jekyll-artisync 0.2 → 0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: bdd1c95dc86f0e265d5e748dd9559cff17968ad55570e7d610ec2441a7339382
-  data.tar.gz: 184d76a6fe6174d9c0888dbe8126b2bf16a1e6306f28dcd836b347dd2698527a
+  metadata.gz: 31d37fc6ffdd79cb4d259d1db001b0abea1f2f4d8be3d17b035e2db5e1a7cc80
+  data.tar.gz: 1b5b26500b61c8e271d0c2a2605b28662c041572a1428469e9d30209c718a0c5
 SHA512:
-  metadata.gz: 6eead02cfc58bfb5bb24d3860e222a6e04ba3883a9a40a8a9c8e673efc94b29b22f851631e103dbc93ca1ff0a9e43ccefeca1eec1bbf0839c395baa0d2c86a94
-  data.tar.gz: 7c34719be25ff7b9ab5501c2a2ef5512d71678a4c3ea918ff7b6d53ae8f874adc7683790d8468e1afaf177e167584ae7fe111d8d35f9ea1f432665aa1244cb97
+  metadata.gz: 5d643be4d2d2c4831251e06e6f092ff419148b7c21fbdc1d274642d7dad5cf7de741b635af6e556221e47090b61dd0a28cc4614b9adae8a829062c354ce71468
+  data.tar.gz: 7334aeea98d30013f3ba030caaeafbb3fbd542d74998f00efb152d1873bbdd83542b9213b0cc041836b2e9becb62622d01e894da0e5fc003b44dbc2a576cb79b

data/lib/jekyll-artisync.rb

@@ -6,8 +6,8 @@ require "jekyll"
 # user agent is necessary otherwise certain sites such as Zhihu throws 400
 USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.86 Safari/537.36'
 
-SITE_TO_ARTICLE_XPATH = {
-  'zhihu' => '//div[contains(@class, "Post-RichText") and contains(@class, "ztext")]',
+HOST_TO_ARTICLE_XPATH = {
+  'zhuanlan.zhihu.com' => '//div[contains(@class, "Post-RichText") and contains(@class, "ztext")]',
 }
 
 class ArticleSyncEmbed < Liquid::Tag
@@ -17,8 +17,7 @@ class ArticleSyncEmbed < Liquid::Tag
     @content = content
   end
 
-  def _fetch_html(url)
-    uri = URI(url)
+  def _fetch_html(uri)
     res = Net::HTTP.start(uri.hostname, uri.port, :use_ssl => true) do |http|
       # :use_ssl => true for the uri is https
       http.request(Net::HTTP::Get.new(uri, {'User-Agent' => USER_AGENT}))
@@ -42,9 +41,12 @@ class ArticleSyncEmbed < Liquid::Tag
   end
 
   def render(context)
-    url, site = @content.strip.split
-    page_html = self._fetch_html(url)
-    article = Nokogiri::HTML(page_html).xpath(SITE_TO_ARTICLE_XPATH[site])
+    url = @content.strip
+    uri = URI(url)
+    page_host = uri.hostname
+    puts page_host
+    page_html = self._fetch_html(uri)
+    article = Nokogiri::HTML(page_html).xpath(HOST_TO_ARTICLE_XPATH[page_host])
     content = []
     article.children.each do |node|
       content.append self._handle_node(node)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-artisync
 version: !ruby/object:Gem::Version
-  version: '0.2'
+  version: '0.4'
 platform: ruby
 authors:
 - Junhan