|
import pandas as pd |
|
import streamlit as st |
|
import time |
|
import random |
|
import os |
|
|
|
|
|
|
|
DATA_FILENAME = "Dr-En-space-test.csv" |
|
DATA_PATH = os.path.join("data", DATA_FILENAME) |
|
|
|
st.set_page_config(layout="wide") |
|
|
|
|
|
def load_data(): |
|
return pd.read_csv(DATA_PATH) |
|
|
|
def save_data(data): |
|
data.to_csv(DATA_PATH, index=False) |
|
|
|
def skip_correction(): |
|
noncorrected_sentences = st.session_state.data[(st.session_state.data.translated == True) & (st.session_state.data.corrected == False)]['sentence'].tolist() |
|
if noncorrected_sentences: |
|
st.session_state.orig_sentence = random.choice(noncorrected_sentences) |
|
st.session_state.orig_translation = st.session_state.data[st.session_state.data.sentence == st.session_state.orig_sentence]['translation'] |
|
else: |
|
st.session_state.orig_sentence = "No more sentences to be corrected" |
|
st.session_state.orig_translation = "No more sentences to be corrected" |
|
|
|
st.title("Darija Translation Corpus Collection") |
|
|
|
if "data" not in st.session_state: |
|
st.session_state.data = load_data() |
|
|
|
if "sentence" not in st.session_state: |
|
untranslated_sentences = st.session_state.data[st.session_state.data['translated'] == False]['sentence'].tolist() |
|
if untranslated_sentences: |
|
st.session_state.sentence = random.choice(untranslated_sentences) |
|
else: |
|
st.session_state.sentence = "No more sentences to translate" |
|
|
|
if "orig_translation" not in st.session_state: |
|
noncorrected_sentences = st.session_state.data[(st.session_state.data.translated == True) & (st.session_state.data.corrected == False)]['sentence'].tolist() |
|
noncorrected_translations = st.session_state.data[(st.session_state.data.translated == True) & (st.session_state.data.corrected == False)]['translation'].tolist() |
|
|
|
if noncorrected_sentences: |
|
st.session_state.orig_sentence = random.choice(noncorrected_sentences) |
|
st.session_state.orig_translation = st.session_state.data.loc[st.session_state.data.sentence == st.session_state.orig_sentence]['translation'].values[0] |
|
else: |
|
st.session_state.orig_sentence = "No more sentences to be corrected" |
|
st.session_state.orig_translation = "No more sentences to be corrected" |
|
|
|
|
|
|
|
if "user_translation" not in st.session_state: |
|
st.session_state.user_translation = "" |
|
|
|
|
|
with st.sidebar: |
|
st.subheader("About") |
|
st.markdown("""This is app is designed to collect Darija translation corpus.""") |
|
|
|
tab1, tab2 = st.tabs(["Translation", "Correction"]) |
|
|
|
with tab1: |
|
with st.container(): |
|
st.subheader("Original Text:") |
|
|
|
st.write('<div style="height: 150px; overflow: auto; border: 2px solid #ddd; padding: 10px; border-radius: 5px;">{}</div>'.format(st.session_state.sentence), unsafe_allow_html=True) |
|
|
|
|
|
st.subheader("Translation:") |
|
st.session_state.user_translation = st.text_area("Enter your translation here:", value=st.session_state.user_translation) |
|
|
|
if st.button("💾 Save"): |
|
if st.session_state.user_translation: |
|
st.session_state.data.loc[st.session_state.data['sentence'] == st.session_state.sentence, 'translation'] = st.session_state.user_translation |
|
st.session_state.data.loc[st.session_state.data['sentence'] == st.session_state.sentence, 'translated'] = True |
|
save_data(st.session_state.data) |
|
|
|
st.session_state.user_translation = "" |
|
|
|
|
|
st.success("Saved!") |
|
|
|
|
|
untranslated_sentences = st.session_state.data[st.session_state.data['translated'] == False]['sentence'].tolist() |
|
if untranslated_sentences: |
|
st.session_state.sentence = random.choice(untranslated_sentences) |
|
|
|
else: |
|
st.session_state.sentence = "No more sentences to translate" |
|
|
|
time.sleep(0.5) |
|
|
|
st.rerun() |
|
|
|
with tab2: |
|
with st.container(): |
|
st.subheader("Original Darija Text:") |
|
st.write('<div style="height: 150px; overflow: auto; border: 2px solid #ddd; padding: 10px; border-radius: 5px;">{}</div>'.format(st.session_state.orig_sentence), unsafe_allow_html=True) |
|
|
|
with st.container(): |
|
st.subheader("Original English Translation:") |
|
st.write('<div style="height: 150px; overflow: auto; border: 2px solid #ddd; padding: 10px; border-radius: 5px;">{}</div>'.format(st.session_state.orig_translation), unsafe_allow_html=True) |
|
|
|
st.subheader("Corrected Darija Translation:") |
|
corrected_translation = st.text_area("Enter the corrected Darija translation here:") |
|
|
|
if st.button("💾 Save Translation"): |
|
if corrected_translation: |
|
st.session_state.data.loc[st.session_state.data['sentence'] == st.session_state.orig_sentence, 'translation'] = corrected_translation |
|
st.session_state.data.loc[st.session_state.data['sentence'] == st.session_state.orig_sentence, 'correction'] = corrected_translation |
|
st.session_state.data.loc[st.session_state.data['sentence'] == st.session_state.orig_sentence, 'corrected'] = True |
|
save_data(st.session_state.data) |
|
|
|
st.success("Saved!") |
|
|
|
|
|
noncorrected_sentences = st.session_state.data[(st.session_state.data.translated == True) & (st.session_state.data.corrected == False)]['sentence'].tolist() |
|
|
|
if noncorrected_sentences: |
|
st.session_state.orig_sentence = random.choice(noncorrected_sentences) |
|
st.session_state.orig_translation = st.session_state.data[st.session_state.data.sentence == st.session_state.orig_sentence]['translation'] |
|
|
|
else: |
|
st.session_state.orig_translation = "No more sentences to be corrected" |
|
|
|
corrected_translation = "" |
|
|
|
st.button("⏩ Skip to the Next Pair", key="skip_button", on_click=skip_correction) |