miniflux-v2/reader/feed/parser.go

92 lines
1.9 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.
package feed
import (
"bytes"
"encoding/xml"
"errors"
2017-11-21 03:34:11 +01:00
"io"
"strings"
"time"
2017-12-13 06:48:13 +01:00
"github.com/miniflux/miniflux/model"
"github.com/miniflux/miniflux/reader/atom"
"github.com/miniflux/miniflux/reader/json"
"github.com/miniflux/miniflux/reader/rdf"
"github.com/miniflux/miniflux/reader/rss"
2018-01-03 04:15:08 +01:00
"github.com/miniflux/miniflux/timer"
2017-11-20 06:10:04 +01:00
"golang.org/x/net/html/charset"
)
2017-11-21 03:34:11 +01:00
// List of feed formats.
2017-11-20 06:10:04 +01:00
const (
2017-11-21 03:34:11 +01:00
FormatRDF = "rdf"
FormatRSS = "rss"
2017-11-20 06:10:04 +01:00
FormatAtom = "atom"
2017-11-21 03:34:11 +01:00
FormatJSON = "json"
2017-11-20 06:10:04 +01:00
FormatUnknown = "unknown"
)
2017-11-21 03:34:11 +01:00
// DetectFeedFormat detect feed format from input data.
2017-11-20 06:10:04 +01:00
func DetectFeedFormat(data io.Reader) string {
2018-01-03 04:15:08 +01:00
defer timer.ExecutionTime(time.Now(), "[Feed:DetectFeedFormat]")
2017-11-20 06:10:04 +01:00
var buffer bytes.Buffer
tee := io.TeeReader(data, &buffer)
decoder := xml.NewDecoder(tee)
decoder.CharsetReader = charset.NewReaderLabel
for {
token, _ := decoder.Token()
if token == nil {
break
}
if element, ok := token.(xml.StartElement); ok {
switch element.Name.Local {
case "rss":
2017-11-21 03:34:11 +01:00
return FormatRSS
2017-11-20 06:10:04 +01:00
case "feed":
return FormatAtom
2017-11-21 03:34:11 +01:00
case "RDF":
return FormatRDF
2017-11-20 06:10:04 +01:00
}
}
}
if strings.HasPrefix(strings.TrimSpace(buffer.String()), "{") {
2017-11-21 03:34:11 +01:00
return FormatJSON
2017-11-20 06:10:04 +01:00
}
return FormatUnknown
}
func parseFeed(data io.Reader) (*model.Feed, error) {
2018-01-03 04:15:08 +01:00
defer timer.ExecutionTime(time.Now(), "[Feed:ParseFeed]")
2017-11-20 06:10:04 +01:00
var buffer bytes.Buffer
io.Copy(&buffer, data)
reader := bytes.NewReader(buffer.Bytes())
format := DetectFeedFormat(reader)
reader.Seek(0, io.SeekStart)
switch format {
case FormatAtom:
return atom.Parse(reader)
2017-11-21 03:34:11 +01:00
case FormatRSS:
2017-11-20 06:10:04 +01:00
return rss.Parse(reader)
2017-11-21 03:34:11 +01:00
case FormatJSON:
2017-11-20 06:10:04 +01:00
return json.Parse(reader)
2017-11-21 03:34:11 +01:00
case FormatRDF:
return rdf.Parse(reader)
2017-11-20 06:10:04 +01:00
default:
return nil, errors.New("Unsupported feed format")
}
}