How to?

[Aparna Lalingkar]

Hi Aki,

Thanks for the quick reply. I disabled the firewall network protection and tried to run the API. Still the same error is appearing.

Is this problem due to accessing one system through other system which is on different network?

What would be the solution to resolve this problem?

thanks and regards,

aparna

[akivela]

OpenCalais' endpoint address is

http://api.opencalais.com/enlighten/

Try to access the address with your web browser. If your browser shows something reasonable instead of Server not found message, then you should be able to access the OpenCalais web service. If you get an error, then your network connection is somehow not allowing you to access the OpenCalais.

Same applies to SemanticHacker. Their service endpoint address is

http://api.semantichacker.com/

and returns No Such Service message if accessed with web browser.

Kind Regards,
Aki / Wandora Team

[Aparna Lalingkar]

I tried to access http://api.opencalais.com/enlighten/ through firefox but it is displaying the following message "enlighten/

Hi there, this is an AXIS service!
Perhaps there will be a form for invoking the service here..."

Same thing about SemanticHacker It is displaying the following message.

"This XML file does not appear to have any style information associated with it. The document tree is shown below.

−
<response>
<message code="200" string="No Such Service"/>
</response>"

What would be the possible problem?



regards,

aparna

[akivela]

Your responses from actual service endpoints look right and both services are accessible. For some reason Wandora (Java) is unable to access those endpoints. I still tend to believe that the problem has something to do with your firewall or network configuration.

You could try to exit Wandora, close your firewall, start Wandora application, and try extractions again. Would that help? Also, I suggest that you try to classify some other text, if the actual problem is in the text format or something similar.

Unfortunately I can't figure out any other solutions with the information I have.

Kind Regards,
Aki / Wandora Team

[Aparna Lalingkar]

Hi Aki,

I followed your suggestions but still it is not working. Even I tried to extract a simple txt file. I don't understand what is wrong!!

regards,

aparna

[Aparna Lalingkar]

Hi Aki,

For my research, I want to create TM for a given document automatically. The document can be in Web page format and/ html and/ word doc and/ a pdf and/ jpeg and/video clip. Can I use Wandora for that? How can I use it in a simple way?

When I followed the documentation still it is not clear how to do it though it claims that one can extract a Topic Map out of the above category files.

Can you clarify and guide?

thanks,

aparna

[akivela]

Hello Aparna

I am a bit amazed you didn't get the extractor working. If you have another machine, maybe on another network, you might give it a try.

About automatic document extraction. Unfortunately all extractions in Wandora require manual user actions at the moment. User has to point out which extractor is used and which document is extracted. Most automatic extraction user reaches using Wandora's Drag and Drop Extractor described at [1].

PDF extraction is similar to simple text file extraction where text in PDF document is stored as an occurrence to a document topic. JPG metadata extractor reads metadata out of JPG images and constructs occurrences for the metadata fields. Moreover, Wandora doesn't support image-data (looking at image or video pixels) extraction at all at the moment. In general extractors can't read text out of MS doc files. However, dropping MS doc file into a Raw text area converts doc to text and you can then continue extraction with raw text.

Kind Regards,
Aki / Wandora Team

[1] http://www.wandora.org/wandora/wiki/ind ... _extractor

-----Edit-----

Oh, I nearly forgot. If you need to extract WWW pages, you might also try Wandora's Firefox/Thunderbird plugin allowing extractions directly in Firefox WWW browser (or Thunderbird email client). See details at

http://www.wandora.org/wandora/wiki/ind ... fox_plugin

[Aparna Lalingkar]

Many thanks Aki for your detailed clarification and explaination. I am going to try it from my home network through my laptop machine. I guess, in that maschine extractor will work.

For future work, (just asking) I guess, it will be possible to write Java programs to call these extractors as per the input data and make the process completely automatic. what do u say?

Also, nowadays in the US there is a law (which eventually get spread all over the world) that if you are uploading any video or image on the Web then you have to annotate English translation or transcript or metadata about that video or image. So, even in that case it will be possible to extract topic map. what do you say?

Nice interacting with you and your team. Will be in touch by time to time.

thanks and regards,

aparna

[akivela]

Ok, please report your extraction success or problems at your home laptop.

Yes, it is possible to write Java application based on Wandora that would do the extraction automatically. Or it might be possible to write Wandora an extension that would do user specific extractions automatically. Although the idea is very strong, we have no plans for such a Wandora feature at the moment.

Oh, I haven't heard about the US law of obligatory metadata. Do you have any references? I would like to read more about it. It's hard to say what happens. People are rather lazy adding metadata voluntarily if they don't see obvious benefits. It's also rather difficult to monitor. However, if all images and videos would contain useful metadata, it would be a glorious moment for applications like Wandora.

Kind Regards,
Aki / Wandora Team

[akivela]

Hi Vondess

You'll find documentation of Wandora here: http://wandora.org/wiki/Documentation

The Documentation of extractors is little outdated. Many api-end-points Wandora uses, have closed or changed radically.

Kind Regards,
Aki