Re: [Tracker] chm mime type
- From: Martyn Russell <martyn lanedo com>
- To: Rustom Mody <rustompmody gmail com>
- Cc: tracker-list gnome org
- Subject: Re: [Tracker] chm mime type
- Date: Mon, 21 Mar 2011 17:41:58 +0000
On 21/03/11 16:55, Rustom Mody wrote:
On Mon, Mar 21, 2011 at 3:30 PM, Martyn Russell <martyn lanedo com
<mailto:martyn lanedo com>> wrote:
On 21/03/11 08:26, ivan frade nokia com
<mailto:ivan frade nokia com> wrote:
Hi Rusi,
Hi,
Fairly easy. You need to implement a extractor module for that
mimetype. It means to implement a couple of mandatory functions,
build it and install it in the right location.
<<snipped>>
Great advise from Ivan there. I just wanted to warn you, these APIs
and examples have changed in master recently. So depending on where
you're targeting, you might want to look at the source.
Thanks both for the answers.
I had a first look and see that it needs to be in glib/gdk etc. I must
admit this will be a bit of a learning curve.
On the other hand what needs to be done is (seems) quite trivial
1. Run extract_chmLib on the chm file (opening into /tmp/newdirname )
2. Run the tracker on the dir
3. Integrate the results of 2 back into the database
1 and 2 seem trivial and would likely be literally 1-liners in something
like python
Make use of other extractors, they use functions available to make
things easier (like tracker_open()). Note the APIs in libtracker-extract
too.
About 3 I dont know...
There are a few things you can do here.
1. Look at the ontology to see what properties you can export and what's
not supported. The tracker-sparql command comes in handy for this, e.g.:
$ tracker-sparql -s document
Classes:
http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#Document
http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#TextDocument
http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#PlainTextDocument
http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#HtmlDocument
http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#PaginatedTextDocument
http://www.semanticdesktop.org/ontologies/2007/03/22/nfo#HelpDocument
Properties:
None
--
You can use tracker-sparql to get properties and other information about
classes to know what to use.
Anyhow I think its still too premature for me to talk before I can find
my way around.
For example I have a man page for tracker-extract but
$ tracker-extract
gives command not found
Commands instantiated by d-bus are installed into $prefix/libexec, this
includes tracker-store, tracker-miner-fs, tracker-miner-rss,
tracker-miner-flickr, tracker-writeback and of course tracker-extract.
--
Regards,
Martyn
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]