Re: Searching inside of attachments (pdf, odt)?

["Tim O'Callaghan" <timo@dspsrv.com>]

FWIW

I think this might be handled easier if all that happened would be a
grep on the attachments, or directories.

The usual grep interface can be used and then it becomes a fast
general purpose data mining extension.

I can see it being used to search a codebase or website for a text string.

I guess it could be further refined with some kind of dispatcher -
like the file dispatcher that invokes a specific tool to view an
attachment, except it uses an attachment specific search or defaults
to grep if its not an emacs editable file.

Possibly an extension fo the current "file:<text-file>::<in buffer
search >", but uses this grep or whatever if it comes up against
something un-emacs-editable.

An added bonus of a search dispatcher type approach: it would give
users the chance to extend the search into whatever tool(s)/file
format(s) they are using without having to become core to org.

Just my 2eurocents worth:

Tim.
2009/10/13 Karl Maihofer <ignoramus@gmx.de>:
> Hi Samuel,
>
> Samuel Wales <samologist@gmail.com> schrieb:
>>
>> My idea is to use ordinary agenda search like this:
>>  1) agenda search displays the headline that has the
>>     attachment.
>>  2) org uses an alist to determine the correct textifier
>>     according to extension.  e.g. '((".pdf" . "pdf2text")).
>>  3) agenda searches normally (as if the contents of the
>>     attachment were body text).
>
> correct me if i'm wrong, but your approach is to search inside (an)
> already identified attachment(s)?
>
> I'd like to find attachments by searching inside the whole set of
> attachments. I do have many articles (pdf-files) to deal with. When i
> write a report on a special topic i have to find articles that are
> relevant to the topic i'm working on at the moment.
>
> If we use the standard textifiers the procedure will probably get very
> slow if there are many attachments. I think using an index would be a
> good idea.
>
> To describe what i'm looking for:
> My first step is to create an entry for each article, define tags
> (describing the content) and add some notes.
>
> * Title of the article                           :tag:tag:tag:
>  :PROPERTIES:
>  :Attachments: article.pdf
>  :ID: 387HJGJD78-758GZFHF87-JKHKJ57dfd9
>  :END:
>  - Very good explanation of X.
>  - New view on Y.
>
> But it would be much more powerful to be able not only to find an
> entry by searching for tags but to search inside the attachments.
>
> I'm not a programmer, so sorry if my ideas are stupid. ;-) But i thing
> the following questions have to be answered:
>
> 1) Is there a tool like Lucene that can index pdf-files as they are
>   stored by orgmode (directory structure)?
> 2) Is it possible to send a query to this tool from within emacs?
> 3) Is it possible to "import" the answer of the tool into emacs and
>   combine it with orgmode so that the result looks somehow like this:
>   "Search string 'XX' found in file 'article.pdf' attached to task
>   'Title of the article'". A click on the name of the attachment
>   should open the pdf-file in the pdf-reader; a click on the task
>   name should show the task in the org-buffer.
>
> Karl
>
>
>
>
>
>
> _______________________________________________
> Emacs-orgmode mailing list
> Remember: use `Reply All' to send replies to the list.
> Emacs-orgmode@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-orgmode
>