What's New
Search Scoring Overhaul
This release introduces a completely revamped search scoring system that dramatically improves search result accuracy and relevance.
New Features
-
Stopword Removal: Common words like "the", "a", "an" are now filtered from searches. Searching for "the Ramones" will now correctly find "Ramones".
-
Fuzzy Matching: Typo tolerance using the Jaro-Winkler algorithm. Searches like "beetles" will find "Beatles", and "metalica" will find "Metallica".
-
N-gram Matching: Partial string matching using bigrams for better substring detection.
-
Reverse Matching: When your search query contains the result name, it now gets properly scored (e.g., searching "the ramones punk" finds "Ramones").
Improved Scoring Tiers
Results are now ranked with more nuanced scoring:
- Exact matches rank highest
- "Starts with" matches rank high
- Word boundary matches (query at start of a word)
- Contains matches
- Fuzzy matches (typo tolerance)
- Partial n-gram matches
Bug Fixes
- Fixed radio stream metadata not displaying correctly when switching between stations
- Fixed metadata "downgrade" issue where good radio metadata was being overwritten by incomplete data
Full Changelog: v2.8.3-beta...v2.8.4-beta