scratchbox/internal/config/config.go

package config

import (
	"errors"
	"fmt"
	"net/netip"
	"os"
	"path/filepath"
	"strconv"
	"strings"
	"time"

	"gopkg.in/yaml.v3"
)

const (
	defaultListenAddr         = ":8080"
	defaultReadHeaderTimeout  = "5s"
	defaultReadTimeout        = "30s"
	defaultWriteTimeout       = "30s"
	defaultIdleTimeout        = "120s"
	defaultMaxHeaderBytes     = 1 << 20
	defaultMaxUploadSize      = "100MiB"
	defaultTTL                = "1h"
	defaultRawCacheMaxSize    = "200MiB"
	defaultRawCacheMaxEntries = 32
	defaultDataDir            = "./data"
	defaultCleanupInterval    = "1m"
	defaultRateLimitEnabled   = true
	defaultRequestsPerMinute  = 30
	defaultRateLimitBurst     = 10
	maxDefaultTTLLimit        = 24 * time.Hour
	minCleanupInterval        = 10 * time.Second
)

type Config struct {
	Server   ServerConfig   `yaml:"server" comment:"HTTP listener settings."`
	Limits   LimitsConfig   `yaml:"limits" comment:"Upload, expiration, and raw cache defaults."`
	Storage  StorageConfig  `yaml:"storage" comment:"Filesystem storage location and cleanup cadence."`
	Security SecurityConfig `yaml:"security" comment:"Write-route IP restrictions and rate limiting."`
}

type ServerConfig struct {
	ListenAddr           string        `yaml:"listen_addr" comment:"Address to bind the HTTP server to."`
	ReadHeaderTimeout    string        `yaml:"read_header_timeout" comment:"Maximum duration for reading request headers (for slowloris protection). Must be > 0."`
	ReadTimeout          string        `yaml:"read_timeout" comment:"Maximum duration for reading the full request, including body. Must be > 0."`
	WriteTimeout         string        `yaml:"write_timeout" comment:"Maximum duration before timing out writes of a response. Must be > 0."`
	IdleTimeout          string        `yaml:"idle_timeout" comment:"Maximum amount of time to wait for the next request when keep-alives are enabled. Must be > 0."`
	MaxHeaderBytes       int           `yaml:"max_header_bytes" comment:"Maximum size of request headers in bytes. Must be > 0."`
	ReadHeaderTimeoutDur time.Duration `yaml:"-"`
	ReadTimeoutDur       time.Duration `yaml:"-"`
	WriteTimeoutDur      time.Duration `yaml:"-"`
	IdleTimeoutDur       time.Duration `yaml:"-"`
}

type LimitsConfig struct {
	MaxUploadSize      string        `yaml:"max_upload_size" comment:"Maximum upload request body size. Supports B, KB, MB, GB and KiB, MiB, GiB (case-insensitive). Defaults to bytes when unit is omitted (example: 1048576)."`
	DefaultTTL         string        `yaml:"default_ttl" comment:"Default expiration for new scratches. Must be > 0 and <= 24h."`
	RawCacheMaxSize    string        `yaml:"raw_cache_max_size" comment:"Maximum total decompressed bytes cached in memory for /api/raw range requests. Supports B, KB, MB, GB and KiB, MiB, GiB. Set this above limits.max_upload_size based on your expected concurrent large downloads."`
	RawCacheMaxEntries int           `yaml:"raw_cache_max_entries" comment:"Maximum number of decompressed /api/raw entries retained in memory cache. Use this as a secondary cap alongside raw_cache_max_size."`
	MaxUploadSizeBytes int64         `yaml:"-"`
	DefaultTTLDuration time.Duration `yaml:"-"`
	RawCacheMaxBytes   int64         `yaml:"-"`
}

type StorageConfig struct {
	DataDir               string        `yaml:"data_dir" comment:"Directory where scratch files and metadata are stored."`
	CleanupInterval       string        `yaml:"cleanup_interval" comment:"How often expired scratches are deleted (minimum 10s)."`
	CleanupIntervalParsed time.Duration `yaml:"-"`
}

type SecurityConfig struct {
	AllowedIPs           []string        `yaml:"allowed_ips" comment:"Allowlist for POST /api/scratch as IPs or CIDRs (examples: 127.0.0.1, 192.168.1.0/24). Keep 127.0.0.1 for local access; add trusted ranges as needed. Empty means unrestricted."`
	TrustProxyHeaders    bool            `yaml:"trust_proxy_headers" comment:"Honor X-Forwarded-For / X-Real-IP only when request source IP matches security.trusted_proxy_ips."`
	TrustedProxyIPs      []string        `yaml:"trusted_proxy_ips" comment:"IPs/CIDRs of reverse proxies allowed to supply forwarded client IP headers. Required when trust_proxy_headers is true."`
	RateLimit            RateLimitConfig `yaml:"rate_limit" comment:"Per-client token bucket on write and scratch read routes."`
	AllowedPrefixes      []netip.Prefix  `yaml:"-"`
	TrustedProxyPrefixes []netip.Prefix  `yaml:"-"`
}

type RateLimitConfig struct {
	Enabled           bool `yaml:"enabled" comment:"Enable per-client rate limiting for write and scratch read routes."`
	RequestsPerMinute int  `yaml:"requests_per_minute" comment:"Steady refill rate per client IP."`
	Burst             int  `yaml:"burst" comment:"Maximum immediate burst capacity per client IP."`
}

func Load(path string) (Config, error) {
	cfg := defaults()
	if path != "" {
		b, err := os.ReadFile(path)
		if err != nil {
			return Config{}, fmt.Errorf("read config: %w", err)
		}
		if err := yaml.Unmarshal(b, &cfg); err != nil {
			return Config{}, fmt.Errorf("parse config yaml: %w", err)
		}
	}

	if err := cfg.Validate(); err != nil {
		return Config{}, err
	}
	return cfg, nil
}

func Default() Config {
	return defaults()
}

func defaults() Config {
	return Config{
		Server: ServerConfig{
			ListenAddr:        defaultListenAddr,
			ReadHeaderTimeout: defaultReadHeaderTimeout,
			ReadTimeout:       defaultReadTimeout,
			WriteTimeout:      defaultWriteTimeout,
			IdleTimeout:       defaultIdleTimeout,
			MaxHeaderBytes:    defaultMaxHeaderBytes,
		},
		Limits: LimitsConfig{
			MaxUploadSize:      defaultMaxUploadSize,
			DefaultTTL:         defaultTTL,
			RawCacheMaxSize:    defaultRawCacheMaxSize,
			RawCacheMaxEntries: defaultRawCacheMaxEntries,
		},
		Storage: StorageConfig{
			DataDir:         defaultDataDir,
			CleanupInterval: defaultCleanupInterval,
		},
		Security: SecurityConfig{
			AllowedIPs: []string{
				"127.0.0.1",
			},
			RateLimit: RateLimitConfig{
				Enabled:           defaultRateLimitEnabled,
				RequestsPerMinute: defaultRequestsPerMinute,
				Burst:             defaultRateLimitBurst,
			},
		},
	}
}

func (c *Config) Validate() error {
	if strings.TrimSpace(c.Server.ListenAddr) == "" {
		return errors.New("server.listen_addr is required")
	}
	readHeaderTimeout, err := time.ParseDuration(strings.TrimSpace(c.Server.ReadHeaderTimeout))
	if err != nil {
		return fmt.Errorf("server.read_header_timeout invalid: %w", err)
	}
	if readHeaderTimeout <= 0 {
		return errors.New("server.read_header_timeout must be > 0")
	}
	c.Server.ReadHeaderTimeoutDur = readHeaderTimeout

	readTimeout, err := time.ParseDuration(strings.TrimSpace(c.Server.ReadTimeout))
	if err != nil {
		return fmt.Errorf("server.read_timeout invalid: %w", err)
	}
	if readTimeout <= 0 {
		return errors.New("server.read_timeout must be > 0")
	}
	c.Server.ReadTimeoutDur = readTimeout

	writeTimeout, err := time.ParseDuration(strings.TrimSpace(c.Server.WriteTimeout))
	if err != nil {
		return fmt.Errorf("server.write_timeout invalid: %w", err)
	}
	if writeTimeout <= 0 {
		return errors.New("server.write_timeout must be > 0")
	}
	c.Server.WriteTimeoutDur = writeTimeout

	idleTimeout, err := time.ParseDuration(strings.TrimSpace(c.Server.IdleTimeout))
	if err != nil {
		return fmt.Errorf("server.idle_timeout invalid: %w", err)
	}
	if idleTimeout <= 0 {
		return errors.New("server.idle_timeout must be > 0")
	}
	c.Server.IdleTimeoutDur = idleTimeout

	if c.Server.MaxHeaderBytes <= 0 {
		return errors.New("server.max_header_bytes must be > 0")
	}

	sizeBytes, err := parseHumanSize(c.Limits.MaxUploadSize)
	if err != nil {
		return fmt.Errorf("limits.max_upload_size invalid: %w", err)
	}
	c.Limits.MaxUploadSizeBytes = sizeBytes

	ttl, err := time.ParseDuration(strings.TrimSpace(c.Limits.DefaultTTL))
	if err != nil {
		return fmt.Errorf("limits.default_ttl invalid: %w", err)
	}
	if ttl <= 0 || ttl > maxDefaultTTLLimit {
		return fmt.Errorf("limits.default_ttl must be in (0, %s]", maxDefaultTTLLimit)
	}
	c.Limits.DefaultTTLDuration = ttl

	rawCacheBytes, err := parseHumanSize(c.Limits.RawCacheMaxSize)
	if err != nil {
		return fmt.Errorf("limits.raw_cache_max_size invalid: %w", err)
	}
	c.Limits.RawCacheMaxBytes = rawCacheBytes
	if c.Limits.RawCacheMaxEntries <= 0 {
		return errors.New("limits.raw_cache_max_entries must be > 0")
	}

	cleanupInterval, err := time.ParseDuration(strings.TrimSpace(c.Storage.CleanupInterval))
	if err != nil {
		return fmt.Errorf("storage.cleanup_interval invalid: %w", err)
	}
	if cleanupInterval < minCleanupInterval {
		return fmt.Errorf("storage.cleanup_interval must be at least %s", minCleanupInterval)
	}
	c.Storage.CleanupIntervalParsed = cleanupInterval

	if strings.TrimSpace(c.Storage.DataDir) == "" {
		return errors.New("storage.data_dir is required")
	}
	if err := ensureWritableDir(c.Storage.DataDir); err != nil {
		return fmt.Errorf("storage.data_dir not writable: %w", err)
	}

	prefixes, err := parseAllowedPrefixes(c.Security.AllowedIPs)
	if err != nil {
		return fmt.Errorf("security.allowed_ips invalid: %w", err)
	}
	c.Security.AllowedPrefixes = prefixes
	trustedProxyPrefixes, err := parseAllowedPrefixes(c.Security.TrustedProxyIPs)
	if err != nil {
		return fmt.Errorf("security.trusted_proxy_ips invalid: %w", err)
	}
	c.Security.TrustedProxyPrefixes = trustedProxyPrefixes
	if c.Security.TrustProxyHeaders && len(c.Security.TrustedProxyPrefixes) == 0 {
		return errors.New("security.trusted_proxy_ips is required when security.trust_proxy_headers is true")
	}

	if c.Security.RateLimit.RequestsPerMinute <= 0 {
		return errors.New("security.rate_limit.requests_per_minute must be > 0")
	}
	if c.Security.RateLimit.Burst <= 0 {
		return errors.New("security.rate_limit.burst must be > 0")
	}

	return nil
}

func ensureWritableDir(dir string) error {
	if err := os.MkdirAll(dir, 0o755); err != nil {
		return err
	}

	probe := filepath.Join(dir, ".write_probe")
	if err := os.WriteFile(probe, []byte("ok"), 0o644); err != nil {
		return err
	}
	return os.Remove(probe)
}

func parseAllowedPrefixes(values []string) ([]netip.Prefix, error) {
	if len(values) == 0 {
		return nil, nil
	}

	out := make([]netip.Prefix, 0, len(values))
	for _, value := range values {
		item := strings.TrimSpace(value)
		if item == "" {
			continue
		}

		if strings.Contains(item, "/") {
			prefix, err := netip.ParsePrefix(item)
			if err != nil {
				return nil, fmt.Errorf("parse cidr %q: %w", item, err)
			}
			out = append(out, prefix)
			continue
		}

		ip, err := netip.ParseAddr(item)
		if err != nil {
			return nil, fmt.Errorf("parse ip %q: %w", item, err)
		}
		out = append(out, netip.PrefixFrom(ip, ip.BitLen()))
	}
	return out, nil
}

func parseHumanSize(raw string) (int64, error) {
	value := strings.TrimSpace(strings.ToUpper(raw))
	if value == "" {
		return 0, errors.New("size is empty")
	}

	idx := 0
	for idx < len(value) && (value[idx] == '.' || (value[idx] >= '0' && value[idx] <= '9')) {
		idx++
	}
	if idx == 0 {
		return 0, fmt.Errorf("size %q does not start with a number", raw)
	}

	numPart := value[:idx]
	unitPart := strings.TrimSpace(value[idx:])
	if unitPart == "" {
		unitPart = "B"
	}

	number, err := strconv.ParseFloat(numPart, 64)
	if err != nil {
		return 0, fmt.Errorf("parse number %q: %w", numPart, err)
	}
	if number <= 0 {
		return 0, errors.New("size must be > 0")
	}

	multipliers := map[string]float64{
		"B":   1,
		"KB":  1_000,
		"MB":  1_000_000,
		"GB":  1_000_000_000,
		"KIB": 1024,
		"MIB": 1024 * 1024,
		"GIB": 1024 * 1024 * 1024,
	}

	multiplier, ok := multipliers[unitPart]
	if !ok {
		return 0, fmt.Errorf("unsupported unit %q", unitPart)
	}

	size := int64(number * multiplier)
	if size <= 0 {
		return 0, errors.New("calculated size must be > 0")
	}
	return size, nil
}