Krishna086's picture
Update app.py
35cccff verified
raw
history blame
5.99 kB
import streamlit as st
import importlib
from io import BytesIO
import docx
from PyPDF2 import PdfReader
st.set_page_config(page_title="Multilingual Translator", page_icon="🌐", layout="centered")
# Import LANGUAGES from translation.py
try:
from translation import LANGUAGES
except ImportError as e:
st.error(f"Failed to import translation module: {e}")
st.stop()
def extract_text_from_file(file):
try:
if file.type == "application/pdf":
pdf_reader = PdfReader(file)
text = ""
for page in pdf_reader.pages:
text += page.extract_text() or ""
return text.encode().decode('utf-8', errors='ignore').strip()
elif file.type == "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
doc = docx.Document(file)
text = "\n".join([para.text for para in doc.paragraphs])
return text.encode().decode('utf-8', errors='ignore').strip()
elif file.type == "text/plain":
return file.read().decode('utf-8', errors='ignore').strip()
return ""
except Exception:
return ""
def main():
try:
translation = importlib.import_module("translation")
lang_detect = importlib.import_module("lang_detect")
audio_processor = importlib.import_module("audio_processor")
# Header
st.markdown("<h1 style='text-align: center; color: #4285F4;'>Multilingual Translator</h1>", unsafe_allow_html=True)
st.markdown("<p style='text-align: center; color: #666;'>Translate text like Google Translate</p>", unsafe_allow_html=True)
# Language and Input/Output Layout
col1, col2 = st.columns(2)
with col1:
detected_options = lang_detect.detect_language(st.session_state.get("input_text", "")) if st.session_state.get("input_text", "").strip() else [("Auto-detect", 1.0, "Auto-detect")]
source_lang = detected_options[0][2] if detected_options[0][0] != "Auto-detect" else "Auto-detect"
source_lang_code = next((k for k, v in LANGUAGES.items() if v[1] == source_lang), "hi") if source_lang != "Auto-detect" else "auto"
source_options = ["Auto-detect"] + [f"{v[0]} ({v[1]})" for v in LANGUAGES.values()]
st.selectbox("Source", options=source_options, index=0 if source_lang == "Auto-detect" else source_options.index(f"{LANGUAGES[source_lang_code][0]} ({source_lang})"), key="source_lang")
input_text = st.text_area("", height=300, key="input_text", label_visibility="hidden")
input_type = st.radio("", ["Text", "File"], horizontal=True, label_visibility="hidden")
if input_type == "File":
uploaded_file = st.file_uploader("", type=["txt", "docx", "pdf"], key="file_input", label_visibility="hidden")
if uploaded_file and uploaded_file.size < 1024*1024: # 1 MB limit
st.session_state.input_text = extract_text_from_file(uploaded_file)
elif uploaded_file and uploaded_file.size >= 1024*1024:
st.error("File size must be less than 1 MB")
st.button("Translate", key="translate_btn", on_click=trigger_translation, args=(translation, lang_detect, audio_processor,))
with col2:
source_lang_display = st.session_state.source_lang.split(" (")[0] if " (" in st.session_state.source_lang else st.session_state.source_lang
target_options = [f"{v[0]} ({v[1]})" for v in LANGUAGES.values() if v[0] != source_lang_display and v[1] != source_lang_display]
st.selectbox("Target", options=target_options, index=target_options.index(f"{LANGUAGES['en'][0]} ({LANGUAGES['en'][1]})") if "English" not in source_lang_display else 0, key="target_lang")
if "translated_text" in st.session_state:
st.text_area("", value=st.session_state.translated_text, height=300, key="output_text", disabled=True, label_visibility="hidden")
if st.button("πŸ”Š", key="audio_btn", on_click=play_audio, args=(audio_processor,), help="Play audio", use_container_width=False):
pass
# Footer
if "translated_text" in st.session_state:
st.markdown("<p style='text-align: center; color: #666;'>Developed by Krishna Prakash <a href='https://www.linkedin.com/in/krishna-prakash-123456' target='_blank'>LinkedIn</a></p>", unsafe_allow_html=True)
except Exception as e:
st.error(f"App error: {e}")
def trigger_translation(translation, lang_detect, audio_processor):
text = st.session_state.get("input_text", "").strip()
if text:
source_lang = st.session_state.source_lang.split(" (")[0] if " (" in st.session_state.source_lang else st.session_state.source_lang
target_lang = st.session_state.target_lang.split(" (")[0] if " (" in st.session_state.target_lang else st.session_state.target_lang
if source_lang == "Auto-detect":
detected_options = lang_detect.detect_language(text)
source_lang_code = next((k for k, v in LANGUAGES.items() if v[1] == detected_options[0][0]), "hi")
else:
source_lang_code = next((k for k, v in LANGUAGES.items() if v[0] == source_lang), "hi")
target_lang_code = next((k for k, v in LANGUAGES.items() if v[0] == target_lang), "en")
translated_text = translation.translate(text, source_lang_code, target_lang_code)
st.session_state.translated_text = translated_text or text
def play_audio(audio_processor):
if "translated_text" in st.session_state and st.session_state.translated_text:
target_lang = next((k for k, v in LANGUAGES.items() if v[0] == st.session_state.target_lang.split(" (")[0]), "en")
audio = audio_processor.text_to_speech(st.session_state.translated_text, target_lang)
if audio and audio.getbuffer().nbytes > 0:
st.audio(audio, format="audio/mp3")
if __name__ == "__main__":
main()