src/main.py

import display_gloss as dg
import synonyms_preprocess as sp
from NLP_Spacy_base_translator import NlpSpacyBaseTranslator 
from flask import Flask, render_template, Response, request, send_file
import io
import cv2
import numpy as np
import os
import requests
from urllib.parse import quote, unquote
import tempfile
import re

app = Flask(__name__, static_folder='static')
app.config['TITLE'] = 'Sign Language Translate'

nlp, dict_docs_spacy = sp.load_spacy_values()
dataset, list_2000_tokens = dg.load_data()

def clean_quotes(text):
    """따옴표 정리 함수"""
    # 연속된 따옴표 제거
    text = re.sub(r"'+", "'", text)
    # 단어 중간의 따옴표 제거
    text = re.sub(r"(\w)'(\w)", r"\1\2", text)
    return text

def normalize_quotes(text):
    """따옴표 형식을 정규화하는 함수"""
    # 먼저 모든 따옴표를 정리
    text = clean_quotes(text)
    
    # 한글 또는 영어 단어를 찾아서 처리
    pattern = r'([가-힣A-Za-z]+)'
    
    def process_match(match):
        word = match.group(1)
        # 이미 따옴표로 둘러싸인 경우는 처리하지 않음
        if not re.match(r"'.*'", word):
            return f"'{word}'"
        return word
    
    # 단어 단위로 처리
    words = text.split()
    processed_words = []
    
    for word in words:
        if re.search(pattern, word):
            # 이미 따옴표가 있는 경우는 그대로 두고, 없는 경우만 추가
            if not word.startswith("'") and not word.endswith("'"):
                word = f"'{word}'"
        processed_words.append(word)
    
    return ' '.join(processed_words)

def find_quoted_words(text):
    """작은따옴표로 묶인 단어들을 찾는 함수"""
    return re.findall(r"'([^']*)'", text)

def spell_out_word(word):
    """단어를 개별 알파벳으로 분리하는 함수"""
    return ' '.join(list(word.lower()))

def is_english(text):
    """텍스트가 영어인지 확인하는 함수"""
    english_pattern = re.compile(r'^[A-Za-z\s\'".,!?-]+$')
    return bool(english_pattern.match(text.replace("'", "")))

def translate_korean_to_english(text):
    """전체 텍스트 번역 함수"""
    try:
        # 입력 텍스트 정규화
        text = normalize_quotes(text)
        
        # 영어 입력 확인
        if is_english(text):
            return text

        # 따옴표로 묶인 단어들 찾기
        quoted_words = re.findall(r"'([^']*)'", text)
        translated_quoted = {}

        # 따옴표 안의 단어들 먼저 번역
        for word in quoted_words:
            if not word.strip():  # 빈 문자열 건너뛰기
                continue
            url = "https://translate.googleapis.com/translate_a/single"
            params = {
                "client": "gtx",
                "sl": "ko",
                "tl": "en",
                "dt": "t",
                "q": word
            }
            response = requests.get(url, params=params)
            if response.status_code == 200:
                translated = response.json()[0][0][0].upper()
                translated_quoted[word] = translated
                # 임시 마커로 대체
                text = text.replace(f"'{word}'", f"QUOTED_{len(translated_quoted)}_")

        # 전체 문장 번역
        params = {
            "client": "gtx",
            "sl": "ko",
            "tl": "en",
            "dt": "t",
            "q": text
        }
        response = requests.get(url, params=params)
        
        if response.status_code == 200:
            translated_text = ' '.join(item[0] for item in response.json()[0] if item[0])
            
            # 번역된 텍스트에서 마커를 번역된 단어로 대체
            for i, (original, translated) in enumerate(translated_quoted.items(), 1):
                translated_text = translated_text.replace(f"QUOTED_{i}_", f"'{translated}'")
            
            return translated_text
        else:
            raise Exception(f"Translation API returned status code: {response.status_code}")
    except Exception as e:
        print(f"Translation error: {e}")
        return text

@app.route('/')
def index():
    return render_template('index.html', title=app.config['TITLE'])

@app.route('/translate/', methods=['POST'])
def result():
    if request.method == 'POST':
        input_text = request.form['inputSentence'].strip()
        if not input_text:
            return render_template('error.html', error="Please enter text to translate")
        
        try:
            # 입력 텍스트 정규화
            input_text = normalize_quotes(input_text)
            
            # 번역 수행
            english_text = translate_korean_to_english(input_text)
            if not english_text:
                raise Exception("Translation failed")
            
            # 따옴표로 묶인 단어 추출
            quoted_words = [word.strip("'") for word in re.findall(r"'([^']*)'", english_text)]
            
            # ASL 변환을 위해 따옴표 제거
            clean_english = re.sub(r"'([^']*)'", r"\1", english_text)
            eng_to_asl_translator = NlpSpacyBaseTranslator(sentence=clean_english)
            generated_gloss = eng_to_asl_translator.translate_to_gloss()
            
            # 단어 처리
            processed_gloss = []
            words = generated_gloss.split()
            
            for word in words:
                word_upper = word.upper()
                if any(quoted.upper() == word_upper for quoted in quoted_words):
                    # 고유명사인 경우 철자를 하나씩 분리
                    spelled_word = spell_out_word(word)
                    processed_gloss.extend(['FINGERSPELL-START'] + spelled_word.split() + ['FINGERSPELL-END'])
                else:
                    # 일반 단어는 기존 방식대로 처리
                    word_lower = word.lower()
                    if word_lower.isalnum():
                        processed_gloss.append(word_lower)
            
            gloss_sentence_before_synonym = " ".join(processed_gloss)
            
            # 고유명사가 아닌 단어들만 동의어 처리
            final_gloss = []
            i = 0
            while i < len(processed_gloss):
                if processed_gloss[i] == 'FINGERSPELL-START':
                    final_gloss.extend(processed_gloss[i:i+2])
                    i += 2
                    while i < len(processed_gloss) and processed_gloss[i] != 'FINGERSPELL-END':
                        final_gloss.append(processed_gloss[i])
                        i += 1
                    if i < len(processed_gloss):
                        final_gloss.append(processed_gloss[i])
                        i += 1
                else:
                    word = processed_gloss[i]
                    final_gloss.append(sp.find_synonyms(word, nlp, dict_docs_spacy, list_2000_tokens))
                    i += 1
            
            gloss_sentence_after_synonym = " ".join(final_gloss)
            
            return render_template('result.html',
                                title=app.config['TITLE'],
                                original_sentence=input_text,
                                english_translation=english_text,
                                gloss_sentence_before_synonym=gloss_sentence_before_synonym,
                                gloss_sentence_after_synonym=gloss_sentence_after_synonym)
        except Exception as e:
            return render_template('error.html', error=f"Translation error: {str(e)}")

def generate_complete_video(gloss_list, dataset, list_2000_tokens):
    try:
        frames = []
        is_spelling = False
        
        for gloss in gloss_list:
            if gloss == 'FINGERSPELL-START':
                is_spelling = True
                continue
            elif gloss == 'FINGERSPELL-END':
                is_spelling = False
                continue
                
            for frame in dg.generate_video([gloss], dataset, list_2000_tokens):
                frame_data = frame.split(b'\r\n\r\n')[1]
                nparr = np.frombuffer(frame_data, np.uint8)
                img = cv2.imdecode(nparr, cv2.IMREAD_COLOR)
                frames.append(img)
        
        if not frames:
            raise Exception("No frames generated")

        height, width = frames[0].shape[:2]
        fourcc = cv2.VideoWriter_fourcc(*'mp4v')
        
        with tempfile.NamedTemporaryFile(suffix='.mp4', delete=False) as temp_file:
            temp_path = temp_file.name
            
        out = cv2.VideoWriter(temp_path, fourcc, 25, (width, height))
        
        for frame in frames:
            out.write(frame)
        out.release()
        
        with open(temp_path, 'rb') as f:
            video_bytes = f.read()
        
        os.remove(temp_path)
        return video_bytes
    except Exception as e:
        print(f"Error generating video: {str(e)}")
        raise

@app.route('/video_feed')
def video_feed():
    sentence = request.args.get('gloss_sentence_to_display', '')
    gloss_list = sentence.split()
    return Response(dg.generate_video(gloss_list, dataset, list_2000_tokens), 
                   mimetype='multipart/x-mixed-replace; boundary=frame')

@app.route('/download_video/<path:gloss_sentence>')
def download_video(gloss_sentence):
    try:
        decoded_sentence = unquote(gloss_sentence)
        gloss_list = decoded_sentence.split()
        
        if not gloss_list:
            return "No gloss provided", 400
            
        video_bytes = generate_complete_video(gloss_list, dataset, list_2000_tokens)
        
        if not video_bytes:
            return "Failed to generate video", 500
            
        return send_file(
            io.BytesIO(video_bytes),
            mimetype='video/mp4',
            as_attachment=True,
            download_name='sign_language.mp4'
        )
    except Exception as e:
        print(f"Download error: {str(e)}")
        return f"Error downloading video: {str(e)}", 500

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=7860, debug=True)