Re: Conversion of tei.2 files to TEI P5 (LOC American Memory DTD)

Hi all,

Just a brief update: I've completed a basic conversion of the data in
question and posted the conversion script, along with the resulting data
and generated eXist app (generated with TEI Publisher), at
https://github.com/joewiz/adst with some screenshots and directions.

The direct link to the conversion script is:
https://github.com/joewiz/adst/blob/master/import-scripts/04-convert-to-tei-p5.xq
.

If you haven't seen TEI Publisher, its project homepage is
http://teipublisher.com.

Thanks again for everyone's help,
Joe


On Tue, Feb 7, 2017 at 4:12 PM, Joe Wicentowski <[email protected]> wrote:

> Thanks, James, Stuart, and those who replied off list.  These look like
> great resources.
>
> Having only started working with TEI just after the release of P5 in late
> 2007, I blindly assumed that these LOC documents, whose root element was
> <TEI2>, were necessarily circa P2.  But I gather from the comments that
> these appear P4-esque.
>
> I now see from the Vault that P4 used <TEI.2>, P3 used <tei.2>, P2 used
> <TEI.2>, and P1 used <teidoc> (!).  Some guideposts for the TEI
> forensicists...
>
> The DTD comments note that the DTD was updated to P3, but P3 didn't use
> <TEI.2>.  The American Memory DTD notes: "LC staff renamed TEI element
> <TEI.2> as <TEI2>."  And, to boot, the samples I've looked at use <tei2>,
> not <TEI2>; the entity files referenced in each TEI file are all missing;
> and the DOCTYPE declarations are encased in comments.
>
> So it seems these LOC files represent a rather significant customization -
> some kind of hybrid.
>
> On the plus side, the DTD is well documented and the TEI files themselves
> appear to be all well-formed XML - so assuming they're internally
> consistent, it shouldn't be too hard to pull the data out into a basic
> P5-compliant form.
>
> Thanks again!
> Joe
>
> On Tue, Feb 7, 2017 at 2:54 PM, Stuart A. Yeates <[email protected]>
> wrote:
>
>> There is an entire category of xsl scripts in the wiki for handling
>> different aspects of this transition. http://wiki.tei-c.
>> org/index.php/Category:P4toP5
>>
>> The one I personally used is at http://wiki.tei-c.org/index
>> .php/P4toP5NZETC and is a mix of strict migration issues, fixing local
>> issues and fixing things in the text that weren't checked for by the
>> previous schema.
>>
>> cheers
>> stuart
>>
>> --
>> ...let us be heard from red core to black sky
>>
>> On 8 February 2017 at 04:20, James Cummings <[email protected]>
>> wrote:
>>
>>> Hi Joe,
>>>
>>> Walking back to my hotel at the TEI Council meeting in Prague.  There is
>>> a p4top5 stylesheet at
>>> https://github.com/TEIC/Stylesheets/blob/dev/profiles/defaul
>>> t/p4/from.xsl
>>>
>>> It isn't perfect especially where ppl have customised and the p5 it
>>> produces might be slightly out of date, but that is where is start if
>>> someone hasn't converted it already.
>>>
>>> James
>>>
>>>
>>> --
>>> Dr James Cummings, Academic IT Services, University of Oxford
>>>
>>>
>>> On 7 Feb 2017 4:10 p.m., Joe Wicentowski <[email protected]> wrote:
>>>
>>> Hi everyone,
>>>
>>> Has anyone converted tei.2 files (ideally, those from the Library of
>>> Congress's American Memory project, DTD developed circa 1997[1]) to TEI
>>> P5?  I'm looking at a batch of ~1,700 files with transcripts of oral
>>> history interviews [2], and while I could rig up a script to perform a
>>> basic conversion, I thought I'd ask if there was any previous work in this
>>> area.
>>>
>>> Thanks,
>>> Joe
>>>
>>> [1] ftp://ftp.loc.gov/pub/american.memory/sgml/ammem2/INDEX
>>> [2] Collection homepage: https://www.loc.gov/collection
>>> s/foreign-affairs-oral-history/about-this-collection/. Sample specimen:
>>> https://cdn.loc.gov/service/mss/mfdip/2004/2004fre01/2004fre01.xml.
>>>
>>>
>>>
>>
>