[LUCENE-1702] Thai token type() bug

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 3.1, 4.0-ALPHA
• Component/s: modules/analysis
• Lucene Fields: New
• Assignee: Unassigned
• Reporter: Robert Muir
• Created: 19/Jun/09 15:33
• Updated: 28/Aug/22 12:02
• Resolved: 29/Sep/10 05:51

Description

While adding tests for offsets & type to ThaiAnalyzer, i discovered it does not type Thai numeric digits correctly.
ThaiAnalyzer uses StandardTokenizer, and this is really an issue with the grammar, which adds the entire [:Thai:] block to ALPHANUM.

i propose that alphanum be described a little bit differently in the grammar.
Instead, [:letter:] should be allowed to have diacritics/signs/combining marks attached to it.

this would allow the [:thai:] hack to be completely removed, would allow StandardTokenizer to parse complex writing systems such as Indian languages, and would fix LUCENE-1545.

Issue Links

• is part of: LUCENE-2167 Implement StandardTokenizer with the UAX#29 Standard