handling translation errors

[Ashok]

Hi Luca,

How do i use this functionality ?

I dont see them being used in the unit tests ... Can you give me an
example ?

Ashok

[Luca Cervone]

Sorry Ashok, but I didn't understand the question ... 

Ciao 

Luca

Luca Cervone

Web and XML solutions designer

e-mail:     cervo...@gmail.com

  luca.c...@unibo.it

lcer...@cs.unibo.it

mobile phone:    0039 348 26 27 545

home   phone:  0039 051 199 82 854

skype:   cervoneluca

[Ashok Hariharan]

I mean how do I trap and present validation errors ?

There is a function call in translate() to :

SchemaValidator.getInstance().validate(fileToReturn,
this.akomantosoSchemaPath);

How do i use the errors it may return ?

[Luca Cervone]

Dear Ashok, 

Apologies for the delay. 

Sorry but I still can't understand what you are meaning. 

The functionality that you are talking about is built in in the translator mechanism. 

It throws an exception every time the validator founds an error.

The exception contains the line of the error. If you remember we had talk about the mechanism to link the error to the line of the original ODF document (that one that the user sees on the screen).

I have prepared the mechanism but I need the file that I requested to you. 

Let me know it is clearer. 

Ciao

Luca

[Ashok Hariharan]

Hi Luca:

Please find attached a small script that outputs the file as you had specified :

section_id: first line of text from section

it works like this :

java -jar ExtractSections.jar <input odf file with sections> <path to
output text file with section info>

there is some sample output in the gzip file ... in out.txt ...

Is that the input your system is expecting ?

[Luca Cervone]

Ashok, 

Probably it is exactly what I need.

Let me test it and then I will reply to you. 

Ciao

Luca

<xtractsectioninfo.tar.gz>

[Ashok Hariharan]

Hello Luca,

Were you able to test this ?

thanks

Ashok

[Luca Cervone]

Dear Ashok, 

I have an error with the extractor

The launched exception is this

Exception in thread "main" java.lang.NoClassDefFoundError: org/apache/log4j/Logger

at XtractSections.<clinit>(XtractSections.java:25)

Caused by: java.lang.ClassNotFoundException: org.apache.log4j.Logger

at java.net.URLClassLoader$1.run(URLClassLoader.java:200)

at java.security.AccessController.doPrivileged(Native Method)

at java.net.URLClassLoader.findClass(URLClassLoader.java:188)

at java.lang.ClassLoader.loadClass(ClassLoader.java:316)

at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:288)

at java.lang.ClassLoader.loadClass(ClassLoader.java:251)

at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:374)

... 1 more

Can You fix it? 

However I seen the sample output that you sent to me. It looks like this: 

debaterecord:NATIONAL ASSEMBLYOFFICIAL REPORTThursday, 9th December, 2004The House met at 2.3

masthead:NATIONAL ASSEMBLYOFFICIAL REPORTThursday, 9th December, 2004The House met at 2.3

comment1:[Mr. Deputy Speaker in the Chair] 

comment2:PRAYERS

But in the part after the ":" I should have the number of the line in witch the section start. 

But for me is the same If you want we can also display the first words of the section 

Ciao

Luca

 

<xtractsectioninfo.tar.gz>

[Ashok Hariharan]

On Tue, May 5, 2009 at 1:51 PM, Luca Cervone <cervo...@gmail.com> wrote:

Dear Ashok, 

I have an error with the extractor

I just tried on mac os x and it works fine for me. Are you sure extracted also the 'lib' folder  ?

the application looks for log4j in the ./lib folder ... i.e. if the jar is in a folder called 'xtract' the lib files must be in a folder called 'lib' within 'xtract'

 

The launched exception is this

Exception in thread "main" java.lang.NoClassDefFoundError: org/apache/log4j/Logger

at XtractSections.<clinit>(XtractSections.java:25)

Caused by: java.lang.ClassNotFoundException: org.apache.log4j.Logger

at java.net.URLClassLoader$1.run(URLClassLoader.java:200)

at java.security.AccessController.doPrivileged(Native Method)

at java.net.URLClassLoader.findClass(URLClassLoader.java:188)

at java.lang.ClassLoader.loadClass(ClassLoader.java:316)

at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:288)

at java.lang.ClassLoader.loadClass(ClassLoader.java:251)

at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:374)

... 1 more

Can You fix it? 

However I seen the sample output that you sent to me. It looks like this: 

debaterecord:NATIONAL ASSEMBLYOFFICIAL REPORTThursday, 9th December, 2004The House met at 2.3

masthead:NATIONAL ASSEMBLYOFFICIAL REPORTThursday, 9th December, 2004The House met at 2.3

comment1:[Mr. Deputy Speaker in the Chair] 

comment2:PRAYERS

But in the part after the ":" I should have the number of the line in witch the section start. 

But for me is the same If you want we can also display the first words of the section 

Line number is really a 'virtual concept' for the document since it depends on the application rendering the  file...or the orientation with which the file is being viewed (landscape or portrait)...

Ashok

--
++++ Ashok Hariharan ++++

[Luca Cervone]

Dear Ashok, 

The Extractor now works well. 

I'm setting the ODTTranslator in order to map the error. 

So, when an error occurs we will have something like: "some data are missing in the section starting with ... " (of course the message will be more specific). 

But remember that the input that you have to give to the translator will be the ODT file and the section file. 

So what do you think is the better solution? 

1) you give in input the ODT and the file produced to by the Extractor

2) I use the Extractor from my side and you give to the translator only the ODT file (in this case you have to give me the source code or the documentation of the extractor)

Let me know. 

Ciao

Luca

[Ashok Hariharan]

Hello Luca,

See attached the source of the extractor ---

On the issue of line-number (since we cannot provide a line number...) ... would paragraph number  be more relevant ?

How exactly does the error reporter work ? ... Does it identify the text fragment which has the error ... or does it simply report that the section has an error ... ? I am asking becuase you could have a section spanning 2-3 pages... so simply reporting an error at the section level may be inadequate ...

Ashok