oss/miniflux-v2
oss/miniflux-v2
1
0
Fork 0
mirror of https://github.com/miniflux/v2.git synced 2025-07-22 17:18:37 +00:00
Code Wiki Activity

Improve EstimateReadingTime's speed by a factor 7

Browse source 
- Refactorise the tests and add some
- Use 250 signs instead of the whole text
- Only check for Korean, Chinese and Japanese script
- Add a benchmark
- Use a more idiomatic control flow

```console
$ # main branch
$ go test -bench=.
goos: linux
goarch: amd64
pkg: miniflux.app/v2/internal/reader/readingtime
BenchmarkEstimateReadingTime-12              267           4821268 ns/op
PASS
ok      miniflux.app/v2/internal/reader/readingtime     1.754s
$ # speed_up_reading_time branch
$ go test -bench=.
goos: linux
goarch: amd64
pkg: miniflux.app/v2/internal/reader/readingtime
cpu: 12th Gen Intel(R) Core(TM) i7-1265U
BenchmarkEstimateReadingTime-12             1941            653312 ns/op
PASS
ok      miniflux.app/v2/internal/reader/readingtime     1.342s
$
```
This commit is contained in:
jvoisin 2024-02-29 14:11:46 +01:00 • committed by Frédéric Guillot
parent 31ac62f410
commit ab85d4d678
2 changed files with 63 additions and 28 deletions
Download patch file Download diff file Expand all files Collapse all files

24
internal/reader/readingtime/readingtime.go
View file

@ -17,15 +17,23 @@ import (
// EstimateReadingTime returns the estimated reading time of an article in minute.
func EstimateReadingTime(content string, defaultReadingSpeed, cjkReadingSpeed int) int {
sanitizedContent := sanitizer.StripTags(content)
langInfo := whatlanggo.Detect(sanitizedContent)
var timeToReadInt int
if langInfo.IsReliable() && (langInfo.Lang == whatlanggo.Jpn || langInfo.Lang == whatlanggo.Cmn || langInfo.Lang == whatlanggo.Kor) {
timeToReadInt = int(math.Ceil(float64(utf8.RuneCountInString(sanitizedContent)) / float64(cjkReadingSpeed)))
} else {
nbOfWords := len(strings.Fields(sanitizedContent))
timeToReadInt = int(math.Ceil(float64(nbOfWords) / float64(defaultReadingSpeed)))
// Litterature on language detection says that around 100 signes is enough, we're safe here.
truncationPoint := int(math.Min(float64(len(sanitizedContent)), 250))
// We're only interested in identifying Japanse/Chinese/Korean
options := whatlanggo.Options{
Whitelist: map[whatlanggo.Lang]bool{
whatlanggo.Jpn: true,
whatlanggo.Cmn: true,
whatlanggo.Kor: true,
},
}
langInfo := whatlanggo.DetectWithOptions(sanitizedContent[:truncationPoint], options)
return timeToReadInt
if langInfo.IsReliable() {
return int(math.Ceil(float64(utf8.RuneCountInString(sanitizedContent)) / float64(cjkReadingSpeed)))
}
nbOfWords := len(strings.Fields(sanitizedContent))
return int(math.Ceil(float64(nbOfWords) / float64(defaultReadingSpeed)))
}