miniflux-v2/reader/subscription/finder.go

89 lines
2.3 KiB
Go
Raw Normal View History

2017-11-20 06:10:04 +01:00
// Copyright 2017 Frédéric Guillot. All rights reserved.
// Use of this source code is governed by the Apache 2.0
// license that can be found in the LICENSE file.
2018-08-25 06:51:50 +02:00
package subscription // import "miniflux.app/reader/subscription"
2017-11-20 06:10:04 +01:00
import (
"io"
"strings"
2018-08-25 06:51:50 +02:00
"miniflux.app/errors"
"miniflux.app/http/client"
"miniflux.app/reader/browser"
"miniflux.app/reader/parser"
2018-08-25 06:51:50 +02:00
"miniflux.app/url"
2017-11-20 06:10:04 +01:00
"github.com/PuerkitoBio/goquery"
)
var (
errUnreadableDoc = "Unable to analyze this page: %v"
)
// FindSubscriptions downloads and try to find one or more subscriptions from an URL.
func FindSubscriptions(websiteURL, userAgent, username, password string) (Subscriptions, *errors.LocalizedError) {
request := client.New(websiteURL)
request.WithCredentials(username, password)
request.WithUserAgent(userAgent)
response, err := browser.Exec(request)
2017-11-20 06:10:04 +01:00
if err != nil {
return nil, err
}
body := response.String()
if format := parser.DetectFeedFormat(body); format != parser.FormatUnknown {
2017-11-20 06:10:04 +01:00
var subscriptions Subscriptions
subscriptions = append(subscriptions, &Subscription{
Title: response.EffectiveURL,
URL: response.EffectiveURL,
Type: format,
})
return subscriptions, nil
}
return parseDocument(response.EffectiveURL, strings.NewReader(body))
2017-11-20 06:10:04 +01:00
}
func parseDocument(websiteURL string, data io.Reader) (Subscriptions, *errors.LocalizedError) {
2017-11-20 06:10:04 +01:00
var subscriptions Subscriptions
queries := map[string]string{
"link[type='application/rss+xml']": "rss",
"link[type='application/atom+xml']": "atom",
"link[type='application/json']": "json",
}
doc, err := goquery.NewDocumentFromReader(data)
if err != nil {
return nil, errors.NewLocalizedError(errUnreadableDoc, err)
}
for query, kind := range queries {
doc.Find(query).Each(func(i int, s *goquery.Selection) {
subscription := new(Subscription)
subscription.Type = kind
if title, exists := s.Attr("title"); exists {
subscription.Title = title
} else {
subscription.Title = "Feed"
}
if feedURL, exists := s.Attr("href"); exists {
2017-12-02 07:29:18 +01:00
subscription.URL, _ = url.AbsoluteURL(websiteURL, feedURL)
2017-11-20 06:10:04 +01:00
}
if subscription.Title == "" {
subscription.Title = subscription.URL
}
if subscription.URL != "" {
subscriptions = append(subscriptions, subscription)
}
})
}
return subscriptions, nil
}