import { type FileUpload, type UploadScannerOptions } from '../security/upload-scanner.js';
import { type IngestionOptions, type IngestionResult } from './ingestion.js';
/** Turns one kind of file into text, such as PDF or an image through OCR. */
export interface FileTextExtractor {
    /** Whether it handles this file. */
    supports(file: FileUpload): boolean;
    /** Extracts the file's text. */
    extract(file: FileUpload): Promise<string> | string;
}
/** Options for `ingestFilesAfterScan()`. */
export interface FileIngestionOptions extends IngestionOptions {
    /** Upload scan settings. Every file is scanned before any is read. */
    scan?: UploadScannerOptions;
    /** Extractors tried in order. Text files need none. */
    extractors?: FileTextExtractor[];
}
/** The chunks produced from a set of files, and the files that could not be read. */
export interface FileIngestionResult extends IngestionResult {
    /** Files scanned. */
    scannedFiles: number;
    /** Files skipped because no extractor handles them. */
    skippedFiles: Array<{
        name: string;
        reason: string;
    }>;
}
/**
 * Scans uploads, extracts their text, and splits it into chunks. Throws when the scan finds
 * anything `high` or `critical`.
 */
export declare function ingestFilesAfterScan(files: FileUpload[], options?: FileIngestionOptions): Promise<FileIngestionResult>;
/** An extractor for PDF files, around your own PDF-to-text function. */
export declare function createPdfExtractor(extractPdfText: (file: FileUpload) => Promise<string> | string): FileTextExtractor;
/** An extractor for images, around your own OCR function. Defaults to PNG, JPEG, and WebP. */
export declare function createOcrExtractor(ocrImage: (file: FileUpload) => Promise<string> | string, mimeTypes?: string[]): FileTextExtractor;