2017-11-20 00:10:04 -05:00
|
|
|
// 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 00:51:50 -04:00
|
|
|
package json // import "miniflux.app/reader/json"
|
2017-11-20 00:10:04 -05:00
|
|
|
|
|
|
|
import (
|
2017-11-20 21:57:54 -05:00
|
|
|
"strings"
|
|
|
|
"time"
|
|
|
|
|
2018-08-25 00:51:50 -04:00
|
|
|
"miniflux.app/crypto"
|
|
|
|
"miniflux.app/logger"
|
|
|
|
"miniflux.app/model"
|
|
|
|
"miniflux.app/reader/date"
|
|
|
|
"miniflux.app/reader/sanitizer"
|
|
|
|
"miniflux.app/url"
|
2017-11-20 00:10:04 -05:00
|
|
|
)
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
type jsonFeed struct {
|
2017-11-20 00:10:04 -05:00
|
|
|
Version string `json:"version"`
|
|
|
|
Title string `json:"title"`
|
|
|
|
SiteURL string `json:"home_page_url"`
|
|
|
|
FeedURL string `json:"feed_url"`
|
2017-11-20 21:57:54 -05:00
|
|
|
Author jsonAuthor `json:"author"`
|
|
|
|
Items []jsonItem `json:"items"`
|
2017-11-20 00:10:04 -05:00
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
type jsonAuthor struct {
|
2017-11-20 00:10:04 -05:00
|
|
|
Name string `json:"name"`
|
|
|
|
URL string `json:"url"`
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
type jsonItem struct {
|
2017-11-20 00:10:04 -05:00
|
|
|
ID string `json:"id"`
|
|
|
|
URL string `json:"url"`
|
|
|
|
Title string `json:"title"`
|
|
|
|
Summary string `json:"summary"`
|
|
|
|
Text string `json:"content_text"`
|
2017-11-20 21:57:54 -05:00
|
|
|
HTML string `json:"content_html"`
|
2017-11-20 00:10:04 -05:00
|
|
|
DatePublished string `json:"date_published"`
|
|
|
|
DateModified string `json:"date_modified"`
|
2017-11-20 21:57:54 -05:00
|
|
|
Author jsonAuthor `json:"author"`
|
|
|
|
Attachments []jsonAttachment `json:"attachments"`
|
2017-11-20 00:10:04 -05:00
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
type jsonAttachment struct {
|
2017-11-20 00:10:04 -05:00
|
|
|
URL string `json:"url"`
|
|
|
|
MimeType string `json:"mime_type"`
|
|
|
|
Title string `json:"title"`
|
2018-03-14 23:09:06 -04:00
|
|
|
Size int64 `json:"size_in_bytes"`
|
2017-11-20 00:10:04 -05:00
|
|
|
Duration int `json:"duration_in_seconds"`
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonFeed) GetAuthor() string {
|
2017-11-20 00:10:04 -05:00
|
|
|
return getAuthor(j.Author)
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonFeed) Transform() *model.Feed {
|
2017-11-20 00:10:04 -05:00
|
|
|
feed := new(model.Feed)
|
|
|
|
feed.FeedURL = j.FeedURL
|
|
|
|
feed.SiteURL = j.SiteURL
|
2017-11-22 17:52:31 -05:00
|
|
|
feed.Title = strings.TrimSpace(j.Title)
|
2017-11-20 00:10:04 -05:00
|
|
|
|
|
|
|
if feed.Title == "" {
|
|
|
|
feed.Title = feed.SiteURL
|
|
|
|
}
|
|
|
|
|
|
|
|
for _, item := range j.Items {
|
|
|
|
entry := item.Transform()
|
2017-12-13 23:16:15 -05:00
|
|
|
entryURL, err := url.AbsoluteURL(feed.SiteURL, entry.URL)
|
|
|
|
if err == nil {
|
|
|
|
entry.URL = entryURL
|
|
|
|
}
|
|
|
|
|
2017-11-20 00:10:04 -05:00
|
|
|
if entry.Author == "" {
|
|
|
|
entry.Author = j.GetAuthor()
|
|
|
|
}
|
|
|
|
|
|
|
|
feed.Entries = append(feed.Entries, entry)
|
|
|
|
}
|
|
|
|
|
|
|
|
return feed
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) GetDate() time.Time {
|
2017-11-20 00:10:04 -05:00
|
|
|
for _, value := range []string{j.DatePublished, j.DateModified} {
|
|
|
|
if value != "" {
|
|
|
|
d, err := date.Parse(value)
|
|
|
|
if err != nil {
|
2017-12-15 21:55:57 -05:00
|
|
|
logger.Error("json: %v", err)
|
2017-11-20 00:10:04 -05:00
|
|
|
return time.Now()
|
|
|
|
}
|
|
|
|
|
|
|
|
return d
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return time.Now()
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) GetAuthor() string {
|
2017-11-20 00:10:04 -05:00
|
|
|
return getAuthor(j.Author)
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) GetHash() string {
|
|
|
|
for _, value := range []string{j.ID, j.URL, j.Text + j.HTML + j.Summary} {
|
2017-11-20 00:10:04 -05:00
|
|
|
if value != "" {
|
2018-01-02 22:15:08 -05:00
|
|
|
return crypto.Hash(value)
|
2017-11-20 00:10:04 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return ""
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) GetTitle() string {
|
|
|
|
for _, value := range []string{j.Title, j.Summary, j.Text, j.HTML} {
|
2017-11-20 00:10:04 -05:00
|
|
|
if value != "" {
|
2017-11-22 17:52:31 -05:00
|
|
|
return truncate(sanitizer.StripTags(value))
|
2017-11-20 00:10:04 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return j.URL
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) GetContent() string {
|
|
|
|
for _, value := range []string{j.HTML, j.Text, j.Summary} {
|
2017-11-20 00:10:04 -05:00
|
|
|
if value != "" {
|
|
|
|
return value
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return ""
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) GetEnclosures() model.EnclosureList {
|
2017-11-20 00:10:04 -05:00
|
|
|
enclosures := make(model.EnclosureList, 0)
|
|
|
|
|
|
|
|
for _, attachment := range j.Attachments {
|
|
|
|
enclosures = append(enclosures, &model.Enclosure{
|
|
|
|
URL: attachment.URL,
|
|
|
|
MimeType: attachment.MimeType,
|
|
|
|
Size: attachment.Size,
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
|
|
|
return enclosures
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func (j *jsonItem) Transform() *model.Entry {
|
2017-11-20 00:10:04 -05:00
|
|
|
entry := new(model.Entry)
|
|
|
|
entry.URL = j.URL
|
|
|
|
entry.Date = j.GetDate()
|
2017-11-22 17:52:31 -05:00
|
|
|
entry.Author = j.GetAuthor()
|
2017-11-20 00:10:04 -05:00
|
|
|
entry.Hash = j.GetHash()
|
2017-12-12 01:16:32 -05:00
|
|
|
entry.Content = j.GetContent()
|
2017-11-22 17:52:31 -05:00
|
|
|
entry.Title = strings.TrimSpace(j.GetTitle())
|
2017-11-20 00:10:04 -05:00
|
|
|
entry.Enclosures = j.GetEnclosures()
|
|
|
|
return entry
|
|
|
|
}
|
|
|
|
|
2017-11-20 21:57:54 -05:00
|
|
|
func getAuthor(author jsonAuthor) string {
|
2017-11-20 00:10:04 -05:00
|
|
|
if author.Name != "" {
|
2017-11-22 17:52:31 -05:00
|
|
|
return strings.TrimSpace(author.Name)
|
2017-11-20 00:10:04 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
return ""
|
|
|
|
}
|
|
|
|
|
|
|
|
func truncate(str string) string {
|
|
|
|
max := 100
|
2017-11-22 17:52:31 -05:00
|
|
|
str = strings.TrimSpace(str)
|
2017-11-20 00:10:04 -05:00
|
|
|
if len(str) > max {
|
|
|
|
return str[:max] + "..."
|
|
|
|
}
|
|
|
|
|
|
|
|
return str
|
|
|
|
}
|