Re: [CODE4LIB] Zotero, unapi, and formats?

[Owen Stephens]

At the moment Zotero development seem to be focussing on the use of RDFa
using the Bibo ontology for picking up bib details from within pages (see
discussion on the Bibo Google group)

Owen

On Tue, Apr 6, 2010 at 4:17 PM, Chad Fennell  wrote:

> > It's still a LOT better than COinS for Zotero, I assume though.
>
> Yes, if only because you get more complete metadata with things like
> RIS than COinS does via OpenURL.  I do like the theoretical benefit of
> a metadata format request API , but the promise of richer metadata
> (primarily for Zotero) was ultimately why I chose unAPI over COinS.
> And yeah, better documentation would be nice, thanks for looking into
> it.
>
> -Chad
>



-- 
Owen Stephens
Owen Stephens Consulting
Web: http://www.ostephens.com
Email: o...@ostephens.com

Re: [CODE4LIB] Zotero, unapi, and formats?

[Chad Fennell]

> It's still a LOT better than COinS for Zotero, I assume though.

Yes, if only because you get more complete metadata with things like
RIS than COinS does via OpenURL.  I do like the theoretical benefit of
a metadata format request API , but the promise of richer metadata
(primarily for Zotero) was ultimately why I chose unAPI over COinS.
And yeah, better documentation would be nice, thanks for looking into
it.

-Chad

Re: [CODE4LIB] Zotero, unapi, and formats?

[Jonathan Rochkind]

It's still a LOT better than COinS for Zotero, I assume though. I'd like 
there to be better documentation to encourage more people to use unAPI 
instead of COinS; even the limited Zotero documentation there is admits 
that COinS is pretty limited.


Jonathan

Bill Dueber wrote:

The unAPI support is also...non-ideal...in that you can't present
preferences for the best format to use. For example, the Refworks
Tagged format just plain has more tags (and hence more or
more-finely-grained information) than other formats (e.g., Endnote),
but Zotero will prefer Endnote just because it does. My RIS output is
better than my endnote output, but there's no way for me to tell
Zotero that.  For Mirlyn I ended up just having exactly one format
listed in my unapi-server file. Which is dumb. But I'm not sure what
else to do.

On Tue, Apr 6, 2010 at 10:16 AM, Jonathan Rochkind  wrote:
  

Yeah, we need some actual documentation on Zotero's use of unAPI in general.
Maybe if I can figure it out (perhaps by asking the developer(s)) I'll write
some for them.

Robert Forkel wrote:


well, looks like a combination:
in case of mods it checks for the namespace URL, in case of rdf, it
looks for a format name of rdf_dc, ...
and yes, endnote export would have to have a name of "endnote" (i ran
into this problem as well with names like endnote-utf-8, ...). i think
unapi would be more usable if there were at least a recommendation of
common format names.

On Tue, Apr 6, 2010 at 4:07 PM, Jonathan Rochkind 
wrote:

  

Wait, does it actually recognize the format by the format _name_ used,
and
not by a mime content-type?  Like unless my unAPI server calls the
endnote
format "endnote", it won't recognize it?  That would be odd, and good to
know. I thought the unAPI format names were purely arbitrary, but
recognized
by their association with a mime content-type like "application/x-
/endnote/-refer".   But no, at least as far as Zotero is concerned, you
have
to pick format shortnames that match what Zotero expects?


Robert Forkel wrote:



from looking at line 14 here
https://www.zotero.org/trac/browser/extension/trunk/translators/unAPI.js
i'd say:
ad 1. RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris",
"bibtex", "rdf"] also see function checkFormats
ad 2. the order listed above
ad 4.: from my experience the unapi scraper takes precedence over coins

On Tue, Apr 6, 2010 at 3:48 PM, Jonathan Rochkind 
wrote:


  

Anyone know if there's any developer documentation for Zotero on it's
use
of
unAPI?  Alternately, anyone know where I can find the answers to these
questions, or know the answers to these questions themselves?

1. What formats will Zotero use via unAPI. What mime content-types does
it
use to recognize those formats (sometimes a format has several in use,
or
no
official content-type).

2. What is Zotero's order of preference when multiple formats via unAPI
are
available?

3. Will Zotero get confused if different documents on the page have
different formats available?  This can be described with unAPI, but it
seems
atypical, so not sure if it will confuse Zotero.

4. If both unAPI and COinS are on a given page -- will Zotero use both
(resulting in possible double-import for citations exposed both ways).
Or
only one? Or depends on how you set up the HTML?

5. Somewhere that now I can't find I saw a mention of a "Zotero RDF"
format
that Zotero would consume via unAPI. Is there any documentation of this
format/vocabulary, how can I find out how to write it?




  
  




  

Re: [CODE4LIB] Zotero, unapi, and formats?

[Bill Dueber]

The unAPI support is also...non-ideal...in that you can't present
preferences for the best format to use. For example, the Refworks
Tagged format just plain has more tags (and hence more or
more-finely-grained information) than other formats (e.g., Endnote),
but Zotero will prefer Endnote just because it does. My RIS output is
better than my endnote output, but there's no way for me to tell
Zotero that.  For Mirlyn I ended up just having exactly one format
listed in my unapi-server file. Which is dumb. But I'm not sure what
else to do.

On Tue, Apr 6, 2010 at 10:16 AM, Jonathan Rochkind  wrote:
> Yeah, we need some actual documentation on Zotero's use of unAPI in general.
> Maybe if I can figure it out (perhaps by asking the developer(s)) I'll write
> some for them.
>
> Robert Forkel wrote:
>>
>> well, looks like a combination:
>> in case of mods it checks for the namespace URL, in case of rdf, it
>> looks for a format name of rdf_dc, ...
>> and yes, endnote export would have to have a name of "endnote" (i ran
>> into this problem as well with names like endnote-utf-8, ...). i think
>> unapi would be more usable if there were at least a recommendation of
>> common format names.
>>
>> On Tue, Apr 6, 2010 at 4:07 PM, Jonathan Rochkind 
>> wrote:
>>
>>>
>>> Wait, does it actually recognize the format by the format _name_ used,
>>> and
>>> not by a mime content-type?  Like unless my unAPI server calls the
>>> endnote
>>> format "endnote", it won't recognize it?  That would be odd, and good to
>>> know. I thought the unAPI format names were purely arbitrary, but
>>> recognized
>>> by their association with a mime content-type like "application/x-
>>> /endnote/-refer".   But no, at least as far as Zotero is concerned, you
>>> have
>>> to pick format shortnames that match what Zotero expects?
>>>
>>>
>>> Robert Forkel wrote:
>>>

 from looking at line 14 here
 https://www.zotero.org/trac/browser/extension/trunk/translators/unAPI.js
 i'd say:
 ad 1. RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris",
 "bibtex", "rdf"] also see function checkFormats
 ad 2. the order listed above
 ad 4.: from my experience the unapi scraper takes precedence over coins

 On Tue, Apr 6, 2010 at 3:48 PM, Jonathan Rochkind 
 wrote:


>
> Anyone know if there's any developer documentation for Zotero on it's
> use
> of
> unAPI?  Alternately, anyone know where I can find the answers to these
> questions, or know the answers to these questions themselves?
>
> 1. What formats will Zotero use via unAPI. What mime content-types does
> it
> use to recognize those formats (sometimes a format has several in use,
> or
> no
> official content-type).
>
> 2. What is Zotero's order of preference when multiple formats via unAPI
> are
> available?
>
> 3. Will Zotero get confused if different documents on the page have
> different formats available?  This can be described with unAPI, but it
> seems
> atypical, so not sure if it will confuse Zotero.
>
> 4. If both unAPI and COinS are on a given page -- will Zotero use both
> (resulting in possible double-import for citations exposed both ways).
> Or
> only one? Or depends on how you set up the HTML?
>
> 5. Somewhere that now I can't find I saw a mention of a "Zotero RDF"
> format
> that Zotero would consume via unAPI. Is there any documentation of this
> format/vocabulary, how can I find out how to write it?
>
>
>


>>
>>
>



-- 
Bill Dueber
Library Systems Programmer
University of Michigan Library

Re: [CODE4LIB] Zotero, unapi, and formats?

[Jonathan Rochkind]

Yeah, we need some actual documentation on Zotero's use of unAPI in 
general. Maybe if I can figure it out (perhaps by asking the 
developer(s)) I'll write some for them.


Robert Forkel wrote:

well, looks like a combination:
in case of mods it checks for the namespace URL, in case of rdf, it
looks for a format name of rdf_dc, ...
and yes, endnote export would have to have a name of "endnote" (i ran
into this problem as well with names like endnote-utf-8, ...). i think
unapi would be more usable if there were at least a recommendation of
common format names.

On Tue, Apr 6, 2010 at 4:07 PM, Jonathan Rochkind  wrote:
  

Wait, does it actually recognize the format by the format _name_ used, and
not by a mime content-type?  Like unless my unAPI server calls the endnote
format "endnote", it won't recognize it?  That would be odd, and good to
know. I thought the unAPI format names were purely arbitrary, but recognized
by their association with a mime content-type like "application/x-
/endnote/-refer".   But no, at least as far as Zotero is concerned, you have
to pick format shortnames that match what Zotero expects?


Robert Forkel wrote:


from looking at line 14 here
https://www.zotero.org/trac/browser/extension/trunk/translators/unAPI.js
i'd say:
ad 1. RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris",
"bibtex", "rdf"] also see function checkFormats
ad 2. the order listed above
ad 4.: from my experience the unapi scraper takes precedence over coins

On Tue, Apr 6, 2010 at 3:48 PM, Jonathan Rochkind 
wrote:

  

Anyone know if there's any developer documentation for Zotero on it's use
of
unAPI?  Alternately, anyone know where I can find the answers to these
questions, or know the answers to these questions themselves?

1. What formats will Zotero use via unAPI. What mime content-types does
it
use to recognize those formats (sometimes a format has several in use, or
no
official content-type).

2. What is Zotero's order of preference when multiple formats via unAPI
are
available?

3. Will Zotero get confused if different documents on the page have
different formats available?  This can be described with unAPI, but it
seems
atypical, so not sure if it will confuse Zotero.

4. If both unAPI and COinS are on a given page -- will Zotero use both
(resulting in possible double-import for citations exposed both ways). Or
only one? Or depends on how you set up the HTML?

5. Somewhere that now I can't find I saw a mention of a "Zotero RDF"
format
that Zotero would consume via unAPI. Is there any documentation of this
format/vocabulary, how can I find out how to write it?



  


  

Re: [CODE4LIB] Zotero, unapi, and formats?

[Robert Forkel]

well, looks like a combination:
in case of mods it checks for the namespace URL, in case of rdf, it
looks for a format name of rdf_dc, ...
and yes, endnote export would have to have a name of "endnote" (i ran
into this problem as well with names like endnote-utf-8, ...). i think
unapi would be more usable if there were at least a recommendation of
common format names.

On Tue, Apr 6, 2010 at 4:07 PM, Jonathan Rochkind  wrote:
> Wait, does it actually recognize the format by the format _name_ used, and
> not by a mime content-type?  Like unless my unAPI server calls the endnote
> format "endnote", it won't recognize it?  That would be odd, and good to
> know. I thought the unAPI format names were purely arbitrary, but recognized
> by their association with a mime content-type like "application/x-
> /endnote/-refer".   But no, at least as far as Zotero is concerned, you have
> to pick format shortnames that match what Zotero expects?
>
>
> Robert Forkel wrote:
>>
>> from looking at line 14 here
>> https://www.zotero.org/trac/browser/extension/trunk/translators/unAPI.js
>> i'd say:
>> ad 1. RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris",
>> "bibtex", "rdf"] also see function checkFormats
>> ad 2. the order listed above
>> ad 4.: from my experience the unapi scraper takes precedence over coins
>>
>> On Tue, Apr 6, 2010 at 3:48 PM, Jonathan Rochkind 
>> wrote:
>>
>>>
>>> Anyone know if there's any developer documentation for Zotero on it's use
>>> of
>>> unAPI?  Alternately, anyone know where I can find the answers to these
>>> questions, or know the answers to these questions themselves?
>>>
>>> 1. What formats will Zotero use via unAPI. What mime content-types does
>>> it
>>> use to recognize those formats (sometimes a format has several in use, or
>>> no
>>> official content-type).
>>>
>>> 2. What is Zotero's order of preference when multiple formats via unAPI
>>> are
>>> available?
>>>
>>> 3. Will Zotero get confused if different documents on the page have
>>> different formats available?  This can be described with unAPI, but it
>>> seems
>>> atypical, so not sure if it will confuse Zotero.
>>>
>>> 4. If both unAPI and COinS are on a given page -- will Zotero use both
>>> (resulting in possible double-import for citations exposed both ways). Or
>>> only one? Or depends on how you set up the HTML?
>>>
>>> 5. Somewhere that now I can't find I saw a mention of a "Zotero RDF"
>>> format
>>> that Zotero would consume via unAPI. Is there any documentation of this
>>> format/vocabulary, how can I find out how to write it?
>>>
>>>
>>
>>
>

Re: [CODE4LIB] Zotero, unapi, and formats?

[Jonathan Rochkind]

Wait, does it actually recognize the format by the format _name_ used, 
and not by a mime content-type?  Like unless my unAPI server calls the 
endnote format "endnote", it won't recognize it?  That would be odd, and 
good to know. I thought the unAPI format names were purely arbitrary, 
but recognized by their association with a mime content-type like 
"application/x-
/endnote/-refer".   But no, at least as far as Zotero is concerned, you 
have to pick format shortnames that match what Zotero expects?



Robert Forkel wrote:

from looking at line 14 here
https://www.zotero.org/trac/browser/extension/trunk/translators/unAPI.js
i'd say:
ad 1. RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris",
"bibtex", "rdf"] also see function checkFormats
ad 2. the order listed above
ad 4.: from my experience the unapi scraper takes precedence over coins

On Tue, Apr 6, 2010 at 3:48 PM, Jonathan Rochkind  wrote:
  

Anyone know if there's any developer documentation for Zotero on it's use of
unAPI?  Alternately, anyone know where I can find the answers to these
questions, or know the answers to these questions themselves?

1. What formats will Zotero use via unAPI. What mime content-types does it
use to recognize those formats (sometimes a format has several in use, or no
official content-type).

2. What is Zotero's order of preference when multiple formats via unAPI are
available?

3. Will Zotero get confused if different documents on the page have
different formats available?  This can be described with unAPI, but it seems
atypical, so not sure if it will confuse Zotero.

4. If both unAPI and COinS are on a given page -- will Zotero use both
(resulting in possible double-import for citations exposed both ways). Or
only one? Or depends on how you set up the HTML?

5. Somewhere that now I can't find I saw a mention of a "Zotero RDF" format
that Zotero would consume via unAPI. Is there any documentation of this
format/vocabulary, how can I find out how to write it?




  

Re: [CODE4LIB] Zotero, unapi, and formats?

[Jonathan Rochkind]

Thanks Ere. Yeah, I found that information too , but it's not enough. 
Maybe I should contact the Zotero developer(s)?  Anyone know the best 
way to do that?


Ere Maijala wrote:
This was the only information I found when I developed unAPI support for 
our MetaLib installation: 
. Based on my 
experimentation and looking at the code, if my memory serves:


1. At least the formats mentioned in the forum post. I believe it uses 
the docs attribute to distinguish formats, as type can be e.g. 
application/xml for multiple formats.


2. Weird, but I don't remember how. I ended up providing only MARCXML, 
DC and RIS, because it chose MODS over MARCXML if it was available and 
did something that sucked. Things may have changed, this was in 2008.


3. Didn't test this one, we only provide a single record at a time.

4. It chose COinS over unAPI at least at the time, and I found that to 
be a bit problematic.


5. Dunno.

--Ere

On 6.4.2010 16:48, Jonathan Rochkind wrote:
  

Anyone know if there's any developer documentation for Zotero on it's
use of unAPI? Alternately, anyone know where I can find the answers to
these questions, or know the answers to these questions themselves?

1. What formats will Zotero use via unAPI. What mime content-types does
it use to recognize those formats (sometimes a format has several in
use, or no official content-type).

2. What is Zotero's order of preference when multiple formats via unAPI
are available?

3. Will Zotero get confused if different documents on the page have
different formats available? This can be described with unAPI, but it
seems atypical, so not sure if it will confuse Zotero.

4. If both unAPI and COinS are on a given page -- will Zotero use both
(resulting in possible double-import for citations exposed both ways).
Or only one? Or depends on how you set up the HTML?

5. Somewhere that now I can't find I saw a mention of a "Zotero RDF"
format that Zotero would consume via unAPI. Is there any documentation
of this format/vocabulary, how can I find out how to write it?

Re: [CODE4LIB] Zotero, unapi, and formats?

[Ere Maijala]

This was the only information I found when I developed unAPI support for 
our MetaLib installation: 
. Based on my 
experimentation and looking at the code, if my memory serves:


1. At least the formats mentioned in the forum post. I believe it uses 
the docs attribute to distinguish formats, as type can be e.g. 
application/xml for multiple formats.


2. Weird, but I don't remember how. I ended up providing only MARCXML, 
DC and RIS, because it chose MODS over MARCXML if it was available and 
did something that sucked. Things may have changed, this was in 2008.


3. Didn't test this one, we only provide a single record at a time.

4. It chose COinS over unAPI at least at the time, and I found that to 
be a bit problematic.


5. Dunno.

--Ere

On 6.4.2010 16:48, Jonathan Rochkind wrote:

Anyone know if there's any developer documentation for Zotero on it's
use of unAPI? Alternately, anyone know where I can find the answers to
these questions, or know the answers to these questions themselves?

1. What formats will Zotero use via unAPI. What mime content-types does
it use to recognize those formats (sometimes a format has several in
use, or no official content-type).

2. What is Zotero's order of preference when multiple formats via unAPI
are available?

3. Will Zotero get confused if different documents on the page have
different formats available? This can be described with unAPI, but it
seems atypical, so not sure if it will confuse Zotero.

4. If both unAPI and COinS are on a given page -- will Zotero use both
(resulting in possible double-import for citations exposed both ways).
Or only one? Or depends on how you set up the HTML?

5. Somewhere that now I can't find I saw a mention of a "Zotero RDF"
format that Zotero would consume via unAPI. Is there any documentation
of this format/vocabulary, how can I find out how to write it?

Re: [CODE4LIB] Zotero, unapi, and formats?

[Robert Forkel]

from looking at line 14 here
https://www.zotero.org/trac/browser/extension/trunk/translators/unAPI.js
i'd say:
ad 1. RECOGNIZABLE_FORMATS = ["mods", "marc", "endnote", "ris",
"bibtex", "rdf"] also see function checkFormats
ad 2. the order listed above
ad 4.: from my experience the unapi scraper takes precedence over coins

On Tue, Apr 6, 2010 at 3:48 PM, Jonathan Rochkind  wrote:
> Anyone know if there's any developer documentation for Zotero on it's use of
> unAPI?  Alternately, anyone know where I can find the answers to these
> questions, or know the answers to these questions themselves?
>
> 1. What formats will Zotero use via unAPI. What mime content-types does it
> use to recognize those formats (sometimes a format has several in use, or no
> official content-type).
>
> 2. What is Zotero's order of preference when multiple formats via unAPI are
> available?
>
> 3. Will Zotero get confused if different documents on the page have
> different formats available?  This can be described with unAPI, but it seems
> atypical, so not sure if it will confuse Zotero.
>
> 4. If both unAPI and COinS are on a given page -- will Zotero use both
> (resulting in possible double-import for citations exposed both ways). Or
> only one? Or depends on how you set up the HTML?
>
> 5. Somewhere that now I can't find I saw a mention of a "Zotero RDF" format
> that Zotero would consume via unAPI. Is there any documentation of this
> format/vocabulary, how can I find out how to write it?
>