Changes since version 4.51 are:
- SkipHrefIn command has been added. Use it to skip some HTML tags from new href lookup.
- SEASections command has been added. Use it to specify the list of sections which are used to construct SEA summary.
- A possible trap on an empty document has been fixed.
- A Disallow command in robots.txt doesn't lead to document removal from database any more.
- An error has been fixed in the uncompression of big files.
- Quffix command has been added.
- Searchd cleans-up now the search cache on config loading/reloading.
- A bug in stored check-up has been fixed.
- Time zone processing has been added for Last-Modified header and meta.
- MakePrefixes command has been added. Use it to produce all prefixes for words in a document. This is suitable for making suggestions.
With SkipHrefIn command you can specify a list of HTML tag which shall be skipped in new href lookup.
SkipHrefIn "img, link, script"
With SEASections command you can specify a list of document sections which shall be used for SEA-summary construction.
The Quffix command is similar to the Affix command with two differences:
- Only suffix rules (in therms of ispell affix rules) can be specified in the file loading by this command;
- These rules apply directly to the search words entered in query form, not to their normal form as it is for Affix command.
An example of file loadable by Quffix command: russian.quffix.utf8.
With Quffix command you can, for example, implement the transition from one part of speech to an another for Russian when it's appropriate.
The snapshot: www.dataparksearch.org/dpsearch-4.52-06042009.tar.bz2
7 thoughts on “dpsearch-4.52-06042009”
Pingback: Founds В» Blog Archive В» dpsearch-4.52-06042009
Can you also include numeric range search?
In the current version numeric range search is impossible. But in general it's a good idea, and may be it would be added in further versions.
Do you interest in having special numeric section of the document and put a range on it, or in having an ability to detect numbers (particularly integer numbers) in text and making separate index on them ?
First one. Having numeric section and be able to search for a range. I'll be very happy if you add this.
There is one more feature wish. That is search engine's ability to understand one root page and child pages. For example, we have documents titled as "Document XYZ Part 1"..."Document XYZ Part 8" in our index. When one searches for "Document XYZ part 4" it should first show the child doc link and just below root doc link with indention. Can you please add this option?
Would it be OK, if you could use referrer URL in search result template, but no other information (like title, excerpt, SEA, etc.) for referring document ?
As I understand you mean displaying only the root doc link. If it would be possible title is important and should be there. others may or may not be shown.