finder.go 2.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102
  1. // Copyright 2017 Frédéric Guillot. All rights reserved.
  2. // Use of this source code is governed by the Apache 2.0
  3. // license that can be found in the LICENSE file.
  4. package subscription
  5. import (
  6. "bytes"
  7. "fmt"
  8. "io"
  9. "time"
  10. "github.com/miniflux/miniflux/errors"
  11. "github.com/miniflux/miniflux/http"
  12. "github.com/miniflux/miniflux/logger"
  13. "github.com/miniflux/miniflux/reader/feed"
  14. "github.com/miniflux/miniflux/timer"
  15. "github.com/miniflux/miniflux/url"
  16. "github.com/PuerkitoBio/goquery"
  17. )
  18. var (
  19. errConnectionFailure = "Unable to open this link: %v"
  20. errUnreadableDoc = "Unable to analyze this page: %v"
  21. )
  22. // FindSubscriptions downloads and try to find one or more subscriptions from an URL.
  23. func FindSubscriptions(websiteURL string) (Subscriptions, error) {
  24. defer timer.ExecutionTime(time.Now(), fmt.Sprintf("[FindSubscriptions] url=%s", websiteURL))
  25. client := http.NewClient(websiteURL)
  26. response, err := client.Get()
  27. if err != nil {
  28. return nil, errors.NewLocalizedError(errConnectionFailure, err)
  29. }
  30. body, err := response.NormalizeBodyEncoding()
  31. if err != nil {
  32. return nil, err
  33. }
  34. var buffer bytes.Buffer
  35. io.Copy(&buffer, body)
  36. reader := bytes.NewReader(buffer.Bytes())
  37. if format := feed.DetectFeedFormat(reader); format != feed.FormatUnknown {
  38. var subscriptions Subscriptions
  39. subscriptions = append(subscriptions, &Subscription{
  40. Title: response.EffectiveURL,
  41. URL: response.EffectiveURL,
  42. Type: format,
  43. })
  44. return subscriptions, nil
  45. }
  46. reader.Seek(0, io.SeekStart)
  47. return parseDocument(response.EffectiveURL, bytes.NewReader(buffer.Bytes()))
  48. }
  49. func parseDocument(websiteURL string, data io.Reader) (Subscriptions, error) {
  50. var subscriptions Subscriptions
  51. queries := map[string]string{
  52. "link[type='application/rss+xml']": "rss",
  53. "link[type='application/atom+xml']": "atom",
  54. "link[type='application/json']": "json",
  55. }
  56. doc, err := goquery.NewDocumentFromReader(data)
  57. if err != nil {
  58. return nil, errors.NewLocalizedError(errUnreadableDoc, err)
  59. }
  60. for query, kind := range queries {
  61. doc.Find(query).Each(func(i int, s *goquery.Selection) {
  62. subscription := new(Subscription)
  63. subscription.Type = kind
  64. if title, exists := s.Attr("title"); exists {
  65. subscription.Title = title
  66. } else {
  67. subscription.Title = "Feed"
  68. }
  69. if feedURL, exists := s.Attr("href"); exists {
  70. subscription.URL, _ = url.AbsoluteURL(websiteURL, feedURL)
  71. }
  72. if subscription.Title == "" {
  73. subscription.Title = subscription.URL
  74. }
  75. if subscription.URL != "" {
  76. logger.Debug("[FindSubscriptions] %s", subscription)
  77. subscriptions = append(subscriptions, subscription)
  78. }
  79. })
  80. }
  81. return subscriptions, nil
  82. }