import { SearchResult } from "../../models/SearchResult";
import { InternalProviderResult } from "../../providers/types/InternalProviderResult";
import { HtmlCleaner } from "./HtmlCleaner";
import { TextNormalizer } from "./TextNormalizer";
import { UrlValidator } from "./UrlValidator";

/**
 * Engine-owned immutable builder for canonical SearchResult (Discovery Object) instances.
 * Guarantees that there is EXACTLY ONE PLACE in the codebase where public SearchResult instances are constructed.
 *
 * Rules:
 * - Immutably maps InternalProviderResult to brand new SearchResult.
 * - Strips HTML/XML tags and decodes entities.
 * - Normalizes whitespace.
 * - Validates URL and title requirements.
 * - Truncates snippet at maxSnippetLength without inventing synthetic text.
 * - Omits internal raw payloads (rawItem) from the returned public SearchResult.
 */
export class CanonicalSearchResultBuilder {
  /**
   * Immutably constructs a canonical SearchResult from an InternalProviderResult candidate.
   *
   * @param candidate Internal candidate item returned by search provider extractors.
   * @param maxSnippetLength Configured maximum snippet character limit (default: 400).
   * @returns Clean, canonical SearchResult, or undefined if title or URL are invalid.
   */
  public static build(
    candidate: InternalProviderResult,
    maxSnippetLength: number = 400
  ): SearchResult | undefined {
    if (!candidate) return undefined;

    // 1. Validate & Normalize URL
    const cleanUrl = UrlValidator.validateUrl(candidate.url);
    if (!cleanUrl) {
      return undefined;
    }

    // 2. Clean & Normalize Title
    const rawTitle = candidate.title ?? "";
    const cleanTitle = TextNormalizer.truncate(
      TextNormalizer.normalizeWhitespace(HtmlCleaner.stripHtml(rawTitle)),
      200
    );
    if (!cleanTitle) {
      return undefined;
    }

    // 3. Clean & Truncate Snippet (Non-inventive rule: empty candidate remains empty)
    const rawSnippet = candidate.snippetCandidate ?? "";
    const cleanSnippet = TextNormalizer.truncate(
      TextNormalizer.normalizeWhitespace(HtmlCleaner.stripHtml(rawSnippet)),
      maxSnippetLength
    );

    // 4. Construct Fresh Canonical SearchResult (Omit raw / internal data)
    return {
      title: cleanTitle,
      url: cleanUrl,
      snippet: cleanSnippet,
      source: candidate.source ? candidate.source.toLowerCase().trim() : "unknown",
      ...(candidate.publishedDate ? { publishedDate: candidate.publishedDate } : {}),
      ...(typeof candidate.score === "number" ? { score: candidate.score } : {}),
    };
  }
}

import { SearchResult } from "../../models/SearchResult";
import { InternalProviderResult } from "../../providers/types/InternalProviderResult";
import { HtmlCleaner } from "./HtmlCleaner";
import { TextNormalizer } from "./TextNormalizer";
import { UrlValidator } from "./UrlValidator";

/**
 * Engine-owned immutable builder for canonical SearchResult (Discovery Object) instances.
 * Guarantees that there is EXACTLY ONE PLACE in the codebase where public SearchResult instances are constructed.
 *
 * Rules:
 * - Immutably maps InternalProviderResult to brand new SearchResult.
 * - Strips HTML/XML tags and decodes entities.
 * - Normalizes whitespace.
 * - Validates URL and title requirements.
 * - Truncates snippet at maxSnippetLength without inventing synthetic text.
 * - Omits internal raw payloads (rawItem) from the returned public SearchResult.
 */
export class CanonicalSearchResultBuilder {
  /**
   * Immutably constructs a canonical SearchResult from an InternalProviderResult candidate.
   *
   * @param candidate Internal candidate item returned by search provider extractors.
   * @param maxSnippetLength Configured maximum snippet character limit (default: 400).
   * @returns Clean, canonical SearchResult, or undefined if title or URL are invalid.
   */
  public static build(
    candidate: InternalProviderResult,
    maxSnippetLength: number = 400
  ): SearchResult | undefined {
    if (!candidate) return undefined;

    // 1. Validate & Normalize URL
    const cleanUrl = UrlValidator.validateUrl(candidate.url);
    if (!cleanUrl) {
      return undefined;
    }

    // 2. Clean & Normalize Title
    const rawTitle = candidate.title ?? "";
    const cleanTitle = TextNormalizer.truncate(
      TextNormalizer.normalizeWhitespace(HtmlCleaner.stripHtml(rawTitle)),
      200
    );
    if (!cleanTitle) {
      return undefined;
    }

    // 3. Clean & Truncate Snippet (Non-inventive rule: empty candidate remains empty)
    const rawSnippet = candidate.snippetCandidate ?? "";
    const cleanSnippet = TextNormalizer.truncate(
      TextNormalizer.normalizeWhitespace(HtmlCleaner.stripHtml(rawSnippet)),
      maxSnippetLength
    );

    // 4. Construct Fresh Canonical SearchResult (Omit raw / internal data)
    return {
      title: cleanTitle,
      url: cleanUrl,
      snippet: cleanSnippet,
      source: candidate.source ? candidate.source.toLowerCase().trim() : "unknown",
      ...(candidate.publishedDate ? { publishedDate: candidate.publishedDate } : {}),
      ...(typeof candidate.score === "number" ? { score: candidate.score } : {}),
    };
  }
}