Update app.py

app.py

@@ -8,10 +8,10 @@ import numpy as np
 import librosa
 import soundfile as sf
 import matplotlib.pyplot as plt
-import matplotlib.gridspec as gridspec
 import scipy.signal as sps
 
-from io_utils import read_audio_info, load_audio_mono
+# Local modules
+from io_utils import read_audio_info
 from time_domain import compute_time_domain_stats
 from spectral import compute_spectral_analysis
 from loudness_calc import compute_loudness, LOUDNESS_AVAILABLE
@@ -19,84 +19,85 @@ from issue_detection import detect_audio_issues
 from synthetic_detector import detect_synthetic_voice
 from report_generator import create_report
 
+
 # ============================================================
-# MAIN ANALYSIS FUNCTION
+# MAIN FORENSIC ANALYSIS FUNCTION
 # ============================================================
 
 def analyze_audio(audio_file, progress=gr.Progress()):
-    """Main Gradio callback — performs full forensic analysis."""
+    """Full forensic audio analysis pipeline."""
 
     if audio_file is None:
         return None, "⚠️ Please upload an audio file."
 
     try:
-        progress(0.1, desc="Reading audio file...")
-
         path = Path(audio_file)
+
+        # -----------------------------
+        # FILE INFO + LOAD AUDIO
+        # -----------------------------
+        progress(0.10, desc="Reading file info...")
         info = read_audio_info(str(path))
 
         progress(0.25, desc="Loading waveform...")
         y, sr = librosa.load(str(path), sr=None, mono=True)
 
-        # ======================================================
+        # -----------------------------
         # TIME DOMAIN
-        # ======================================================
-        progress(0.35, desc="Analyzing time-domain...")
+        # -----------------------------
+        progress(0.35, desc="Time-domain analysis...")
         time_stats = compute_time_domain_stats(y)
 
-        # ======================================================
-        # SPECTRAL
-        # ======================================================
-        progress(0.50, desc="Computing spectral analysis...")
+        # -----------------------------
+        # SPECTRAL ANALYSIS
+        # -----------------------------
+        progress(0.50, desc="Spectral analysis...")
         spectral = compute_spectral_analysis(y, sr)
 
-        # ======================================================
+        # -----------------------------
         # LOUDNESS
-        # ======================================================
+        # -----------------------------
         progress(0.60, desc="Computing loudness...")
         lufs = compute_loudness(y, sr) if LOUDNESS_AVAILABLE else None
 
-        # ======================================================
-        # ISSUES
-        # ======================================================
-        progress(0.70, desc="Detecting issues...")
+        # -----------------------------
+        # ISSUE DETECTION
+        # -----------------------------
+        progress(0.70, desc="Detecting anomalies...")
         issues = detect_audio_issues(spectral, time_stats)
 
-        # ======================================================
-        # SYNTHETIC DETECTION (informational only)
-        # ======================================================
+        # -----------------------------
+        # SYNTHETIC VOICE (INFO ONLY)
+        # -----------------------------
         progress(0.78, desc="Synthetic voice estimation...")
         synthetic = detect_synthetic_voice(y, sr, spectral)
 
         # ======================================================
-        # SCORING
+        # SCORING — CLEAN VERSION (NO DUPLICATE ASSIGNMENT)
         # ======================================================
         progress(0.82, desc="Scoring...")
-        
-        # Count issues by severity
+
         critical = sum(1 for _, sev, _ in issues if sev == "CRITICAL")
         high = sum(1 for _, sev, _ in issues if sev == "HIGH")
         medium = sum(1 for _, sev, _ in issues if sev == "MEDIUM")
         low = sum(1 for _, sev, _ in issues if sev == "LOW")
-        
-        # Score formula
-        score = 100 - (critical * 30) - (high * 15) - (medium * 5)
-        score = max(0, score)
-        
-        # Grade + Quality Label (same logic as your report)
-        if score >= 90:
+
+        score_value = 100 - (critical * 30) - (high * 15) - (medium * 5)
+        score_value = max(0, score_value)
+
+        if score_value >= 90:
             grade, quality = "A", "EXCELLENT"
             color = "🟢"
             recommendation = "Excellent for TTS dataset"
-        elif score >= 75:
+        elif score_value >= 75:
             grade, quality = "B", "GOOD"
             color = "🟢"
             recommendation = "Very good quality; suitable for TTS"
-        elif score >= 60:
+        elif score_value >= 60:
             grade, quality = "C", "FAIR"
             color = "🟡"
             recommendation = "Usable but may contain processing artifacts"
-        elif score >= 40:
+        elif score_value >= 40:
             grade, quality = "D", "POOR"
             color = "🟠"
             recommendation = "Not recommended for TTS (heavy processing)"
@@ -104,15 +105,12 @@ def analyze_audio(audio_file, progress=gr.Progress()):
             grade, quality = "F", "CRITICAL"
             color = "🔴"
             recommendation = "Severely degraded or processed; avoid for TTS"
-        
-        # Cleanliness score (needed by report_generator)
+
         cleanliness_score = max(0, 100 - (medium * 5 + low * 3))
-        
-        # Processing severity index (needed by report_generator)
         processing_severity = (critical * 3) + (high * 2) + medium
-        
-        audio_data["score"] = {
-            "score": score,
+
+        score_dict = {
+            "score": score_value,
             "grade": grade,
             "quality": quality,
             "recommendation": recommendation,
@@ -124,16 +122,10 @@ def analyze_audio(audio_file, progress=gr.Progress()):
             "low": low,
             "color": color
         }
-        
 
         # ======================================================
-        # CREATE REPORT PNG
+        # BUILD FINAL AUDIO_DATA PAYLOAD (SAFE ORDER)
         # ======================================================
-        output_dir = Path("reports")
-        output_dir.mkdir(exist_ok=True)
-
-        output_file = output_dir / (path.stem + "_report.png")
-
         audio_data = {
             "filename": path.name,
             "info": info,
@@ -141,73 +133,70 @@ def analyze_audio(audio_file, progress=gr.Progress()):
             "spectral": spectral,
             "lufs": lufs,
             "issues": issues,
-            # Use the full score dictionary (correct)
-            "score": {
-                "score": score,
-                "grade": grade,
-                "quality": quality,
-                "recommendation": recommendation,
-                "cleanliness_score": cleanliness_score,
-                "processing_severity": processing_severity,
-                "critical": critical,
-                "high": high,
-                "medium": medium,
-                "low": low,
-                "color": color
-            },
+            "score": score_dict,
             "synthetic": synthetic,
             "timestamp": datetime.now().strftime("%Y-%m-%d %H:%M:%S")
         }
 
-
+        # ======================================================
+        # REPORT GENERATION
+        # ======================================================
         progress(0.92, desc="Rendering PNG report...")
+
+        output_dir = Path("reports")
+        output_dir.mkdir(exist_ok=True)
+
+        output_file = output_dir / f"{path.stem}_report.png"
+
         create_report(audio_data, str(output_file))
 
         # ======================================================
-        # SUMMARY MARKDOWN
+        # MARKDOWN SUMMARY (CORRECTED)
         # ======================================================
 
-        progress(1.0, desc="Done!")
+        progress(1.00, desc="Complete!")
+
+        s = score_dict
 
         md = f"""
-# 🎵 Analysis Complete — Audio Forensic Analyzer
+# 🎵 Audio Forensic Analysis — Summary Report
 
-## File Information
-- **Filename:** `{audio_data['filename']}`
+## 📁 File
+- **Name:** `{audio_data['filename']}`
 - **Duration:** {info['duration']:.2f}s  
 - **Sample Rate:** {info['samplerate']} Hz  
-- **Channels:** {info['channels']}  
+- **Channels:** {info['channels']}
 
 ---
 
-## 🔍 Quality Assessment
-- **Score:** {score['score']}/100  
-- **Grade:** {score['grade']}  
-- **Quality:** {score['quality']}  
-- **Recommendation:** {score['recommendation']}
+## 🧪 Quality Assessment
+- **Score:** {s['score']}/100  
+- **Grade:** {s['grade']}  
+- **Quality:** {s['quality']}  
+- **Recommendation:** {s['recommendation']}
 
 ---
 
-## 🔧 Time-Domain Stats
+## 🔧 Time-Domain Metrics
 | Metric | Value |
 |--------|--------|
-| Peak Level | {time_stats['peak_db']:.2f} dBFS |
-| RMS Level | {time_stats['rms_db']:.2f} dBFS |
-| Crest Factor | {time_stats['crest_factor_db']:.2f} dB |
+| Peak | {time_stats['peak_db']:.2f} dBFS |
+| RMS | {time_stats['rms_db']:.2f} dBFS |
+| Crest | {time_stats['crest_factor_db']:.2f} dB |
 | SNR | {time_stats['snr_db']:.1f} dB |
 | ZCR | {time_stats['zero_crossing_rate']:.4f} |
 
 """
 
         if lufs is not None:
-            md += f"| Integrated LUFS | {lufs:.2f} LUFS |\n"
+            md += f"- **Integrated LUFS:** {lufs:.2f} LUFS\n"
 
         md += f"""
 ---
 
-## 🎚 Synthetic Voice Estimate (Informational Only)
-- **Probability:** {synthetic['synthetic_probability']:.2f}
-- **Label:** **{synthetic['synthetic_label']}**
+## 🤖 Synthetic Voice (Informational Only)
+- **Probability:** {synthetic['synthetic_probability']:.3f}
+- **Label:** {synthetic['synthetic_label']}
 
 ---
 
@@ -215,18 +204,17 @@ def analyze_audio(audio_file, progress=gr.Progress()):
 """
 
         if issues:
-            icons = {"CRITICAL": "🔴", "HIGH": "🟠", "MEDIUM": "🟡", "LOW": "🟢"}
+            icon_map = {"CRITICAL": "🔴", "HIGH": "���", "MEDIUM": "🟡", "LOW": "🟢"}
             for issue, sev, desc in issues:
-                md += f"- {icons.get(sev,'⚪')} **[{sev}] {issue}** — {desc}\n"
+                md += f"- {icon_map.get(sev,'⚪')} **[{sev}] {issue}** — {desc}\n"
         else:
-            md += "- ✅ No significant issues\n"
+            md += "- ✅ No significant issues detected\n"
 
         md += f"""
-
 ---
 
-📊 **Report PNG saved:** `{output_file.name}`
-
+📊 **PNG Report Saved:** `{output_file.name}`  
+🕒 Generated: {audio_data['timestamp']}
 """
 
         return str(output_file), md
@@ -245,26 +233,23 @@ with gr.Blocks(title="Audio Forensic Analyzer", theme="soft") as demo:
 
     gr.Markdown("""
 # 🎧 Audio Forensic Analyzer
-Upload an audio file and generate a **full forensic report**:
-- HF/LF rolloff  
-- Filtering (LPF/HPF/Brickwall)  
-- Compression & clipping  
+Upload an audio file to generate a forensic-quality report including:
+- HF/LF analysis  
+- Filtering detection (HPF / LPF / Brickwall)  
 - Noise reduction artifacts  
+- Compression indicators  
 - Spectral notches  
 - Loudness (LUFS)  
-- Synthetic Voice Probability (Informational Only)
-
-Report includes a PNG + formatted summary.
+- Synthetic voice estimation  
+The tool outputs a **PNG forensic report** + **Summary Markdown**.
 """)
 
     with gr.Row():
         with gr.Column(scale=1):
             audio_input = gr.Audio(
                 label="📁 Upload Audio",
-                type="filepath",
-                sources=["upload"]
+                type="filepath"
             )
-
             analyze_button = gr.Button("🔍 Analyze Audio", variant="primary")
 
         with gr.Column(scale=2):
@@ -283,6 +268,5 @@ Report includes a PNG + formatted summary.
     )
 
 
-# Run in HuggingFace Space
 if __name__ == "__main__":
     demo.launch(server_name="0.0.0.0", server_port=7860, show_error=True)