otel/attribute/set.go

// Copyright The OpenTelemetry Authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

package attribute // import "go.opentelemetry.io/otel/attribute"

import (
	"encoding/json"
	"reflect"
	"sort"
	"sync"
)

type (
	// Set is the representation for a distinct label set.  It
	// manages an immutable set of labels, with an internal cache
	// for storing label encodings.
	//
	// This type supports the `Equivalent` method of comparison
	// using values of type `Distinct`.
	//
	// This type is used to implement:
	// 1. Metric labels
	// 2. Resource sets
	// 3. Correlation map (TODO)
	Set struct {
		equivalent Distinct

		lock     sync.Mutex
		encoders [maxConcurrentEncoders]EncoderID
		encoded  [maxConcurrentEncoders]string
	}

	// Distinct wraps a variable-size array of `KeyValue`,
	// constructed with keys in sorted order.  This can be used as
	// a map key or for equality checking between Sets.
	Distinct struct {
		iface interface{}
	}

	// Filter supports removing certain labels from label sets.
	// When the filter returns true, the label will be kept in
	// the filtered label set.  When the filter returns false, the
	// label is excluded from the filtered label set, and the
	// label instead appears in the `removed` list of excluded labels.
	Filter func(KeyValue) bool

	// Sortable implements `sort.Interface`, used for sorting
	// `KeyValue`.  This is an exported type to support a
	// memory optimization.  A pointer to one of these is needed
	// for the call to `sort.Stable()`, which the caller may
	// provide in order to avoid an allocation.  See
	// `NewSetWithSortable()`.
	Sortable []KeyValue
)

var (
	// keyValueType is used in `computeDistinctReflect`.
	keyValueType = reflect.TypeOf(KeyValue{})

	// emptySet is returned for empty label sets.
	emptySet = &Set{
		equivalent: Distinct{
			iface: [0]KeyValue{},
		},
	}
)

const maxConcurrentEncoders = 3

// EmptySet returns a reference to a Set with no elements.
//
// This is a convenience provided for optimized calling utility.
func EmptySet() *Set {
	return emptySet
}

// reflect abbreviates `reflect.ValueOf`.
func (d Distinct) reflect() reflect.Value {
	return reflect.ValueOf(d.iface)
}

// Valid returns true if this value refers to a valid `*Set`.
func (d Distinct) Valid() bool {
	return d.iface != nil
}

// Len returns the number of labels in this set.
func (l *Set) Len() int {
	if l == nil || !l.equivalent.Valid() {
		return 0
	}
	return l.equivalent.reflect().Len()
}

// Get returns the KeyValue at ordered position `idx` in this set.
func (l *Set) Get(idx int) (KeyValue, bool) {
	if l == nil {
		return KeyValue{}, false
	}
	value := l.equivalent.reflect()

	if idx >= 0 && idx < value.Len() {
		// Note: The Go compiler successfully avoids an allocation for
		// the interface{} conversion here:
		return value.Index(idx).Interface().(KeyValue), true
	}

	return KeyValue{}, false
}

// Value returns the value of a specified key in this set.
func (l *Set) Value(k Key) (Value, bool) {
	if l == nil {
		return Value{}, false
	}
	rValue := l.equivalent.reflect()
	vlen := rValue.Len()

	idx := sort.Search(vlen, func(idx int) bool {
		return rValue.Index(idx).Interface().(KeyValue).Key >= k
	})
	if idx >= vlen {
		return Value{}, false
	}
	keyValue := rValue.Index(idx).Interface().(KeyValue)
	if k == keyValue.Key {
		return keyValue.Value, true
	}
	return Value{}, false
}

// HasValue tests whether a key is defined in this set.
func (l *Set) HasValue(k Key) bool {
	if l == nil {
		return false
	}
	_, ok := l.Value(k)
	return ok
}

// Iter returns an iterator for visiting the labels in this set.
func (l *Set) Iter() Iterator {
	return Iterator{
		storage: l,
		idx:     -1,
	}
}

// ToSlice returns the set of labels belonging to this set, sorted,
// where keys appear no more than once.
func (l *Set) ToSlice() []KeyValue {
	iter := l.Iter()
	return iter.ToSlice()
}

// Equivalent returns a value that may be used as a map key.  The
// Distinct type guarantees that the result will equal the equivalent
// Distinct value of any label set with the same elements as this,
// where sets are made unique by choosing the last value in the input
// for any given key.
func (l *Set) Equivalent() Distinct {
	if l == nil || !l.equivalent.Valid() {
		return emptySet.equivalent
	}
	return l.equivalent
}

// Equals returns true if the argument set is equivalent to this set.
func (l *Set) Equals(o *Set) bool {
	return l.Equivalent() == o.Equivalent()
}

// Encoded returns the encoded form of this set, according to
// `encoder`.  The result will be cached in this `*Set`.
func (l *Set) Encoded(encoder Encoder) string {
	if l == nil || encoder == nil {
		return ""
	}

	id := encoder.ID()
	if !id.Valid() {
		// Invalid IDs are not cached.
		return encoder.Encode(l.Iter())
	}

	var lookup *string
	l.lock.Lock()
	for idx := 0; idx < maxConcurrentEncoders; idx++ {
		if l.encoders[idx] == id {
			lookup = &l.encoded[idx]
			break
		}
	}
	l.lock.Unlock()

	if lookup != nil {
		return *lookup
	}

	r := encoder.Encode(l.Iter())

	l.lock.Lock()
	defer l.lock.Unlock()

	for idx := 0; idx < maxConcurrentEncoders; idx++ {
		if l.encoders[idx] == id {
			return l.encoded[idx]
		}
		if !l.encoders[idx].Valid() {
			l.encoders[idx] = id
			l.encoded[idx] = r
			return r
		}
	}

	// TODO: This is a performance cliff.  Find a way for this to
	// generate a warning.
	return r
}

func empty() Set {
	return Set{
		equivalent: emptySet.equivalent,
	}
}

// NewSet returns a new `Set`.  See the documentation for
// `NewSetWithSortableFiltered` for more details.
//
// Except for empty sets, this method adds an additional allocation
// compared with calls that include a `*Sortable`.
func NewSet(kvs ...KeyValue) Set {
	// Check for empty set.
	if len(kvs) == 0 {
		return empty()
	}
	s, _ := NewSetWithSortableFiltered(kvs, new(Sortable), nil)
	return s //nolint
}

// NewSetWithSortable returns a new `Set`.  See the documentation for
// `NewSetWithSortableFiltered` for more details.
//
// This call includes a `*Sortable` option as a memory optimization.
func NewSetWithSortable(kvs []KeyValue, tmp *Sortable) Set {
	// Check for empty set.
	if len(kvs) == 0 {
		return empty()
	}
	s, _ := NewSetWithSortableFiltered(kvs, tmp, nil)
	return s //nolint
}

// NewSetWithFiltered returns a new `Set`.  See the documentation for
// `NewSetWithSortableFiltered` for more details.
//
// This call includes a `Filter` to include/exclude label keys from
// the return value.  Excluded keys are returned as a slice of label
// values.
func NewSetWithFiltered(kvs []KeyValue, filter Filter) (Set, []KeyValue) {
	// Check for empty set.
	if len(kvs) == 0 {
		return empty(), nil
	}
	return NewSetWithSortableFiltered(kvs, new(Sortable), filter)
}

// NewSetWithSortableFiltered returns a new `Set`.
//
// Duplicate keys are eliminated by taking the last value.  This
// re-orders the input slice so that unique last-values are contiguous
// at the end of the slice.
//
// This ensures the following:
//
// - Last-value-wins semantics
// - Caller sees the reordering, but doesn't lose values
// - Repeated call preserve last-value wins.
//
// Note that methods are defined on `*Set`, although this returns `Set`.
// Callers can avoid memory allocations by:
//
// - allocating a `Sortable` for use as a temporary in this method
// - allocating a `Set` for storing the return value of this
//   constructor.
//
// The result maintains a cache of encoded labels, by attribute.EncoderID.
// This value should not be copied after its first use.
//
// The second `[]KeyValue` return value is a list of labels that were
// excluded by the Filter (if non-nil).
func NewSetWithSortableFiltered(kvs []KeyValue, tmp *Sortable, filter Filter) (Set, []KeyValue) {
	// Check for empty set.
	if len(kvs) == 0 {
		return empty(), nil
	}

	*tmp = kvs

	// Stable sort so the following de-duplication can implement
	// last-value-wins semantics.
	sort.Stable(tmp)

	*tmp = nil

	position := len(kvs) - 1
	offset := position - 1

	// The requirements stated above require that the stable
	// result be placed in the end of the input slice, while
	// overwritten values are swapped to the beginning.
	//
	// De-duplicate with last-value-wins semantics.  Preserve
	// duplicate values at the beginning of the input slice.
	for ; offset >= 0; offset-- {
		if kvs[offset].Key == kvs[position].Key {
			continue
		}
		position--
		kvs[offset], kvs[position] = kvs[position], kvs[offset]
	}
	if filter != nil {
		return filterSet(kvs[position:], filter)
	}
	return Set{
		equivalent: computeDistinct(kvs[position:]),
	}, nil
}

// filterSet reorders `kvs` so that included keys are contiguous at
// the end of the slice, while excluded keys precede the included keys.
func filterSet(kvs []KeyValue, filter Filter) (Set, []KeyValue) {
	var excluded []KeyValue

	// Move labels that do not match the filter so
	// they're adjacent before calling computeDistinct().
	distinctPosition := len(kvs)

	// Swap indistinct keys forward and distinct keys toward the
	// end of the slice.
	offset := len(kvs) - 1
	for ; offset >= 0; offset-- {
		if filter(kvs[offset]) {
			distinctPosition--
			kvs[offset], kvs[distinctPosition] = kvs[distinctPosition], kvs[offset]
			continue
		}
	}
	excluded = kvs[:distinctPosition]

	return Set{
		equivalent: computeDistinct(kvs[distinctPosition:]),
	}, excluded
}

// Filter returns a filtered copy of this `Set`.  See the
// documentation for `NewSetWithSortableFiltered` for more details.
func (l *Set) Filter(re Filter) (Set, []KeyValue) {
	if re == nil {
		return Set{
			equivalent: l.equivalent,
		}, nil
	}

	// Note: This could be refactored to avoid the temporary slice
	// allocation, if it proves to be expensive.
	return filterSet(l.ToSlice(), re)
}

// computeDistinct returns a `Distinct` using either the fixed- or
// reflect-oriented code path, depending on the size of the input.
// The input slice is assumed to already be sorted and de-duplicated.
func computeDistinct(kvs []KeyValue) Distinct {
	iface := computeDistinctFixed(kvs)
	if iface == nil {
		iface = computeDistinctReflect(kvs)
	}
	return Distinct{
		iface: iface,
	}
}

// computeDistinctFixed computes a `Distinct` for small slices.  It
// returns nil if the input is too large for this code path.
func computeDistinctFixed(kvs []KeyValue) interface{} {
	switch len(kvs) {
	case 1:
		ptr := new([1]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 2:
		ptr := new([2]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 3:
		ptr := new([3]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 4:
		ptr := new([4]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 5:
		ptr := new([5]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 6:
		ptr := new([6]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 7:
		ptr := new([7]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 8:
		ptr := new([8]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 9:
		ptr := new([9]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	case 10:
		ptr := new([10]KeyValue)
		copy((*ptr)[:], kvs)
		return *ptr
	default:
		return nil
	}
}

// computeDistinctReflect computes a `Distinct` using reflection,
// works for any size input.
func computeDistinctReflect(kvs []KeyValue) interface{} {
	at := reflect.New(reflect.ArrayOf(len(kvs), keyValueType)).Elem()
	for i, keyValue := range kvs {
		*(at.Index(i).Addr().Interface().(*KeyValue)) = keyValue
	}
	return at.Interface()
}

// MarshalJSON returns the JSON encoding of the `*Set`.
func (l *Set) MarshalJSON() ([]byte, error) {
	return json.Marshal(l.equivalent.iface)
}

// Len implements `sort.Interface`.
func (l *Sortable) Len() int {
	return len(*l)
}

// Swap implements `sort.Interface`.
func (l *Sortable) Swap(i, j int) {
	(*l)[i], (*l)[j] = (*l)[j], (*l)[i]
}

// Less implements `sort.Interface`.
func (l *Sortable) Less(i, j int) bool {
	return (*l)[i].Key < (*l)[j].Key
}