import { z } from 'zod';

/**
 * Search schemas and views (issue #49/#50, ADR 0010). Full-text search runs on
 * PostgreSQL behind the `SearchProvider` interface. Diacritic-insensitive
 * matching ('Baume' finds 'Bäume') is achieved by normalizing both the indexed
 * text and the query in application code — so no Postgres `unaccent` extension
 * is required (which keeps the schema-pushed test databases working).
 */

// Combining diacritical marks (U+0300–U+036F), removed after NFKD decomposition.
const COMBINING_MARKS = /[̀-ͯ]/g;

/**
 * Fold a string for search: strip diacritics (NFKD + drop combining marks) and
 * lowercase, leaving tokenization to Postgres `to_tsvector('simple', …)`. Used
 * for BOTH the stored search vector and the query, so they always agree.
 */
export function normalizeForSearch(text: string): string {
  return text.normalize('NFKD').replace(COMBINING_MARKS, '').toLowerCase();
}

/** Sentinels wrapping a matched span in a result snippet (`ts_headline`), split
 * by the UI to render highlights without trusting HTML from the content.
 * Private-use codepoints that never occur in page text. */
export const SEARCH_HIGHLIGHT_START = String.fromCodePoint(0xe000);
export const SEARCH_HIGHLIGHT_END = String.fromCodePoint(0xe001);

/** How many results one search returns (v1, ADR 0010 target scale). */
export const SEARCH_RESULT_LIMIT = 30;

export const searchQuerySchema = z.object({
  q: z.string().trim().min(1, 'validation.required').max(200, 'validation.tooLong'),
  /** Restrict to one pond; omitted = all ponds the user may read. */
  pondId: z.string().min(1).optional(),
  /** Restrict to pages carrying any of these label ids. */
  labels: z.array(z.string().min(1)).optional(),
});
export type SearchQuery = z.infer<typeof searchQuerySchema>;

/** One hit: the page, its pond, its labels, and a highlighted snippet. */
export interface SearchResultView {
  pageId: string;
  title: string;
  slug: string;
  pondId: string;
  pondSlug: string;
  pondName: string;
  labelIds: string[];
  /** Snippet with matches wrapped in the highlight sentinels above. */
  snippet: string;
}