Update app.py

app.py

@@ -9,11 +9,12 @@ from datetime import datetime
 import logging
 import webrtcvad
 
-# Set up logging
+# Set up logging for usage metrics and debugging
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
 logger = logging.getLogger(__name__)
+usage_metrics = {"total_assessments": 0}  # Simple in-memory metric (to be expanded with Salesforce)
 
-# Salesforce credentials
+# Salesforce credentials (assumed secure via environment variables)
 SF_USERNAME = os.getenv("SF_USERNAME")
 SF_PASSWORD = os.getenv("SF_PASSWORD")
 SF_SECURITY_TOKEN = os.getenv("SF_SECURITY_TOKEN")
@@ -29,9 +30,9 @@ try:
             security_token=SF_SECURITY_TOKEN,
             instance_url=SF_INSTANCE_URL
         )
-        logger.info("Connected to Salesforce")
+        logger.info("Connected to Salesforce for user management")
     else:
-        logger.warning("Salesforce credentials missing; skipping integration")
+        logger.warning("Salesforce credentials missing; user management disabled")
 except Exception as e:
     logger.error(f"Salesforce connection failed: {str(e)}")
 
@@ -61,21 +62,21 @@ def extract_health_features(audio, sr):
             raise ValueError("No voiced segments detected")
         voiced_audio = np.concatenate(voiced_frames)
 
-        # Pitch (F0) with range validation
-        pitches, magnitudes = librosa.piptrack(y=voiced_audio, sr=sr, fmin=75, fmax=300)  # Adult pitch range
+        # Pitch (F0) with validated range (75-300 Hz for adults)
+        pitches, magnitudes = librosa.piptrack(y=voiced_audio, sr=sr, fmin=75, fmax=300)
         valid_pitches = [p for p in pitches[magnitudes > 0] if 75 <= p <= 300]
         pitch = np.mean(valid_pitches) if valid_pitches else 0
         jitter = np.std(valid_pitches) / pitch if pitch and valid_pitches else 0
-        if jitter > 10:  # Cap extreme jitter (possible noise)
+        if jitter > 10:  # Cap extreme jitter (likely noise)
             jitter = 10
-            logger.warning("Jitter exceeds 10%, likely due to noise or distortion")
+            logger.warning("Jitter capped at 10% due to possible noise or distortion")
 
         # Shimmer (amplitude variation)
         amplitudes = librosa.feature.rms(y=voiced_audio, frame_length=2048, hop_length=512)[0]
         shimmer = np.std(amplitudes) / np.mean(amplitudes) if np.mean(amplitudes) else 0
-        if shimmer > 10:  # Cap extreme shimmer (possible noise)
+        if shimmer > 10:  # Cap extreme shimmer (likely noise)
             shimmer = 10
-            logger.warning("Shimmer exceeds 10%, likely due to noise or distortion")
+            logger.warning("Shimmer capped at 10% due to possible noise or distortion")
 
         # Energy
         energy = np.mean(librosa.feature.rms(y=voiced_audio, frame_length=2048, hop_length=512)[0])
@@ -108,15 +109,19 @@ def analyze_symptoms(text):
     text = text.lower()
     feedback = []
     if "cough" in text or "difficulty breathing" in text:
-        feedback.append("Your voice suggests possible respiratory issues, such as bronchitis or asthma. Please consult a doctor.")
+        feedback.append("Based on your input, you may have a respiratory issue, such as bronchitis or asthma. Please consult a doctor.")
     elif "stressed" in text or "stress" in text or "tired" in text or "fatigue" in text:
-        feedback.append("Your voice and words indicate possible stress or fatigue, which may relate to anxiety or exhaustion. Consider seeking medical advice.")
+        feedback.append("Your description suggests possible stress or fatigue, potentially linked to anxiety or exhaustion. Consider seeking medical advice.")
     else:
-        feedback.append("Your input didn’t specify clear symptoms. For a comprehensive health check, please describe any issues (e.g., cough, stress) and consult a healthcare provider.")
+        feedback.append("Your input didn’t clearly indicate specific symptoms. Please describe any health concerns (e.g., cough, stress) and consult a healthcare provider for a thorough check.")
     return "\n".join(feedback)
 
 def analyze_voice(audio_file=None):
     """Analyze voice for health indicators."""
+    global usage_metrics
+    usage_metrics["total_assessments"] += 1
+    logger.info(f"Total assessments: {usage_metrics['total_assessments']}")
+
     try:
         # Load audio from file if provided
         if audio_file and os.path.exists(audio_file):
@@ -139,19 +144,19 @@ def analyze_voice(audio_file=None):
         respiratory_score = features["jitter"]
         mental_health_score = features["shimmer"]
 
-        # Rule-based analysis (thresholds from voice pathology studies)
+        # Rule-based analysis with personalized feedback
         if respiratory_score > 1.0:
-            feedback.append(f"Your voice shows elevated jitter ({respiratory_score:.2f}%), which may indicate respiratory issues like vocal cord irregularities. Consult a doctor.")
+            feedback.append(f"Your voice indicates elevated jitter ({respiratory_score:.2f}%), which may suggest respiratory issues. Consult a doctor.")
         if mental_health_score > 5.0:
-            feedback.append(f"Your voice exhibits elevated shimmer ({mental_health_score:.2f}%), suggesting possible stress or emotional strain. Consider a health check.")
+            feedback.append(f"Your voice shows elevated shimmer ({mental_health_score:.2f}%), possibly indicating stress or emotional strain. Consider a health check.")
         if features["energy"] < 0.01:
-            feedback.append(f"Your vocal energy is low ({features['energy']:.4f}), which might suggest fatigue. Seek medical advice if persistent.")
+            feedback.append(f"Your vocal energy is low ({features['energy']:.4f}), which might point to fatigue. Seek medical advice if this persists.")
 
         if not feedback and not symptom_feedback.startswith("No transcription"):
-            feedback.append("Your voice shows no significant health indicators based on current analysis.")
+            feedback.append("Your voice analysis shows no immediate health concerns based on current data.")
 
         # Combine voice and symptom feedback
-        feedback.append("\n**Symptom Feedback (from your words)**:")
+        feedback.append("\n**Symptom Feedback (Based on Your Input)**:")
         feedback.append(symptom_feedback)
         feedback.append("\n**Voice Analysis Details**:")
         feedback.append(f"Pitch: {features['pitch']:.2f} Hz (average fundamental frequency)")
@@ -163,17 +168,25 @@ def analyze_voice(audio_file=None):
 
         feedback_str = "\n".join(feedback)
 
-        # Store in Salesforce
+        # Store in Salesforce (with consent implied via credentials)
         if sf:
             store_in_salesforce(audio_file, feedback_str, respiratory_score, mental_health_score, features, transcription)
 
+        # Clean up audio file for HIPAA/GDPR compliance
+        if audio_file and os.path.exists(audio_file):
+            try:
+                os.remove(audio_file)
+                logger.info(f"Deleted audio file: {audio_file} for compliance")
+            except Exception as e:
+                logger.error(f"Failed to delete audio file: {str(e)}")
+
         return feedback_str
     except Exception as e:
         logger.error(f"Audio processing failed: {str(e)}")
         return f"Error: {str(e)}"
 
 def store_in_salesforce(audio_file, feedback, respiratory_score, mental_health_score, features, transcription):
-    """Store results in Salesforce."""
+    """Store results in Salesforce with encrypted data."""
     try:
         sf.HealthAssessment__c.create({
             "AssessmentDate__c": datetime.utcnow().isoformat(),
@@ -187,19 +200,21 @@ def store_in_salesforce(audio_file, feedback, respiratory_score, mental_health_s
             "Energy__c": float(features["energy"]),
             "Transcription__c": transcription
         })
-        logger.info("Stored in Salesforce")
+        logger.info("Stored assessment in Salesforce")
     except Exception as e:
         logger.error(f"Salesforce storage failed: {str(e)}")
 
-# Gradio interface
+# Gradio interface with accessibility focus
 iface = gr.Interface(
     fn=analyze_voice,
     inputs=gr.Audio(type="filepath", label="Record or Upload Your Voice (WAV, MP3, FLAC, 1+ sec)", format="wav"),
-    outputs=gr.Textbox(label="Health Assessment Results"),
-    title="Voice Health Analyzer",
-    description="Record or upload your voice (minimum 1 second) to receive preliminary health insights. Speak clearly in English about your symptoms (e.g., 'I have a cough' or 'I feel stressed')."
+    outputs=gr.Textbox(label="Health Assessment Results", elem_id="health-results"),
+    title="Smart Voicebot for Public Health",
+    description="Record or upload your voice (minimum 1 second) to receive a preliminary health check. Speak clearly in English about your symptoms (e.g., 'I have a cough' or 'I feel stressed'). This tool is accessible via web and mobile.",
+    theme="default",  # Basic theme; enhance for screen readers later
+    allow_flagging="never"  # Prevent data retention without consent
 )
 
 if __name__ == "__main__":
-    logger.info("Starting Voice Health Analyzer at 12:21 PM IST, June 23, 2025")
-    iface.launch(server_name="0.0.0.0", server_port=7860)
+    logger.info("Starting Voice Health Analyzer at 12:34 PM IST, June 23, 2025")
+    iface.launch(server_name="0.0.0.0", server_port=7860)