Read Excel Content

const XLSX = require("xlsx");

/*

 * Reads content from a specified or latest downloaded Excel file.

 *

 * @param {string|null} [fileName=null] - The name of the Excel file.

*If null, the latest downloaded file will be used.

 * @param {number|null} [rowToCheck=null] - The row number in the

 *Excel content to retrieve (1-based index). If null, all rows will

 *be returned.

 * @returns {Promise<Object|Object[]>} - The content of the

*specified row or all rows if rowToCheck is not provided.

 */

async function getExcelContent(fileName = null, rowToCheck = null) {

    // Define the download directory path

    const downloadDir = CONSTANTS.FILE.FILE_PATH;

    let excelFilePath;

    if (fileName !== null) {

        // Use the provided filename to construct the file path

        excelFilePath = `${downloadDir}/${fileName}`;

        console.log(`Reading content from the specified file:

                        ${fileName}`);

    } else {

        // Get the latest downloaded file name

        const latestDownloadedFileName =

                    await this.getLatestDownloadedFileName();

      // Construct the full path of the latest downloaded Excel file

        excelFilePath = `${downloadDir}/${latestDownloadedFileName}`;

        console.log(

            `Reading content from the latest downloaded file:

                        ${latestDownloadedFileName}`

        );

    }

    // Read the Excel file

    const workbook = XLSX.readFile(excelFilePath);

    const sheetName = workbook.SheetNames[0]; // Get the first sheet

    const sheet = workbook.Sheets[sheetName];

    const data = XLSX.utils.sheet_to_json(sheet);

    // Convert sheet to JSON

    // Print each row of the Excel content in a prettier format

    console.log(`Content of the file ${fileName ||

                                "latest downloaded Excel"}:`);

    data.forEach((row, index) => {

        console.log(`Row ${index + 1}:`);

        for (const [key, value] of Object.entries(row)) {

            console.log(`  ${key}: ${value}`);

        }

        console.log(""); // Add a blank line between rows for

                // readability

    });

    // Return data based on the specified row or all rows

    //if rowToCheck is not provided

    if (rowToCheck !== null) {

        if (rowToCheck < 1 || rowToCheck > data.length) {

           console.warn(`Row number ${rowToCheck} is out of range.`);

           return {}; // Return an empty object if the row number

                // is out of range

        }

        return data[rowToCheck - 1]; // Return the specified row

    }

    return data; // Return all rows

}

Get Pdf Page Counts

 

const pdfParse = require("pdf-parse"); 

* Retrieves the total number of pages in a specified or latest

downloaded PDF file.

 *

 * @param {string|null} [fileName=null] - The name of the PDF file.

*If null, the latest downloaded file will be used.

 * @returns {Promise<number>} - The total number of pages in the

*PDF file.

 */

async function getPdfPageCount(fileName = null) {

    // Define the download directory path

    const downloadDir = CONSTANTS.FILE.FILE_PATH;

    let pdfFilePath;

    if (fileName !== null) {

        // Use the provided filename to construct the file path

        pdfFilePath = `${downloadDir}/${fileName}`;

        console.log(`Getting page count from the specified file:

                        ${fileName}`);

    } else {

        // Get the latest downloaded file name

        const latestDownloadedFileName = await

                        this.getLatestDownloadedFileName();

        // Construct the full path of the latest downloaded PDF file

        pdfFilePath = `${downloadDir}/${latestDownloadedFileName}`;

        console.log(

            `Getting page count from the latest downloaded file:

                            ${latestDownloadedFileName}`

        );

    }

    // Read the PDF file

    const dataBuffer = fs.readFileSync(pdfFilePath);

    // Parse the PDF and extract data

    const data = await pdfParse(dataBuffer);

    // Return the total number of pages

    return data.numpages; // This will return the number of pages

    // starting from 1

}

Get Pdf Text Content Line Based

 

const pdfParse = require("pdf-parse"); 

* Reads the text content from a specified or latest downloaded

PDF file.

 *

 * @param {string|null} [fileName=null] - The name of the PDF file.

If null, the latest downloaded file will be used.

 * @param {number|null} [lineToCheck=null] - The line number in the

PDF content to retrieve (1-based index). If null, all lines will

be returned.

 * @returns {Promise<string|string[]>} - The text content of

the specified line or all lines if lineToCheck is not provided.

 */

async function getPdfTextContentLineBased(fileName = null,

                lineToCheck = null) {

    // Define the download directory path

    const downloadDir = CONSTANTS.FILE.FILE_PATH;

    let pdfFilePath;

    if (fileName !== null) {

        // Use the provided filename to construct the file path

        pdfFilePath = `${downloadDir}/${fileName}`;

        console.log(`Reading content from the specified file:

        ${fileName}`);

    } else {

        // Get the latest downloaded file name

        const latestDownloadedFileName =

                await this.getLatestDownloadedFileName();

        // Construct the full path of the latest downloaded PDF file

        pdfFilePath = `${downloadDir}/${latestDownloadedFileName}`;

        console.log(

            `Reading content from the latest downloaded file:

            ${latestDownloadedFileName}`

        );

    }

    // Read the PDF file

    const dataBuffer = fs.readFileSync(pdfFilePath);

    // Parse the PDF and extract text

    const data = await pdfParse(dataBuffer);

    // Split text into lines

    const lines = data.text.split("\n");

    // Print all lines of the PDF text content

    console.log(`Content of the file ${fileName ||

                    "latest downloaded PDF"}:`);

    console.table(

        lines.map((line, index) => ({ Line: index + 1,

                        Content: line.trim() }))

    );

    // Return text based on the specified line or all lines

if lineToCheck is not provided

    if (lineToCheck !== null) {

        if (lineToCheck < 1 || lineToCheck > lines.length) {

            console.warn(`Line number ${lineToCheck} is out of

                                           range.`);

            return ""; // Return an empty string if

the line number is out of range

        }

        return lines[lineToCheck - 1].trim(); // Return the

specified line

    }

    return lines.map((line) => line.trim()); // Return all lines

}