yamlconv.go

// Copyright (c) 2021 Terminus, Inc.
//
// This program is free software: you can use, redistribute, and/or modify
// it under the terms of the GNU Affero General Public License, version 3
// or later ("AGPL"), as published by the Free Software Foundation.
//
// This program is distributed in the hope that it will be useful, but WITHOUT
// ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
// FITNESS FOR A PARTICULAR PURPOSE.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.

package oasconv

import (
	"bytes"
	"encoding"
	"encoding/json"
	"fmt"
	"io"
	"reflect"
	"sort"
	"strconv"
	"strings"
	"sync"
	"unicode"
	"unicode/utf8"

	"gopkg.in/yaml.v2"
)

// Marshals the object into JSON then converts JSON to YAML and returns the
// YAML.
func Marshal(o interface{}) ([]byte, error) {
	j, err := json.Marshal(o)
	if err != nil {
		return nil, fmt.Errorf("error marshaling into JSON: %v", err)
	}

	y, err := JSONToYAML(j)
	if err != nil {
		return nil, fmt.Errorf("error converting JSON to YAML: %v", err)
	}

	return y, nil
}

// JSONOpt is a decoding option for decoding from JSON format.
type JSONOpt func(*json.Decoder) *json.Decoder

// Unmarshal converts YAML to JSON then uses JSON to unmarshal into an object,
// optionally configuring the behavior of the JSON unmarshal.
func Unmarshal(y []byte, o interface{}, opts ...JSONOpt) error {
	return unmarshal(yaml.Unmarshal, y, o, opts)
}

// UnmarshalStrict is like Unmarshal except that any mapping keys that are
// duplicates will result in an error.
// To also be strict about unknown fields, add the DisallowUnknownFields option.
func UnmarshalStrict(y []byte, o interface{}, opts ...JSONOpt) error {
	return unmarshal(yaml.UnmarshalStrict, y, o, opts)
}

func unmarshal(f func(in []byte, out interface{}) (err error), y []byte, o interface{}, opts []JSONOpt) error {
	vo := reflect.ValueOf(o)
	j, err := yamlToJSON(y, &vo, f)
	if err != nil {
		return fmt.Errorf("error converting YAML to JSON: %v", err)
	}

	err = jsonUnmarshal(bytes.NewReader(j), o, opts...)
	if err != nil {
		return fmt.Errorf("error unmarshaling JSON: %v", err)
	}

	return nil
}

// jsonUnmarshal unmarshals the JSON byte stream from the given reader into the
// object, optionally applying decoder options prior to decoding.  We are not
// using json.Unmarshal directly as we want the chance to pass in non-default
// options.
func jsonUnmarshal(r io.Reader, o interface{}, opts ...JSONOpt) error {
	d := json.NewDecoder(r)
	for _, opt := range opts {
		d = opt(d)
	}
	if err := d.Decode(&o); err != nil {
		return fmt.Errorf("while decoding JSON: %v", err)
	}
	return nil
}

// Convert JSON to YAML.
func JSONToYAML(j []byte) ([]byte, error) {
	// Convert the JSON to an object.
	var jsonObj interface{}
	// We are using yaml.Unmarshal here (instead of json.Unmarshal) because the
	// Go JSON library doesn't try to pick the right number type (int, float,
	// etc.) when unmarshalling to interface{}, it just picks float64
	// universally. go-yaml does go through the effort of picking the right
	// number type, so we can preserve number type throughout this process.
	err := yaml.Unmarshal(j, &jsonObj)
	if err != nil {
		return nil, err
	}

	// Marshal this object into YAML.
	return yaml.Marshal(jsonObj)
}

// YAMLToJSON converts YAML to JSON. Since JSON is a subset of YAML,
// passing JSON through this method should be a no-op.
//
// Things YAML can do that are not supported by JSON:
// * In YAML you can have binary and null keys in your maps. These are invalid
//   in JSON. (int and float keys are converted to strings.)
// * Binary data in YAML with the !!binary tag is not supported. If you want to
//   use binary data with this library, encode the data as base64 as usual but do
//   not use the !!binary tag in your YAML. This will ensure the original base64
//   encoded data makes it all the way through to the JSON.
//
// For strict decoding of YAML, use YAMLToJSONStrict.
func YAMLToJSON(y []byte) ([]byte, error) {
	return yamlToJSON(y, nil, yaml.Unmarshal)
}

// 返回格式化的 json
func YAMLToJSONIndent(y []byte, prefix, indent string) ([]byte, error) {
	j, err := YAMLToJSON(y)
	if err != nil {
		return nil, err
	}
	var buf = bytes.NewBuffer(nil)
	if err = json.Indent(buf, j, prefix, indent); err != nil {
		return j, nil
	}
	return buf.Bytes(), nil
}

// YAMLToJSONStrict is like YAMLToJSON but enables strict YAML decoding,
// returning an error on any duplicate field names.
func YAMLToJSONStrict(y []byte) ([]byte, error) {
	return yamlToJSON(y, nil, yaml.UnmarshalStrict)
}

func yamlToJSON(y []byte, jsonTarget *reflect.Value, yamlUnmarshal func([]byte, interface{}) error) ([]byte, error) {
	// Convert the YAML to an object.
	var yamlObj interface{}
	err := yamlUnmarshal(y, &yamlObj)
	if err != nil {
		return nil, err
	}

	// YAML objects are not completely compatible with JSON objects (e.g. you
	// can have non-string keys in YAML). So, convert the YAML-compatible object
	// to a JSON-compatible object, failing with an error if irrecoverable
	// incompatibilities happen along the way.
	jsonObj, err := convertToJSONableObject(yamlObj, jsonTarget)
	if err != nil {
		return nil, err
	}

	// Convert this object to JSON and return the data.
	return json.Marshal(jsonObj)
}

func convertToJSONableObject(yamlObj interface{}, jsonTarget *reflect.Value) (interface{}, error) {
	var err error

	// Resolve jsonTarget to a concrete value (i.e. not a pointer or an
	// interface). We pass decodingNull as false because we're not actually
	// decoding into the value, we're just checking if the ultimate target is a
	// string.
	if jsonTarget != nil {
		ju, tu, pv := indirect(*jsonTarget, false)
		// We have a JSON or Text Umarshaler at this level, so we can't be trying
		// to decode into a string.
		if ju != nil || tu != nil {
			jsonTarget = nil
		} else {
			jsonTarget = &pv
		}
	}

	// If yamlObj is a number or a boolean, check if jsonTarget is a string -
	// if so, coerce.  Else return normal.
	// If yamlObj is a map or array, find the field that each key is
	// unmarshaling to, and when you recurse pass the reflect.Value for that
	// field back into this function.
	switch typedYAMLObj := yamlObj.(type) {
	case map[interface{}]interface{}:
		// JSON does not support arbitrary keys in a map, so we must convert
		// these keys to strings.
		//
		// From my reading of go-yaml v2 (specifically the resolve function),
		// keys can only have the types string, int, int64, float64, binary
		// (unsupported), or null (unsupported).
		strMap := make(map[string]interface{})
		for k, v := range typedYAMLObj {
			// Resolve the key to a string first.
			var keyString string
			switch typedKey := k.(type) {
			case string:
				keyString = typedKey
			case int:
				keyString = strconv.Itoa(typedKey)
			case int64:
				// go-yaml will only return an int64 as a key if the system
				// architecture is 32-bit and the key's value is between 32-bit
				// and 64-bit. Otherwise the key type will simply be int.
				keyString = strconv.FormatInt(typedKey, 10)
			case float64:
				// Stolen from go-yaml to use the same conversion to string as
				// the go-yaml library uses to convert float to string when
				// Marshaling.
				s := strconv.FormatFloat(typedKey, 'g', -1, 32)
				switch s {
				case "+Inf":
					s = ".inf"
				case "-Inf":
					s = "-.inf"
				case "NaN":
					s = ".nan"
				}
				keyString = s
			case bool:
				if typedKey {
					keyString = "true"
				} else {
					keyString = "false"
				}
			default:
				return nil, fmt.Errorf("Unsupported map key of type: %s, key: %+#v, value: %+#v",
					reflect.TypeOf(k), k, v)
			}

			// jsonTarget should be a struct or a map. If it's a struct, find
			// the field it's going to map to and pass its reflect.Value. If
			// it's a map, find the element type of the map and pass the
			// reflect.Value created from that type. If it's neither, just pass
			// nil - JSON conversion will error for us if it's a real issue.
			if jsonTarget != nil {
				t := *jsonTarget
				if t.Kind() == reflect.Struct {
					keyBytes := []byte(keyString)
					// Find the field that the JSON library would use.
					var f *field
					fields := cachedTypeFields(t.Type())
					for i := range fields {
						ff := &fields[i]
						if bytes.Equal(ff.nameBytes, keyBytes) {
							f = ff
							break
						}
						// Do case-insensitive comparison.
						if f == nil && ff.equalFold(ff.nameBytes, keyBytes) {
							f = ff
						}
					}
					if f != nil {
						// Find the reflect.Value of the most preferential
						// struct field.
						jtf := t.Field(f.index[0])
						strMap[keyString], err = convertToJSONableObject(v, &jtf)
						if err != nil {
							return nil, err
						}
						continue
					}
				} else if t.Kind() == reflect.Map {
					// Create a zero value of the map's element type to use as
					// the JSON target.
					jtv := reflect.Zero(t.Type().Elem())
					strMap[keyString], err = convertToJSONableObject(v, &jtv)
					if err != nil {
						return nil, err
					}
					continue
				}
			}
			strMap[keyString], err = convertToJSONableObject(v, nil)
			if err != nil {
				return nil, err
			}
		}
		return strMap, nil
	case []interface{}:
		// We need to recurse into arrays in case there are any
		// map[interface{}]interface{}'s inside and to convert any
		// numbers to strings.

		// If jsonTarget is a slice (which it really should be), find the
		// thing it's going to map to. If it's not a slice, just pass nil
		// - JSON conversion will error for us if it's a real issue.
		var jsonSliceElemValue *reflect.Value
		if jsonTarget != nil {
			t := *jsonTarget
			if t.Kind() == reflect.Slice {
				// By default slices point to nil, but we need a reflect.Value
				// pointing to a value of the slice type, so we create one here.
				ev := reflect.Indirect(reflect.New(t.Type().Elem()))
				jsonSliceElemValue = &ev
			}
		}

		// Make and use a new array.
		arr := make([]interface{}, len(typedYAMLObj))
		for i, v := range typedYAMLObj {
			arr[i], err = convertToJSONableObject(v, jsonSliceElemValue)
			if err != nil {
				return nil, err
			}
		}
		return arr, nil
	default:
		// If the target type is a string and the YAML type is a number,
		// convert the YAML type to a string.
		if jsonTarget != nil && (*jsonTarget).Kind() == reflect.String {
			// Based on my reading of go-yaml, it may return int, int64,
			// float64, or uint64.
			var s string
			switch typedVal := typedYAMLObj.(type) {
			case int:
				s = strconv.FormatInt(int64(typedVal), 10)
			case int64:
				s = strconv.FormatInt(typedVal, 10)
			case float64:
				s = strconv.FormatFloat(typedVal, 'g', -1, 32)
			case uint64:
				s = strconv.FormatUint(typedVal, 10)
			case bool:
				if typedVal {
					s = "true"
				} else {
					s = "false"
				}
			}
			if len(s) > 0 {
				yamlObj = interface{}(s)
			}
		}
		return yamlObj, nil
	}

	return nil, nil
}

func indirect(v reflect.Value, decodingNull bool) (json.Unmarshaler, encoding.TextUnmarshaler, reflect.Value) {
	// If v is a named type and is addressable,
	// start with its address, so that if the type has pointer methods,
	// we find them.
	if v.Kind() != reflect.Ptr && v.Type().Name() != "" && v.CanAddr() {
		v = v.Addr()
	}
	for {
		// Load value from interface, but only if the result will be
		// usefully addressable.
		if v.Kind() == reflect.Interface && !v.IsNil() {
			e := v.Elem()
			if e.Kind() == reflect.Ptr && !e.IsNil() && (!decodingNull || e.Elem().Kind() == reflect.Ptr) {
				v = e
				continue
			}
		}

		if v.Kind() != reflect.Ptr {
			break
		}

		if v.Elem().Kind() != reflect.Ptr && decodingNull && v.CanSet() {
			break
		}
		if v.IsNil() {
			if v.CanSet() {
				v.Set(reflect.New(v.Type().Elem()))
			} else {
				v = reflect.New(v.Type().Elem())
			}
		}
		if v.Type().NumMethod() > 0 {
			if u, ok := v.Interface().(json.Unmarshaler); ok {
				return u, nil, reflect.Value{}
			}
			if u, ok := v.Interface().(encoding.TextUnmarshaler); ok {
				return nil, u, reflect.Value{}
			}
		}
		v = v.Elem()
	}
	return nil, nil, v
}

// A field represents a single field found in a struct.
type field struct {
	name      string
	nameBytes []byte                 // []byte(name)
	equalFold func(s, t []byte) bool // bytes.EqualFold or equivalent

	tag       bool
	index     []int
	typ       reflect.Type
	omitEmpty bool
	quoted    bool
}

func fillField(f field) field {
	f.nameBytes = []byte(f.name)
	f.equalFold = foldFunc(f.nameBytes)
	return f
}

// byName sorts field by name, breaking ties with depth,
// then breaking ties with "name came from json tag", then
// breaking ties with index sequence.
type byName []field

func (x byName) Len() int { return len(x) }

func (x byName) Swap(i, j int) { x[i], x[j] = x[j], x[i] }

func (x byName) Less(i, j int) bool {
	if x[i].name != x[j].name {
		return x[i].name < x[j].name
	}
	if len(x[i].index) != len(x[j].index) {
		return len(x[i].index) < len(x[j].index)
	}
	if x[i].tag != x[j].tag {
		return x[i].tag
	}
	return byIndex(x).Less(i, j)
}

// byIndex sorts field by index sequence.
type byIndex []field

func (x byIndex) Len() int { return len(x) }

func (x byIndex) Swap(i, j int) { x[i], x[j] = x[j], x[i] }

func (x byIndex) Less(i, j int) bool {
	for k, xik := range x[i].index {
		if k >= len(x[j].index) {
			return false
		}
		if xik != x[j].index[k] {
			return xik < x[j].index[k]
		}
	}
	return len(x[i].index) < len(x[j].index)
}

// typeFields returns a list of fields that JSON should recognize for the given type.
// The algorithm is breadth-first search over the set of structs to include - the top struct
// and then any reachable anonymous structs.
func typeFields(t reflect.Type) []field {
	// Anonymous fields to explore at the current level and the next.
	current := []field{}
	next := []field{{typ: t}}

	// Count of queued names for current level and the next.
	count := map[reflect.Type]int{}
	nextCount := map[reflect.Type]int{}

	// Types already visited at an earlier level.
	visited := map[reflect.Type]bool{}

	// Fields found.
	var fields []field

	for len(next) > 0 {
		current, next = next, current[:0]
		count, nextCount = nextCount, map[reflect.Type]int{}

		for _, f := range current {
			if visited[f.typ] {
				continue
			}
			visited[f.typ] = true

			// Scan f.typ for fields to include.
			for i := 0; i < f.typ.NumField(); i++ {
				sf := f.typ.Field(i)
				if sf.PkgPath != "" { // unexported
					continue
				}
				tag := sf.Tag.Get("json")
				if tag == "-" {
					continue
				}
				name, opts := parseTag(tag)
				if !isValidTag(name) {
					name = ""
				}
				index := make([]int, len(f.index)+1)
				copy(index, f.index)
				index[len(f.index)] = i

				ft := sf.Type
				if ft.Name() == "" && ft.Kind() == reflect.Ptr {
					// Follow pointer.
					ft = ft.Elem()
				}

				// Record found field and index sequence.
				if name != "" || !sf.Anonymous || ft.Kind() != reflect.Struct {
					tagged := name != ""
					if name == "" {
						name = sf.Name
					}
					fields = append(fields, fillField(field{
						name:      name,
						tag:       tagged,
						index:     index,
						typ:       ft,
						omitEmpty: opts.Contains("omitempty"),
						quoted:    opts.Contains("string"),
					}))
					if count[f.typ] > 1 {
						// If there were multiple instances, add a second,
						// so that the annihilation code will see a duplicate.
						// It only cares about the distinction between 1 or 2,
						// so don't bother generating any more copies.
						fields = append(fields, fields[len(fields)-1])
					}
					continue
				}

				// Record new anonymous struct to explore in next round.
				nextCount[ft]++
				if nextCount[ft] == 1 {
					next = append(next, fillField(field{name: ft.Name(), index: index, typ: ft}))
				}
			}
		}
	}

	sort.Sort(byName(fields))

	// Delete all fields that are hidden by the Go rules for embedded fields,
	// except that fields with JSON tags are promoted.

	// The fields are sorted in primary order of name, secondary order
	// of field index length. Loop over names; for each name, delete
	// hidden fields by choosing the one dominant field that survives.
	out := fields[:0]
	for advance, i := 0, 0; i < len(fields); i += advance {
		// One iteration per name.
		// Find the sequence of fields with the name of this first field.
		fi := fields[i]
		name := fi.name
		for advance = 1; i+advance < len(fields); advance++ {
			fj := fields[i+advance]
			if fj.name != name {
				break
			}
		}
		if advance == 1 { // Only one field with this name
			out = append(out, fi)
			continue
		}
		dominant, ok := dominantField(fields[i : i+advance])
		if ok {
			out = append(out, dominant)
		}
	}

	fields = out
	sort.Sort(byIndex(fields))

	return fields
}

// dominantField looks through the fields, all of which are known to
// have the same name, to find the single field that dominates the
// others using Go's embedding rules, modified by the presence of
// JSON tags. If there are multiple top-level fields, the boolean
// will be false: This condition is an error in Go and we skip all
// the fields.
func dominantField(fields []field) (field, bool) {
	// The fields are sorted in increasing index-length order. The winner
	// must therefore be one with the shortest index length. Drop all
	// longer entries, which is easy: just truncate the slice.
	length := len(fields[0].index)
	tagged := -1 // Index of first tagged field.
	for i, f := range fields {
		if len(f.index) > length {
			fields = fields[:i]
			break
		}
		if f.tag {
			if tagged >= 0 {
				// Multiple tagged fields at the same level: conflict.
				// Return no field.
				return field{}, false
			}
			tagged = i
		}
	}
	if tagged >= 0 {
		return fields[tagged], true
	}
	// All remaining fields have the same length. If there's more than one,
	// we have a conflict (two fields named "X" at the same level) and we
	// return no field.
	if len(fields) > 1 {
		return field{}, false
	}
	return fields[0], true
}

var fieldCache struct {
	sync.RWMutex
	m map[reflect.Type][]field
}

// cachedTypeFields is like typeFields but uses a cache to avoid repeated work.
func cachedTypeFields(t reflect.Type) []field {
	fieldCache.RLock()
	f := fieldCache.m[t]
	fieldCache.RUnlock()
	if f != nil {
		return f
	}

	// Compute fields without lock.
	// Might duplicate effort but won't hold other computations back.
	f = typeFields(t)
	if f == nil {
		f = []field{}
	}

	fieldCache.Lock()
	if fieldCache.m == nil {
		fieldCache.m = map[reflect.Type][]field{}
	}
	fieldCache.m[t] = f
	fieldCache.Unlock()
	return f
}

func isValidTag(s string) bool {
	if s == "" {
		return false
	}
	for _, c := range s {
		switch {
		case strings.ContainsRune("!#$%&()*+-./:<=>?@[]^_{|}~ ", c):
			// Backslash and quote chars are reserved, but
			// otherwise any punctuation chars are allowed
			// in a tag name.
		default:
			if !unicode.IsLetter(c) && !unicode.IsDigit(c) {
				return false
			}
		}
	}
	return true
}

const (
	caseMask     = ^byte(0x20) // Mask to ignore case in ASCII.
	kelvin       = '\u212a'
	smallLongEss = '\u017f'
)

// foldFunc returns one of four different case folding equivalence
// functions, from most general (and slow) to fastest:
//
// 1) bytes.EqualFold, if the key s contains any non-ASCII UTF-8
// 2) equalFoldRight, if s contains special folding ASCII ('k', 'K', 's', 'S')
// 3) asciiEqualFold, no special, but includes non-letters (including _)
// 4) simpleLetterEqualFold, no specials, no non-letters.
//
// The letters S and K are special because they map to 3 runes, not just 2:
//  * S maps to s and to U+017F 'ſ' Latin small letter long s
//  * k maps to K and to U+212A 'K' Kelvin sign
// See http://play.golang.org/p/tTxjOc0OGo
//
// The returned function is specialized for matching against s and
// should only be given s. It's not curried for performance reasons.
func foldFunc(s []byte) func(s, t []byte) bool {
	nonLetter := false
	special := false // special letter
	for _, b := range s {
		if b >= utf8.RuneSelf {
			return bytes.EqualFold
		}
		upper := b & caseMask
		if upper < 'A' || upper > 'Z' {
			nonLetter = true
		} else if upper == 'K' || upper == 'S' {
			// See above for why these letters are special.
			special = true
		}
	}
	if special {
		return equalFoldRight
	}
	if nonLetter {
		return asciiEqualFold
	}
	return simpleLetterEqualFold
}

// equalFoldRight is a specialization of bytes.EqualFold when s is
// known to be all ASCII (including punctuation), but contains an 's',
// 'S', 'k', or 'K', requiring a Unicode fold on the bytes in t.
// See comments on foldFunc.
func equalFoldRight(s, t []byte) bool {
	for _, sb := range s {
		if len(t) == 0 {
			return false
		}
		tb := t[0]
		if tb < utf8.RuneSelf {
			if sb != tb {
				sbUpper := sb & caseMask
				if 'A' <= sbUpper && sbUpper <= 'Z' {
					if sbUpper != tb&caseMask {
						return false
					}
				} else {
					return false
				}
			}
			t = t[1:]
			continue
		}
		// sb is ASCII and t is not. t must be either kelvin
		// sign or long s; sb must be s, S, k, or K.
		tr, size := utf8.DecodeRune(t)
		switch sb {
		case 's', 'S':
			if tr != smallLongEss {
				return false
			}
		case 'k', 'K':
			if tr != kelvin {
				return false
			}
		default:
			return false
		}
		t = t[size:]

	}
	if len(t) > 0 {
		return false
	}
	return true
}

// asciiEqualFold is a specialization of bytes.EqualFold for use when
// s is all ASCII (but may contain non-letters) and contains no
// special-folding letters.
// See comments on foldFunc.
func asciiEqualFold(s, t []byte) bool {
	if len(s) != len(t) {
		return false
	}
	for i, sb := range s {
		tb := t[i]
		if sb == tb {
			continue
		}
		if ('a' <= sb && sb <= 'z') || ('A' <= sb && sb <= 'Z') {
			if sb&caseMask != tb&caseMask {
				return false
			}
		} else {
			return false
		}
	}
	return true
}

// simpleLetterEqualFold is a specialization of bytes.EqualFold for
// use when s is all ASCII letters (no underscores, etc) and also
// doesn't contain 'k', 'K', 's', or 'S'.
// See comments on foldFunc.
func simpleLetterEqualFold(s, t []byte) bool {
	if len(s) != len(t) {
		return false
	}
	for i, b := range s {
		if b&caseMask != t[i]&caseMask {
			return false
		}
	}
	return true
}

// tagOptions is the string following a comma in a struct field's "json"
// tag, or the empty string. It does not include the leading comma.
type tagOptions string

// parseTag splits a struct field's json tag into its name and
// comma-separated options.
func parseTag(tag string) (string, tagOptions) {
	if idx := strings.Index(tag, ","); idx != -1 {
		return tag[:idx], tagOptions(tag[idx+1:])
	}
	return tag, tagOptions("")
}

// Contains reports whether a comma-separated list of options
// contains a particular substr flag. substr must be surrounded by a
// string boundary or commas.
func (o tagOptions) Contains(optionName string) bool {
	if len(o) == 0 {
		return false
	}
	s := string(o)
	for s != "" {
		var next string
		i := strings.Index(s, ",")
		if i >= 0 {
			s, next = s[:i], s[i+1:]
		}
		if s == optionName {
			return true
		}
		s = next
	}
	return false
}