[open-bibliography] getting a personal bib library out

[Mark MacGillivray]

On Wed, Jan 18, 2012 at 10:54 PM, Peter Murray-Rust <pm286 at cam.ac.uk> wrote:
>
>
> On Wed, Jan 18, 2012 at 8:18 PM, Jim Pitman <pitman at stat.berkeley.edu>
> wrote:
>>
>> Great, thanks for pointer!
>> Then here is the dataset on my Berkeley BibServer
>>
>>
>> http://bibserver.berkeley.edu/cgi-bin/bibs7?https://raw.github.com/langner/library/master/library.small.bib
>>
>> and here it is on the BibSoup BibServer
>>
>> http://bibsoup.net/pitman/karol_m_langer_s_bibliography

Thanks Jim, we had intermittent connection at the time so could not
push it remotely, and by the time that was solved you had already
uploaded.


> Brilliant everyone! we have a technology which seems to be deployable in
> very short time.

Yes.


>> It is a deficiency of the present BibSoup BibServer compared to the legacy
>> BibServer that there
>> is no obvious way to query BibSoup to see if it already has a copy of
>> https://raw.github.com/langner/library/master/library.small.bib
>> and if not to ingest one.
>> e.g. searching BibSoup for the source url returns nothing.
>> Mark, can you see how to remedy this?
>> It may be as simple as putting the source url in the right search field, I
>> just dont see how.

You can view all the collections at http://bibsoup.net/collections,
and in the search bar there you can put in
"http://bibserver.berkeley.edu/tmp/testbib.bib"
to return any collection that was created from that source. There are
search options on the main search pages where you can select exact
match (equivalent to adding the quotes), but I have not put them on
the collections page search yet. Will add on the next update.


> I suspect this is solved by convention-over-configuration. We can recommend
> that Bibsoupers use URLs and URIs in specified ways and if they do I suspect
> this is soluble. It's the clash between addresses and names. We probably
> need a Bibsoup URI Mint
>
>>
>> It is a serious problem for any user who wishes to programmatically upload
>> many bibliographies to
>> BibSoup that there is no obvious way to retrieve them after programmatic
>> upload without user
>> retaining a dictionary to e.g. map
>> https://raw.github.com/langner/library/master/library.small.bib to
>> karol_m_langer_s_bibliography

We did originally allow the collection ID to be created from the
source URL if a collection ID were not provided, but this created long
and unwieldy collection IDs so we opted for requiring one is provided.
If you are doing programmatic upload, you must have to define the
collection IDs in advance anyway, so do you not already know what they
are for subsequent queries? If you are meaning the case where you
upload a collection from a particular URL a second time, then you do
not actually have to worry about matching the collection IDs - any
collection that you previously created will be overwritten if it has
the same source URL, and the collection ID will be altered to match
whichever new collection ID you provide (or will be the same if you
use the same collection ID).

As Peter says, I am really glad that overall this is working well for
people, and that we can create these collections rapidly. The work
this week on updating the websites and issue lists is going well, so
we are in a good position to do some final tweaking and put up more
examples.

Mark


>>
> Do you mean find the Bibservers or retrieve their content? I fully agree
> that a Bibserver needs something like .dump()
>
>>
>> many thanks
>>
>>
>>
>> _______________________________________________
>> open-bibliography mailing list
>> open-bibliography at lists.okfn.org
>> http://lists.okfn.org/mailman/listinfo/open-bibliography
>
>
>
>
> --
> Peter Murray-Rust
> Reader in Molecular Informatics
> Unilever Centre, Dep. Of Chemistry
> University of Cambridge
> CB2 1EW, UK
> +44-1223-763069
>
> _______________________________________________
> open-bibliography mailing list
> open-bibliography at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/open-bibliography
>