tampermonkey_export_lark_transcript.js

// ==UserScript==
// @name         Export Full Meeting Transcripts (Lark)
// @namespace    https://example.com
// @version      1.1
// @description  Export all visible and virtualized transcript data
// @match        https://*.larksuite.com/minutes/*
// @grant        none
// ==/UserScript==

(function () {
    'use strict';

    const PARAGRAPH_SELECTOR = '.paragraph-editor-wrapper';
    const SPEAKER_SELECTOR = '.p-user-name';
    const TIME_SELECTOR = '.p-time';
    const TEXT_LINE_SELECTOR = '.ace-line';

    const exportButton = document.createElement('button');
    exportButton.textContent = 'Export Full Transcripts';
    Object.assign(exportButton.style, {
        position: 'fixed',
        top: '10px',
        right: '10px',
        zIndex: 99999,
        padding: '6px 12px',
        backgroundColor: '#007bff',
        color: '#fff',
        border: 'none',
        borderRadius: '4px',
        cursor: 'pointer'
    });
    document.body.appendChild(exportButton);

    const seenParagraphs = new Set();

    function extractParagraphData(paragraph) {
        const key = paragraph.getAttribute('data-zone-id');
        if (key && seenParagraphs.has(key)) return null;
        if (key) seenParagraphs.add(key);

        const speakerEl = paragraph.querySelector(SPEAKER_SELECTOR);
        const timeEl = paragraph.querySelector(TIME_SELECTOR);
        const textEls = paragraph.querySelectorAll(TEXT_LINE_SELECTOR);

        const speaker = speakerEl?.getAttribute('user-name-content') || speakerEl?.textContent.trim() || '';
        const time = timeEl?.getAttribute('time-content') || timeEl?.textContent.trim() || '';

        let text = '';
        textEls.forEach(line => {
            text += line.innerText.trim() + ' ';
        });

        return {
            speaker,
            time,
            text: text.trim()
        };
    }

    async function scrollAndCollectTranscripts() {
        const container = document.querySelector('.rc-virtual-list-holder');
        if (!container) {
            alert("Transcript container not found.");
            return;
        }

        const results = [];
        let previousScrollTop = -1;

        while (true) {
            const paragraphs = document.querySelectorAll(PARAGRAPH_SELECTOR);
            for (const p of paragraphs) {
                const data = extractParagraphData(p);
                if (data) results.push(data);
            }

            container.scrollTop += 500;
            await new Promise(r => setTimeout(r, 300)); // wait for content to load

            // Check if we're at the bottom (no new scroll movement)
            if (container.scrollTop === previousScrollTop) break;
            previousScrollTop = container.scrollTop;
        }

        return results;
    }

    function downloadFile(filename, content) {
        const blob = new Blob([content], { type: 'application/json' });
        const url = URL.createObjectURL(blob);
        const link = document.createElement('a');
        link.href = url;
        link.download = filename;
        link.click();
        URL.revokeObjectURL(url);
    }

    exportButton.onclick = async () => {
        exportButton.textContent = 'Exporting...';
        const results = await scrollAndCollectTranscripts();
        downloadFile('full_transcript.json', JSON.stringify(results, null, 2));
        exportButton.textContent = 'Export Full Transcripts';
    };
})();