jekyll-artisync 0.4 → 0.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 31d37fc6ffdd79cb4d259d1db001b0abea1f2f4d8be3d17b035e2db5e1a7cc80
-  data.tar.gz: 1b5b26500b61c8e271d0c2a2605b28662c041572a1428469e9d30209c718a0c5
+  metadata.gz: '0606097a6fb1063a731eb03353da74882e9d5aca8edf40143f7c6182e189c91a'
+  data.tar.gz: 7bac92d7cae8be329f1f31c782efa8faced9f280450164f110f81f7f96865e36
 SHA512:
-  metadata.gz: 5d643be4d2d2c4831251e06e6f092ff419148b7c21fbdc1d274642d7dad5cf7de741b635af6e556221e47090b61dd0a28cc4614b9adae8a829062c354ce71468
-  data.tar.gz: 7334aeea98d30013f3ba030caaeafbb3fbd542d74998f00efb152d1873bbdd83542b9213b0cc041836b2e9becb62622d01e894da0e5fc003b44dbc2a576cb79b
+  metadata.gz: 897c6f5173ed4af5541c0784609a136e7f09d3a22e174ea38f4fbc908471a7228a311c670d61557611bbce5d62cb62831f667506f25188f4938562d989d8b207
+  data.tar.gz: 6b51c4520cb6a6005994184a478073dac60387e3e363f11432363bd9f38a59312f47763488d2fbdc54b4bb03db4933394ab0852fc5d73f72d75dd5e9d911d977

data/lib/jekyll-artisync.rb

@@ -1,14 +1,6 @@
-require 'net/http'
-require 'nokogiri'
-
+require 'syncers/syncer_factory'
 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'
-
-HOST_TO_ARTICLE_XPATH = {
-  'zhuanlan.zhihu.com' => '//div[contains(@class, "Post-RichText") and contains(@class, "ztext")]',
-}
 
 class ArticleSyncEmbed < Liquid::Tag
 
@@ -17,42 +9,12 @@ class ArticleSyncEmbed < Liquid::Tag
     @content = content
   end
 
-  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}))
-    end
-
-    res.body
-  end
-
-  def _handle_node(node)
-    case node.name
-    when 'figure'
-      img_node = node.children[1]
-
-      img_url = img_node['data-actualsrc']
-      img_url['_b.jpg'] = '_720w.jpg'
-      img_url['/v2'] = "/80/v2"
-
-      img_node['src'] = img_url
-    end
-    node.to_html
-  end
-
   def render(context)
     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)
-    end
-
-    content.join("\n")
+    syncer = SyncerFactory.get_syncer(uri)
+    return syncer.gen_html
   end
 
   Liquid::Template.register_tag "artisync", self

data/lib/syncers/abstract_syncer.rb

@@ -0,0 +1,31 @@
+require 'net/http'
+
+
+# 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'
+
+MESSAGE = "AbstractSyncer::[%s]: override required."
+
+class AbstractSyncer
+  def initialize(uri)
+    @uri = uri
+  end
+
+  def _fetch_html
+    uri = @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}))
+    end
+
+    res.body
+  end
+
+  def get_article_xpath
+    raise MESSAGE % [__method__]
+  end
+
+  def gen_html
+    raise MESSAGE % [__method__]
+  end
+end

data/lib/syncers/syncer_factory.rb

@@ -0,0 +1,18 @@
+require_relative 'zhihu_syncer'
+
+
+ZHIHU_ZHUANLAN = 'zhuanlan.zhihu.com'
+
+class SyncerFactory
+  def self.get_syncer(uri)
+    host_name = uri.host
+
+    case host_name
+    when ZHIHU_ZHUANLAN
+      return ZhihuSyncer.new(uri)
+    else
+      raise "Not supported website for host: #{host_name}"
+    end
+  end
+end
+

data/lib/syncers/zhihu_syncer.rb

@@ -0,0 +1,35 @@
+require_relative 'abstract_syncer'
+require 'nokogiri'
+
+
+class ZhihuSyncer < AbstractSyncer
+  def get_article_xpath
+    '//div[contains(@class, "Post-RichText") and contains(@class, "ztext")]'
+  end
+
+  def _handle_node(node)
+    case node.name
+    when 'figure'
+      img_node = node.children[1]
+
+      img_url = img_node['data-actualsrc']
+      img_url['_b.jpg'] = '_720w.jpg'
+      img_url['/v2'] = "/80/v2"
+
+      img_node['src'] = img_url
+    end
+    node.to_html
+  end
+
+  def gen_html
+    page_html = self._fetch_html
+    article = Nokogiri::HTML(page_html).xpath(self.get_article_xpath)
+    content = []
+    article.children.each do |node|
+      content.append self._handle_node(node)
+    end
+
+    content.join("\n")
+  end
+
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: jekyll-artisync
 version: !ruby/object:Gem::Version
-  version: '0.4'
+  version: '0.5'
 platform: ruby
 authors:
 - Junhan
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-05-24 00:00:00.000000000 Z
+date: 2020-05-27 00:00:00.000000000 Z
 dependencies: []
 description: Take an article from a given site and display on current Jekyll page.
 email:
@@ -18,6 +18,9 @@ extensions: []
 extra_rdoc_files: []
 files:
 - lib/jekyll-artisync.rb
+- lib/syncers/abstract_syncer.rb
+- lib/syncers/syncer_factory.rb
+- lib/syncers/zhihu_syncer.rb
 homepage: https://github.com/junhan-z/jekyll-artisync
 licenses:
 - MIT