[openbiblio-dev] model.base.local_path

[William Waites]

* [2011-03-17 14:12:12 +0000] Edmund Chamberlain <emc59 at cam.ac.uk> écrit:

] 
] >The URIs of our graphs not necessarily are their location on the
] >database, as I understand. For example I could import triples from other
] >bibliographica instances, and their URI will still be the original one,
] >but I want to browse them in my instance.
] >
] 
] Just a quick +1 for this, having finished a local install recently and 
] imported a graph from bibliographica, (as in the install instructions) 
] it links directly to the view on bibliograhpica, rather than the 
] imported record in a local system.

Hi Edmund, 

We have discussed this quite a bit in IRC and there seems to
be two strategies for doing this.

Firstly it is important to understand that the ordf(1) tool
is not part of openbiblio but part of the ordf library and
is analogous to tools like pg_dump and pg_restore - it just
takes data and puts it in or takes it out of the database.
It has no concept of openbiblio's models or representations,
nor should it.

So what you are doing is just copying data, when really what
we want to do is import data.

This can be accomplished with a specialised tool which might
take some time to write and this might be a bit premature 
at this stage.

The other way this can be accomplished is to give ordf another
flag that might work something like sed(1). The relevant
parameters could equally be specified in the config. What
you would do then is to import data as opposed to copying
it is run something like,

    ordf -c config.ini -s --nocs -i http://example.com/entry/abc123 \
        -e s@^http://.*bibliographica.org/@http://example.com/@ \
        http://bibliographica.org/entry/abc123

If this seems a little low-level an approach you're right. 
But I think we can get 90% of the way to what we want very
quickly this way whereas a specialised tool would be harder
and take longer.

The important thing to remember is that the web interface
represents the data in the store and the URIs are integral
parts of that data. Because of this there is a difference
between copying the data and importing it into your own
namespace.

Make sense?

Cheers,
-w

-- 
William Waites                <mailto:ww at styx.org>
http://river.styx.org/ww/        <sip:ww at styx.org>
F4B3 39BF E775 CF42 0BAB  3DF0 BE40 A6DF B06F FD45