Finish TensorFlow scraper

assets/javascripts/news.json

@@ -1,5 +1,8 @@
 [
   [
+    "2016-03-06",
+    "New documentation: <a href=\"/tensorflow/\">TensorFlow</a>"
+  ], [
     "2016-02-28",
     "New documentations: <a href=\"/codeigniter/\">CodeIgniter</a>, <a href=\"/nginx_lua_module/\">nginx Lua Module</a> and <a href=\"/influxdata/\">InfluxData</a>"
   ], [

assets/javascripts/templates/pages/about_tmpl.coffee

@@ -414,6 +414,11 @@ credits = [
     'The Regents of the University of California, Sun Microsystems, Inc., Scriptics Corporation, and other parties',
     'Tcl/Tk',
     'http://tcl.tk/software/tcltk/license.html'
+  ], [
+    'TensorFlow',
+    '2015 The TensorFlow Authors',
+    'Apache',
+    'https://raw.githubusercontent.com/tensorflow/tensorflow/master/LICENSE'
   ], [
     'Underscore.js',
     '2009-2015 Jeremy Ashkenas, DocumentCloud and Investigative Reporters & Editors',

assets/javascripts/views/pages/tensorflow.coffee

@@ -0,0 +1,7 @@
+#= require views/pages/base
+
+class app.views.TensorflowPage extends app.views.BasePage
+  prepare: ->
+    @highlightCode @findAll('pre[class*="lang-c++"]'), 'cpp'
+    @highlightCode @findAll('pre.lang-python'), 'python'
+    return

assets/stylesheets/global/_icons.scss

@@ -129,3 +129,4 @@
 ._icon-ramda:before         { background-position: -3rem -10rem; @extend %darkIconFix !optional; }
 ._icon-codeigniter:before   { background-position: -4rem -10rem; @extend %darkIconFix !optional; }
 ._icon-influxdata:before    { background-position: -5rem -10rem; @extend %darkIconFix !optional; }
+._icon-tensorflow:before    { background-position: -6rem -10rem; }

assets/stylesheets/pages/_tensorflow.scss

@@ -1,12 +1,6 @@
 ._tensorflow {
-  h2, h3, h4 { @extend %block-heading, %label-blue; }
+  @extend %simple;
-  p > code, li > code { @extend %label; }
 
-  // These are used for lists of arguments and return values in the docs.
+  h4 { @extend %block-label; }
-  b > code {
+  > .toc ul ul { margin: .25rem 0; }
-    @extend %label;
-    background-color: initial;
-    font-weight: 500;
-    font-size: 1.1em;
-  }
 }

lib/docs/core/doc.rb

@@ -48,7 +48,7 @@ module Docs
 
       def slug
         slug = @slug || name.try(:downcase)
-        version? ? "#{slug}~#{version.downcase.gsub(/[^a-z0-9\_\.]/, '_')}" : slug
+        version? ? "#{slug}~#{version.downcase.gsub('+', 'p').gsub(/[^a-z0-9\_\.]/, '_')}" : slug
       end
 
       def path

lib/docs/filters/tensorflow/clean_html.rb

@@ -3,6 +3,16 @@ module Docs
     class CleanHtmlFilter < Filter
       def call
         css('hr').remove
+
+        css('pre > code').each do |node|
+          node.parent['class'] = node['class']
+          node.parent.content = node.content
+        end
+
+        css('b').each do |node|
+          node.before(node.children).remove
+        end
+
         doc
       end
     end

lib/docs/filters/tensorflow/entries.rb

@@ -2,31 +2,26 @@ module Docs
   class Tensorflow
     class EntriesFilter < Docs::EntriesFilter
       def get_name
-        at_css('h1').content
+        name = at_css('h1').content.strip
+        name.remove! 'class '
+        name.remove! 'struct '
+        name
       end
 
       def get_type
-        at_css('h1').content
+        type = name.dup
-      end
+        type.remove! %r{\ \(.*\)}
-
+        type.remove! 'tensorflow::'
-      def include_default_entry?
+        type
-        false
       end
 
       def additional_entries
-        entries = []
+        css('h2 code', 'h3 code', 'h4 code', 'h5 code').map do |node|
-
+          name = node.content
-        # Just get everything that is a code tag inside a header tag. I haven't
+          name.sub! %r{\(.*}, '()'
-        # checked if all of these are necessary.
+          name = name.split(' ').last
-        ents = css('h5 code') + css('h4 code') + css('h3 code') + css('h2 code')
+          [name, node.parent['id']]
-
-        ents.each do |node|
-          name = node.content.sub(/\(.*\)/, '()')
-          id = node.parent['id']
-          entries << [name, id, get_name]
         end
-
-        entries
       end
     end
   end

lib/docs/scrapers/tensorflow.rb

@@ -1,18 +1,31 @@
 module Docs
   class Tensorflow < UrlScraper
     self.name = 'TensorFlow'
-    self.slug = 'tensorflow'
     self.type = 'tensorflow'
-    self.release = '0.6.0-py'
-    self.base_url = 'https://www.tensorflow.org/versions/0.6.0/api_docs/python/'
 
-    options[:container] = '#content'
+    html_filters.push 'tensorflow/entries', 'tensorflow/clean_html'
 
-    html_filters.push 'tensorflow/entries', 'tensorflow/clean_html', 'clean_html'
+    options[:container] = '#content'
 
     options[:attribution] = <<-HTML
-      &copy; The TensorFlow Authors.  All rights reserved.<br>
+      &copy; 2015 The TensorFlow Authors. All rights reserved.<br>
       Licensed under the Apache 2.0 License.
     HTML
+
+    version 'Python' do
+      self.base_url = 'https://www.tensorflow.org/versions/r0.7/api_docs/python/'
+      self.release = '0.7'
+    end
+
+    version 'C++' do
+      self.base_url = 'https://www.tensorflow.org/versions/r0.7/api_docs/cc/'
+      self.release = '0.7'
+
+      options[:fix_urls] = ->(url) {
+        url.sub! '/api_docs/cc/class', '/api_docs/cc/Class'
+        url.sub! '/api_docs/cc/struct', '/api_docs/cc/Struct'
+        url
+      }
+    end
   end
 end