPublicURL, cfg.ClientName, cfg.GatewayClientName,
		cfg.UpstreamTimeout,
	)
	// The bench measures the path the service actually takes, so it must read bytes from the
	// same address the service reads them from. Benchmarking the public route and deploying
	// the internal one would report a number belonging to neither.
	embyClient.SetMediaURL(cfg.EmbyMediaURL)
	resolver := credits.NewEmbyResolver(embyClient, emby.Credentials{
		UserID: cfg.SyncUserID, Token: cfg.SyncAPIKey,
		DeviceID: "memby-credits-bench", DeviceName: "MbyGateway Credits", Gateway: true,
	})
	resolved, err := resolver.Resolve(ctx, itemID)
	if err != nil {
		return fmt.Errorf("resolve %s: %w", itemID, err)
	}
	if resolved.RuntimeMs <= 0 {
		return fmt.Errorf("item %s reports no runtime; nothing to scan", itemID)
	}

	// The database is optional here on purpose. A benchmark that could not run without a
	// Postgres would be one nobody runs, and the two things it contributes — season history
	// and behavioural evidence — are exactly the two the report is meant to show the value
	// of, so their absence is worth being able to measure as the baseline.
	var (
		history   []credits.Marker
		runtimeOf func(credits.Marker) int64
		evidence  credits.BehaviourEvidence
	)
	if st, err := store.Open(ctx, cfg.DatabaseURL); err == nil {
		defer st.Close()
		database := credits.Postgres{Store: st}
		if resolved.SeriesID != "" && resolved.Season > 0 {
			if markers, err := database.SeasonMarkers(
				ctx, resolved.SeriesID, resolved.Season, 6,
			); err == nil {
				history = markers
				episodeRuntime := resolved.RuntimeMs
				runtimeOf = func(credits.Marker) int64 { return episodeRuntime }
			}
		}
		if stops, err := database.Stops(ctx, itemID); err == nil {
			evidence = credits.AnalyseStops(stops, resolved.RuntimeMs)
		}
	} else {
		fmt.Println("note: no database; measuring the un-narrowed baseline")
	}

	window := credits.NarrowWindow(resolved.RuntimeMs, history, runtimeOf, evidence)
	sampler := &credits.Sampler{Binary: cfg.CreditsFFmpeg}
	if !sampler.Available() {
		return fmt.Errorf("ffmpeg not found; set MEMBY_CREDITS_FFMPEG")
	}

	var before, after runtime.MemStats
	runtime.GC()
	runtime.ReadMemStats(&before)

	started := time.Now()
	detector := &credits.VisualDetector{Sampler: sampler}
	detection, err := detector.Detect(ctx, credits.MediaInfo{
		URL: resolved.URL, RuntimeMs: resolved.RuntimeMs, Window: window,
	})
	elapsed := time.Since(started)
	if err != nil {
		return err
	}
	runtime.ReadMemStats(&after)

	combined, acceptable := credits.Combine(detection, evidence)

	fmt.Printf("Episode runtime:        %s\n", clock(resolved.RuntimeMs))
	if evidence.Found {
		fmt.Printf("Tracearr cluster:       %s  (%d viewers, spread %ds)\n",
			clock(evidence.StartMs), evidence.UserCount, evidence.SpreadMs/1000)
	}
	fmt.Printf("Scan region:            %s – %s  (%s)\n",
		clock(window.StartMs), clock(window.EndMs), window.Source)
	fmt.Printf("Season history:         %d marker(s)\n", len(history))
	// Estimated rather than measured: ffmpeg does not report how much of its input it read,
	// and a proxy to find out would cost more than the number is worth. Derived from the
	// window as a fraction of the file, which is the ratio the whole design turns on.
	fmt.Printf("Bytes read:             ~%s (estimated, %.1f%% of file)\n",
		bytesLabel(estimateBytes(resolved.Version.SizeBytes, window, resolved.RuntimeMs)),
		100*float64(window.DurationMs())/float64(resolved.RuntimeMs))
	fmt.Printf("Frames sampled:         %d\n", detection.FramesSampled)
	fmt.Printf("Peak memory:            ~%s\n", bytesLabel(int64(after.TotalAlloc-before.TotalAlloc)))
	fmt.Printf("Analysis time:          %.1fs\n", elapsed.Seconds())

	if !detection.Found {
		fmt.Println("Detected marker:        none (visual)")
	} else {
		fmt.Printf("Detected marker:        %s\n", clock(detection.StartMs))
		fmt.Printf("Confidence:             %.2f\n", detection.Confidence)
	}
	if acceptable {
		fmt.Printf("Stored marker would be: %s  confidence %.2f  method %s\n",
			clock(combined.StartMs), combined.Confidence, combined.Method)
	} else {
		fmt.Println("Stored marker would be: none — below the confidence threshold")
	}
	return nil
}

// estimateBytes is the window as a share of the file. Crude, and honest about being crude:
// the point of the figure is the order of magnitude, and a scan reading two minutes of a
// forty-four minute episode reads about five percent of it whatever the container does.
func estimateBytes(sizeBytes int64, window credits.ScanWindow, runtimeMs int64) int64 {
	if sizeBytes <= 0 || runtimeMs <= 0 {
		return 0
	}
	return int64(float64(sizeBytes) * float64(window.DurationMs()) / float64(runtimeMs))
}

func clock(ms int64) string {
	total := ms / 1000
	if hours := total / 3600; hours > 0 {
		return fmt.Sprintf("%dh %02dm %02ds", hours, (total%3600)/60, total%60)
	}
	return fmt.Sprintf("%dm %02ds", total/60, total%60)
}

func bytesLabel(value int64) string {
	switch {
	case value <= 0:
		return "unknown"
	case value > 1<<20:
		return fmt.Sprintf("%.1f MB", float64(value)/(1<<20))
	case value > 1<<10:
		return fmt.Sprintf("%.1f KB", float64(value)/(1<<10))
	default:
		return fmt.Sprintf("%d B", value)
	}
}