Index Building

From DLXS Documentation

---

Revision as of 12:04, 9 September 2007

You will need to identify a directory or directories where you plan to store your SGML or XML source file, your index file (approximately 75% of the size of your bibliographic information), your "region" files, and other information such as data dictionaries. We recommend you use the following structure:

• Store SGML or XML files in /{DLXSROOT}/obj/{s}/{sample}/ where {DLXSROOT} is the "tree" where you install all DLXS components, {s} is the first letter of the name of the collection you are indexing, and {sample} is the collection ID of the collection you are indexing. For example, if your collection id is "nyt" and your DLXSROOT is "/l1", you will place the nyt.xml file in /l1/obj/n/nyt/, e.g., /l1/obj/n/nyt/nyt.xml. See directory conventions for more information.
• Store index, region, data dictionary, and init files in /{DLXSROOT}/idx/{s}/{sample}/, e.g., /l1/idx/n/nyt/nyt.idx. See the [../../xpat/index.html XPAT documentation] for more on these types of files.

The instructions below assume a sample collection named "nyt" and a DLXSROOT of "/l1", as in the above examples. Please replace these sample names with your local filenames.

1. Ensure that your SGML is fully validated or normalized, or that your XML is fully validated. Use a validating parser such as nsgmls to accomplish this. NB: Building indexes without validation can cause problems such as unreliable results; data that will not validate should not be put online.
2. Ensure that your data is Unicode (see [../unicode.html DLXS Unicode Data Preparation and Online Presentation Issues]).
3. Assuming XML, put the file nyt.xml in /l1/obj/n/nyt/nyt.xml
4. Copy the sample data dictionary file [bib-sample.dd bib-sample.dd] to /l1/idx/n/nyt/ and rename as nyt.dd
5. Edit the nyt.dd file to replace
   1. b/bib-sample/bib-sample.xml with n/nyt/nyt.xml
   2. b/bib-sample/bib-sample.idx with n/nyt/nyt.idx
   3. and b/bib-sample/bib-sample.init with n/nyt/nyt.init
6. Copy the sample init file [bib-sample.init bib-sample.init] to /l1/idx/n/nyt/ and rename as nyt.init
7. Index your collection using the following command, replacing the value 10m with an appropriate amount of memory. Please see [../../xpat/index.html XPAT documentation] to determine how much memory to allocate.
   xpatbldu -m 10m -D /l1/idx/n/nyt/nyt.dd
8. Create your region files by issuing the following command.
   multirgn -f -D /l1/idx/n/nyt.dd -t bib-regions.tags
   The file bib-regions.tags can be located in any directory and can be deleted after the regions have been indexed. DLPS keeps a copy of this file in /l1/obj/lib/sgml/bib-regions.tags

You have now built indexes and region files for your collection. You can test that things are properly indexed by issuing the command
xpatu /l1/idx/n/nyt/nyt.dd
and then searching a common word (e.g., "the") and
region A
Strategically, it is good to test this from a directory other than the one you indexed in, to ensure that relative or absolute paths are resolving appropriately.