You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@nutch.apache.org by "Håvard W. Kongsgård" <h....@niap.no> on 2006/02/20 14:12:39 UTC
Pdf document title in nutch search
When searching with nutch the title of pdf documents is a url to the
file like:
http://www.ists.dartmouth.edu/library/wse0901.pdf
I have noticed that google and ultraseek creates a normal title like:
WebALPS: A Survey of E-Commerce Privacy and Security Applications
Is it possible to make nutch do the same?
Re: Pdf document title in nutch search
Posted by Jérôme Charron <je...@gmail.com>.
> Must I have index-more enabled to get the pdf titles to work.
> I did a test with some pdf files, all pdf titles were ignored (nutch 0.7.1
> ).
I used nutch-0.8-dev for my tests.
Are you sure that the PDF documents you are indexing contain a Title ?
Jérôme
--
http://motrech.free.fr/
http://www.frutch.org/
Re: Pdf document title in nutch search
Posted by "Håvard W. Kongsgård" <h....@niap.no>.
Must I have index-more enabled to get the pdf titles to work.
I did a test with some pdf files, all pdf titles were ignored (nutch 0.7.1).
Håvard W. Kongsgård wrote:
> It'd be nice if this was changed so that if a PDF has no title then
> the first xx words become the new title.
> (but it seems that the Google title process is more advanced that this)
>
>
>
> Jérôme Charron wrote:
>
>>> When searching with nutch the title of pdf documents is a url to the
>>> file like:
>>> http://www.ists.dartmouth.edu/library/wse0901.pdf
>>>
>>
>>
>> In Nutch, the title of PDF file is displayed if a title is available,
>> otherwise the URL
>> of the document is displayed.
>>
>> Regards
>>
>> Jérôme
>>
>> --
>> http://motrech.free.fr/
>> http://www.frutch.org/
>>
>>
>>
>
>
Re: Pdf document title in nutch search
Posted by Jérôme Charron <je...@gmail.com>.
> The pdf document (RAND_DB468-1.sum.pdf) has no pdf title,
This document has a pdf title (Near-Term Options for Improving Security at
Los Angeles International Airport).
You can get it by looking at the PDF file properties, and Google simply
display this title (like nutch).
But you are right, sometimes Google use the Title of the pdf documents (not
always the first words).
Jérôme
--
http://motrech.free.fr/
http://www.frutch.org/
Re: Pdf document title in nutch search
Posted by "Håvard W. Kongsgård" <h....@niap.no>.
Take a look at the Google search result of this rand publication
http://www.google.com/search?hs=z0n&hl=en&lr=&client=firefox-a&rls=org.mozilla%3Aen-US%3Aofficial&q=Implementing+Security+Improvement+Options+at+Los+Angeles+International+Airport+&btnG=Search
The pdf document (RAND_DB468-1.sum.pdf) has no pdf title, and google
don't use the first 2 pages of the document for a title!
Jérôme Charron wrote:
>>It'd be nice if this was changed so that if a PDF has no title then the
>>first xx words become the new title.
>>
>>
>
>I agree with that.
>Please create a JIRA issue for this point.
>
>
>
>
>>(but it seems that the Google title process is more advanced that this)
>>
>>
>
>Really?
>Take a look at this :
>http://www.google.com/search?num=100&hl=fr&safe=off&c2coff=1&as_qdr=all&q=http%3A%2F%2Fwww.trellix.com%2Fproducts%2Fdownloads%2Fsearchengines_siteopt.pdf++++&btnG=Rechercher&lr=
>In fact Google always take the first characters of the document as the
>title.
>Google never use the Title property of the document.
>So, when there is some shaded characters in the first characters of the pdf
>document, you get a "TTTiiitttllleee llliiikkkeee ttthhhaaattt" ... is it
>really an advanced title processing?
>
>Regards
>
>Jérôme
>
>--
>http://motrech.free.fr/
>http://www.frutch.org/
>
>
>
Re: Pdf document title in nutch search
Posted by Jérôme Charron <je...@gmail.com>.
> It'd be nice if this was changed so that if a PDF has no title then the
> first xx words become the new title.
I agree with that.
Please create a JIRA issue for this point.
> (but it seems that the Google title process is more advanced that this)
Really?
Take a look at this :
http://www.google.com/search?num=100&hl=fr&safe=off&c2coff=1&as_qdr=all&q=http%3A%2F%2Fwww.trellix.com%2Fproducts%2Fdownloads%2Fsearchengines_siteopt.pdf++++&btnG=Rechercher&lr=
In fact Google always take the first characters of the document as the
title.
Google never use the Title property of the document.
So, when there is some shaded characters in the first characters of the pdf
document, you get a "TTTiiitttllleee llliiikkkeee ttthhhaaattt" ... is it
really an advanced title processing?
Regards
Jérôme
--
http://motrech.free.fr/
http://www.frutch.org/
Re: Pdf document title in nutch search
Posted by "Håvard W. Kongsgård" <h....@niap.no>.
It'd be nice if this was changed so that if a PDF has no title then the
first xx words become the new title.
(but it seems that the Google title process is more advanced that this)
Jérôme Charron wrote:
>>When searching with nutch the title of pdf documents is a url to the
>>file like:
>>http://www.ists.dartmouth.edu/library/wse0901.pdf
>>
>>
>
>In Nutch, the title of PDF file is displayed if a title is available,
>otherwise the URL
>of the document is displayed.
>
>Regards
>
>Jérôme
>
>--
>http://motrech.free.fr/
>http://www.frutch.org/
>
>
>
Re: Pdf document title in nutch search
Posted by Jérôme Charron <je...@gmail.com>.
> When searching with nutch the title of pdf documents is a url to the
> file like:
> http://www.ists.dartmouth.edu/library/wse0901.pdf
In Nutch, the title of PDF file is displayed if a title is available,
otherwise the URL
of the document is displayed.
Regards
Jérôme
--
http://motrech.free.fr/
http://www.frutch.org/