cfedermann / Appraise

Appraise evaluation system for manual evaluation of machine translation output
http://www.appraise.cf/
BSD 3-Clause "New" or "Revised" License
73 stars 37 forks source link

XML building script, new feature #18

Closed mjpost closed 12 years ago

mjpost commented 12 years ago

I wrote a script to build the XML file. It is self-documenting.

I also added a feature which I hope you'll accept. When displaying left and right context for the source-language sentence, the document IDs of the sentences are queried, and the context is not displayed if the documents are different. This allows you to do evaluations and not print the context if it doesn't make sense in an easy, general way.

cfedermann commented 12 years ago

Hi Matt,

I had already merged in your previous pull request, hence you might need to update build-xml.py as I renamed it to a (more Python-friendly) name of build_xml.py...

Sorry for that, thought you were already done with it :)

The context patch is cool and will be merged as soon as you have merged in the current master code. Thanks, by the way!

Cheers, Christian

Christian Federmann Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

On Tuesday, July 10, 2012 at 4:51 PM, Matt Post wrote:

I wrote a script to build the XML file. It is self-documenting.

I also added a feature which I hope you'll accept. When displaying left and right context for the source-language sentence, the document IDs of the sentences are queried, and the context is not displayed if the documents are different. This allows you to do evaluations and not print the context if it doesn't make sense in an easy, general way.

You can merge this Pull Request by running:

git pull https://github.com/mjpost/Appraise master

Or you can view, comment on it, or merge it online at:

https://github.com/cfedermann/Appraise/pull/18

-- Commit Summary --

  • added script to build XML files
  • feature: context is only displayed if the context sentences have the same doc-id

-- File Changes --

M appraise/evaluation/views.py (6) M scripts/build-xml.py (11)

-- Patch Links --

https://github.com/cfedermann/Appraise/pull/18.patch https://github.com/cfedermann/Appraise/pull/18.diff


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18

mjpost commented 12 years ago

Okay, I fixed the request.

mjpost commented 12 years ago

Great, I updated the pull request

Out of curiosity, why are underscores preferred to dashes in script names?

matt

On Jul 10, 2012, at 11:00 AM, Christian Federmann wrote:

Hi Matt,

I had already merged in your previous pull request, hence you might need to update build-xml.py as I renamed it to a (more Python-friendly) name of build_xml.py...

Sorry for that, thought you were already done with it :)

The context patch is cool and will be merged as soon as you have merged in the current master code. Thanks, by the way!

Cheers, Christian

Christian Federmann Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

On Tuesday, July 10, 2012 at 4:51 PM, Matt Post wrote:

I wrote a script to build the XML file. It is self-documenting.

I also added a feature which I hope you'll accept. When displaying left and right context for the source-language sentence, the document IDs of the sentences are queried, and the context is not displayed if the documents are different. This allows you to do evaluations and not print the context if it doesn't make sense in an easy, general way.

You can merge this Pull Request by running:

git pull https://github.com/mjpost/Appraise master

Or you can view, comment on it, or merge it online at:

https://github.com/cfedermann/Appraise/pull/18

-- Commit Summary --

  • added script to build XML files
  • feature: context is only displayed if the context sentences have the same doc-id

-- File Changes --

M appraise/evaluation/views.py (6) M scripts/build-xml.py (11)

-- Patch Links --

https://github.com/cfedermann/Appraise/pull/18.patch https://github.com/cfedermann/Appraise/pull/18.diff


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18#issuecomment-6878070

cfedermann commented 12 years ago

Because a script is a module so you could import from it; if the name contains a dash, it is not a valid Python name which would prevent import from a script...

It's all about PEP8, the Python style guide ;)

Christian Federmann Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

On Tuesday, July 10, 2012 at 5:11 PM, Matt Post wrote:

Great, I updated the pull request

Out of curiosity, why are underscores preferred to dashes in script names?

matt

On Jul 10, 2012, at 11:00 AM, Christian Federmann wrote:

Hi Matt,

I had already merged in your previous pull request, hence you might need to update build-xml.py as I renamed it to a (more Python-friendly) name of build_xml.py...

Sorry for that, thought you were already done with it :)

The context patch is cool and will be merged as soon as you have merged in the current master code. Thanks, by the way!

Cheers, Christian

Christian Federmann Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

On Tuesday, July 10, 2012 at 4:51 PM, Matt Post wrote:

I wrote a script to build the XML file. It is self-documenting.

I also added a feature which I hope you'll accept. When displaying left and right context for the source-language sentence, the document IDs of the sentences are queried, and the context is not displayed if the documents are different. This allows you to do evaluations and not print the context if it doesn't make sense in an easy, general way.

You can merge this Pull Request by running:

git pull https://github.com/mjpost/Appraise master

Or you can view, comment on it, or merge it online at:

https://github.com/cfedermann/Appraise/pull/18

-- Commit Summary --

  • added script to build XML files
  • feature: context is only displayed if the context sentences have the same doc-id

-- File Changes --

M appraise/evaluation/views.py (6) M scripts/build-xml.py (11)

-- Patch Links --

https://github.com/cfedermann/Appraise/pull/18.patch https://github.com/cfedermann/Appraise/pull/18.diff


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18#issuecomment-6878070


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18#issuecomment-6878390

mjpost commented 12 years ago

Ah, I see.

Are you interested in a pull that displays the sentences for the three-way task side-by-side? I found the vertical text-comparison a little awkward against the horizontal buttons.

matt

On Jul 10, 2012, at 11:48 AM, Christian Federmann wrote:

Because a script is a module so you could import from it; if the name contains a dash, it is not a valid Python name which would prevent import from a script...

It's all about PEP8, the Python style guide ;)

Christian Federmann Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

On Tuesday, July 10, 2012 at 5:11 PM, Matt Post wrote:

Great, I updated the pull request

Out of curiosity, why are underscores preferred to dashes in script names?

matt

On Jul 10, 2012, at 11:00 AM, Christian Federmann wrote:

Hi Matt,

I had already merged in your previous pull request, hence you might need to update build-xml.py as I renamed it to a (more Python-friendly) name of build_xml.py...

Sorry for that, thought you were already done with it :)

The context patch is cool and will be merged as soon as you have merged in the current master code. Thanks, by the way!

Cheers, Christian

Christian Federmann Sent with Sparrow (http://www.sparrowmailapp.com/?sig)

On Tuesday, July 10, 2012 at 4:51 PM, Matt Post wrote:

I wrote a script to build the XML file. It is self-documenting.

I also added a feature which I hope you'll accept. When displaying left and right context for the source-language sentence, the document IDs of the sentences are queried, and the context is not displayed if the documents are different. This allows you to do evaluations and not print the context if it doesn't make sense in an easy, general way.

You can merge this Pull Request by running:

git pull https://github.com/mjpost/Appraise master

Or you can view, comment on it, or merge it online at:

https://github.com/cfedermann/Appraise/pull/18

-- Commit Summary --

  • added script to build XML files
  • feature: context is only displayed if the context sentences have the same doc-id

-- File Changes --

M appraise/evaluation/views.py (6) M scripts/build-xml.py (11)

-- Patch Links --

https://github.com/cfedermann/Appraise/pull/18.patch https://github.com/cfedermann/Appraise/pull/18.diff


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18#issuecomment-6878070


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18#issuecomment-6878390


Reply to this email directly or view it on GitHub: https://github.com/cfedermann/Appraise/pull/18#issuecomment-6879471