249 lines
6.8 KiB
Go
249 lines
6.8 KiB
Go
package main
|
|
|
|
import (
|
|
"context"
|
|
"flag"
|
|
"fmt"
|
|
"log"
|
|
"os"
|
|
"sync"
|
|
"sync/atomic"
|
|
"time"
|
|
|
|
"github.com/jackc/pgx/v5/pgxpool"
|
|
)
|
|
|
|
type Config struct {
|
|
DBUrl string
|
|
IconsDir string
|
|
SiteBucket string
|
|
EntriesPerBundle int
|
|
Concurrency int
|
|
DryRun bool
|
|
OutputDir string
|
|
Limit int
|
|
LogFile string
|
|
LogErrors bool
|
|
}
|
|
|
|
type Stats struct {
|
|
TotalHosts int
|
|
HostsWithIcon int
|
|
HostsNoIcon int
|
|
BundlesCreated int
|
|
ConvertErrors atomic.Int64
|
|
TotalBytes int64
|
|
StartedAt time.Time
|
|
}
|
|
|
|
func main() {
|
|
cfg := Config{}
|
|
flag.StringVar(&cfg.DBUrl, "db", "", "Postgres connection string (required)")
|
|
flag.StringVar(&cfg.IconsDir, "icons-dir", "icons", "Directory with downloaded icons")
|
|
flag.StringVar(&cfg.SiteBucket, "site-bucket", "everytab-site", "S3 bucket for the static site")
|
|
flag.IntVar(&cfg.EntriesPerBundle, "entries-per-bundle", 120, "Tabs per bundle JSON file")
|
|
flag.IntVar(&cfg.Concurrency, "concurrency", 200, "Concurrent icon conversions")
|
|
flag.BoolVar(&cfg.DryRun, "dry-run", false, "Write bundles to local disk instead of S3")
|
|
flag.StringVar(&cfg.OutputDir, "output-dir", "bundles", "Local output dir for dry-run mode")
|
|
flag.IntVar(&cfg.Limit, "limit", 0, "Max hosts to process (0 = all)")
|
|
flag.StringVar(&cfg.LogFile, "log-file", "", "Mirror log lines to this file")
|
|
flag.BoolVar(&cfg.LogErrors, "log-errors-only", false, "Only write errors to log file")
|
|
flag.Parse()
|
|
|
|
if cfg.DBUrl == "" {
|
|
fmt.Println("Usage: bundle_gen --db DATABASE_URL [OPTIONS]")
|
|
flag.PrintDefaults()
|
|
os.Exit(1)
|
|
}
|
|
|
|
ctx := context.Background()
|
|
|
|
// Init S3 (for uploading bundles)
|
|
if err := initS3(); err != nil {
|
|
log.Fatalf("Failed to init S3: %v", err)
|
|
}
|
|
|
|
// Init DB
|
|
pool, err := pgxpool.New(ctx, cfg.DBUrl)
|
|
if err != nil {
|
|
log.Fatalf("Failed to connect to database: %v", err)
|
|
}
|
|
defer pool.Close()
|
|
|
|
// Setup log file
|
|
var logWriter *LogWriter
|
|
if cfg.LogFile != "" {
|
|
logWriter, err = NewLogWriter(cfg.LogFile, cfg.LogErrors)
|
|
if err != nil {
|
|
log.Fatalf("Failed to open log file: %v", err)
|
|
}
|
|
defer logWriter.Close()
|
|
}
|
|
|
|
stats := &Stats{StartedAt: time.Now()}
|
|
|
|
// Count hosts
|
|
fmt.Println("=== Bundle Generator ===")
|
|
var totalHosts, hostsWithIcon int
|
|
err = pool.QueryRow(ctx, `SELECT COUNT(*) FROM hosts WHERE html_title IS NOT NULL`).Scan(&totalHosts)
|
|
if err != nil {
|
|
log.Fatalf("Failed to count hosts: %v", err)
|
|
}
|
|
err = pool.QueryRow(ctx, `SELECT COUNT(*) FROM hosts WHERE html_title IS NOT NULL AND best_icon_s3_key IS NOT NULL`).Scan(&hostsWithIcon)
|
|
if err != nil {
|
|
log.Fatalf("Failed to count icons: %v", err)
|
|
}
|
|
|
|
stats.TotalHosts = totalHosts
|
|
stats.HostsWithIcon = hostsWithIcon
|
|
stats.HostsNoIcon = totalHosts - hostsWithIcon
|
|
|
|
fmt.Printf("Total hosts: %d (with icon: %d, no icon: %d)\n", totalHosts, hostsWithIcon, totalHosts-hostsWithIcon)
|
|
fmt.Printf("Entries per bundle: %d\n", cfg.EntriesPerBundle)
|
|
fmt.Printf("Concurrency: %d\n", cfg.Concurrency)
|
|
fmt.Printf("Dry run: %v\n\n", cfg.DryRun)
|
|
|
|
if cfg.DryRun {
|
|
os.MkdirAll(cfg.OutputDir, 0755)
|
|
}
|
|
|
|
// Clean old bundles before writing new ones
|
|
if !cfg.DryRun {
|
|
fmt.Println("Cleaning old bundles from S3...")
|
|
if err := s3DeletePrefix(cfg.SiteBucket, "tabs/"); err != nil {
|
|
log.Fatalf("Failed to clean old bundles: %v", err)
|
|
}
|
|
}
|
|
|
|
// Stream hosts from DB in pages, convert icons, write bundles incrementally
|
|
fmt.Println("Processing hosts and writing bundles...")
|
|
|
|
bundleCount := 0
|
|
var totalBytes int64
|
|
var lastRandom float64 = -1
|
|
pageSize := cfg.EntriesPerBundle * 50 // fetch 50 bundles worth at a time
|
|
var entryBuf []BundleEntry
|
|
hostsProcessed := 0
|
|
|
|
for {
|
|
// Fetch a page of hosts
|
|
limit := pageSize
|
|
if cfg.Limit > 0 {
|
|
remaining := cfg.Limit - hostsProcessed
|
|
if remaining <= 0 {
|
|
break
|
|
}
|
|
if limit > remaining {
|
|
limit = remaining
|
|
}
|
|
}
|
|
|
|
hosts, err := fetchHostsPage(ctx, pool, lastRandom, limit)
|
|
if err != nil {
|
|
log.Fatalf("Failed to fetch hosts: %v", err)
|
|
}
|
|
if len(hosts) == 0 {
|
|
break
|
|
}
|
|
lastRandom = hosts[len(hosts)-1].RandomOrder
|
|
hostsProcessed += len(hosts)
|
|
|
|
// Convert icons concurrently for this page
|
|
pageEntries := make([]BundleEntry, len(hosts))
|
|
var wg sync.WaitGroup
|
|
sem := make(chan struct{}, cfg.Concurrency)
|
|
|
|
for i, host := range hosts {
|
|
wg.Add(1)
|
|
sem <- struct{}{}
|
|
go func(idx int, h HostRow) {
|
|
defer wg.Done()
|
|
defer func() { <-sem }()
|
|
pageEntries[idx] = buildEntry(h, cfg.IconsDir, logWriter, stats)
|
|
}(i, host)
|
|
}
|
|
wg.Wait()
|
|
|
|
entryBuf = append(entryBuf, pageEntries...)
|
|
|
|
// Write complete bundles from the buffer
|
|
for len(entryBuf) >= cfg.EntriesPerBundle {
|
|
chunk := entryBuf[:cfg.EntriesPerBundle]
|
|
entryBuf = entryBuf[cfg.EntriesPerBundle:]
|
|
|
|
data, err := serializeBundle(chunk)
|
|
if err != nil {
|
|
log.Fatalf("Failed to serialize bundle %d: %v", bundleCount, err)
|
|
}
|
|
|
|
if cfg.DryRun {
|
|
err = writeBundleLocal(cfg.OutputDir, bundleCount, data)
|
|
} else {
|
|
err = writeBundleS3(cfg.SiteBucket, bundleCount, data)
|
|
}
|
|
if err != nil {
|
|
log.Fatalf("Failed to write bundle %d: %v", bundleCount, err)
|
|
}
|
|
|
|
logLine := fmt.Sprintf("bundle: %04d.json %d entries %dKB", bundleCount, len(chunk), len(data)/1024)
|
|
fmt.Println(logLine)
|
|
if logWriter != nil {
|
|
logWriter.Write(logLine, false)
|
|
}
|
|
|
|
totalBytes += int64(len(data))
|
|
bundleCount++
|
|
}
|
|
}
|
|
|
|
// Write final partial bundle
|
|
if len(entryBuf) > 0 {
|
|
data, err := serializeBundle(entryBuf)
|
|
if err != nil {
|
|
log.Fatalf("Failed to serialize final bundle: %v", err)
|
|
}
|
|
|
|
if cfg.DryRun {
|
|
err = writeBundleLocal(cfg.OutputDir, bundleCount, data)
|
|
} else {
|
|
err = writeBundleS3(cfg.SiteBucket, bundleCount, data)
|
|
}
|
|
if err != nil {
|
|
log.Fatalf("Failed to write final bundle: %v", err)
|
|
}
|
|
|
|
logLine := fmt.Sprintf("bundle: %04d.json %d entries %dKB", bundleCount, len(entryBuf), len(data)/1024)
|
|
fmt.Println(logLine)
|
|
if logWriter != nil {
|
|
logWriter.Write(logLine, false)
|
|
}
|
|
|
|
totalBytes += int64(len(data))
|
|
bundleCount++
|
|
}
|
|
|
|
stats.BundlesCreated = bundleCount
|
|
stats.TotalBytes = totalBytes
|
|
|
|
// Summary
|
|
duration := time.Since(stats.StartedAt)
|
|
fmt.Printf("\n=== Summary ===\n")
|
|
fmt.Printf("Duration: %s\n", duration.Round(time.Second))
|
|
fmt.Printf("Total hosts: %d\n", stats.TotalHosts)
|
|
fmt.Printf("Hosts with icon: %d\n", stats.HostsWithIcon)
|
|
fmt.Printf("Hosts without icon: %d\n", stats.HostsNoIcon)
|
|
fmt.Printf("Convert errors: %d\n", stats.ConvertErrors.Load())
|
|
fmt.Printf("Bundles created: %d\n", stats.BundlesCreated)
|
|
fmt.Printf("Total size: %.1f MB\n", float64(stats.TotalBytes)/(1024*1024))
|
|
fmt.Printf("Avg bundle size: %.0f KB\n", float64(stats.TotalBytes)/float64(max(stats.BundlesCreated, 1))/1024)
|
|
fmt.Printf("TOTAL_BUNDLES = %d (bake this into the frontend)\n", stats.BundlesCreated)
|
|
|
|
writeStats(stats)
|
|
}
|
|
|
|
func max(a, b int) int {
|
|
if a > b {
|
|
return a
|
|
}
|
|
return b
|
|
}
|