From 21b4c031983f2b102c79b33b2d64f575e10912db Mon Sep 17 00:00:00 2001
From: Travis Ralston <travpc@gmail.com>
Date: Mon, 29 Jan 2018 20:40:10 -0700
Subject: [PATCH] Limit the URL preview title length by words
Fixes #62
---
config.sample.yaml | 1 +
.../turt2live/matrix-media-repo/config/config.go | 2 ++
.../services/url_service/opengraph_previewer.go | 9 ++++++---
3 files changed, 9 insertions(+), 3 deletions(-)
diff --git a/config.sample.yaml b/config.sample.yaml
index 6cb6258b..70a8b7c4 100644
--- a/config.sample.yaml
+++ b/config.sample.yaml
@@ -91,6 +91,7 @@ urlPreviews:
enabled: true # If enabled, the preview_url routes will be accessible
maxPageSizeBytes: 10485760 # 10MB default, 0 to disable
numWords: 30 # The number of words to include in a preview (maximum)
+ numTitleWords: 30 # The maximum number of words to include in a preview's title
# The number of workers to use when generating url previews. Raise this number if url
# previews are slow or timing out.
diff --git a/src/github.com/turt2live/matrix-media-repo/config/config.go b/src/github.com/turt2live/matrix-media-repo/config/config.go
index 69207837..edeea26f 100644
--- a/src/github.com/turt2live/matrix-media-repo/config/config.go
+++ b/src/github.com/turt2live/matrix-media-repo/config/config.go
@@ -62,6 +62,7 @@ type ThumbnailSize struct {
type UrlPreviewsConfig struct {
Enabled bool `yaml:"enabled"`
NumWords int `yaml:"numWords"`
+ NumTitleWords int `yaml:"numTitleWords"`
MaxPageSizeBytes int64 `yaml:"maxPageSizeBytes"`
NumWorkers int `yaml:"numWorkers"`
DisallowedNetworks []string `yaml:"disallowedNetworks,flow"`
@@ -204,6 +205,7 @@ func NewDefaultConfig() *MediaRepoConfig {
UrlPreviews: &UrlPreviewsConfig{
Enabled: true,
NumWords: 30,
+ NumTitleWords: 30,
MaxPageSizeBytes: 10485760, // 10mb
NumWorkers: 10,
DisallowedNetworks: []string{
diff --git a/src/github.com/turt2live/matrix-media-repo/services/url_service/opengraph_previewer.go b/src/github.com/turt2live/matrix-media-repo/services/url_service/opengraph_previewer.go
index 1cab61ca..a397395d 100644
--- a/src/github.com/turt2live/matrix-media-repo/services/url_service/opengraph_previewer.go
+++ b/src/github.com/turt2live/matrix-media-repo/services/url_service/opengraph_previewer.go
@@ -71,11 +71,15 @@ func (p *openGraphUrlPreviewer) GeneratePreview(urlStr string) (openGraphResult,
og.Images = calcImages(html)
}
+ // Be sure to trim the title and description
+ og.Title = summarize(og.Title, config.Get().UrlPreviews.NumTitleWords)
+ og.Description = summarize(og.Description, config.Get().UrlPreviews.NumWords)
+
graph := &openGraphResult{
Type: og.Type,
Url: og.URL,
Title: og.Title,
- Description: summarize(og.Description),
+ Description: og.Description,
SiteName: og.SiteName,
}
@@ -260,7 +264,7 @@ func calcImages(html string) []*opengraph.Image {
return []*opengraph.Image{&img}
}
-func summarize(text string) (string) {
+func summarize(text string, maxWords int) (string) {
// Normalize the whitespace to be something useful (crush it to one giant line)
surroundingWhitespace := regexp.MustCompile(`^[\s\p{Zs}]+|[\s\p{Zs}]+$`)
interiorWhitespace := regexp.MustCompile(`[\s\p{Zs}]{2,}`)
@@ -269,7 +273,6 @@ func summarize(text string) (string) {
text = interiorWhitespace.ReplaceAllString(text, " ")
text = newlines.ReplaceAllString(text, " ")
- maxWords := config.Get().UrlPreviews.NumWords
words := strings.Split(text, " ")
if len(words) < maxWords {
return text
--
GitLab