/**
 * Code-Aware Chunking Engine
 *
 * Implements heuristic-based code-aware chunking that splits code at semantic
 * boundaries (functions, classes, methods) rather than fixed character counts.
 *
 * Benefits:
 * - Chunks align with code structure (no mid-function splits)
 * - Search results are more coherent and complete
 * - Reduced overlap requirement (from 20% to ~5%)
 *
 * Supported languages (20+):
 * - TypeScript/JavaScript: function declarations, class declarations, exports
 * - Python: def statements, class statements, decorators
 * - Java: class, interface, enum, method declarations
 * - Go: func, type struct, type interface declarations
 * - Rust: fn, struct, enum, impl, trait, macro declarations
 * - C#: class, struct, interface, enum, method declarations
 * - C/C++: function definitions, struct, class, namespace
 * - Kotlin: fun, class, interface, object declarations
 * - Swift: func, class, struct, enum, protocol, extension
 * - Ruby: def, class, module declarations
 * - PHP: function, class, interface, trait declarations
 * - Scala: def, class, trait, object declarations
 * - Shell/Bash: function declarations
 * - CSS/SCSS/LESS: rule blocks, @media, @keyframes
 * - HTML: structural elements
 * - Vue/Svelte: template, script, style blocks
 * - SQL: CREATE, ALTER, SELECT statements
 * - YAML/JSON/XML: structural markers
 * - GraphQL: type, query, mutation, subscription
 * - Terraform/HCL: resource, variable, output blocks
 * - Dockerfile: FROM, RUN, COPY instructions
 *
 * Falls back to character-based chunking for unsupported languages or on errors.
 *
 * @module codeAwareChunking
 */
import { ChunkWithLines } from './chunking.js';
/**
 * Supported programming languages for code-aware chunking (22 languages)
 *
 * Tier 1 - High Priority:
 * - typescript, javascript, python (existing)
 * - java, go, rust, csharp, c, cpp, kotlin, swift
 *
 * Tier 2 - Medium Priority:
 * - ruby, php, scala, shell
 *
 * Tier 3 - Markup/Config:
 * - css, scss, less, html, vue, svelte, sql, yaml, json, xml, graphql
 *
 * Tier 4 - Infrastructure:
 * - terraform, hcl, dockerfile
 */
export type SupportedLanguage = 'typescript' | 'javascript' | 'python' | 'java' | 'go' | 'rust' | 'csharp' | 'c' | 'cpp' | 'kotlin' | 'swift' | 'ruby' | 'php' | 'scala' | 'shell' | 'css' | 'scss' | 'less' | 'html' | 'vue' | 'svelte' | 'sql' | 'yaml' | 'json' | 'xml' | 'graphql' | 'terraform' | 'hcl' | 'dockerfile' | 'unknown';
/**
 * Configuration for code-aware chunking
 */
export interface CodeAwareChunkOptions {
    /** Target chunk size in characters (~4000 for ~1000 tokens) */
    chunkSize: number;
    /** Minimum overlap size in characters (~200 for ~50 tokens - reduced from 800) */
    chunkOverlap: number;
    /** Maximum chunk size before forcing a split */
    maxChunkSize: number;
}
/**
 * Default options for code-aware chunking
 *
 * Reduced overlap compared to character-based chunking since we're
 * splitting at semantic boundaries.
 */
export declare const DEFAULT_CODE_AWARE_OPTIONS: CodeAwareChunkOptions;
/**
 * Detect the programming language from a file path
 *
 * @param filePath - Path to the file (can be relative or absolute)
 * @returns The detected language or 'unknown'
 */
export declare function detectLanguage(filePath: string): SupportedLanguage;
/**
 * Split code at semantic boundaries with line number tracking
 *
 * Algorithm:
 * 1. Detect language from file path
 * 2. Find all semantic boundaries (functions, classes, etc.)
 * 3. Group boundaries into chunks that fit within the size limit
 * 4. If a single semantic unit exceeds maxChunkSize, split it at line boundaries
 * 5. Add minimal overlap at chunk boundaries
 *
 * @param text - Source code text
 * @param filePath - Path to the file (for language detection)
 * @param options - Chunking options
 * @returns Array of chunks with line number information
 */
export declare function splitCodeWithLineNumbers(text: string, filePath: string, options?: Partial<CodeAwareChunkOptions>): ChunkWithLines[];
/**
 * Check if a file can be processed with code-aware chunking
 *
 * @param filePath - Path to the file
 * @returns true if the file's language is supported
 */
export declare function supportsCodeAwareChunking(filePath: string): boolean;
/**
 * Get the language name for a file path
 *
 * @param filePath - Path to the file
 * @returns Human-readable language name
 */
export declare function getLanguageName(filePath: string): string;
/**
 * Get the human-readable display name for a language
 *
 * @param language - Language identifier
 * @returns Human-readable name
 */
export declare function getLanguageDisplayName(language: SupportedLanguage): string;
/**
 * Get a list of all supported languages
 *
 * @returns Array of supported language identifiers (excluding 'unknown')
 */
export declare function getSupportedLanguages(): SupportedLanguage[];
//# sourceMappingURL=codeAwareChunking.d.ts.map