Integrate Bleve search engine for full-text search with multi-word query support

[Copilot]

Replaces naive substring matching with Bleve (https://blevesearch.com/) for the search API endpoint.

Changes

• Store: Added in-memory Bleve index, built during data load. Documents indexed with lowercased content using keyword analyzer for case-insensitive wildcard matching.

• Search: Wildcard queries (*query*) replace linear scans. Maintains pagination, sheet filtering, and API contract.

• Multi-Word Search: Query terms separated by spaces are combined with AND logic. For example, searching "悲伤 表情" (sad expression) will find text containing both "悲伤" AND "表情", such as "英雄可不能露出悲伤的表情啊".

• Document IDs: Use |@| delimiter format (sheet|@|rowId|@|index) to handle rowIds containing underscores or special characters.

CJK Language Support

Keyword analyzer preserves full text without tokenization, enabling wildcard substring matching across Chinese, Japanese, and Korean:

// Single-word queries
/api/search?lang=ja&q=ブレード  // Japanese
/api/search?lang=chs&q=战斗    // Chinese  
/api/search?lang=en&q=blade    // English

// Multi-word queries (AND logic)
/api/search?lang=chs&q=悲伤%20表情  // Finds text with both terms
/api/search?lang=en&q=sad%20expression  // Finds text with both terms

Performance

Index-based search vs. O(n) iteration over all items. Relevant for larger datasets.

Original prompt

搜索部分接入 https://blevesearch.com/

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[oott123]

@copilot 加入分词，比如搜索“悲伤 表情” 的时候要能搜出 “英雄可不能露出悲伤的表情啊”

[Copilot]

@copilot 加入分词，比如搜索“悲伤 表情” 的时候要能搜出 “英雄可不能露出悲伤的表情啊”

已实现。现在搜索 "悲伤 表情" 可以找到包含这两个词的文本（如 "英雄可不能露出悲伤的表情啊"）。多个词用空格分隔，会使用 AND 逻辑匹配。

Commit: 075e799