format.go 1.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960
  1. // SPDX-FileCopyrightText: Copyright The Miniflux Authors. All rights reserved.
  2. // SPDX-License-Identifier: Apache-2.0
  3. package parser // import "miniflux.app/v2/internal/reader/parser"
  4. import (
  5. "bytes"
  6. "encoding/xml"
  7. "io"
  8. rxml "miniflux.app/v2/internal/reader/xml"
  9. )
  10. // List of feed formats.
  11. const (
  12. FormatRDF = "rdf"
  13. FormatRSS = "rss"
  14. FormatAtom = "atom"
  15. FormatJSON = "json"
  16. FormatUnknown = "unknown"
  17. )
  18. // DetectFeedFormat tries to guess the feed format from input data.
  19. func DetectFeedFormat(r io.ReadSeeker) (string, string) {
  20. var dataArray = [32]byte{}
  21. data := dataArray[:]
  22. r.Read(data)
  23. if bytes.HasPrefix(bytes.TrimSpace(data), []byte("{")) {
  24. return FormatJSON, ""
  25. }
  26. r.Seek(0, io.SeekStart)
  27. decoder := rxml.NewXMLDecoder(r)
  28. for {
  29. token, _ := decoder.Token()
  30. if token == nil {
  31. break
  32. }
  33. if element, ok := token.(xml.StartElement); ok {
  34. switch element.Name.Local {
  35. case "rss":
  36. return FormatRSS, ""
  37. case "feed":
  38. for _, attr := range element.Attr {
  39. if attr.Name.Local == "version" && attr.Value == "0.3" {
  40. return FormatAtom, "0.3"
  41. }
  42. }
  43. return FormatAtom, "1.0"
  44. case "RDF":
  45. return FormatRDF, ""
  46. }
  47. }
  48. }
  49. return FormatUnknown, ""
  50. }