Hi Craig,

Trying nslookup on our hosted contentDM server [1] I come up with cdml02.contentdm.oclc.org, which gives this error when attempting to access the API: "The web server is not configured to handle requests for the virtual host named cdml02.contentdm.oclc.org".

One thing I see on our contentDM homepage is that the "Skip to main content" link contains the numerical part of the server ID as part of the anchor for the first collection in the list, in our case href="#collection_link_p16232coll18". This should work most of the time, although because we've had contentDM for so long one of our collections actually has a different number as part of its collection ID.

Hope this helps and kudos on Tensor, what looks to be quite a useful application!

[1] http://wcudigitalcollection.contentdm.oclc.org/cdm/

Joel Marchesoni
Tech Support Analyst
Hunter Library at Western Carolina University
828-227-2860
http://www.wcu.edu/hunter-library/



-----Original Message-----
From: Code for Libraries [mailto:[log in to unmask]] On Behalf Of Kyle Banerjee
Sent: Tuesday, July 3, 2018 18:30
To: [log in to unmask]
Subject: Re: [CODE4LIB] Trustworthy way to get server ID# from CONTENTdm

Hi Craig,

DISCLAIMER: I don't work with CONTENTdm, so everything I'm telling you is being pulled out of my hat.

Eyeballing your specific examples, it appears that the cdm-applnxprd1.library.ubc.ca is locally hosted -- in which case it wouldn't follow the pattern you seek. As such, it appears the most reliable way to do it would be to do a regex match that treats OCLC and locally hosted installations differently.

Having said that, in your examples, the first cname entry appears to be what you need. If that pattern holds, something along the lines of:

nslookup digitalcollections.library.ubc.ca |tr '\n' ' ' | grep -o "canonical name = [^ ]\+" |head -1 |sed 's/canonical name = //'|sed 's/\( \|\.$\)//'

will grab the first canonical name, though I'm sure there's a more elegant way to do it. HTH

kyle


On Tue, Jul 3, 2018 at 12:56 PM, Craig Dietrich <[log in to unmask]>
wrote:

> Hi Kyle,
>
> Thanks for your quick reply!
>
> That worked!  When I run a nslookup for 
> "digitalcollections.library.ubc.ca
> "
> it returns "cdm-applnxprd1.library.ubc.ca" that, while it doesn't 
> follow the "server12345" pattern, does return results when I append API fields:
>
> http://cdm-applnxprd1.library.ubc.ca/dmwebservices/index.
> php?q=dmQuery/all/title^%20Victoria%20Cruz%20and%20Maxwell%20Maxwell%2
> 0at% 20the%20CiTR%20booth%20at%20Clubs%20Days^all^and/title!
> subjec!descri!thumb/nosort/10/1/1/0/0/0/xml
>
> Like all things CONTENTdm I'll need to code around a few things, for 
> example, doing a nslookup for "digitalcollections.lib.washington.edu"
> returns a couple URLs including "cdml02.contentdm.oclc.org" which 
> doesn't return results when similarly adding API fields.  However, for 
> the Washington site I'm able to get a working URL from their HTML.  So 
> I suppose a combination of those things will work sufficiently.  
> Thanks again!
>
> Re the why:  I'm a Scalar co-creator and have spun Scalar's import 
> system (whereby authors can import content into their projects from 
> various digital archives) into its own application, Tensor.  
> Architected from the ground up Tensor can do many more things than 
> Scalar's antiquated import system can do, including linking to 
> archives that have complex or multi-dimensional APIs.  All that's 
> needed is a parser file then one can use Tensor to move content in 
> Scalar, and actually my goal is for Tensor to not only hook into many 
> archives but also many publishing systems, not just Scalar.
>
> Here is a list of the parsers I've already created:
> https://github.com/craigdietrich/tensor-profiles/tree/master/parsers
>
> Once a parser is created, any archive that uses that platform can be 
> accessed.
>
> I've attached a couple screengrabs, not sure if they'll come through 
> the list server but provides a general sense of Tensor, hopefully.
>
> All the best,
> Craig
>
> On Tue, Jul 3, 2018 at 10:48 AM Kyle Banerjee 
> <[log in to unmask]>
> wrote:
>
> > Hi Craig,
> >
> > Have you tried using nslookup? That should return the CNAME entry 
> > containing what you seek.
> >
> > Also, curiosity is killing me as to why knowing server numbers for
> machines
> > you don't control would be useful.
> >
> > kyle
> >
> > On Tue, Jul 3, 2018 at 10:17 AM, Craig Dietrich 
> > <[log in to unmask]
> >
> > wrote:
> >
> > > Hi all,
> > >
> > > This has been driving me crazy and OCLC hasn't written me back: is
> there
> > a
> > > trustworthy way to get the server ID# from CONTENTdm?
> > >
> > > I've been relying on
> > >
> > > '//'.$hostname.'/utils/diagnostics'
> > >
> > > ... which under some cases returns this string:
> > >
> > > server12345.contentdm.oclc.org
> > >
> > > ... with 12345 in this case being the server ID#.
> > >
> > > But for many CONTENTdm sites this string isn't present.
> > >
> > > The home page HTML for CONTENTdm has a lot of useful info in it, 
> > > but
> not
> > > the ID# that I can find.
> > >
> > > Anyone had any luck?
> > >
> > > Thanks!
> > > Craig
> > >
> >
>