Ingest "Postcards" (pudl0009)

[tpendragon]

Note: MODS

@jpstroop I assume "MODS" means there's no bib records? Depends on #1674 Depends on #1675

[tpendragon]

Files: https://drive.google.com/drive/u/0/folders/0B4Wo5hgOEFY3ZlZGUU5OS3FIbDA

[tpendragon]

http://pudl.princeton.edu/collections/pudl0009

[tpendragon]

The METS is separated into 5 directories. Each directory is a "series", which is referenced by the MODS. Each series is described in a finding aid: https://findingaids.princeton.edu/collections/AC045.xml

There's no item-level descriptions in the finding aid - there's "parts" of boxes.

Field	Example
title (non-Sort)	The
title	Chapel, Princeton University
relatedItem	Princeton University Historical postcard collection Series 1 Buildings, circa 1900-1960 (separated into a title and two sub-titles
publisher	Published by H. M. Hinkson, Stationer, Princeton, N. J. & The Albertype Co., Brooklyn, N.Y.
publisher date	1941
language	eng
abstract	Exterior view of the chapel in black and white. Divided back postcard.
genre (aat)	Picture postcards
subject (aat)	Eye-level views, Exterior views
subject (lcsh)	Princeton University, Buildings, Princeton University. Chapel
location	Princeton University Library. Department of Rare Books and Special Collections. Seeley G. Mudd Manuscript Library.

[tpendragon]

When migrating sort title, migrate into two fields: "Sort Title" and "Title." Title would be "The Chapel, Princeton University" and sort title would be "Chapel, Princeton University."

[tpendragon]

Map location to our controlled vocab.

[tpendragon]

This one seems pretty straight forward, as long as we don't have to link publisher to published date.

[jpstroop]

Are we controlling publisher? I could imagine a single publication statement that's just a string (could include both publishers, dates and locations in whatever format, all in one field) and then separate controlled fields for publisher (just another name with a role of publisher), date(s), and location.

[christinach]

Some files include more than one originInfo:

<originInfo>
      <publisher>Published by H. M. Hinkson, Stationer, Princeton, N. J.</publisher>
 </originInfo>
 <originInfo>
      <publisher>The Albertype Co., Brooklyn, N. Y.</publisher>
</originInfo>

In this case do we want to pull publication information from the first node or from all?

[jpstroop]

Definitely need both. Is publisher (or what ever it is) a repeatable field?

[christinach]

@jpstroop based on your previous comment my understanding was to add a publication field and pull the publisher, dates and locations information from the originInfo node. I'm not sure I understand '..Is publisher a repeatable field..'

[jpstroop]

I don't know how things are being modelled, but this record essentially had two publication statements, which are really just strings, and I think could either store two strings: publisher: ["s1, "s2"] or if that isn't possible then mash them together as one string and separate them with ;, i.e.: publisher: ["{s1} ; {s2}"].

I would think that the date would be stored separately?

[christinach]

I can add only the publisher information. The dateCreated is stored separately.

[tpendragon]

@christinach For now I'd just make two fields "publisher" and "published date." Extract the publisher name part to publisher and the date part of the publication to published date. (There's also a dateCreated, but that should be handled already right?)

[christinach]

yes, dateCreated is stored already. I'm assuming dateOther is the published Date.

[tpendragon]

@christinach I think that's accurate. There's also a <copyrightDate> it looks like.

[tpendragon]

1/0196.mets also has a dateIssued

[christinach]

yes there is a copyrightDate

[christinach]

@tpendragon thank you, I will add all the relevant dates.

[christinach]

The folder pudl0009 includes a folder 'uncat' with tif, img and png files. @roelmoon said that the png and jpg can be ignored (or deleted).