Local collections
Introduction
A local collection comprises documents from a directory/folder which is present on the Funnelback server. Use this collection for indexing text content (such as HTML, TXT) only. The files located in the directory are not copied or pre-processed and are indexed as is. Thus any binary document types (e.g. Word, PDF etc.) will not normally be returned in search results, as they will not have had their text extracted for indexing. Because of this restriction you may wish to consider using file-copy collections instead.
The folder to index must be specified in the data_root collection.cfg parameter.
Local collection result URLs
The result URLs in a local collection will all be local to a specific machine. This means they will need to be processed at query time to make them usable. Options for doing this include:
- Using a user interface hook script to modify the URL
- Editing the search template to use a cache link instead
For a worked example of the last option, see details on result presentation for XML documents.