Use Odysee video duration as read time

This feature works by scraping the Odysee website. To enable it, set the FETCH_ODYSEE_WATCH_TIME environment variable to 1.
miniflux · Mar 21, 2023 · 8a0a08c · 8a0a08c
1 parent 6eed037
commit 8a0a08c
Show file tree

Hide file tree

Showing 2 changed files with 46 additions and 0 deletions.
diff --git a/config/options.go b/config/options.go
@@ -51,6 +51,7 @@ const (
 	defaultProxyMediaTypes                    = "image"
 	defaultProxyUrl                           = ""
 	defaultFetchYouTubeWatchTime              = false
+	defaultFetchOdyseeWatchTime               = false
 	defaultCreateAdmin                        = false
 	defaultAdminUsername                      = ""
 	defaultAdminPassword                      = ""
@@ -127,6 +128,7 @@ type Options struct {
 	proxyMediaTypes                    []string
 	proxyUrl                           string
 	fetchYouTubeWatchTime              bool
+	fetchOdyseeWatchTime               bool
 	oauth2UserCreationAllowed          bool
 	oauth2ClientID                     string
 	oauth2ClientSecret                 string
@@ -196,6 +198,7 @@ func NewOptions() *Options {
 		proxyMediaTypes:                    []string{defaultProxyMediaTypes},
 		proxyUrl:                           defaultProxyUrl,
 		fetchYouTubeWatchTime:              defaultFetchYouTubeWatchTime,
+		fetchOdyseeWatchTime:               defaultFetchOdyseeWatchTime,
 		oauth2UserCreationAllowed:          defaultOAuth2UserCreation,
 		oauth2ClientID:                     defaultOAuth2ClientID,
 		oauth2ClientSecret:                 defaultOAuth2ClientSecret,
@@ -429,6 +432,12 @@ func (o *Options) FetchYouTubeWatchTime() bool {
 	return o.fetchYouTubeWatchTime
 }
 
+// FetchOdyseeWatchTime returns true if the Odysee video duration
+// should be fetched and used as a reading time.
+func (o *Options) FetchOdyseeWatchTime() bool {
+	return o.fetchOdyseeWatchTime
+}
+
 // ProxyOption returns "none" to never proxy, "http-only" to proxy non-HTTPS, "all" to always proxy.
 func (o *Options) ProxyOption() string {
 	return o.proxyOption
@@ -574,6 +583,7 @@ func (o *Options) SortedOptions(redactSecret bool) []*Option {
 		"DISABLE_SCHEDULER_SERVICE":              !o.schedulerService,
 		"DISABLE_HTTP_SERVICE":                   !o.httpService,
 		"FETCH_YOUTUBE_WATCH_TIME":               o.fetchYouTubeWatchTime,
+		"FETCH_ODYSEE_WATCH_TIME":                o.fetchOdyseeWatchTime,
 		"HTTPS":                                  o.HTTPS,
 		"HTTP_CLIENT_MAX_BODY_SIZE":              o.httpClientMaxBodySize,
 		"HTTP_CLIENT_PROXY":                      o.httpClientProxy,

diff --git a/reader/processor/processor.go b/reader/processor/processor.go
@@ -33,6 +33,7 @@ import (
 
 var (
 	youtubeRegex           = regexp.MustCompile(`youtube\.com/watch\?v=(.*)`)
+	odyseeRegex            = regexp.MustCompile(`odysee\.com/@(.*)/(.*)`)
 	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)?)?$`)
 	customReplaceRuleRegex = regexp.MustCompile(`rewrite\("(.*)"\|"(.*)"\)`)
 )
@@ -220,6 +221,15 @@ func shouldFetchYouTubeWatchTime(entry *model.Entry) bool {
 	return urlMatchesYouTubePattern
 }
 
+func shouldFetchOdyseeWatchTime(entry *model.Entry) bool {
+	if !config.Opts.FetchOdyseeWatchTime() {
+		return false
+	}
+	matches := odyseeRegex.FindStringSubmatch(entry.URL)
+	urlMatchesOdyseePattern := len(matches) == 3
+	return urlMatchesOdyseePattern
+}
+
 func fetchYouTubeWatchTime(url string) (int, error) {
 	clt := client.NewClientWithConfig(url, config.Opts)
 	response, browserErr := browser.Exec(clt)
@@ -245,6 +255,32 @@ func fetchYouTubeWatchTime(url string) (int, error) {
 	return int(dur.Minutes()), nil
 }
 
+func fetchOdyseeWatchTime(url string) (int, error) {
+	clt := client.NewClientWithConfig(url, config.Opts)
+	response, browserErr := browser.Exec(clt)
+	if browserErr != nil {
+		return 0, browserErr
+	}
+
+	doc, docErr := goquery.NewDocumentFromReader(response.Body)
+	if docErr != nil {
+		return 0, docErr
+	}
+
+	durs, exists := doc.Find(`meta[property="og:video:duration"]`).First().Attr("content")
+	// durs contains video watch time in seconds
+	if !exists {
+		return 0, errors.New("duration has not found")
+	}
+
+	dur, err := strconv.ParseInt(durs, 10, 64)
+	if err != nil {
+		return 0, fmt.Errorf("unable to parse duration %s: %v", durs, err)
+	}
+
+	return int(dur / 60), nil
+}
+
 // parseISO8601 parses an ISO 8601 duration string.
 func parseISO8601(from string) (time.Duration, error) {
 	var match []string