المحلّلات (Analyzers) والصلة

في هذا الدرس

understand how text is processed and ranked.

الموضوعات المخطط لها

• An analyzer = character filters, a tokenizer (split into words), and token filters (lowercase, remove stop words, stemming, synonyms).
• The same analyzer should be applied at index time and query time, unless you deliberately differ (e.g. autocomplete).
• Relevance scoring: BM25 (verify default) rewards rare terms and frequent matches in short fields; field boosts raise importance (title over description).
• Analyzer choice changes results: stemming makes "running" match "run", but can over-match.
• Language matters: use the analyzer for the content language.