Add set of benchmarks to compare performance of different handlers

.gitignore

@@ -10,3 +10,6 @@ fetch/results.txt
 others/others-benchmark.png
 others/results.csv
 others/results.txt
+handlers/handlers-benchmark.png
+handlers/results.csv
+handlers/results.txt

Gemfile

@@ -1,8 +1,11 @@
 source 'https://rubygems.org'
 
 gem 'curb'
+gem 'feedjira'
+gem 'ox'
+gem 'oga'
+
 gem 'feed-normalizer'
 gem 'feed_me'
 gem 'feed_parser'
-gem 'feedjira'
 gem 'simple-rss'

Gemfile.lock

@@ -1,26 +1,34 @@
 GEM
  remote: https://rubygems.org/
  specs:
- curb (0.8.5)
+ ansi (1.5.0)
+ ast (2.0.0)
+ curb (0.8.8)
  feed-normalizer (1.5.2)
  hpricot (>= 0.6)
  simple-rss (>= 1.1)
  feed_me (0.7.0)
  nokogiri
  feed_parser (0.3.5)
  nokogiri
- feedjira (1.1.0)
- curb (~> 0.8.1)
- loofah (~> 1.2.1)
- sax-machine (~> 0.2.1)
+ feedjira (1.6.0)
+ curb (~> 0.8)
+ loofah (~> 2.0)
+ sax-machine (~> 1.0)
  hpricot (0.8.6)
- loofah (1.2.1)
- nokogiri (>= 1.4.4)
- mini_portile (0.5.3)
- nokogiri (1.6.1)
- mini_portile (~> 0.5.0)
- sax-machine (0.2.1)
- nokogiri (~> 1.6.0)
+ loofah (2.0.1)
+ nokogiri (>= 1.5.9)
+ mini_portile (0.6.2)
+ nokogiri (1.6.6.2)
+ mini_portile (~> 0.6.0)
+ oga (0.3.4)
+ ast
+ ruby-ll (~> 2.1)
+ ox (2.1.8)
+ ruby-ll (2.1.1)
+ ansi
+ ast
+ sax-machine (1.3.2)
  simple-rss (1.3.1)
 
 PLATFORMS
@@ -32,4 +40,6 @@ DEPENDENCIES
  feed_me
  feed_parser
  feedjira
+ oga
+ ox
  simple-rss

README.md

@@ -80,3 +80,18 @@ This command will output three files:
 * others/results.txt - the raw numbers from the benchmark
 * others/results.csv - the totals from the raw numbers in CSV
 * others/others-benchmark.png - that CSV converted to chart form
+
+## Running the handlers benchmark
+
+This benchmark compares Feedjira with different `sax-machine`
+handlers:
+
+```
+$ handlers/run
+```
+
+This command will output three files:
+
+* handlers/results.txt - the raw numbers from the benchmark
+* handlers/results.csv - the totals from the raw numbers in CSV
+* handlers/handlers-benchmark.png - that CSV converted to chart form

feed_xml/loop.xml

@@ -21,7 +21,7 @@
 <atom:link rel='hub' href='http://www.loopinsight.com/?pushpress=hub'/>
 <item>
 <title>The Wisdom of Slash</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www2.gibson.com/News-Lifestyle/Features/en-us/The-Wisdom-of-Slash.aspx</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www2.gibson.com/News-Lifestyle/Features/en-us/The-Wisdom-of-Slash.aspx</link>
 <comments>http://www.loopinsight.com/2013/11/29/the-wisdom-of-slash/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 17:21:09 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -40,7 +40,7 @@
 </item>
 <item>
 <title>CSS alignment and sizing</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://timseverien.nl/2013/10/css-alignment-and-sizing/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://timseverien.nl/2013/10/css-alignment-and-sizing/</link>
 <comments>http://www.loopinsight.com/2013/11/29/css-alignment-and-sizing/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 15:56:33 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -59,7 +59,7 @@
 </item>
 <item>
 <title>∞ Zakk Wylde shows you his live setup</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.loopinsight.com/2013/11/29/zakk-wylde-shows-you-his-live-setup/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.loopinsight.com/2013/11/29/zakk-wylde-shows-you-his-live-setup/</link>
 <comments>http://www.loopinsight.com/2013/11/29/zakk-wylde-shows-you-his-live-setup/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 15:35:49 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -77,7 +77,7 @@
 </item>
 <item>
 <title>Pedal Jeanie Pedalboard</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.pedaljeanie.com/pedalboard/pedal-jeanie/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.pedaljeanie.com/pedalboard/pedal-jeanie/</link>
 <comments>http://www.loopinsight.com/2013/11/29/pedal-jeanie-pedalboard/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 15:21:50 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -96,7 +96,7 @@
 </item>
 <item>
 <title>Tesla takes on lawmakers with loaner cars and star power</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.mercurynews.com/business/ci_24612314/tesla-faces-uphill-fight-against-dealers-lawmakers</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.mercurynews.com/business/ci_24612314/tesla-faces-uphill-fight-against-dealers-lawmakers</link>
 <comments>http://www.loopinsight.com/2013/11/29/tesla-takes-on-lawmakers-with-loaner-cars-and-star-power/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 13:01:11 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -115,7 +115,7 @@
 </item>
 <item>
 <title>How to Travel Around the World for $418</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://gizmodo.com/5710654/how-to-fly-35000-miles-visit-4-continents-9-countries-and-15-cities-for-418/1473283510/@andrewliszewski</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://gizmodo.com/5710654/how-to-fly-35000-miles-visit-4-continents-9-countries-and-15-cities-for-418/1473283510/@andrewliszewski</link>
 <comments>http://www.loopinsight.com/2013/11/29/how-to-travel-around-the-world-for-418/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 12:47:10 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -134,7 +134,7 @@
 </item>
 <item>
 <title>World Beard and Moustache Championships</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.worldbeardchampionships.com</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.worldbeardchampionships.com</link>
 <comments>http://www.loopinsight.com/2013/11/29/world-beard-and-moustache-championships/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 12:41:14 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -153,7 +153,7 @@
 </item>
 <item>
 <title>∞ Incredible a cappella version of Royals</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.loopinsight.com/2013/11/29/incredible-a-cappella-version-of-royals/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.loopinsight.com/2013/11/29/incredible-a-cappella-version-of-royals/</link>
 <comments>http://www.loopinsight.com/2013/11/29/incredible-a-cappella-version-of-royals/#comments</comments>
 <pubDate>Fri, 29 Nov 2013 12:34:44 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -172,7 +172,7 @@
 </item>
 <item>
 <title>A brainteaser by Albert Einstein</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.coudal.com/thefish.php</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.coudal.com/thefish.php</link>
 <comments>http://www.loopinsight.com/2013/11/28/a-brainteaser-by-albert-einstein/#comments</comments>
 <pubDate>Thu, 28 Nov 2013 16:16:27 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -191,7 +191,7 @@
 </item>
 <item>
 <title>Man sends action figures of himself to potential employers</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.core77.com/blog/strategy_research/forget_resumes_jens_lennartsson_sends_action_figures_of_himself_to_potential_employers_25988.asp</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.core77.com/blog/strategy_research/forget_resumes_jens_lennartsson_sends_action_figures_of_himself_to_potential_employers_25988.asp</link>
 <comments>http://www.loopinsight.com/2013/11/28/man-sends-action-figures-of-himself-to-potential-employers/#comments</comments>
 <pubDate>Thu, 28 Nov 2013 15:58:02 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -210,7 +210,7 @@
 </item>
 <item>
 <title>Samsung caught cheating on benchmarks&#8230; again</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.theregister.co.uk/2013/11/26/samsung_and_htc_face_android_3d_graphics_test_delisting_shame/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.theregister.co.uk/2013/11/26/samsung_and_htc_face_android_3d_graphics_test_delisting_shame/</link>
 <comments>http://www.loopinsight.com/2013/11/28/samsung-caught-cheating-on-benchmarks-again/#comments</comments>
 <pubDate>Thu, 28 Nov 2013 15:54:22 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -237,7 +237,7 @@ I'm shocked Samsung would stoop to such levels.<a href="http://www.loopinsight.c
 </item>
 <item>
 <title>Man tosses hard drive with more than $7.5 million of bitcoins</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.theguardian.com/technology/2013/nov/27/hard-drive-bitcoin-landfill-site</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.theguardian.com/technology/2013/nov/27/hard-drive-bitcoin-landfill-site</link>
 <comments>http://www.loopinsight.com/2013/11/28/man-tosses-hard-drive-with-more-than-7-5-million-of-bitcoins/#comments</comments>
 <pubDate>Thu, 28 Nov 2013 13:54:15 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -256,7 +256,7 @@ I'm shocked Samsung would stoop to such levels.<a href="http://www.loopinsight.c
 </item>
 <item>
 <title>Excerpt from &#8220;Jony Ive: The Genius Behind Apple&#8217;s Greatest Products&#8221;</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=https://medium.com/book-excerpts/7bb11b0284c2</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=https://medium.com/book-excerpts/7bb11b0284c2</link>
 <comments>http://www.loopinsight.com/2013/11/28/excerpt-from-jony-ive-the-genius-behind-apples-greatest-products/#comments</comments>
 <pubDate>Thu, 28 Nov 2013 12:40:39 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -276,7 +276,7 @@ I'm shocked Samsung would stoop to such levels.<a href="http://www.loopinsight.c
 </item>
 <item>
 <title>The forgotten Apple Newsstand</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://pandodaily.com/2013/11/27/who-hates-jony-ives-ios7-publishers-thats-who/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://pandodaily.com/2013/11/27/who-hates-jony-ives-ios7-publishers-thats-who/</link>
 <comments>http://www.loopinsight.com/2013/11/27/the-forgotten-apple-newsstand/#comments</comments>
 <pubDate>Thu, 28 Nov 2013 00:10:01 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -303,7 +303,7 @@ I'm shocked Samsung would stoop to such levels.<a href="http://www.loopinsight.c
 </item>
 <item>
 <title>Amplified: A Metric Thanksgiving</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://5by5.tv/amplified/81</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://5by5.tv/amplified/81</link>
 <comments>http://www.loopinsight.com/2013/11/27/amplified-a-metric-thanksgiving/#comments</comments>
 <pubDate>Wed, 27 Nov 2013 23:05:54 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -322,7 +322,7 @@ I'm shocked Samsung would stoop to such levels.<a href="http://www.loopinsight.c
 </item>
 <item>
 <title>My Paper feed reader for iOS</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://loopu.in/1aW84YA</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://loopu.in/1aW84YA</link>
 <comments>http://www.loopinsight.com/2013/11/27/my-paper-feed-reader-for-ios/#comments</comments>
 <pubDate>Wed, 27 Nov 2013 23:02:45 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -349,7 +349,7 @@ And it's free.<a href="http://www.loopinsight.com/redirect?s=/feed/&amp;u=http:/
 </item>
 <item>
 <title>What is Android?</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://techpinions.com/android-is-eating-the-world/25276</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://techpinions.com/android-is-eating-the-world/25276</link>
 <comments>http://www.loopinsight.com/2013/11/27/what-is-android/#comments</comments>
 <pubDate>Wed, 27 Nov 2013 20:58:06 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -377,7 +377,7 @@ And it's free.<a href="http://www.loopinsight.com/redirect?s=/feed/&amp;u=http:/
 </item>
 <item>
 <title>7 things you had no idea the world was running out of</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://gizmodo.com/7-things-you-had-no-idea-the-world-is-running-out-of-1467868161</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://gizmodo.com/7-things-you-had-no-idea-the-world-is-running-out-of-1467868161</link>
 <comments>http://www.loopinsight.com/2013/11/27/7-things-you-had-no-idea-the-world-was-running-out-of/#comments</comments>
 <pubDate>Wed, 27 Nov 2013 20:23:24 +0000</pubDate>
 <dc:creator><![CDATA[Dave Mark]]></dc:creator>
@@ -396,7 +396,7 @@ And it's free.<a href="http://www.loopinsight.com/redirect?s=/feed/&amp;u=http:/
 </item>
 <item>
 <title>Paul Stewart: Capturing the movie &#8220;Slap Shot&#8221;</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.hockeybuzz.com/blog/Eklund/Paul-Stewart-A-Slap-Shot-Story---My-Day-as-a-Long-Island-Duck/1/55993</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.hockeybuzz.com/blog/Eklund/Paul-Stewart-A-Slap-Shot-Story---My-Day-as-a-Long-Island-Duck/1/55993</link>
 <comments>http://www.loopinsight.com/2013/11/27/paul-stewart-capturing-the-movie-slap-shot/#comments</comments>
 <pubDate>Wed, 27 Nov 2013 16:20:37 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>
@@ -415,7 +415,7 @@ And it's free.<a href="http://www.loopinsight.com/redirect?s=/feed/&amp;u=http:/
 </item>
 <item>
 <title>∞ Slash: &#8220;Sweet Child O&#8217; Mine&#8221; acoustic</title>
-<link>http://www.loopinsight.com/redirect?s=/feed/&#038;u=http://www.loopinsight.com/2013/11/27/slash-sweet-child-o-mine-acoustic/</link>
+<link>http://www.loopinsight.com/redirect?s=/feed/&#38;u=http://www.loopinsight.com/2013/11/27/slash-sweet-child-o-mine-acoustic/</link>
 <comments>http://www.loopinsight.com/2013/11/27/slash-sweet-child-o-mine-acoustic/#comments</comments>
 <pubDate>Wed, 27 Nov 2013 15:23:57 +0000</pubDate>
 <dc:creator><![CDATA[Jim Dalrymple]]></dc:creator>

feed_xml/white_house.xml

@@ -305,7 +305,7 @@
  <guid isPermaLink="false">247101 at http://www.whitehouse.gov</guid>
 </item>
 <item>
- <title>A Woman&#039;s Health Care Decisions Should Be in Her Own Hands, Not Her Boss&#039;s </title>
+ <title>A Woman&#39;s Health Care Decisions Should Be in Her Own Hands, Not Her Boss&#39;s </title>
  <link>http://www.whitehouse.gov/blog/2013/11/26/womans-health-care-decisions-should-be-her-own-hands-not-her-bosss</link>
  <description><![CDATA[<p>Ed. note: This is cross-posted from The Huffington Post. See the original post <a href="http://www.huffingtonpost.com/valerie-jarrett/supreme-court-birth-control_b_4346143.html">here</a>.</p>
 <p>Ensuring the full freedom of women as health care consumers to access essential preventative health services is a vital component of the Affordable Care Act (ACA). And nowhere are health decisions more personal or essential to keep in their hands, than those regarding reproductive health. The ACA was designed to ensure that health care decisions are made between a woman and her doctor, and not by her boss, or Washington politicians.</p>
@@ -544,7 +544,7 @@
  <guid isPermaLink="false">246546 at http://www.whitehouse.gov</guid>
 </item>
 <item>
- <title>President Obama Delivers Remarks on Iran&#039;s Nuclear Program</title>
+ <title>President Obama Delivers Remarks on Iran&#39;s Nuclear Program</title>
  <link>http://www.whitehouse.gov/blog/2013/11/23/president-obama-delivers-remarks-iran</link>
  <description><![CDATA[<div>
 
@@ -672,7 +672,7 @@
  <guid isPermaLink="false">245816 at http://www.whitehouse.gov</guid>
 </item>
 <item>
- <title>Here&#039;s Why Today&#039;s Filibuster Rule Change is a Big Deal</title>
+ <title>Here&#39;s Why Today&#39;s Filibuster Rule Change is a Big Deal</title>
  <link>http://www.whitehouse.gov/blog/2013/11/21/heres-why-todays-filibuster-rule-change-really-big-deal</link>
  <description><![CDATA[<p></p>
 <p class="rteright"><a href="https://www.youtube.com/watch?v=joZFPU6nhQU">Watch on YouTube.</a></p>
@@ -882,7 +882,7 @@
  <guid isPermaLink="false">245376 at http://www.whitehouse.gov</guid>
 </item>
 <item>
- <title>President Obama&#039;s Handwritten Essay Marking the 150th Anniversary of the Gettysburg Address</title>
+ <title>President Obama&#39;s Handwritten Essay Marking the 150th Anniversary of the Gettysburg Address</title>
  <link>http://www.whitehouse.gov/blog/2013/11/19/president-obamas-handwritten-essay-marking-150th-anniversary-gettysburg-address</link>
  <description><![CDATA[<p>One hundred and fifty years ago today, <a href="http://www.whitehouse.gov/about/presidents/abrahamlincoln">President Abraham Lincoln</a> delivered one of the most memorable speeches in U.S. history from Gettysburg, PA. In dedicating the military cemetery where thousands of soldiers were buried following the Battle of Gettysburg, fought just four months earlier, Lincoln described &quot;a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.&quot;</p>
 <p>Paying tribute to the historic speech, President Obama hand wrote an essay for an exhibit at the <a href="http://www.alplm.org/">Lincoln Presidential Library</a>. President Obama joins former Presidents Clinton, Carter and H.W. Bush who have submitted their contributions, along with other notable essayists including Justice Sandra Day O&rsquo;Connor, General Colin Powell, and Martin Luther King III, among hundreds of others.</p>

handlers/benchmark.rb

@@ -0,0 +1,46 @@
+require 'benchmark'
+require 'csv'
+
+require 'feedjira'
+
+# suppress warnings
+$VERBOSE = nil
+
+def files
+ @files ||= Dir.glob(File.dirname(__FILE__) + '/../feed_xml/*.xml')
+end
+
+def xmls
+ @xmls ||= files.map { |file| File.read file }
+end
+
+def parse_bench(&block)
+ 10.times { xmls.each &block }
+end
+
+benchmarks = Benchmark.bmbm do |b|
+ b.report('nokogiri') do
+ SAXMachine.handler = :nokogiri
+
+ parse_bench { |xml| Feedjira::Feed.parse xml }
+ end
+
+ b.report('ox') do
+ SAXMachine.handler = :ox
+
+ parse_bench { |xml| Feedjira::Feed.parse xml }
+ end
+
+ b.report('oga') do
+ SAXMachine.handler = :oga
+
+ parse_bench { |xml| Feedjira::Feed.parse xml }
+ end
+end
+
+data = benchmarks.map { |b| { b.label => b.real } }
+
+CSV.open('handlers/results.csv', 'w') do |c|
+ c << data.map(&:keys).flatten
+ c << data.map(&:values).flatten
+end

handlers/chart.r

@@ -0,0 +1,6 @@
+data <- read.csv("results.csv")
+y_max <- as.integer(max(data)) + 1
+png(file="handlers-benchmark.png",width=600,height=400)
+par(mar=c(4, 4, 4, 2))
+barplot(as.matrix(data), main="Comparison of sax-machine handlers", col="#0e8292", ylim=c(0, 3))
+dev.off()

handlers/run

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+results="$PWD/handlers/results.txt"
+
+rm -f $results
+
+log() {
+ echo "$1" 1>&2
+}
+
+bundle exec ruby handlers/benchmark.rb >> $results
+
+log "Done running the handlers benchmark."
+
+cd handlers
+
+R --vanilla < chart.r
+log "Done producing the handlers chart."