Nameresolver2

Overview and Design

Nameresolver2 was developed in conjunction with the move away from broker20 and Bibliographic Class in DLPS. Instead of what old nameresolver offered-- a method for creating short URLs using xpat searches-- we developed a MySQL database that would store all the URLs and avoid using xpat. It also avoided the need for creating Bibliographic Class metadata in order to gather the long URLs that would be made into short URLs, which old nameresolver needed and new nameresolver2 does not. New nameresolver2 acts very similarly to idresolver, so we'll compare the differences below.

idresolver vs. nameresolver2 (idresolver-nr)

The idresolver is a simple mechanism for resolving a unique ID to a URL. The DLXS Classes can use it to, for example, link digital object identifiers in an Electronic Finding Aid to the actual digital object in another class.

idresolver	There is always just one URL for each ID in the database.
nameresolver2	Records are keyed on id / collection pairs so more than one URL for a given ID can exist.

idresolver	A URL (marked up in XML) for an ID is returned (see example below).
nameresolver2	If a single entry for an ID is found, the user is redirected to that URL.
nameresolver2	If more than one entry for an ID is found, the user is prompted with links to the item in different collections.

idresolver example:

    http://quod.lib.umich.edu/cgi/i/idresolver/idresolver?id=is.blake.0001

    <ITEM MTIME="2008-07-24 16:28:00">
      <ID>is.blake.0001</ID>
      <URI>http://images.umdl.umich.edu/cgi/i/image/image-idx?view=entry;subview=detail;
        cc=blakeic;entryid=X-1;viewid=1</URI>
    </ITEM>

Nameresolver2 Usage

Table structure:

Database: dlxs
Table name: nameresolver

id	coll	url	modified (auto- generated)
0599998.0001.001	alajournals	http://quod.lib.umich.edu/cgi/t/text/text-idx?c=alajournals;idno=0599998.0001.001	2008-01-15 09:54:32
0599998	alajournals	http://quod.lib.umich.edu/cgi/t/text/text-idx?c=alajournals;idno=0599998	2008-01-15 09:54:32
BAD1447.0002.001	genpub	http://quod.lib.umich.edu/cgi/t/text/text-idx?c=genpub;idno=BAD1447.0002.001	2008-01-15 10:01:30
BAD1447.0002.001	micounty	http://quod.lib.umich.edu/cgi/t/text/text-idx?c=micounty;idno=BAD1447.0002.001	2008-01-15 10:04:00

Long ID vs. Short ID

The nameresolver table contains the full or long ID for a particular record as well as a short ID. For our naming conventions, a long ID points to an individual record. It has the format 0599998.0001.002. The first part (0599998) represents the title. The second part is the volume (0001) and the third is the issue (002). This naming convention is useful for multi-volume works but it is not required for nameresolver2. If your IDs do not take this format, only the full ID will be used.

Browser Display

Long ID - User is forwarded directly to the record with no interim page

http://name.umdl.umich.edu/0599998.0001.001

Short ID that corresponds to only one long ID - User is forwarded directly to the record with no interim page

http://name.umdl.umich.edu/B1942529 (long id B1942529.0001.001)

Short ID that links to multiple long IDs - User is forwarded to the collection which presents the volumes for that short id

http://name.umdl.umich.edu/0599998

ID in multiple collections - User is provided a list of collections in which the record resides and can choose to view it in any of the collections

http://name.umdl.umich.edu/BAD1447.0002.001

Location and Modification

The nameresolver2 files are stored at $DLXSROOT/bin/n/nameresolver. Files of interest are:

NRTable.sql - SQL script to create the database table. This should work without modification.
IdParser.pl - Perl script to that extracts the long and short ids, collection, and URL information for each record and loads it to the database. It uses the $DLXSROOT/lib/LibGlobals.cfg configuration file to connect to the MySQL database.

Populating the Database

The IdParser.pl script stored at $DLXSROOT/bin/n/nameresolver has a well-documented usage message. If you run IdParser.pl -h, you will see a list of available options. There are several ways to update nameresolver2.

Option 1: Bib Class files

This method can be used for the original migration from old nameresolver to nameresolver2. Consistent with the old nameresolver, it uses the ID (not IDNO) tag for the identifier.
Example: To add all Bib Class files from a given directory do:
```
./IdParser.pl -d fullDirectoryPath
```
To add a specific Bib Class file, add the -f option:
```
./IdParser.pl -d fullDirectoryPath -f myColl-bib.xml
```

Option 2: Simplified XML format

If you would like to add a small number of records to nameresolver2, there is a simpler format than Bib Class. This could be created using a list of records, such as from a database query. The format is:
```
<ITEMS><ITEM><ID>123</ID><COLL>collid</COLL><URL>http://</URL></ITEM><ITEM>...</ITEM></ITEMS>
```
Add -u to the examples from option 1.
To add all Bib Class files from a given directory do:
```
./IdParser.pl -d fullDirectoryPath -u
```

To add a specific Bib Class file, add the -f option:

./IdParser.pl -d fullDirectoryPath -f myRecords.xml -u

Option 3: Single ID from the command line

If you have only a small number of records to add, you can add them individually from the command line. This will process the long and short ID for the given ID.

Example:

./IdParser.pl -c collName -i yourIdentifier -l "yourUrl"

Option 4: Comma-separated value (.csv) file (spreadsheet or flat text file)

This may be the easiest option if you have a large number of ids to add which have the same base URL.
Example: LoadLLMCIDs.pl. This script is provided as an example and would have be to altered for your use.

Option 5: From UMProvider

The UMProvider suite synchronizes with nameresolver to add new nameresolver records along with new records to the oai database. When the UMProvider update is run, new DLXS text and image class records that are added to the oai database are also added to nameresolver2.