import pdfParse from "pdf-parse";
import { ExtractedWebpage } from "../cache/WebpageCache";
import { normalizeWhitespace } from "./htmlExtractor";

export interface PdfExtractResult {
  text: string;
  pageCount: number;
  info: {
    title?: string;
    author?: string;
    creationDate?: string;
  };
}

/**
 * Evaluates whether a buffer starts with the PDF magic byte signature (%PDF-).
 */
export function isPdfBuffer(buffer: Buffer): boolean {
  if (!buffer || buffer.length < 5) return false;
  const header = buffer.subarray(0, 1024).toString("latin1");
  return header.includes("%PDF-");
}

/**
 * Battle-tested digital PDF text extractor powered by pdf-parse / Mozilla pdfjs-dist.
 * Handles embedded fonts, CMap/ToUnicode tables, object streams, XRef tables, and all compression filters.
 */
export async function parsePdfBuffer(buffer: Buffer): Promise<PdfExtractResult> {
  if (!buffer || buffer.length === 0) {
    throw new Error("Empty PDF buffer");
  }

  if (!isPdfBuffer(buffer)) {
    throw new Error("Invalid PDF file (missing %PDF- header signature)");
  }

  try {
    const data = await pdfParse({ data: new Uint8Array(buffer) } as any, {
      pagerender: (pageData: any) => {
        return pageData.getTextContent().then((textContent: any) => {
          let lastY: number | null = null;
          let text = "";
          for (const item of textContent.items) {
            if (lastY !== null && Math.abs(item.transform[5] - lastY) > 5) {
              text += "\n";
            } else if (text.length > 0 && !text.endsWith(" ") && !text.endsWith("\n")) {
              text += " ";
            }
            text += item.str;
            lastY = item.transform[5];
          }
          return text;
        });
      },
    });

    const info: { title?: string; author?: string; creationDate?: string } = {};
    if (data.info) {
      if (data.info.Title) info.title = String(data.info.Title);
      if (data.info.Author) info.author = String(data.info.Author);
      if (data.info.CreationDate) info.creationDate = String(data.info.CreationDate);
    }

    const cleanText = data.text ? data.text.trim() : "";
    if (!cleanText) {
      throw new Error("No readable text content streams found in digital PDF");
    }

    return {
      text: cleanText,
      pageCount: data.numpages || 1,
      info,
    };
  } catch (err) {
    const msg = err instanceof Error ? err.message : String(err);
    throw new Error(`PDF parsing failed: ${msg}`);
  }
}

/**
 * Standard PDF Extractor integration into Browse ExtractedWebpage schema.
 */
export async function extractPdfContent(buffer: Buffer, url: string): Promise<ExtractedWebpage> {
  const rawSizeBytes = buffer.length;
  const pdfRes = await parsePdfBuffer(buffer);

  const cleanBody = normalizeWhitespace(pdfRes.text);
  const title = pdfRes.info.title || "PDF Document";

  const headerLines: string[] = [`# ${title}`];
  const metaDetails: string[] = [];

  if (pdfRes.info.author) {
    metaDetails.push(`**Author:** ${pdfRes.info.author}`);
  }
  metaDetails.push(`**Pages:** ${pdfRes.pageCount}`);

  if (metaDetails.length > 0) {
    headerLines.push(metaDetails.join(" | "));
  }

  const fullMarkdown = normalizeWhitespace(`${headerLines.join("\n\n")}\n\n${cleanBody}`);

  return {
    url,
    metadata: {
      title: pdfRes.info.title || "PDF Document",
      author: pdfRes.info.author,
      publishedDate: pdfRes.info.creationDate,
    },
    fullMarkdown,
    totalChars: fullMarkdown.length,
    isSparse: false,
    rawHtmlLength: rawSizeBytes,
    readabilityLength: cleanBody.length,
    usedJsonLd: false,
    usedOg: false,
    tablesDetected: fullMarkdown.includes("| ---") || fullMarkdown.includes("| --- |"),
    pdfPageCount: pdfRes.pageCount,
    imageMarkers: [],
  };
}

import pdfParse from "pdf-parse";
import { ExtractedWebpage } from "../cache/WebpageCache";
import { normalizeWhitespace } from "./htmlExtractor";

export interface PdfExtractResult {
  text: string;
  pageCount: number;
  info: {
    title?: string;
    author?: string;
    creationDate?: string;
  };
}

/**
 * Evaluates whether a buffer starts with the PDF magic byte signature (%PDF-).
 */
export function isPdfBuffer(buffer: Buffer): boolean {
  if (!buffer || buffer.length < 5) return false;
  const header = buffer.subarray(0, 1024).toString("latin1");
  return header.includes("%PDF-");
}

/**
 * Battle-tested digital PDF text extractor powered by pdf-parse / Mozilla pdfjs-dist.
 * Handles embedded fonts, CMap/ToUnicode tables, object streams, XRef tables, and all compression filters.
 */
export async function parsePdfBuffer(buffer: Buffer): Promise<PdfExtractResult> {
  if (!buffer || buffer.length === 0) {
    throw new Error("Empty PDF buffer");
  }

  if (!isPdfBuffer(buffer)) {
    throw new Error("Invalid PDF file (missing %PDF- header signature)");
  }

  try {
    const data = await pdfParse({ data: new Uint8Array(buffer) } as any, {
      pagerender: (pageData: any) => {
        return pageData.getTextContent().then((textContent: any) => {
          let lastY: number | null = null;
          let text = "";
          for (const item of textContent.items) {
            if (lastY !== null && Math.abs(item.transform[5] - lastY) > 5) {
              text += "\n";
            } else if (text.length > 0 && !text.endsWith(" ") && !text.endsWith("\n")) {
              text += " ";
            }
            text += item.str;
            lastY = item.transform[5];
          }
          return text;
        });
      },
    });

    const info: { title?: string; author?: string; creationDate?: string } = {};
    if (data.info) {
      if (data.info.Title) info.title = String(data.info.Title);
      if (data.info.Author) info.author = String(data.info.Author);
      if (data.info.CreationDate) info.creationDate = String(data.info.CreationDate);
    }

    const cleanText = data.text ? data.text.trim() : "";
    if (!cleanText) {
      throw new Error("No readable text content streams found in digital PDF");
    }

    return {
      text: cleanText,
      pageCount: data.numpages || 1,
      info,
    };
  } catch (err) {
    const msg = err instanceof Error ? err.message : String(err);
    throw new Error(`PDF parsing failed: ${msg}`);
  }
}

/**
 * Standard PDF Extractor integration into Browse ExtractedWebpage schema.
 */
export async function extractPdfContent(buffer: Buffer, url: string): Promise<ExtractedWebpage> {
  const rawSizeBytes = buffer.length;
  const pdfRes = await parsePdfBuffer(buffer);

  const cleanBody = normalizeWhitespace(pdfRes.text);
  const title = pdfRes.info.title || "PDF Document";

  const headerLines: string[] = [`# ${title}`];
  const metaDetails: string[] = [];

  if (pdfRes.info.author) {
    metaDetails.push(`**Author:** ${pdfRes.info.author}`);
  }
  metaDetails.push(`**Pages:** ${pdfRes.pageCount}`);

  if (metaDetails.length > 0) {
    headerLines.push(metaDetails.join(" | "));
  }

  const fullMarkdown = normalizeWhitespace(`${headerLines.join("\n\n")}\n\n${cleanBody}`);

  return {
    url,
    metadata: {
      title: pdfRes.info.title || "PDF Document",
      author: pdfRes.info.author,
      publishedDate: pdfRes.info.creationDate,
    },
    fullMarkdown,
    totalChars: fullMarkdown.length,
    isSparse: false,
    rawHtmlLength: rawSizeBytes,
    readabilityLength: cleanBody.length,
    usedJsonLd: false,
    usedOg: false,
    tablesDetected: fullMarkdown.includes("| ---") || fullMarkdown.includes("| --- |"),
    pdfPageCount: pdfRes.pageCount,
    imageMarkers: [],
  };
}