// Copyright 2022 gorse Project 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 search

import (
	"context"
	"fmt"
	"reflect"
	"sort"

	"github.com/chewxy/math32"
	"github.com/zhenghaoz/gorse/base/floats"
	"github.com/zhenghaoz/gorse/base/log"
	"go.uber.org/zap"
	"modernc.org/sortutil"
)

type Vector interface {
	Distance(vector Vector) float32
	Terms() []string
	IsHidden() bool
	Centroid(vectors []Vector, indices []int32) CentroidVector
}

type DenseVector struct {
	data     []float32
	terms    []string
	isHidden bool
}

func NewDenseVector(data []float32, terms []string, isHidden bool) *DenseVector {
	return &DenseVector{
		data:     data,
		terms:    terms,
		isHidden: isHidden,
	}
}

func (v *DenseVector) Distance(vector Vector) float32 {
	feedbackVector, isFeedback := vector.(*DenseVector)
	if !isFeedback {
		log.Logger().Fatal("vector type mismatch",
			zap.String("expect", reflect.TypeOf(v).String()),
			zap.String("actual", reflect.TypeOf(vector).String()))
	}
	return -floats.Dot(v.data, feedbackVector.data)
}

func (v *DenseVector) Terms() []string {
	return v.terms
}

func (v *DenseVector) IsHidden() bool {
	return v.isHidden
}

func (v *DenseVector) Centroid(_ []Vector, _ []int32) CentroidVector {
	panic("not implemented")
}

type DictionaryVector struct {
	isHidden bool
	terms    []string
	indices  []int32
	values   []float32
	norm     float32
}

func NewDictionaryVector(indices []int32, values []float32, terms []string, isHidden bool) *DictionaryVector {
	sort.Sort(sortutil.Int32Slice(indices))
	var norm float32
	for _, i := range indices {
		norm += values[i]
	}
	norm = math32.Sqrt(norm)
	return &DictionaryVector{
		isHidden: isHidden,
		terms:    terms,
		indices:  indices,
		values:   values,
		norm:     norm,
	}
}

func (v *DictionaryVector) Dot(vector *DictionaryVector) (float32, float32) {
	i, j, sum, common := 0, 0, float32(0), float32(0)
	for i < len(v.indices) && j < len(vector.indices) {
		if v.indices[i] == vector.indices[j] {
			sum += v.values[v.indices[i]]
			common++
			i++
			j++
		} else if v.indices[i] < vector.indices[j] {
			i++
		} else if v.indices[i] > vector.indices[j] {
			j++
		}
	}
	return sum, common
}

const similarityShrink = 100

func (v *DictionaryVector) Distance(vector Vector) float32 {
	var score float32
	if dictVec, isDictVec := vector.(*DictionaryVector); !isDictVec {
		panic(fmt.Sprintf("unexpected vector type: %v", reflect.TypeOf(vector)))
	} else {
		dot, common := v.Dot(dictVec)
		if dot > 0 {
			score = -dot / v.norm / dictVec.norm * common / (common + similarityShrink)
		}
	}
	return score
}

func (v *DictionaryVector) Terms() []string {
	return v.terms
}

func (v *DictionaryVector) IsHidden() bool {
	return v.isHidden
}

type CentroidVector interface {
	Distance(vector Vector) float32
}

type DictionaryCentroidVector struct {
	data map[int32]float32
	norm float32
}

func (v DictionaryVector) Centroid(vectors []Vector, indices []int32) CentroidVector {
	data := make(map[int32]float32)
	for _, i := range indices {
		vector, isDictVector := vectors[i].(*DictionaryVector)
		if !isDictVector {
			panic(fmt.Sprintf("unexpected vector type: %v", reflect.TypeOf(vector)))
		}
		for _, i := range vector.indices {
			data[i] += math32.Sqrt(vector.values[i])
		}
	}
	var norm float32
	for _, val := range data {
		norm += val * val
	}
	norm = math32.Sqrt(norm)
	for i := range data {
		data[i] /= norm
	}
	return &DictionaryCentroidVector{
		data: data,
		norm: norm,
	}
}

func (v *DictionaryCentroidVector) Distance(vector Vector) float32 {
	var sum, common float32
	if dictVector, isDictVec := vector.(*DictionaryVector); !isDictVec {
		panic(fmt.Sprintf("unexpected vector type: %v", reflect.TypeOf(vector)))
	} else {
		for _, i := range dictVector.indices {
			if val, exist := v.data[i]; exist {
				sum += val * math32.Sqrt(v.data[i])
				common++
			}
		}
	}
	return -sum * common / (common + similarityShrink)
}

type VectorIndex interface {
	Build(ctx context.Context)
	Search(q Vector, n int, prune0 bool) ([]int32, []float32)
	MultiSearch(q Vector, terms []string, n int, prune0 bool) (map[string][]int32, map[string][]float32)
}