zapx/v15/new.go

//  Copyright (c) 2018 Couchbase, Inc.
//
// 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 zap

import (
	"bytes"
	"encoding/binary"
	"math"
	"sort"
	"sync"

	"github.com/RoaringBitmap/roaring"
	index "github.com/blevesearch/bleve_index_api"
	segment "github.com/blevesearch/scorch_segment_api/v2"
	"github.com/blevesearch/vellum"
	"github.com/golang/snappy"
)

var NewSegmentBufferNumResultsBump int = 100
var NewSegmentBufferNumResultsFactor float64 = 1.0
var NewSegmentBufferAvgBytesPerDocFactor float64 = 1.0

// ValidateDocFields can be set by applications to perform additional checks
// on fields in a document being added to a new segment, by default it does
// nothing.
// This API is experimental and may be removed at any time.
var ValidateDocFields = func(field index.Field) error {
	return nil
}

// New creates an in-memory zap-encoded SegmentBase from a set of Documents
func (z *ZapPlugin) New(results []index.Document) (
	segment.Segment, uint64, error) {
	return z.newWithChunkMode(results, DefaultChunkMode)
}

func (*ZapPlugin) newWithChunkMode(results []index.Document,
	chunkMode uint32) (segment.Segment, uint64, error) {
	s := interimPool.Get().(*interim)

	var br bytes.Buffer
	if s.lastNumDocs > 0 {
		// use previous results to initialize the buf with an estimate
		// size, but note that the interim instance comes from a
		// global interimPool, so multiple scorch instances indexing
		// different docs can lead to low quality estimates
		estimateAvgBytesPerDoc := int(float64(s.lastOutSize/s.lastNumDocs) *
			NewSegmentBufferNumResultsFactor)
		estimateNumResults := int(float64(len(results)+NewSegmentBufferNumResultsBump) *
			NewSegmentBufferAvgBytesPerDocFactor)
		br.Grow(estimateAvgBytesPerDoc * estimateNumResults)
	}

	s.results = results
	s.chunkMode = chunkMode
	s.w = NewCountHashWriter(&br)

	storedIndexOffset, fieldsIndexOffset, fdvIndexOffset, dictOffsets,
		err := s.convert()
	if err != nil {
		return nil, uint64(0), err
	}

	sb, err := InitSegmentBase(br.Bytes(), s.w.Sum32(), chunkMode,
		s.FieldsMap, s.FieldsInv, uint64(len(results)),
		storedIndexOffset, fieldsIndexOffset, fdvIndexOffset, dictOffsets)

	if err == nil && s.reset() == nil {
		s.lastNumDocs = len(results)
		s.lastOutSize = len(br.Bytes())
		interimPool.Put(s)
	}

	return sb, uint64(len(br.Bytes())), err
}

var interimPool = sync.Pool{New: func() interface{} { return &interim{} }}

// interim holds temporary working data used while converting from
// analysis results to a zap-encoded segment
type interim struct {
	results []index.Document

	chunkMode uint32

	w *CountHashWriter

	// FieldsMap adds 1 to field id to avoid zero value issues
	//  name -> field id + 1
	FieldsMap map[string]uint16

	// FieldsInv is the inverse of FieldsMap
	//  field id -> name
	FieldsInv []string

	// Term dictionaries for each field
	//  field id -> term -> postings list id + 1
	Dicts []map[string]uint64

	// Terms for each field, where terms are sorted ascending
	//  field id -> []term
	DictKeys [][]string

	// Fields whose IncludeDocValues is true
	//  field id -> bool
	IncludeDocValues []bool

	// postings id -> bitmap of docNums
	Postings []*roaring.Bitmap

	// postings id -> freq/norm's, one for each docNum in postings
	FreqNorms        [][]interimFreqNorm
	freqNormsBacking []interimFreqNorm

	// postings id -> locs, one for each freq
	Locs        [][]interimLoc
	locsBacking []interimLoc

	numTermsPerPostingsList []int // key is postings list id
	numLocsPerPostingsList  []int // key is postings list id

	builder    *vellum.Builder
	builderBuf bytes.Buffer

	metaBuf bytes.Buffer

	tmp0 []byte
	tmp1 []byte

	lastNumDocs int
	lastOutSize int
}

func (s *interim) reset() (err error) {
	s.results = nil
	s.chunkMode = 0
	s.w = nil
	s.FieldsMap = nil
	s.FieldsInv = nil
	for i := range s.Dicts {
		s.Dicts[i] = nil
	}
	s.Dicts = s.Dicts[:0]
	for i := range s.DictKeys {
		s.DictKeys[i] = s.DictKeys[i][:0]
	}
	s.DictKeys = s.DictKeys[:0]
	for i := range s.IncludeDocValues {
		s.IncludeDocValues[i] = false
	}
	s.IncludeDocValues = s.IncludeDocValues[:0]
	for _, idn := range s.Postings {
		idn.Clear()
	}
	s.Postings = s.Postings[:0]
	s.FreqNorms = s.FreqNorms[:0]
	for i := range s.freqNormsBacking {
		s.freqNormsBacking[i] = interimFreqNorm{}
	}
	s.freqNormsBacking = s.freqNormsBacking[:0]
	s.Locs = s.Locs[:0]
	for i := range s.locsBacking {
		s.locsBacking[i] = interimLoc{}
	}
	s.locsBacking = s.locsBacking[:0]
	s.numTermsPerPostingsList = s.numTermsPerPostingsList[:0]
	s.numLocsPerPostingsList = s.numLocsPerPostingsList[:0]
	s.builderBuf.Reset()
	if s.builder != nil {
		err = s.builder.Reset(&s.builderBuf)
	}
	s.metaBuf.Reset()
	s.tmp0 = s.tmp0[:0]
	s.tmp1 = s.tmp1[:0]
	s.lastNumDocs = 0
	s.lastOutSize = 0

	return err
}

func (s *interim) grabBuf(size int) []byte {
	buf := s.tmp0
	if cap(buf) < size {
		buf = make([]byte, size)
		s.tmp0 = buf
	}
	return buf[0:size]
}

type interimStoredField struct {
	vals      [][]byte
	typs      []byte
	arrayposs [][]uint64 // array positions
}

type interimFreqNorm struct {
	freq    uint64
	norm    float32
	numLocs int
}

type interimLoc struct {
	fieldID   uint16
	pos       uint64
	start     uint64
	end       uint64
	arrayposs []uint64
}

func (s *interim) convert() (uint64, uint64, uint64, []uint64, error) {
	s.FieldsMap = map[string]uint16{}

	s.getOrDefineField("_id") // _id field is fieldID 0

	for _, result := range s.results {
		result.VisitComposite(func(field index.CompositeField) {
			s.getOrDefineField(field.Name())
		})
		result.VisitFields(func(field index.Field) {
			s.getOrDefineField(field.Name())
		})
	}

	sort.Strings(s.FieldsInv[1:]) // keep _id as first field

	for fieldID, fieldName := range s.FieldsInv {
		s.FieldsMap[fieldName] = uint16(fieldID + 1)
	}

	if cap(s.IncludeDocValues) >= len(s.FieldsInv) {
		s.IncludeDocValues = s.IncludeDocValues[:len(s.FieldsInv)]
	} else {
		s.IncludeDocValues = make([]bool, len(s.FieldsInv))
	}

	s.prepareDicts()

	for _, dict := range s.DictKeys {
		sort.Strings(dict)
	}

	s.processDocuments()

	storedIndexOffset, err := s.writeStoredFields()
	if err != nil {
		return 0, 0, 0, nil, err
	}

	var fdvIndexOffset uint64
	var dictOffsets []uint64

	if len(s.results) > 0 {
		fdvIndexOffset, dictOffsets, err = s.writeDicts()
		if err != nil {
			return 0, 0, 0, nil, err
		}
	} else {
		dictOffsets = make([]uint64, len(s.FieldsInv))
	}

	fieldsIndexOffset, err := persistFields(s.FieldsInv, s.w, dictOffsets)
	if err != nil {
		return 0, 0, 0, nil, err
	}

	return storedIndexOffset, fieldsIndexOffset, fdvIndexOffset, dictOffsets, nil
}

func (s *interim) getOrDefineField(fieldName string) int {
	fieldIDPlus1, exists := s.FieldsMap[fieldName]
	if !exists {
		fieldIDPlus1 = uint16(len(s.FieldsInv) + 1)
		s.FieldsMap[fieldName] = fieldIDPlus1
		s.FieldsInv = append(s.FieldsInv, fieldName)

		s.Dicts = append(s.Dicts, make(map[string]uint64))

		n := len(s.DictKeys)
		if n < cap(s.DictKeys) {
			s.DictKeys = s.DictKeys[:n+1]
			s.DictKeys[n] = s.DictKeys[n][:0]
		} else {
			s.DictKeys = append(s.DictKeys, []string(nil))
		}
	}

	return int(fieldIDPlus1 - 1)
}

// fill Dicts and DictKeys from analysis results
func (s *interim) prepareDicts() {
	var pidNext int

	var totTFs int
	var totLocs int

	visitField := func(field index.Field) {
		fieldID := uint16(s.getOrDefineField(field.Name()))

		dict := s.Dicts[fieldID]
		dictKeys := s.DictKeys[fieldID]

		tfs := field.AnalyzedTokenFrequencies()
		for term, tf := range tfs {
			pidPlus1, exists := dict[term]
			if !exists {
				pidNext++
				pidPlus1 = uint64(pidNext)

				dict[term] = pidPlus1
				dictKeys = append(dictKeys, term)

				s.numTermsPerPostingsList = append(s.numTermsPerPostingsList, 0)
				s.numLocsPerPostingsList = append(s.numLocsPerPostingsList, 0)
			}

			pid := pidPlus1 - 1

			s.numTermsPerPostingsList[pid] += 1
			s.numLocsPerPostingsList[pid] += len(tf.Locations)

			totLocs += len(tf.Locations)
		}

		totTFs += len(tfs)

		s.DictKeys[fieldID] = dictKeys
	}

	for _, result := range s.results {
		// walk each composite field
		result.VisitComposite(func(field index.CompositeField) {
			visitField(field)
		})

		// walk each field
		result.VisitFields(visitField)
	}

	numPostingsLists := pidNext

	if cap(s.Postings) >= numPostingsLists {
		s.Postings = s.Postings[:numPostingsLists]
	} else {
		postings := make([]*roaring.Bitmap, numPostingsLists)
		copy(postings, s.Postings[:cap(s.Postings)])
		for i := 0; i < numPostingsLists; i++ {
			if postings[i] == nil {
				postings[i] = roaring.New()
			}
		}
		s.Postings = postings
	}

	if cap(s.FreqNorms) >= numPostingsLists {
		s.FreqNorms = s.FreqNorms[:numPostingsLists]
	} else {
		s.FreqNorms = make([][]interimFreqNorm, numPostingsLists)
	}

	if cap(s.freqNormsBacking) >= totTFs {
		s.freqNormsBacking = s.freqNormsBacking[:totTFs]
	} else {
		s.freqNormsBacking = make([]interimFreqNorm, totTFs)
	}

	freqNormsBacking := s.freqNormsBacking
	for pid, numTerms := range s.numTermsPerPostingsList {
		s.FreqNorms[pid] = freqNormsBacking[0:0]
		freqNormsBacking = freqNormsBacking[numTerms:]
	}

	if cap(s.Locs) >= numPostingsLists {
		s.Locs = s.Locs[:numPostingsLists]
	} else {
		s.Locs = make([][]interimLoc, numPostingsLists)
	}

	if cap(s.locsBacking) >= totLocs {
		s.locsBacking = s.locsBacking[:totLocs]
	} else {
		s.locsBacking = make([]interimLoc, totLocs)
	}

	locsBacking := s.locsBacking
	for pid, numLocs := range s.numLocsPerPostingsList {
		s.Locs[pid] = locsBacking[0:0]
		locsBacking = locsBacking[numLocs:]
	}
}

func (s *interim) processDocuments() {
	numFields := len(s.FieldsInv)
	reuseFieldLens := make([]int, numFields)
	reuseFieldTFs := make([]index.TokenFrequencies, numFields)

	for docNum, result := range s.results {
		for i := 0; i < numFields; i++ { // clear these for reuse
			reuseFieldLens[i] = 0
			reuseFieldTFs[i] = nil
		}

		s.processDocument(uint64(docNum), result,
			reuseFieldLens, reuseFieldTFs)
	}
}

func (s *interim) processDocument(docNum uint64,
	result index.Document,
	fieldLens []int, fieldTFs []index.TokenFrequencies) {
	visitField := func(field index.Field) {
		fieldID := uint16(s.getOrDefineField(field.Name()))
		fieldLens[fieldID] += field.AnalyzedLength()

		existingFreqs := fieldTFs[fieldID]
		if existingFreqs != nil {
			existingFreqs.MergeAll(field.Name(), field.AnalyzedTokenFrequencies())
		} else {
			fieldTFs[fieldID] = field.AnalyzedTokenFrequencies()
		}
	}

	// walk each composite field
	result.VisitComposite(func(field index.CompositeField) {
		visitField(field)
	})

	// walk each field
	result.VisitFields(visitField)

	// now that it's been rolled up into fieldTFs, walk that
	for fieldID, tfs := range fieldTFs {
		dict := s.Dicts[fieldID]
		norm := math.Float32frombits(uint32(fieldLens[fieldID]))

		for term, tf := range tfs {
			pid := dict[term] - 1
			bs := s.Postings[pid]
			bs.Add(uint32(docNum))

			s.FreqNorms[pid] = append(s.FreqNorms[pid],
				interimFreqNorm{
					freq:    uint64(tf.Frequency()),
					norm:    norm,
					numLocs: len(tf.Locations),
				})

			if len(tf.Locations) > 0 {
				locs := s.Locs[pid]

				for _, loc := range tf.Locations {
					var locf = uint16(fieldID)
					if loc.Field != "" {
						locf = uint16(s.getOrDefineField(loc.Field))
					}
					var arrayposs []uint64
					if len(loc.ArrayPositions) > 0 {
						arrayposs = loc.ArrayPositions
					}
					locs = append(locs, interimLoc{
						fieldID:   locf,
						pos:       uint64(loc.Position),
						start:     uint64(loc.Start),
						end:       uint64(loc.End),
						arrayposs: arrayposs,
					})
				}

				s.Locs[pid] = locs
			}
		}
	}
}

func (s *interim) writeStoredFields() (
	storedIndexOffset uint64, err error) {
	varBuf := make([]byte, binary.MaxVarintLen64)
	metaEncode := func(val uint64) (int, error) {
		wb := binary.PutUvarint(varBuf, val)
		return s.metaBuf.Write(varBuf[:wb])
	}

	data, compressed := s.tmp0[:0], s.tmp1[:0]
	defer func() { s.tmp0, s.tmp1 = data, compressed }()

	// keyed by docNum
	docStoredOffsets := make([]uint64, len(s.results))

	// keyed by fieldID, for the current doc in the loop
	docStoredFields := map[uint16]interimStoredField{}

	for docNum, result := range s.results {
		for fieldID := range docStoredFields { // reset for next doc
			delete(docStoredFields, fieldID)
		}

		var validationErr error
		result.VisitFields(func(field index.Field) {
			fieldID := uint16(s.getOrDefineField(field.Name()))

			if field.Options().IsStored() {
				isf := docStoredFields[fieldID]
				isf.vals = append(isf.vals, field.Value())
				isf.typs = append(isf.typs, field.EncodedFieldType())
				isf.arrayposs = append(isf.arrayposs, field.ArrayPositions())
				docStoredFields[fieldID] = isf
			}

			if field.Options().IncludeDocValues() {
				s.IncludeDocValues[fieldID] = true
			}

			err := ValidateDocFields(field)
			if err != nil && validationErr == nil {
				validationErr = err
			}
		})
		if validationErr != nil {
			return 0, validationErr
		}

		var curr int

		s.metaBuf.Reset()
		data = data[:0]

		// _id field special case optimizes ExternalID() lookups
		idFieldVal := docStoredFields[uint16(0)].vals[0]
		_, err = metaEncode(uint64(len(idFieldVal)))
		if err != nil {
			return 0, err
		}

		// handle non-"_id" fields
		for fieldID := 1; fieldID < len(s.FieldsInv); fieldID++ {
			isf, exists := docStoredFields[uint16(fieldID)]
			if exists {
				curr, data, err = persistStoredFieldValues(
					fieldID, isf.vals, isf.typs, isf.arrayposs,
					curr, metaEncode, data)
				if err != nil {
					return 0, err
				}
			}
		}

		metaBytes := s.metaBuf.Bytes()

		compressed = snappy.Encode(compressed[:cap(compressed)], data)

		docStoredOffsets[docNum] = uint64(s.w.Count())

		_, err := writeUvarints(s.w,
			uint64(len(metaBytes)),
			uint64(len(idFieldVal)+len(compressed)))
		if err != nil {
			return 0, err
		}

		_, err = s.w.Write(metaBytes)
		if err != nil {
			return 0, err
		}

		_, err = s.w.Write(idFieldVal)
		if err != nil {
			return 0, err
		}

		_, err = s.w.Write(compressed)
		if err != nil {
			return 0, err
		}
	}

	storedIndexOffset = uint64(s.w.Count())

	for _, docStoredOffset := range docStoredOffsets {
		err = binary.Write(s.w, binary.BigEndian, docStoredOffset)
		if err != nil {
			return 0, err
		}
	}

	return storedIndexOffset, nil
}

func (s *interim) writeDicts() (fdvIndexOffset uint64, dictOffsets []uint64, err error) {
	dictOffsets = make([]uint64, len(s.FieldsInv))

	fdvOffsetsStart := make([]uint64, len(s.FieldsInv))
	fdvOffsetsEnd := make([]uint64, len(s.FieldsInv))

	buf := s.grabBuf(binary.MaxVarintLen64)

	// these int coders are initialized with chunk size 1024
	// however this will be reset to the correct chunk size
	// while processing each individual field-term section
	tfEncoder := newChunkedIntCoder(1024, uint64(len(s.results)-1))
	locEncoder := newChunkedIntCoder(1024, uint64(len(s.results)-1))

	var docTermMap [][]byte

	if s.builder == nil {
		s.builder, err = vellum.New(&s.builderBuf, nil)
		if err != nil {
			return 0, nil, err
		}
	}

	for fieldID, terms := range s.DictKeys {
		if cap(docTermMap) < len(s.results) {
			docTermMap = make([][]byte, len(s.results))
		} else {
			docTermMap = docTermMap[0:len(s.results)]
			for docNum := range docTermMap { // reset the docTermMap
				docTermMap[docNum] = docTermMap[docNum][:0]
			}
		}

		dict := s.Dicts[fieldID]

		for _, term := range terms { // terms are already sorted
			pid := dict[term] - 1

			postingsBS := s.Postings[pid]

			freqNorms := s.FreqNorms[pid]
			freqNormOffset := 0

			locs := s.Locs[pid]
			locOffset := 0

			chunkSize, err := getChunkSize(s.chunkMode, postingsBS.GetCardinality(), uint64(len(s.results)))
			if err != nil {
				return 0, nil, err
			}
			tfEncoder.SetChunkSize(chunkSize, uint64(len(s.results)-1))
			locEncoder.SetChunkSize(chunkSize, uint64(len(s.results)-1))

			postingsItr := postingsBS.Iterator()
			for postingsItr.HasNext() {
				docNum := uint64(postingsItr.Next())

				freqNorm := freqNorms[freqNormOffset]

				// check if freq/norm is enabled
				if freqNorm.freq > 0 {
					err = tfEncoder.Add(docNum,
						encodeFreqHasLocs(freqNorm.freq, freqNorm.numLocs > 0),
						uint64(math.Float32bits(freqNorm.norm)))
				} else {
					// if disabled, then skip the norm part
					err = tfEncoder.Add(docNum,
						encodeFreqHasLocs(freqNorm.freq, freqNorm.numLocs > 0))
				}
				if err != nil {
					return 0, nil, err
				}

				if freqNorm.numLocs > 0 {
					numBytesLocs := 0
					for _, loc := range locs[locOffset : locOffset+freqNorm.numLocs] {
						numBytesLocs += totalUvarintBytes(
							uint64(loc.fieldID), loc.pos, loc.start, loc.end,
							uint64(len(loc.arrayposs)), loc.arrayposs)
					}

					err = locEncoder.Add(docNum, uint64(numBytesLocs))
					if err != nil {
						return 0, nil, err
					}

					for _, loc := range locs[locOffset : locOffset+freqNorm.numLocs] {
						err = locEncoder.Add(docNum,
							uint64(loc.fieldID), loc.pos, loc.start, loc.end,
							uint64(len(loc.arrayposs)))
						if err != nil {
							return 0, nil, err
						}

						err = locEncoder.Add(docNum, loc.arrayposs...)
						if err != nil {
							return 0, nil, err
						}
					}

					locOffset += freqNorm.numLocs
				}

				freqNormOffset++

				docTermMap[docNum] = append(
					append(docTermMap[docNum], term...),
					termSeparator)
			}

			tfEncoder.Close()
			locEncoder.Close()

			postingsOffset, err :=
				writePostings(postingsBS, tfEncoder, locEncoder, nil, s.w, buf)
			if err != nil {
				return 0, nil, err
			}

			if postingsOffset > uint64(0) {
				err = s.builder.Insert([]byte(term), postingsOffset)
				if err != nil {
					return 0, nil, err
				}
			}

			tfEncoder.Reset()
			locEncoder.Reset()
		}

		err = s.builder.Close()
		if err != nil {
			return 0, nil, err
		}

		// record where this dictionary starts
		dictOffsets[fieldID] = uint64(s.w.Count())

		vellumData := s.builderBuf.Bytes()

		// write out the length of the vellum data
		n := binary.PutUvarint(buf, uint64(len(vellumData)))
		_, err = s.w.Write(buf[:n])
		if err != nil {
			return 0, nil, err
		}

		// write this vellum to disk
		_, err = s.w.Write(vellumData)
		if err != nil {
			return 0, nil, err
		}

		// reset vellum for reuse
		s.builderBuf.Reset()

		err = s.builder.Reset(&s.builderBuf)
		if err != nil {
			return 0, nil, err
		}

		// write the field doc values
		// NOTE: doc values continue to use legacy chunk mode
		chunkSize, err := getChunkSize(LegacyChunkMode, 0, 0)
		if err != nil {
			return 0, nil, err
		}
		fdvEncoder := newChunkedContentCoder(chunkSize, uint64(len(s.results)-1), s.w, false)
		if s.IncludeDocValues[fieldID] {
			for docNum, docTerms := range docTermMap {
				if len(docTerms) > 0 {
					err = fdvEncoder.Add(uint64(docNum), docTerms)
					if err != nil {
						return 0, nil, err
					}
				}
			}
			err = fdvEncoder.Close()
			if err != nil {
				return 0, nil, err
			}

			fdvOffsetsStart[fieldID] = uint64(s.w.Count())

			_, err = fdvEncoder.Write()
			if err != nil {
				return 0, nil, err
			}

			fdvOffsetsEnd[fieldID] = uint64(s.w.Count())

			fdvEncoder.Reset()
		} else {
			fdvOffsetsStart[fieldID] = fieldNotUninverted
			fdvOffsetsEnd[fieldID] = fieldNotUninverted
		}
	}

	fdvIndexOffset = uint64(s.w.Count())

	for i := 0; i < len(fdvOffsetsStart); i++ {
		n := binary.PutUvarint(buf, fdvOffsetsStart[i])
		_, err := s.w.Write(buf[:n])
		if err != nil {
			return 0, nil, err
		}
		n = binary.PutUvarint(buf, fdvOffsetsEnd[i])
		_, err = s.w.Write(buf[:n])
		if err != nil {
			return 0, nil, err
		}
	}

	return fdvIndexOffset, dictOffsets, nil
}

// returns the total # of bytes needed to encode the given uint64's
// into binary.PutUVarint() encoding
func totalUvarintBytes(a, b, c, d, e uint64, more []uint64) (n int) {
	n = numUvarintBytes(a)
	n += numUvarintBytes(b)
	n += numUvarintBytes(c)
	n += numUvarintBytes(d)
	n += numUvarintBytes(e)
	for _, v := range more {
		n += numUvarintBytes(v)
	}
	return n
}

// returns # of bytes needed to encode x in binary.PutUvarint() encoding
func numUvarintBytes(x uint64) (n int) {
	for x >= 0x80 {
		x >>= 7
		n++
	}
	return n + 1
}