Adding new collections to Credo is a multiple step process and each step is necessary to ensure that the collection is properly represented in Credo's collection hierarchy and that our collections play well with Digital Commonwealth and the DPLA.

For an individual item in Credo to be part of a collection, it needs to have a MODS <relatedItem> element pointing to that collection. You can find more information about <relatedItem> use in SCUA's Best Practices for MODS. Here's an example:

<mods:relatedItem type="host">
  <mods:titleInfo>
     <mods:title>W. E. B. Du Bois Papers</mods:title>
  </mods:titleInfo>
  <mods:typeOfResource>mixed material</mods:typeOfResource>
  <mods:identifier type="local">MS 312</mods:identifier>
  <mods:identifier type="uri">mums312</mods:identifier>
</mods:relatedItem>

The three requirements for the item to be properly attached to its collection are 1) the @type=“host” in <mods:relatedItem>, 2) the unique ID for the collection in <mods:identifier type=“uri”>, and 3) the title of the collection in <mods:title>. The title must be consistent across all records.

A collection-level record in Credo creates a page like this for each collection, http://credo.library.umass.edu/view/collection/mums312. The collection page is generated from an EAD/XML finding aid, placed at the top-level of a directory on oubliette. The name of the file must have the following structure:

collection_[collID].xml

For example, collection_mums312.xml in the mums312 directory in oubliette will be used to generate the HTML for the mums312 collection page. In most cases, all you need to do is copy the collection's existing EAD finding aid and rename accordingly. To link to the full finding aid, add <otherfindaid><extref href=“[collID]”>[Coll name]</extref></otherfindaid> to <archdesc>.

Do not add a collection-level record to the ingest folder. Ingesting these records will wreak unfortunate havoc on Credo. It's enough to simply name the file correctly and add it to the collection top-level folder. It might take some time for a new collection record to appear but it should show up after an ingest (but again, these files don't need to be ingested!)

If you want to add a collection page for a collection that does not yet have a full finding aid, it's easy enough to create a stub EAD record using the information in UMarmot. For a good example of a stub collection-level record, look at collection_mums700.xml on oubliette.

There is a template for a collection-level record in spec_collections/credo/metadata/templates

The collection map file helps sort collections in Credo's browse view. You can find the file in the ingest folder named, collmap.txt. Each line of the file contains the display name of the collection followed by the name you want the collection to sort on in the browse view. Typically, this is the authorized form of the collection's primary provenance. Each name is wrapped in double quotes (“) and delimited by a colon (:), e.g.

“W. E. B. Du Bois Papers”:“Du Bois, W. E. B. (William Edward Burghardt), 1868-1963”

The line above will sort the collection under “D” for Du Bois but display the collection title, W. E. B. Du Bois Papers. Changes to this file usually take effect after ingest.

This last step is critical for successful harvest by Digital Commonwealth and DPLA, which depend on the collections appearing in the OAI-PMH endpoint in a specific way in order to properly display the collection hierarchies in those systems.

In OAI-PMH, we represent collections as “Sets” and a complete listing of the collections each of our digital objects belong to can be viewed in the response to the OAI-PMH verb ListSets.

Here is a typical set entry:

<set>
  <setSpec>muph001</setSpec>
  <setName>Frances and Mary Allen Collection of Photographs of Deerfield, 1900-1910</setName>
  <setDescription>
    <oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/"    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
      <dc:description>
Influenced by the arts and crafts movement, Frances and Mary Allen began taking photographs of their native Deerfield, Mass., in the mid-1880s. Displaying a finely honed pictorialist aesthetic, the sisters specialized in views of Deerfield and surrounding towns, posed genre scenes of life in colonial times, and the local scenery, earning a reputation as among the best women photographers of the period. The Allen sisters photograph album contains ten gelatin developing out prints of street scenes in Deerfield, ca.1900-1910. Among these are two shots of the house they inherited from their aunt Kate in 1895, which thereafter became their home and studio.
      </dc:description>
    </oai_dc:dc>
  </setDescription>
</set>

It is critical that the unique ID for the collection that appears in <mods:identifier type=“uri”> of each record in that collection matches the unqiue ID in <setSpec>. A <dc:description> is not required but strongly preferred. All set entries live in a single static XML file kept on the OAI-PMH server.

The server can be accessed via SFTP with the address:

credo-oai.library.umass.edu

user: localadminuser

Please ask Aaron R. for a password.

The file can be found at ~/solr/solr-5.2.1/server/solr/oai and is called ListSets.xml. Download this file and edit to add the new <set>. Once added, re-upload the edited file. In order for these changes to take effect, the OAI-PMH web server will need to be restarted. Please ask Aaron R. to restart the server. In Aaron R.'s absence, create a syshelp ticket in LibWire and Aaron A. will restart.

Series-level landing pages can be created for collections in Credo. These pages allow the user to search within a specific series and includes Title, Collection #, Extent, Series Overview, Language and Location of that series/collection.

Before a series page can be created you must add the following elements to the EAD/XML collection page:

<dsc type="combined" id="contlist">
  <c01 level="series">
    <did>
      <unittitle><extref href="mums312-s01">Series 1. Correspondence</extref></unittitle>
      <unitdate>1877-1965</unitdate>
      <physdesc>195.50 boxes</physdesc>
      <physdesc>
        <extent>(81.52 linear feet)</extent>
      </physdesc>
    </did>
  </c01>

In the series-level record, the following elements must be included to link out to items in the series on Credo:

<dsc type="combined" id="contlist">
  <c01 level="series" id="mums312-s01">
    <did>
      <unittitle>Series 1. Correspondence</unittitle>
      <unitdate>1877-1965</unitdate>
      <physdesc>
        <extent>195.50 Boxes (81.52 linear feet)</extent>
      </physdesc>
    </did>
   </c01>

Information about the arrangement of the series and a scope note for the series can be added to the series-level record as well.

When making a series page, the MODS record for every individual item in the series must include a <relatedItem> element pointing to the series as follows. The identifier must be the same as in the series ID in the collection page.

<mods:relatedItem type="series">
  <mods:titleInfo>
    <mods:title>Series 1. Correspondence</mods:title>
  </mods:titleInfo>
<mods:typeOfResource>mixed material</mods:typeOfResource>
<mods:identifier type="uri">mums312-s01</mods:identifier>
</mods:relatedItem>

This goes after the <relatedItem> element that points to the collection page (the @type=“host”). Like the collection page, the @type=“series” in <mods:relatedItem>, the unique ID for the collection in <mods:identifier type=“uri”>, and the title of the series in <mods:title> must be consistent across all records.

In Fedora, the series-level record should be added to the folder where the items in that series live. Like the collection-level record, do not add the series-level record to the ingest folder, just add the file to the correct series folder.

TIP:

• Make sure to create the file in the spec_collections/EAD folder so the DTD will not throw an error. You can then move the file to a collection folder.
• Watch out for extra elements and XML errors.
• There is a template for a series-level record in spec_collections/credo/metadata/templates

Note: this is only necessary if items from specific subseries in a collection are in Credo.

If a series has subseries, all items in that subseries can be linked to from the series page. To do so, add the following to the series-level record, after the closing </controlaccess> element and before the closing </c01> tag:

<c02 level="subseries">
 <did>
  <unittitle><extref href="mums312-s01A">A. General Correspondence</extref></unittitle><unitdate>1877-1965</unitdate><physdesc/></did>
    </c02>
  <c02 level="subseries">
   <did><unittitle>
    <extref href="mums312-s01B">B. Crisis Correspondence</extref>
   </unittitle>
    <unitdate normal="1911/1934">1911-1934</unitdate></did>
    </c02>

The MODS record for each item in the subseries must include another <relatedItem> element pointing to the subseries:

<mods:relatedItem type="series">
  <mods:titleInfo>
    <mods:title>Series 1A. General Correspondence</mods:title>
  </mods:titleInfo>
<mods:typeOfResource>mixed material</mods:typeOfResource>
<mods:identifier type="uri">mums312-s01A</mods:identifier>
</mods:relatedItem>

Again, the @type=“series” in <mods:relatedItem>, the unique ID for the collection in <mods:identifier type=“uri”>, and the title of the collection in <mods:title> must be consistent across all records.

The subseries <relatedItem> element should be placed after the series <relatedItem> element, and this order does matter for the hierarchy of the links at the top of the item record page in Credo. For example, when the series <relatedItem> element comes before the subseries <relatedItem> element in an item record, the top of the page for the item looks like this:

If the subseries is placed before the series in the item record, it looks like this instead: