app improvement

app.py

@@ -4,6 +4,8 @@ import plotly.express as px
 import plotly.graph_objects as go
 import numpy as np
 import requests
+from datetime import datetime
+import time
 
 ### Config
 st.set_page_config(
@@ -12,44 +14,224 @@ st.set_page_config(
     layout="wide"
 )
 
+# Initialize session state
+if 'history' not in st.session_state:
+    st.session_state.history = []
+if 'api_health' not in st.session_state:
+    st.session_state.api_health = None
+
+# Custom CSS
+st.markdown("""
+    <style>
+    .prediction-box {
+        padding: 20px;
+        border-radius: 5px;
+        margin: 10px 0;
+    }
+    .high-severity {
+        background-color: rgba(255, 0, 0, 0.1);
+        border: 1px solid red;
+    }
+    .medium-severity {
+        background-color: rgba(255, 165, 0, 0.1);
+        border: 1px solid orange;
+    }
+    .low-severity {
+        background-color: rgba(0, 255, 0, 0.1);
+        border: 1px solid green;
+    }
+    </style>
+""", unsafe_allow_html=True)
+
+def check_api_health():
+    """Check if the API is responsive"""
+    try:
+        response = requests.get("https://llepogam-hate-speech-detection-api.hf.space/health")
+        return response.status_code == 200
+    except:
+        return False
 
 def hate_speech_detection(text):
+    """Make API call with error handling"""
     url = "https://llepogam-hate-speech-detection-api.hf.space/predict"
     headers = {
         "accept": "application/json",
         "Content-Type": "application/json"
     }
 
-    # Define the payload
-    Text_to_predict = {
-        "Text": text
-    }
+    try:
+        response = requests.post(
+            url, 
+            headers=headers, 
+            json={"Text": text},
+            timeout=10
+        )
+        response.raise_for_status()
+        return response.json(), None
+    except requests.exceptions.Timeout:
+        return None, "API request timed out. Please try again."
+    except requests.exceptions.RequestException as e:
+        return None, f"API error: {str(e)}"
+    except Exception as e:
+        return None, f"Unexpected error: {str(e)}"
 
-    # Make the POST request
-    response = requests.post(url, headers=headers, json=Text_to_predict)
+def get_severity_class(probability):
+    """Determine severity class based on probability"""
+    if probability > 0.7:
+        return "high-severity"
+    elif probability > 0.4:
+        return "medium-severity"
+    return "low-severity"
 
-    # Process the response
-    if response.status_code == 200:
-        return response.json()
-    else:
-        return f"Failed to get a response. Status code: {response.status_code}, Response: {response.text}"
+# Header Section
+st.title("🚫 Offensive Speech Detection")
+st.markdown("""
+This application helps identify potentially offensive or harmful content in text. 
+It uses a machine learning model to analyze text and determine if it contains offensive speech.
 
+**How it works:**
+1. Enter your text in the input box below
+2. The model will analyze the content and provide a prediction
+3. Results show both the classification and confidence level
+""")
 
+# API Status
+if st.button("Check API Status"):
+    with st.spinner("Checking API health..."):
+        st.session_state.api_health = check_api_health()
+    
+if st.session_state.api_health is not None:
+    status_color = "green" if st.session_state.api_health else "red"
+    status_text = "Online" if st.session_state.api_health else "Offline"
+    st.markdown(f"API Status: :{status_color}[{status_text}]")
 
+# Example inputs
+with st.expander("📝 Example Inputs"):
+    st.markdown("""
+    Try these example texts to test the model:
+    1. "Have a great day!"
+    2. "I disagree with your opinion."
+    3. "You're amazing!"
+    
+    Click on any example to copy it to the input box.
+    """)
+    if st.button("Use Example 1"):
+        st.session_state.user_input = "Have a great day!"
+    if st.button("Use Example 2"):
+        st.session_state.user_input = "I disagree with your opinion."
+    if st.button("Use Example 3"):
+        st.session_state.user_input = "You're amazing!"
 
-st.title("Offensive Speech Detection")
-user_input = st.text_area("Enter a text:")
+# FAQ Section
+with st.expander("❓ Frequently Asked Questions"):
+    st.markdown("""
+    **Q: What is considered offensive speech?**
+    - A: The model identifies content that could be harmful, insulting, or discriminatory.
 
-if user_input:
-    prediction = hate_speech_detection(user_input)
-    st.write(f"Prediction: {prediction['prediction']}")
-    st.write(f"Probability: {prediction['probability']}")
- 
+    **Q: How accurate is the detection?**
+    - A: The model provides a confidence score with each prediction. Higher scores indicate greater confidence.
+
+    **Q: What happens to my input data?**
+    - A: Your text is only used for prediction and temporarily stored in your session history.
+    """)
+
+# Text Input Section
+max_chars = 500
+user_input = st.text_area(
+    "Enter text to analyze:",
+    height=100,
+    key="user_input",
+    help="Enter the text you want to analyze for offensive content. Maximum 500 characters.",
+    max_chars=max_chars
+)
 
+# Character counter
+chars_remaining = max_chars - len(user_input)
+st.caption(f"Characters remaining: {chars_remaining}")
+
+# Clear button
+if st.button("Clear Input"):
+    st.session_state.user_input = ""
+    st.experimental_rerun()
+
+# Process input
+if user_input:
+    if len(user_input.strip()) == 0:
+        st.warning("Please enter some text to analyze.")
+    else:
+        with st.spinner("Analyzing text..."):
+            result, error = hate_speech_detection(user_input)
+            
+            if error:
+                st.error(f"Error: {error}")
+            else:
+                # Format probability as percentage
+                probability_pct = result['probability'] * 100
+                
+                # Create prediction box with appropriate severity class
+                severity_class = get_severity_class(result['probability'])
+                
+                st.markdown(f"""
+                <div class="prediction-box {severity_class}">
+                    <h3>Analysis Results</h3>
+                    <p><strong>Prediction:</strong> {result['prediction']}</p>
+                    <p><strong>Confidence:</strong> {probability_pct:.1f}%</p>
+                </div>
+                """, unsafe_allow_html=True)
+                
+                # Confidence meter using Plotly
+                fig = go.Figure(go.Indicator(
+                    mode = "gauge+number",
+                    value = probability_pct,
+                    title = {'text': "Confidence Level"},
+                    gauge = {
+                        'axis': {'range': [0, 100]},
+                        'bar': {'color': "darkblue"},
+                        'steps': [
+                            {'range': [0, 40], 'color': "lightgreen"},
+                            {'range': [40, 70], 'color': "orange"},
+                            {'range': [70, 100], 'color': "red"}
+                        ]
+                    }
+                ))
+                fig.update_layout(height=300)
+                st.plotly_chart(fig, use_container_width=True)
+                
+                # Add to history
+                st.session_state.history.append({
+                    'timestamp': datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
+                    'text': user_input,
+                    'prediction': result['prediction'],
+                    'confidence': probability_pct
+                })
 
-### Footer 
-empty_space, footer = st.columns([1, 2])
+# History Section
+if st.session_state.history:
+    with st.expander("📜 Analysis History"):
+        history_df = pd.DataFrame(st.session_state.history)
+        st.dataframe(
+            history_df,
+            column_config={
+                "timestamp": "Time",
+                "text": "Input Text",
+                "prediction": "Prediction",
+                "confidence": st.column_config.NumberColumn(
+                    "Confidence",
+                    format="%.1f%%"
+                )
+            },
+            hide_index=True
+        )
+        
+        if st.button("Clear History"):
+            st.session_state.history = []
+            st.experimental_rerun()
 
-with empty_space:
-    st.write("")
+# Footer
+st.markdown("---")
+st.markdown("""
+    <div style='text-align: center'>
+        <p>Developed with ❤️ for safer online communication</p>
+    </div>
+""", unsafe_allow_html=True)