Fixtures need to be updated

[yorickpeterse]

When running the test suite the input/ouput is different and it looks like the fixtures have to be updated. Trimmed output of the tests:

      Diff:
      @@ -6,7 +6,7 @@
             <lp timestamp="2013-04-11T11:11:29CEST" version="1.0" name="Open-nlp tokenizer"/>
           </linguisticProcessors>
           <linguisticProcessor layer="opinions">
      -      <lp timestamp="*" version="1.0" name="Crfsuite machine learning opinion miner"/>
      +      <lp timestamp="2013-11-13T16:32:06CET" version="8nov2013_1.1" name="VUA opinion miner. CRF deluxe"/>
           </linguisticProcessor>
           <linguisticProcessor layer="properties">
             <lp timestamp="2013-04-11T11:12:43CEST" version="1.0" name="Lite PropRecognizer_lookup"/>
      @@ -107,7 +107,7 @@
           </property>
         </properties>
         <opinions>
      -    <opinion oid="o1">
      +    <opinion oid="o_1">
             <opinion_holder>
               <span/>
             </opinion_holder>
      @@ -123,7 +123,7 @@
               </span>
             </opinion_expression>
           </opinion>
      -    <opinion oid="o2">
      +    <opinion oid="o_2">
             <opinion_holder>
               <span/>
             </opinion_holder>
       (RSpec::Expectations::ExpectationNotMetError)
      ./features/step_definitions/kernel_steps.rb:23:in `/^the output should match the fixture "(.*?)"$/'
      features/file_input_and_output.feature:9:in `Then the output should match the fixture "<output_file>"'

Failing Scenarios:
cucumber features/file_input_and_output.feature:6 # Scenario: Extract opinions from KAF
cucumber features/file_input_and_output.feature:6 # Scenario: Extract opinions from KAF
cucumber features/file_input_and_output.feature:6 # Scenario: Extract opinions from KAF

[rubenIzquierdo]

Done

Ruben Izquierdo Bevia Vrije University of Amsterdam ruben.izquierdobevia@vu.nlmailto:ruben.izquierdobevia@vu.nl

On 13 Nov 2013, at 16:33, Yorick Peterse notifications@github.com<mailto:notifications@github.com> wrote:

When running the test suite the input/ouput is different and it looks like the fixtures have to be updated. Trimmed output of the tests:

  Diff:
  @@ -6,7 +6,7 @@
         <lp timestamp="2013-04-11T11:11:29CEST" version="1.0" name="Open-nlp tokenizer"/>
       </linguisticProcessors>
       <linguisticProcessor layer="opinions">
  -      <lp timestamp="*" version="1.0" name="Crfsuite machine learning opinion miner"/>
  +      <lp timestamp="2013-11-13T16:32:06CET" version="8nov2013_1.1" name="VUA opinion miner. CRF deluxe"/>
       </linguisticProcessor>
       <linguisticProcessor layer="properties">
         <lp timestamp="2013-04-11T11:12:43CEST" version="1.0" name="Lite PropRecognizer_lookup"/>
  @@ -107,7 +107,7 @@
       </property>
     </properties>
     <opinions>
  -    <opinion oid="o1">
  +    <opinion oid="o_1">
         <opinion_holder>
           <span/>
         </opinion_holder>
  @@ -123,7 +123,7 @@
           </span>
         </opinion_expression>
       </opinion>
  -    <opinion oid="o2">
  +    <opinion oid="o_2">
         <opinion_holder>
           <span/>
         </opinion_holder>
   (RSpec::Expectations::ExpectationNotMetError)
  ./features/step_definitions/kernel_steps.rb:23:in `/^the output should match the fixture "(.*?)"$/'
  features/file_input_and_output.feature:9:in `Then the output should match the fixture "<output_file>"'

Failing Scenarios: cucumber features/file_input_and_output.feature:6 # Scenario: Extract opinions from KAF cucumber features/file_input_and_output.feature:6 # Scenario: Extract opinions from KAF cucumber features/file_input_and_output.feature:6 # Scenario: Extract opinions from KAF

— Reply to this email directly or view it on GitHubhttps://github.com/opener-project/opinion-detector-base/issues/1.

[yorickpeterse]

After updating the feature helpers to properly specify that no timestamps should be generated there are still some differences in the various files. Can you check those too? You can run rake test to verify the results.

[rubenIzquierdo]

rake test does nothing in my case. After 30 seconds of compilation it ends in:

rm -f Makefile Making distclean in . rm -rf .libs _libs rm -f *.lo test -z "genbinary.sh" || rm -f genbinary.sh test . = "." || test -z "" || rm -f rm -f config.h stamp-h1 rm -f libtool config.lt rm -f TAGS ID GTAGS GRTAGS GSYMS GPATH tags rm -f config.status config.cache config.log configure.lineno config.status.lineno rm -f Makefile Checking for python... yes Checking for python >= 2.7.0... no (python2.6)pruben:opinion-detector-base ruben$

if you let me know what are the differences maybe I can take a look, I wouldn’t like to waste too much time with ruby stuff

Ruben Izquierdo Bevia Vrije University of Amsterdam ruben.izquierdobevia@vu.nlmailto:ruben.izquierdobevia@vu.nl

On 13 Nov 2013, at 17:20, Yorick Peterse notifications@github.com<mailto:notifications@github.com> wrote:

After updating the feature helpers to properly specify that no timestamps should be generated there are still some differences in the various files. Can you check those too? You can run rake test to verify the results.

— Reply to this email directly or view it on GitHubhttps://github.com/opener-project/opinion-detector-base/issues/1#issuecomment-28408613.

[yorickpeterse]

We require Python 2.7 as a minimum everywhere so you'll need that to run the entire setup (note the "Checking for python ..." bit).

[rubenIzquierdo]

Even if I don’t get why you are forcing to use python2.7, but I’m sure there will be a good reason, I switched back to python2.7 with virtualenv. I couldn’t install lxml because of some header files missing —> had to reinstall xcodecs and recompile everything again and finally it worked out. (This took me 40 minutes.)

Then, I ran the “rake test” and...

Examples:
  | language | input_file     | output_file        |
  | English  | input.1.kaf    | input.1.out.kaf    |
  | Dutch    | input.1.nl.kaf | input.1.nl.out.kaf |
  | English  | input.2.kaf    | input.2.out.kaf    |

3 scenarios (3 passed) 9 steps (9 passed) 0m0.330s

So I’m not sure about what do you mean in your last mail with:

"After updating the feature helpers to properly specify that no timestamps should be generated there are still some differences in the various files.”

Ruben Izquierdo Bevia Vrije University of Amsterdam ruben.izquierdobevia@vu.nlmailto:ruben.izquierdobevia@vu.nl

On 13 Nov 2013, at 17:34, Yorick Peterse notifications@github.com<mailto:notifications@github.com> wrote:

We require Python 2.7 as a minimum everywhere so you'll need that to run the entire setup.

— Reply to this email directly or view it on GitHubhttps://github.com/opener-project/opinion-detector-base/issues/1#issuecomment-28409884.

[yorickpeterse]

Even if I don’t get why you are forcing to use python2.7, but I’m sure there will be a good reason

If I recall there were some components that required Python 2.7 so we settled with that. Python 2.6 has also reached its end of life.

Then, I ran the “rake test” and...

I did a clean-up of my local repository (removing compiled files, etc) and that seems to've done the trick for the tests, they now pass again. I guess there were some shady conflicts somewhere.

Thanks!