script.js

document.addEventListener('DOMContentLoaded', () => {
    // UI Elements
    const inputText = document.getElementById('input-text');
    const outputText = document.getElementById('output-text');
    const processBtn = document.getElementById('process-btn');
    const processBtnMobile = document.getElementById('process-btn-mobile');
    const copyBtn = document.getElementById('copy-btn');
    const clearBtn = document.getElementById('clear-btn');
    const blockCount = document.getElementById('block-count');
    const downloadSrtBtn = document.getElementById('download-srt-btn');
    
    // Mode Toggle Elements
    const modeTextBtn = document.getElementById('mode-text');
    const modeAudioBtn = document.getElementById('mode-audio');
    const audioSection = document.getElementById('audio-section');
    const inputSection = document.querySelector('section'); // First section is input
    
    // Audio Elements
    const audioFileInput = document.getElementById('audio-file');
    const apiKeyInput = document.getElementById('api-key');
    const transcribeBtn = document.getElementById('transcribe-btn');
    const fileLabel = document.getElementById('file-label');
    const transcribingStatus = document.getElementById('transcribing-status');

    // State
    let currentSrtData = null;
    let currentTranscriptWords = []; // Stores {word, start, end}
// --- CONSTANTS ---
    const MAX_CHARS = 11;
    
    // Tabu words (Articles, Prepositions, Pronouns, Conjunctions)
    const TABU_WORDS = new Set([
        // PT
        'de', 'do', 'da', 'em', 'no', 'na', 'por', 'para', 'com', 'ao', 'a', 'os', 'as', 
        'e', 'ou', 'mas', 'que', 'se', 'como', 'porque', 'quando', 'então', 'o', 'um', 'uma', 
        'nos', 'nas', 'me', 'te', 'não', 'só', 'já', 'também', 'tipo', 'nosso', 'nossa',
        // EN
        'a', 'an', 'the', 'of', 'in', 'on', 'at', 'to', 'for', 'with', 'from', 'by', 'and', 'or', 'but',
        // ES
        'el', 'la', 'los', 'las', 'un', 'una', 'de', 'en', 'a', 'por', 'para', 'con', 'y', 'o', 'pero', 'que'
    ]);

    const CONNECTIVES = new Set(['e', 'é', 'que', 'and', 'that', 'y']); // Must start new line
    // --- MODE SWITCHING ---
    modeTextBtn.addEventListener('click', () => {
        modeTextBtn.classList.add('mode-active');
        modeAudioBtn.classList.remove('mode-active');
        audioSection.classList.add('hidden');
        inputSection.classList.remove('opacity-50', 'pointer-events-none');
        downloadSrtBtn.classList.add('hidden');
        outputText.value = '';
        blockCount.textContent = '0';
        currentSrtData = null;
    });

    modeAudioBtn.addEventListener('click', () => {
        modeAudioBtn.classList.add('mode-active');
        modeTextBtn.classList.remove('mode-active');
        audioSection.classList.remove('hidden');
        // Optional: Disable manual text input when in audio mode to avoid confusion
        inputSection.classList.add('opacity-50', 'pointer-events-none');
    });

    audioFileInput.addEventListener('change', (e) => {
        if (e.target.files.length > 0) {
            fileLabel.textContent = e.target.files[0].name;
        } else {
            fileLabel.textContent = 'Click or drag audio file here';
        }
    });

    // --- AUDIO TRANSCRIPTION LOGIC ---

    async function handleTranscription() {
        const file = audioFileInput.files[0];
        const apiKey = apiKeyInput.value.trim();

        if (!file) {
            showToast("Please select an audio or video file.", "error");
            return;
        }
        if (!apiKey) {
            showToast("Please enter your OpenAI API Key.", "error");
            return;
        }

        // UI Loading State
        transcribeBtn.disabled = true;
        transcribingStatus.classList.remove('hidden');
        
        try {
            const formData = new FormData();
            formData.append('file', file);
            formData.append('model', 'whisper-1');
            formData.append('response_format', 'verbose_json');
            formData.append('timestamp_granularities', 'word');

            const response = await fetch('https://api.openai.com/v1/audio/transcriptions', {
                method: 'POST',
                headers: {
                    'Authorization': `Bearer ${apiKey}`
                },
                body: formData
            });

            if (!response.ok) {
                const errData = await response.json();
                throw new Error(errData.error?.message || 'Transcription failed');
            }

            const data = await response.json();
            
            // Extract words with timestamps
            // Whisper returns words array when timestamp_granularities is set
            if (!data.words) {
                throw new Error("No word-level timestamps returned. Check API plan.");
            }

            currentTranscriptWords = data.words.map(w => ({
                word: w.word,
                start: w.start,
                end: w.end
            }));

            // Get full text
            const fullText = data.text;

            // Process the text with existing logic
            const processedBlocks = getProcessedBlocks(fullText);
            
            // Map timestamps to processed blocks
            const srtContent = generateSRT(processedBlocks, currentTranscriptWords);
            
            // Display Results
            outputText.value = srtContent; // Show SRT format in textarea or just text? Let's show SRT content so they can see timing
            blockCount.textContent = processedBlocks.length;
            currentSrtData = srtContent;
            downloadSrtBtn.classList.remove('hidden');
            
            showToast("Transcription & Alignment complete!");

        } catch (error) {
            console.error(error);
            showToast(error.message, "error");
        } finally {
            transcribeBtn.disabled = false;
            transcribingStatus.classList.add('hidden');
        }
    }

    transcribeBtn.addEventListener('click', handleTranscription);

    // --- SRT GENERATION ---

    function formatSRTTime(seconds) {
        const date = new Date(0);
        date.setMilliseconds(seconds * 1000);
        const isoString = date.toISOString();
        // Extract HH:MM:SS,ms
        return isoString.substr(11, 8) + ',' + isoString.substr(20, 3);
    }

    function generateSRT(blocks, words) {
        let srtOutput = "";
        let wordIndex = 0;
        let blockIndex = 1;

        // Normalize punctuation in blocks to match Whisper words (roughly)
        // Whisper usually returns words without punctuation attached, or with basic punctuation
        // Our script adds "!" for commas.
        
        for (const block of blocks) {
            // Split block into words (removing our added punctuation for matching)
            // We need to reconstruct the text for display but match based on content
            
            // Simple approach: Count words in the block
            // Calculate how many whisper words correspond to this block text
            const blockWords = block.replace(/[!?.,]/g, '').trim().split(/\s+/).filter(w => w.length > 0);
            const numWords = blockWords.length;
            
            if (numWords === 0) continue;

            if (wordIndex >= words.length) break;

            // Determine start and end time
            // Start is the start of the first word in this chunk
            const startTime = words[wordIndex].start;
            
            // End is the end of the last word in this chunk
            // Look ahead 'numWords - 1'
            let endIndex = wordIndex + numWords - 1;
            if (endIndex >= words.length) endIndex = words.length - 1;
            const endTime = words[endIndex].end;

            // Format Entry
            srtOutput += `${blockIndex}\n`;
            srtOutput += `${formatSRTTime(startTime)} --> ${formatSRTTime(endTime)}\n`;
            srtOutput += `${block}\n\n`;

            wordIndex += numWords;
            blockIndex++;
        }

        return srtOutput;
    }

    downloadSrtBtn.addEventListener('click', () => {
        if (!currentSrtData) return;
        const blob = new Blob([currentSrtData], { type: 'text/plain' });
        const url = URL.createObjectURL(blob);
        const a = document.createElement('a');
        a.href = url;
        a.download = 'capcut_aligned.srt';
        document.body.appendChild(a);
        a.click();
        document.body.removeChild(a);
        URL.revokeObjectURL(url);
        showToast("SRT file downloaded!");
    });

    // --- UTILITIES ---

    function showToast(message, type = 'success') {
const toast = document.createElement('div');
        const bgColor = type === 'success' ? 'bg-green-500' : 'bg-red-500';
        toast.className = `fixed bottom-5 right-5 ${bgColor} text-white px-6 py-3 rounded-lg shadow-lg flex items-center gap-2 z-50 toast`;
        toast.innerHTML = `<i data-feather="${type === 'success' ? 'check-circle' : 'alert-circle'}"></i> ${message}`;
        document.body.appendChild(toast);
        feather.replace();

        setTimeout(() => {
            toast.classList.add('hiding');
            toast.addEventListener('animationend', () => toast.remove());
        }, 3000);
    }

    function cleanText(text) {
        // 1. Remove duplicate spaces
        let cleaned = text.replace(/\s+/g, ' ').trim();
        
        // 2. Comma Transformation: Replace , with ! attached to previous word
        cleaned = cleaned.replace(/,/g, '!');
        
        // 3. Normalize multiple exclamation marks
        cleaned = cleaned.replace(/!+/g, '!');
        
        return cleaned;
    }

    function isWeakEnding(word) {
        if (!word) return false;
        // Remove trailing punctuation for the check, but usually in this flow 
        // the word carries the punctuation from the split logic.
        // We check the alphanumeric part length.
        const cleanWord = word.replace(/[!?.,]/g, '').toLowerCase();
        
        // 2-3 Letter Ban
        if (cleanWord.length <= 3) return true;
        
        // Tabu Words
        if (TABU_WORDS.has(cleanWord)) return true;
        
        return false;
    }

    function countCharsNoSpaces(block) {
        return block.replace(/\s/g, '').length;
    }

    // Separate the text processing logic to be reusable by both text and audio modes
    function getProcessedBlocks(raw) {
        const cleanedText = cleanText(raw);
        const words = cleanedText.split(' ');
        let lines = [];
        let currentLine = [];
        let currentLength = 0;

        for (let i = 0; i < words.length; i++) {
            let word = words[i];
            let nextWord = words[i + 1] || '';

            const endsWithPunctuation = /[!?]|(\.\.)/.test(word.slice(-1)); 
            const isConnective = CONNECTIVES.has(word.toLowerCase().replace(/[!?.,]/g, ''));
            
            let startNewLine = false;

            if (isConnective && currentLine.length > 0) {
                startNewLine = true;
            }

            const proposedLineStr = [...currentLine, word].join(' ');
            const proposedLen = countCharsNoSpaces(proposedLineStr);

            if (currentLine.length > 0 && proposedLen > MAX_CHARS) {
                startNewLine = true;
            }

            if (startNewLine) {
                lines.push(currentLine.join(' '));
                currentLine = [word];
            } else {
                currentLine.push(word);
            }

            if (endsWithPunctuation) {
                 lines.push(currentLine.join(' '));
                 currentLine = [];
            }
        }

        if (currentLine.length > 0) {
            lines.push(currentLine.join(' '));
        }

        // Anti-Weakening
        let changed = true;
        let iterations = 0; 

        while (changed && iterations < 100) {
            changed = false;
            iterations++;

            for (let i = 0; i < lines.length - 1; i++) {
                const lineWords = lines[i].split(' ');
                const lastWord = lineWords[lineWords.length - 1];

                if (isWeakEnding(lastWord)) {
                    const remainingWords = lineWords.slice(0, lineWords.length - 1);
                    
                    if (remainingWords.length === 0) {
                        lines[i] = lines[i+1];
                    } else {
                        lines[i] = remainingWords.join(' ');
                    }

                    const nextLineWords = lines[i+1].split(' ');
                    lines[i+1] = [lastWord, ...nextLineWords].join(' ');

                    if (lines[i].trim() === '') {
                        lines.splice(i, 1);
                        i--; 
                    }
                    
                    changed = true;
                }
            }
        }
        
        return lines.filter(l => l.trim().length > 0);
    }

    function processScript() {
const raw = inputText.value;
        if (!raw.trim()) {
            showToast("Please enter text to process.", "error");
            return;
        }

        const lines = getProcessedBlocks(raw);
blockCount.textContent = lines.length;
        
        showToast("Script processed successfully!");
    }

    // --- Event Listeners ---
    
    const handleProcess = (e) => {
        if(e) e.preventDefault();
        processScript();
    };

    processBtn.addEventListener('click', handleProcess);
    processBtnMobile.addEventListener('click', handleProcess);

    copyBtn.addEventListener('click', () => {
        if (!outputText.value) return;
        navigator.clipboard.writeText(outputText.value).then(() => {
            showToast("Copied to clipboard!");
        });
    });

    clearBtn.addEventListener('click', () => {
        inputText.value = '';
        outputText.value = '';
        blockCount.textContent = '0';
        inputText.focus();
    });

    // Feather icons init backup
    setTimeout(() => feather.replace(), 100);
});