nlp

type EnglishWordsCounterPQ ¶

type EnglishWordsCounterPQ struct {
	Map      map[string]*priority.Item
	Priority *priority.Queue

	sync.Mutex
}

EnglishWordsCounterPQ 分词，统计频数，英文适用，基于堆（优先队列）

func (c *EnglishWordsCounterPQ) AddSentence(s string)

func (c *EnglishWordsCounterPQ) PopMostCommon() string

type EnglishWordsCounterQS ¶

type EnglishWordsCounterQS struct {
	Map  map[string]int // [word]: idx_in_List
	List wordsList

	sync.Mutex
}

EnglishWordsCounterQS 是基于顺序表和快速排序的 counter Benchmark 时空性能不如 PQ

func (c *EnglishWordsCounterQS) AddSentence(s string)

func (c *EnglishWordsCounterQS) PopMostCommon() string

type WordsCounter ¶

type WordsCounter interface {
	// AddSentence 从句子中分词，统计
	AddSentence(s string)
	// PopMostCommon 获取频次最高的词
	PopMostCommon() string
}

WordsCounter 词频统计

目前只有英文的，推荐用 EnglishWordsCounterPQ，这个性能比较好：

$  go test -bench=.  -benchmem
goos: darwin
goarch: amd64
pkg: spotifyplaylist/nlp
cpu: Intel(R) Core(TM) i5-7360U CPU @ 2.30GHz
BenchmarkEnglishWordCounterPQ-4    33    35152096 ns/op    7236090 B/op    57066 allocs/op
BenchmarkEnglishWordCounterQS-4    28    39783905 ns/op    7251620 B/op    57171 allocs/op

func NewEnglishWordCounterPQ() WordsCounter

func NewEnglishWordCounterQS() WordsCounter