via script Tool , how to lookup for the data in the attachement of the KB

[Sundeep Kumar1]

Bellow is the script tool which i wrote and it works perfectly fine to extract the information from a kb article, however  sometime i am not able to get the requested information from the KB article which has an attachment ( PDF, word, csv, excel):

can someone please assist, how to extract information from kb article  which has an attachment ( PDF, word, csv, excel).

Script inputs 

Input name	Description
number	Knowledge article number

Script:
(function(inputs) {

    // only string inputs are allowed

    // return outputs object where the keys in it are understandable by LLM

    // Retrieve KB article

    var kb_gr = new GlideRecord('kb_knowledge');

    kb_gr.addQuery('number', inputs.number);

    kb_gr.orderByDesc('version');

    kb_gr.setLimit(1);

    kb_gr.query();

    if (kb_gr.next()) {

        return kb_gr.text.toString();

    } else {

        throw new Error('No KB found with number ' + inputs.number);

    }

})(inputs);

[rpriyadarshy]

  See if this heslps you . You Can tweak it as per ur need.

(function(inputs) {

    // Retrieve KB article (latest version by number)

    var kb_gr = new GlideRecord('kb_knowledge');

    kb_gr.addQuery('number', inputs.number);

    kb_gr.orderByDesc('version');

    kb_gr.setLimit(1);

    kb_gr.query();

    if (!kb_gr.next()) {

        throw new Error('No KB found with number ' + inputs.number);

    }

    // Start with KB article body

    var outputText = '';

    outputText += (kb_gr.text ? kb_gr.text.toString() : '');

    // ---- Attachments: read text content (or "Summary of : <file>.txt") ----

    var gsa = new GlideSysAttachment(); // Attachment API 

    var attGR = gsa.getAttachments('kb_knowledge', kb_gr.getUniqueValue()); // list attachments 

    var attachmentTextParts = [];

    var MAX_ATTACHMENTS_TO_READ = 10;         

    var MAX_CHARS_PER_ATTACHMENT = 20000;     

    var readCount = 0;

    while (attGR.next() && readCount < MAX_ATTACHMENTS_TO_READ) {

        var fileName = attGR.getValue('file_name') + '';

        var contentType = attGR.getValue('content_type') + '';

        // Skip the auto-generated summary files in the "original loop"

        // We'll include summaries when we either detect they are the file itself OR when we look them up.

        var isSummaryTxt = fileName.indexOf('Summary of :') === 0 && fileName.endsWith('.txt');

        var textContent = '';

        // 1) If attachment itself is readable text, read it directly via getContent()

        // getContent(sys_attachment GR) returns content as string 

        if (contentType.indexOf('text/') === 0 || fileName.match(/\.(txt|md|csv|json|xml|log)$/i) || isSummaryTxt) {

            textContent = gsa.getContent(attGR) + '';

        }

        else {

            // "Summary of : <originalFile>.txt" for readable binary docs (pdf/doc/xlsx etc.) 

            var summaryName = 'Summary of : ' + fileName + '.txt';

            var sumGR = new GlideRecord('sys_attachment');

            sumGR.addQuery('table_name', 'kb_knowledge');

            sumGR.addQuery('table_sys_id', kb_gr.getUniqueValue());

            sumGR.addQuery('file_name', summaryName);

            sumGR.setLimit(1);

            sumGR.query();

            if (sumGR.next()) {

                textContent = gsa.getContent(sumGR) + ''; // extracted text lives here 

            }

        }

        // Normalize + trim

        if (textContent) {

            textContent = textContent.replace(/\u0000/g, ''); // remove nulls if any

            if (textContent.length > MAX_CHARS_PER_ATTACHMENT) {

                textContent = textContent.substring(0, MAX_CHARS_PER_ATTACHMENT) +

                    '\n...[truncated to ' + MAX_CHARS_PER_ATTACHMENT + ' chars]';

            }

            attachmentTextParts.push(

                '\n\n--- Attachment: ' + fileName + ' (' + contentType + ') ---\n' + textContent

            );

            readCount++;

        }

    }

    if (attachmentTextParts.length > 0) {

        outputText += '\n\n====================\nATTACHMENTS CONTENT\n====================';

        outputText += attachmentTextParts.join('');

    }

    return outputText;

})(inputs);

Regards

RP