Skip to content

Commit 66b269e

Browse files
committed
feat(readability): avoid removing elements with content class
1 parent 54abd0a commit 66b269e

File tree

1 file changed

+9
-9
lines changed

1 file changed

+9
-9
lines changed

internal/reader/readability/readability.go

Lines changed: 9 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -18,9 +18,9 @@ import (
1818
const defaultTagsToScore = "section,h2,h3,h4,h5,h6,p,td,pre,div"
1919

2020
var (
21-
strongCandidates = [...]string{"popupbody", "-ad", "g-plus"}
22-
maybeCandidate = [...]string{"and", "article", "body", "column", "main", "shadow"}
23-
unlikelyCandidate = [...]string{"banner", "breadcrumbs", "combx", "comment", "community", "cover-wrap", "disqus", "extra", "foot", "header", "legends", "menu", "modal", "related", "remark", "replies", "rss", "shoutbox", "sidebar", "skyscraper", "social", "sponsor", "supplemental", "ad-break", "agegate", "pagination", "pager", "popup", "yom-remote"}
21+
strongCandidatesToRemove = [...]string{"popupbody", "-ad", "g-plus"}
22+
maybeCandidateToRemove = [...]string{"and", "article", "body", "column", "main", "shadow", "content"}
23+
unlikelyCandidateToRemove = [...]string{"banner", "breadcrumbs", "combx", "comment", "community", "cover-wrap", "disqus", "extra", "foot", "header", "legends", "menu", "modal", "related", "remark", "replies", "rss", "shoutbox", "sidebar", "skyscraper", "social", "sponsor", "supplemental", "ad-break", "agegate", "pagination", "pager", "popup", "yom-remote"}
2424

2525
positiveKeywords = [...]string{"article", "blog", "body", "content", "entry", "h-entry", "hentry", "main", "page", "pagination", "post", "story", "text"}
2626
negativeKeywords = [...]string{"author", "banner", "byline", "com-", "combx", "comment", "contact", "dateline", "foot", "hid", "masthead", "media", "meta", "modal", "outbrain", "promo", "related", "scroll", "share", "shopping", "shoutbox", "sidebar", "skyscraper", "sponsor", "tags", "tool", "widget", "writtenby"}
@@ -185,17 +185,17 @@ func shouldRemoveCandidate(str string) bool {
185185
str = strings.ToLower(str)
186186

187187
// Those candidates have no false-positives, no need to check against `maybeCandidate`
188-
for _, strongCandidate := range strongCandidates {
189-
if strings.Contains(str, strongCandidate) {
188+
for _, strongCandidateToRemove := range strongCandidatesToRemove {
189+
if strings.Contains(str, strongCandidateToRemove) {
190190
return true
191191
}
192192
}
193193

194-
for _, unlikelyCandidate := range unlikelyCandidate {
195-
if strings.Contains(str, unlikelyCandidate) {
194+
for _, unlikelyCandidateToRemove := range unlikelyCandidateToRemove {
195+
if strings.Contains(str, unlikelyCandidateToRemove) {
196196
// Do we have a false positive?
197-
for _, maybe := range maybeCandidate {
198-
if strings.Contains(str, maybe) {
197+
for _, maybeCandidateToRemove := range maybeCandidateToRemove {
198+
if strings.Contains(str, maybeCandidateToRemove) {
199199
return false
200200
}
201201
}

0 commit comments

Comments
 (0)