| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100 |
- // SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
- // SPDX-License-Identifier: Apache-2.0
- package processor
- import (
- "errors"
- "fmt"
- "log/slog"
- "regexp"
- "strconv"
- "time"
- "github.com/PuerkitoBio/goquery"
- "miniflux.app/v2/internal/config"
- "miniflux.app/v2/internal/model"
- "miniflux.app/v2/internal/reader/fetcher"
- )
- var (
- youtubeRegex = regexp.MustCompile(`youtube\.com/watch\?v=(.*)$`)
- iso8601Regex = regexp.MustCompile(`^P((?P<year>\d+)Y)?((?P<month>\d+)M)?((?P<week>\d+)W)?((?P<day>\d+)D)?(T((?P<hour>\d+)H)?((?P<minute>\d+)M)?((?P<second>\d+)S)?)?$`)
- )
- func shouldFetchYouTubeWatchTime(entry *model.Entry) bool {
- if !config.Opts.FetchYouTubeWatchTime() {
- return false
- }
- matches := youtubeRegex.FindStringSubmatch(entry.URL)
- urlMatchesYouTubePattern := len(matches) == 2
- return urlMatchesYouTubePattern
- }
- func fetchYouTubeWatchTime(websiteURL string) (int, error) {
- requestBuilder := fetcher.NewRequestBuilder()
- requestBuilder.WithTimeout(config.Opts.HTTPClientTimeout())
- requestBuilder.WithProxy(config.Opts.HTTPClientProxy())
- responseHandler := fetcher.NewResponseHandler(requestBuilder.ExecuteRequest(websiteURL))
- defer responseHandler.Close()
- if localizedError := responseHandler.LocalizedError(); localizedError != nil {
- slog.Warn("Unable to fetch YouTube page", slog.String("website_url", websiteURL), slog.Any("error", localizedError.Error()))
- return 0, localizedError.Error()
- }
- doc, docErr := goquery.NewDocumentFromReader(responseHandler.Body(config.Opts.HTTPClientMaxBodySize()))
- if docErr != nil {
- return 0, docErr
- }
- durs, exists := doc.Find(`meta[itemprop="duration"]`).First().Attr("content")
- if !exists {
- return 0, errors.New("duration has not found")
- }
- dur, err := parseISO8601(durs)
- if err != nil {
- return 0, fmt.Errorf("unable to parse duration %s: %v", durs, err)
- }
- return int(dur.Minutes()), nil
- }
- func parseISO8601(from string) (time.Duration, error) {
- var match []string
- var d time.Duration
- if iso8601Regex.MatchString(from) {
- match = iso8601Regex.FindStringSubmatch(from)
- } else {
- return 0, errors.New("could not parse duration string")
- }
- for i, name := range iso8601Regex.SubexpNames() {
- part := match[i]
- if i == 0 || name == "" || part == "" {
- continue
- }
- val, err := strconv.ParseInt(part, 10, 64)
- if err != nil {
- return 0, err
- }
- switch name {
- case "hour":
- d += (time.Duration(val) * time.Hour)
- case "minute":
- d += (time.Duration(val) * time.Minute)
- case "second":
- d += (time.Duration(val) * time.Second)
- default:
- return 0, fmt.Errorf("unknown field %s", name)
- }
- }
- return d, nil
- }
|