summercms/modules/lagoon/validate_request.go

package lagoon

import (
	"context"
	"fmt"
	"io"
	"mime/multipart"
	"net/textproto"
	"sort"
	"strconv"
	"strings"
	"unicode/utf8"

	"git.golem15.com/golem15/summercms/modules/phrasebook"
	"gorm.io/gorm"
)

// RequestRule is one attribute of a request rule table: the attribute name,
// which may contain `*` wildcard segments (posts.*.title), and its rules in
// the order Laravel runs them.
type RequestRule struct {
	Field string
	Rules []Rule
}

// Rule is one parsed validation rule: a named Laravel rule with its
// parameters, or a closure built with CustomRule. Build rules with
// ParseRules, In or CustomRule; the zero Rule is ignored.
type Rule struct {
	name   string
	args   []string
	custom func(attribute string, value any) (message string, failed bool)
	re     *compiledRegex
}

// Name returns the snake-case rule name (required, max, regex) or "custom"
// for a CustomRule.
func (r Rule) Name() string {
	if r.custom != nil {
		return "custom"
	}
	return r.name
}

// Args returns the rule parameters as written after the colon.
func (r Rule) Args() []string {
	return append([]string(nil), r.args...)
}

// CustomRule wraps a closure rule, the Go form of a PHP `function
// ($attribute, $value, $fail)` rule. It runs in its declaration position,
// only when the value is present (it is not implicit), and a failure adds
// the returned message as written, after the :attribute placeholder is
// replaced, exactly as Winter adds a closure's $fail message.
func CustomRule(fn func(attribute string, value any) (message string, failed bool)) Rule {
	if fn == nil {
		panic("lagoon: CustomRule with a nil func")
	}
	return Rule{custom: fn}
}

// In is the Go form of Laravel's Rule::in: the value must equal one of
// values. Use it when a value contains a comma or a quote.
func In(values ...string) Rule {
	return Rule{name: "in", args: append([]string(nil), values...)}
}

// UploadedFile is an uploaded file offered to the file rules (file, image,
// mimes and the size rules, which measure it in kilobytes). Open returns the
// content; the rules read at most the first 512 bytes to detect the type.
type UploadedFile struct {
	Filename string
	Size     int64
	Header   textproto.MIMEHeader
	Open     func() (io.ReadCloser, error)
}

// UploadedFileFromHeader adapts a parsed multipart file part.
func UploadedFileFromHeader(fh *multipart.FileHeader) UploadedFile {
	if fh == nil {
		return UploadedFile{}
	}
	return UploadedFile{
		Filename: fh.Filename,
		Size:     fh.Size,
		Header:   fh.Header,
		Open: func() (io.ReadCloser, error) {
			return fh.Open()
		},
	}
}

// ValidateRequest validates decoded request input with Laravel 9 request
// validation semantics and returns Laravel's errors object (attribute to
// ordered messages), or nil when the input passes. The error return is for
// failures that are not the client's fault (a database error in exists:).
//
// The semantics follow Illuminate\Validation\Validator: `*` segments expand
// against the input (posts.0.title, posts.1.title; a wildcard with nothing to
// expand adds no attribute); rules run in order; a non-implicit rule runs only
// when the attribute is present and is skipped for a blank string, for null
// under nullable, and for an absent key under sometimes; an attribute stops
// after a failed implicit rule (required, present, filled, accepted) and,
// under bail, after any failure. Messages come from the lagoon::validation
// catalog in the request locale; size rules pick the numeric, file, array or
// string message by the attribute's type.
func ValidateRequest(ctx context.Context, tx *gorm.DB, input map[string]any, rules []RequestRule, tr *phrasebook.Translator) (map[string][]string, error) {
	v := &requestValidator{ctx: ctx, tx: tx, data: input, tr: tr}
	if v.data == nil {
		v.data = map[string]any{}
	}
	if err := v.explode(rules); err != nil {
		return nil, err
	}
	for _, attr := range v.order {
		ruleset := v.rules[attr]
		for _, rule := range ruleset {
			if err := v.validateAttribute(attr, rule); err != nil {
				return nil, err
			}
			if v.shouldStop(attr) {
				break
			}
		}
	}
	if len(v.messages) == 0 {
		return nil, nil
	}
	return v.messages, nil
}

// ErrorKeys returns the attributes of errs in the order Laravel's message bag
// holds them for rules: explicit attributes in declaration order, then the
// attributes each wildcard rule expanded to, rule by rule, in input order
// (array indexes ascending). Callers that compare a 422 body byte for byte
// use it to emit the errors object in PHP's key order.
func ErrorKeys(errs map[string][]string, rules []RequestRule) []string {
	type ranked struct {
		key   string
		group int
		path  []string
	}
	explicit := map[string]int{}
	for i, rr := range rules {
		if !strings.Contains(rr.Field, "*") {
			if _, ok := explicit[rr.Field]; !ok {
				explicit[rr.Field] = i
			}
		}
	}
	out := make([]ranked, 0, len(errs))
	for key := range errs {
		group := len(rules) * 2
		if i, ok := explicit[key]; ok {
			group = i
		} else {
			for i, rr := range rules {
				if strings.Contains(rr.Field, "*") && wildcardMatches(rr.Field, key) {
					group = len(rules) + i
					break
				}
			}
		}
		out = append(out, ranked{key: key, group: group, path: strings.Split(key, ".")})
	}
	sort.SliceStable(out, func(i, j int) bool {
		if out[i].group != out[j].group {
			return out[i].group < out[j].group
		}
		return pathLess(out[i].path, out[j].path)
	})
	keys := make([]string, len(out))
	for i, r := range out {
		keys[i] = r.key
	}
	return keys
}

func pathLess(a, b []string) bool {
	for i := 0; i < len(a) && i < len(b); i++ {
		if a[i] == b[i] {
			continue
		}
		ai, aerr := strconv.Atoi(a[i])
		bi, berr := strconv.Atoi(b[i])
		if aerr == nil && berr == nil {
			return ai < bi
		}
		return a[i] < b[i]
	}
	return len(a) < len(b)
}

// wildcardMatches reports whether key is an expansion of pattern: every `*`
// stands for one non-empty segment.
func wildcardMatches(pattern, key string) bool {
	ps := strings.Split(pattern, ".")
	ks := strings.Split(key, ".")
	if len(ps) != len(ks) {
		return false
	}
	for i := range ps {
		if ps[i] == "*" {
			if ks[i] == "" {
				return false
			}
			continue
		}
		if ps[i] != ks[i] {
			return false
		}
	}
	return true
}

type requestValidator struct {
	ctx      context.Context
	tx       *gorm.DB
	data     map[string]any
	tr       *phrasebook.Translator
	order    []string
	rules    map[string][]Rule
	primary  map[string]string // expanded attribute -> wildcard pattern
	messages map[string][]string
	failed   map[string]map[string]bool
}

// explode builds the attribute order Laravel's ValidationRuleParser gives:
// explicit attributes keep their declaration position (a later explicit
// declaration of an expanded key replaces its rules), and wildcard
// expansions are appended in expansion order.
func (v *requestValidator) explode(rules []RequestRule) error {
	v.rules = map[string][]Rule{}
	v.primary = map[string]string{}
	seen := map[string]bool{}
	for _, rr := range rules {
		if rr.Field == "" {
			return fmt.Errorf("lagoon: request rule with an empty field")
		}
		if seen[rr.Field] {
			return fmt.Errorf("lagoon: duplicate request rule field %q", rr.Field)
		}
		seen[rr.Field] = true
		v.order = append(v.order, rr.Field)
		v.rules[rr.Field] = cleanRules(rr.Rules)
	}
	for _, rr := range rules {
		if !strings.Contains(rr.Field, "*") {
			v.rules[rr.Field] = cleanRules(rr.Rules)
			continue
		}
		for _, key := range expandWildcard(v.data, rr.Field) {
			if _, ok := v.rules[key]; ok {
				v.rules[key] = append(v.rules[key], cleanRules(rr.Rules)...)
			} else {
				v.order = append(v.order, key)
				v.rules[key] = cleanRules(rr.Rules)
			}
			if _, ok := v.primary[key]; !ok {
				v.primary[key] = rr.Field
			}
		}
		delete(v.rules, rr.Field)
		v.order = removeString(v.order, rr.Field)
	}
	return nil
}

func cleanRules(rules []Rule) []Rule {
	out := make([]Rule, 0, len(rules))
	for _, r := range rules {
		if r.name == "" && r.custom == nil {
			continue
		}
		out = append(out, r)
	}
	return out
}

func removeString(list []string, s string) []string {
	out := list[:0]
	for _, x := range list {
		if x != s {
			out = append(out, x)
		}
	}
	return out
}

type expandedKey struct {
	key  string
	leaf bool
}

// expandWildcard lists the concrete attributes a wildcard pattern stands for
// in data, in the order Laravel's ValidationData gathers them: attributes that
// are leaves of the dotted input first, then the rest, each in input order.
// Literal segments after the last `*` always produce the attribute (its value
// may be missing); a missing or scalar node before a `*` produces nothing.
func expandWildcard(data map[string]any, pattern string) []string {
	segs := strings.Split(pattern, ".")
	last := -1
	for i, s := range segs {
		if s == "*" {
			last = i
		}
	}
	var found []expandedKey
	var walk func(node any, present bool, i int, prefix string)
	walk = func(node any, present bool, i int, prefix string) {
		if i == len(segs) {
			found = append(found, expandedKey{key: prefix, leaf: !present || isDotLeaf(node)})
			return
		}
		seg := segs[i]
		if seg == "*" {
			for _, ch := range children(node) {
				walk(ch.value, true, i+1, joinPath(prefix, ch.key))
			}
			return
		}
		child, ok := childOf(node, seg)
		if i > last {
			walk(child, ok, i+1, joinPath(prefix, seg))
			return
		}
		if !ok {
			return
		}
		walk(child, true, i+1, joinPath(prefix, seg))
	}
	walk(data, true, 0, "")
	out := make([]string, 0, len(found))
	seen := map[string]bool{}
	for _, pass := range []bool{true, false} {
		for _, f := range found {
			if f.leaf != pass || seen[f.key] {
				continue
			}
			seen[f.key] = true
			out = append(out, f.key)
		}
	}
	return out
}

func joinPath(prefix, seg string) string {
	if prefix == "" {
		return seg
	}
	return prefix + "." + seg
}

// isDotLeaf reports whether Arr::dot would emit the value as one key: any
// scalar and any empty array.
func isDotLeaf(v any) bool {
	return len(children(v)) == 0
}

type childEntry struct {
	key   string
	value any
}

// children lists an array's elements or a map's entries. Go maps carry no
// insertion order, so map entries come sorted by key.
func children(v any) []childEntry {
	switch t := v.(type) {
	case []any:
		out := make([]childEntry, len(t))
		for i, x := range t {
			out[i] = childEntry{key: strconv.Itoa(i), value: x}
		}
		return out
	case map[string]any:
		keys := make([]string, 0, len(t))
		for k := range t {
			keys = append(keys, k)
		}
		sort.Strings(keys)
		out := make([]childEntry, len(keys))
		for i, k := range keys {
			out[i] = childEntry{key: k, value: t[k]}
		}
		return out
	case []string:
		out := make([]childEntry, len(t))
		for i, x := range t {
			out[i] = childEntry{key: strconv.Itoa(i), value: x}
		}
		return out
	case []map[string]any:
		out := make([]childEntry, len(t))
		for i, x := range t {
			out[i] = childEntry{key: strconv.Itoa(i), value: x}
		}
		return out
	}
	return nil
}

func childOf(v any, seg string) (any, bool) {
	switch t := v.(type) {
	case map[string]any:
		x, ok := t[seg]
		return x, ok
	case []any:
		i, err := strconv.Atoi(seg)
		if err != nil || i < 0 || i >= len(t) || strconv.Itoa(i) != seg {
			return nil, false
		}
		return t[i], true
	case []string:
		i, err := strconv.Atoi(seg)
		if err != nil || i < 0 || i >= len(t) || strconv.Itoa(i) != seg {
			return nil, false
		}
		return t[i], true
	case []map[string]any:
		i, err := strconv.Atoi(seg)
		if err != nil || i < 0 || i >= len(t) || strconv.Itoa(i) != seg {
			return nil, false
		}
		return t[i], true
	}
	return nil, false
}

// lookup is Arr::get/Arr::has over the dotted attribute.
func (v *requestValidator) lookup(attr string) (any, bool) {
	var node any = v.data
	for _, seg := range strings.Split(attr, ".") {
		x, ok := childOf(node, seg)
		if !ok {
			return nil, false
		}
		node = x
	}
	return node, true
}

func (v *requestValidator) hasRule(attr string, names ...string) bool {
	for _, r := range v.rules[attr] {
		for _, n := range names {
			if r.custom == nil && r.name == n {
				return true
			}
		}
	}
	return false
}

func (v *requestValidator) validateAttribute(attr string, rule Rule) error {
	value, present := v.lookup(attr)
	if !v.isValidatable(rule, attr, value, present) {
		return nil
	}
	if rule.custom != nil {
		msg, failed := rule.custom(attr, value)
		if failed {
			v.addMessage(attr, v.replaceAttribute(msg, attr))
			v.markFailed(attr, "custom")
		}
		return nil
	}
	ok, err := v.passes(rule, attr, value, present)
	if err != nil {
		return err
	}
	if !ok {
		v.addFailure(attr, rule, value)
	}
	return nil
}

func (v *requestValidator) isValidatable(rule Rule, attr string, value any, present bool) bool {
	implicit := rule.isImplicit()
	// presentOrRuleIsImplicit
	if s, ok := value.(string); ok && phpTrim(s) == "" && present {
		if !implicit {
			return false
		}
	} else if !present && !implicit {
		return false
	}
	// passesOptionalCheck
	if v.hasRule(attr, "sometimes") && !present {
		return false
	}
	// isNotNullIfMarkedAsNullable
	if !implicit && v.hasRule(attr, "nullable") && present && value == nil {
		return false
	}
	// hasNotFailedPreviousRuleIfPresenceRule
	if rule.custom == nil && (rule.name == "exists" || rule.name == "unique") && len(v.messages[attr]) > 0 {
		return false
	}
	return true
}

func (v *requestValidator) shouldStop(attr string) bool {
	if v.hasRule(attr, "bail") {
		return len(v.messages[attr]) > 0
	}
	if !v.hasRule(attr, implicitRuleNames...) {
		return false
	}
	for name := range v.failed[attr] {
		if isImplicitName(name) {
			return true
		}
	}
	return false
}

func (v *requestValidator) addMessage(attr, msg string) {
	if v.messages == nil {
		v.messages = map[string][]string{}
	}
	v.messages[attr] = append(v.messages[attr], msg)
}

func (v *requestValidator) markFailed(attr, rule string) {
	if v.failed == nil {
		v.failed = map[string]map[string]bool{}
	}
	if v.failed[attr] == nil {
		v.failed[attr] = map[string]bool{}
	}
	v.failed[attr][rule] = true
}

func (v *requestValidator) addFailure(attr string, rule Rule, value any) {
	msg := v.message(attr, rule.name)
	msg = v.replaceAttribute(msg, attr)
	msg = replaceInput(msg, value)
	msg = replaceIndexes(msg, attr)
	msg = v.replaceRule(msg, attr, rule)
	v.addMessage(attr, msg)
	v.markFailed(attr, rule.name)
}

const catalogPrefix = "lagoon::validation."

// message is Winter's FormatsMessages::getMessage: a custom line for the
// attribute and rule, then the typed line of a size rule, then the rule line.
func (v *requestValidator) message(attr, rule string) string {
	if s, ok := v.line("custom." + attr + "." + rule); ok {
		return s
	}
	if sizeRuleNames[rule] {
		if s, ok := v.line(rule + "." + v.attributeType(attr)); ok {
			return s
		}
		return "validation." + rule + "." + v.attributeType(attr)
	}
	if s, ok := v.line(rule); ok {
		return s
	}
	return "validation." + rule
}

func (v *requestValidator) line(key string) (string, bool) {
	if v.tr == nil {
		return "", false
	}
	full := catalogPrefix + key
	if !v.tr.Has(full) {
		return "", false
	}
	s := v.tr.Get(v.ctx, full, nil)
	if s == full {
		return "", false
	}
	return s, true
}

func (v *requestValidator) attributeType(attr string) string {
	switch {
	case v.hasRule(attr, numericRuleNames...):
		return "numeric"
	case v.hasRule(attr, "array"):
		return "array"
	}
	if val, ok := v.lookup(attr); ok {
		if _, isFile := asUploadedFile(val); isFile {
			return "file"
		}
	}
	return "string"
}

// displayableAttribute is Laravel's getDisplayableAttribute: a catalog
// attribute name for the attribute or its wildcard pattern, the raw name of
// an expanded attribute, else the snake-cased name with spaces.
func (v *requestValidator) displayableAttribute(attr string) string {
	names := []string{attr}
	primary, expanded := v.primary[attr]
	if expanded && primary != attr {
		names = append(names, primary)
	}
	for _, n := range names {
		if s, ok := v.line("attributes." + n); ok {
			return s
		}
	}
	if expanded {
		return attr
	}
	return strings.ReplaceAll(laravelSnake(attr), "_", " ")
}

func (v *requestValidator) replaceAttribute(msg, attr string) string {
	name := v.displayableAttribute(attr)
	return strings.NewReplacer(
		":attribute", name,
		":ATTRIBUTE", strings.ToUpper(name),
		":Attribute", phpUcfirst(name),
	).Replace(msg)
}

func (v *requestValidator) replaceRule(msg, attr string, rule Rule) string {
	switch rule.name {
	case "between":
		return strings.NewReplacer(":min", argAt(rule.args, 0), ":max", argAt(rule.args, 1)).Replace(msg)
	case "min":
		return strings.ReplaceAll(msg, ":min", argAt(rule.args, 0))
	case "max":
		return strings.ReplaceAll(msg, ":max", argAt(rule.args, 0))
	case "size":
		return strings.ReplaceAll(msg, ":size", argAt(rule.args, 0))
	case "in", "not_in", "mimes":
		return strings.ReplaceAll(msg, ":values", strings.Join(rule.args, ", "))
	case "after", "after_or_equal", "before", "before_or_equal":
		arg := argAt(rule.args, 0)
		if _, ok := parseDateArg(arg); !ok {
			return strings.ReplaceAll(msg, ":date", v.displayableAttribute(arg))
		}
		return strings.ReplaceAll(msg, ":date", arg)
	}
	return msg
}

func argAt(args []string, i int) string {
	if i < len(args) {
		return args[i]
	}
	return ""
}

// replaceInput replaces :input with a scalar value, as Laravel does.
func replaceInput(msg string, value any) string {
	if !strings.Contains(msg, ":input") {
		return msg
	}
	s, ok := phpScalarString(value)
	if !ok {
		return msg
	}
	return strings.ReplaceAll(msg, ":input", s)
}

// replaceIndexes replaces :index and :position (zero- and one-based) with
// the first numeric segment of the attribute.
func replaceIndexes(msg, attr string) string {
	if !strings.Contains(msg, ":index") && !strings.Contains(msg, ":position") {
		return msg
	}
	for _, seg := range strings.Split(attr, ".") {
		if n, err := strconv.Atoi(seg); err == nil {
			msg = strings.ReplaceAll(msg, ":index", strconv.Itoa(n))
			msg = strings.ReplaceAll(msg, ":position", strconv.Itoa(n+1))
			break
		}
	}
	return msg
}

// laravelSnake ports Str::snake with the underscore delimiter.
func laravelSnake(s string) string {
	if isCtypeLower(s) {
		return s
	}
	s = phpUcwords(s)
	var b strings.Builder
	first := true
	for _, r := range s {
		if r == ' ' || r == '\t' || r == '\n' || r == '\r' || r == '\f' || r == '\v' {
			continue
		}
		if !first && r >= 'A' && r <= 'Z' {
			b.WriteByte('_')
		}
		b.WriteRune(r)
		first = false
	}
	return strings.ToLower(b.String())
}

func isCtypeLower(s string) bool {
	if s == "" {
		return false
	}
	for i := 0; i < len(s); i++ {
		if s[i] < 'a' || s[i] > 'z' {
			return false
		}
	}
	return true
}

func phpUcwords(s string) string {
	b := []byte(s)
	start := true
	for i, c := range b {
		if start && c >= 'a' && c <= 'z' {
			b[i] = c - 'a' + 'A'
		}
		start = c == ' ' || c == '\t' || c == '\r' || c == '\n' || c == '\f' || c == '\v'
	}
	return string(b)
}

// phpUcfirst ports Str::ucfirst, which upper-cases the first character
// multibyte-safely.
func phpUcfirst(s string) string {
	r, size := utf8.DecodeRuneInString(s)
	if size == 0 || r == utf8.RuneError {
		return s
	}
	return strings.ToUpper(string(r)) + s[size:]
}