Update app.py

app.py

@@ -70,8 +70,17 @@ with st.sidebar:
 ''')
 
 
+st.subheader("YouTube Comments Sentiment Analysis", divider="red")
+tokenizer = transformers.DistilBertTokenizer.from_pretrained("tabularisai/robust-sentiment-analysis")
+model = transformers.DistilBertForSequenceClassification.from_pretrained("tabularisai/robust-sentiment-analysis")
 
+if 'url_count' not in st.session_state:
+    st.session_state['url_count'] = 0
 
+max_attempts = 2
+
+def update_url_count():
+    st.session_state['url_count'] += 1
 
 def clear_question():
     st.session_state["url"] = ""
@@ -98,7 +107,7 @@ if st.button("Sentiment Analysis", type="secondary"):
                 placeholder = st.empty()
                 progress_bar = st.progress(0)
 
-                for item in range(30):
+                for item in range(150):
                     try:
                         body = WebDriverWait(driver, 30).until(EC.visibility_of_element_located((By.TAG_NAME, "body")))
                         body.send_keys(Keys.END)
@@ -112,34 +121,63 @@ if st.button("Sentiment Analysis", type="secondary"):
                 placeholder.text("Scrolling complete.")
                 progress_bar.empty()
 
-                videos = driver.find_elements(By.CLASS_NAME, 'style-scope.ytd-grid-video-renderer')
-                youtube_videos = []
-                for video in videos:
-                    title = video.find_element(By.XPATH, './/*[@id="video-title"]').text
-                    views = video.find_element(By.XPATH, './/*[@id="metadata-line"]/span[1]').text
-                    date = video.find_element(By.XPATH, './/*[@id="metadata-line"]/span[2]').text
-                    video_link = video.find_element(By.XPATH, './/*[@id="thumbnail"]/@href')
-                    vid_items = {
-                        'Title': title,
-                        'Views': views,
-                        'Posted': date,
-                        'Link': video_link}
-                    youtube_videos.append(vid_items)
-                    df = pd.DataFrame(youtube_videos)
-                    st.dataframe(df)
-
-
-
-
-
+                data = []
+                try:
+                    wait.until(EC.presence_of_element_located((By.CSS_SELECTOR, "#contents #contents")))
+                    comments = driver.find_elements(By.CSS_SELECTOR, "#content #content-text")
+                    user_id = 1
+                    for comment in comments:
+                        timestamp = datetime.now().strftime("%Y-%m-%d")
+                        data.append({"User ID": user_id, "Comment": comment.text, "comment_date": timestamp.text})
+                        user_id += 1
+                    data = [dict(t) for t in {tuple(d.items()) for d in data}]
+                except Exception as e:
+                    st.error(f"Exception during comment extraction: {e}")
+                driver.quit()
+                df = pd.DataFrame(data, columns=["User ID", "Comment", "comment_date"])
+                st.dataframe(df)
+
+                if tokenizer and model:
+                    inputs = tokenizer(df['Comment'].tolist(), return_tensors="pt", padding=True, truncation=True)
+                    with torch.no_grad():
+                        logits = model(**inputs).logits
+                        predicted_probabilities = torch.nn.functional.softmax(logits, dim=-1)
+                        predicted_labels = predicted_probabilities.argmax(dim=1)
+                        results = []
+                        for i, label in enumerate(predicted_labels):
+                            results.append({'Review Number': i + 1, 'Sentiment': model.config.id2label[label.item()]})
+                        sentiment_df = pd.DataFrame(results)
+
+                    value_counts1 = sentiment_df['Sentiment'].value_counts().rename_axis('Sentiment').reset_index(name='count')
+                    final_df = value_counts1
+                    tab1, tab2 = st.tabs(["Pie Chart", "Bar Chart"])
+                    with tab1:
+                        fig1 = px.pie(final_df, values='count', names='Sentiment', hover_data=['count'], labels={'count': 'count'})
+                        fig1.update_traces(textposition='inside', textinfo='percent+label')
+                        st.plotly_chart(fig1)
+
+                    result = pd.concat([df, sentiment_df], axis=1)
+                    st.dataframe(result)
+
+                    with tab2:
+                        fig2 = px.bar(result, x="Sentiment", y="comment_date", color="Sentiment")
+                        st.plotly_chart(fig2)
+
+                    csv = result.to_csv(index=False)
+                    st.download_button(
+                        label="Download data as CSV",
+                        data=csv,
+                        file_name='Summary of the results.csv',
+                        mime='text/csv',
+                    )
                 
+        else:
+            st.warning("Please enter a URL.")
+    else:
+        st.warning(f"You have reached the maximum URL attempts ({max_attempts}).")
 
-               
-
-
+if 'url_count' in st.session_state: #added if statement.
+    st.write(f"URL pasted {st.session_state['url_count']} times.")
 
 
             
-
-            
-