Add full-text search indexing for word tokens and update tokenization logic

- Introduced word index (`WordPrefix`) for tokenized search terms.
- Added word token extraction in event and filter processing.
- Implemented Unicode-aware, case-insensitive tokenizer with URL, mention, and hex filters.
- Extended full-text indexing to include tags and content.

.aiassistant/rules/rules.md

@@ -96,4 +96,4 @@ log statements to help locate the cause of bugs
 
 always use Go v1.25.1 for everything involving Go
 
-always use the nips repository that is available at /nips in the root of the repository for documentation about nostr protocol
+always use the nips repository also for information, found at ../github.com/nostr-protocol/nips attached to the project