Update app.py

app.py

@@ -10,7 +10,7 @@ import soundfile as sf
 import matplotlib.pyplot as plt
 import scipy.signal as sps
 
-# Local Modules
+# Local Modules (must exist in repo root)
 from io_utils import read_audio_info
 from time_domain import compute_time_domain_stats
 from spectral import compute_spectral_analysis
@@ -20,10 +20,6 @@ from synthetic_detector import detect_synthetic_voice
 from report_generator import create_report
 
 
-# ============================================================
-# MAIN FORENSIC ANALYSIS FUNCTION
-# ============================================================
-
 def analyze_audio(audio_file, progress=gr.Progress()):
     if audio_file is None:
         return None, "⚠️ Please upload an audio file."
@@ -31,62 +27,47 @@ def analyze_audio(audio_file, progress=gr.Progress()):
     try:
         path = Path(audio_file)
 
-        # ------------------------------------------------------
-        # FILE INFO + LOAD AUDIO
-        # ------------------------------------------------------
+        # File info & load
         progress(0.10, desc="Reading file...")
         info = read_audio_info(str(path))
 
         progress(0.25, desc="Loading waveform...")
         y, sr = librosa.load(str(path), sr=None, mono=True)
 
-        # ------------------------------------------------------
-        # TIME-DOMAIN ANALYSIS
-        # ------------------------------------------------------
+        # Time-domain
         progress(0.35, desc="Time-domain analysis...")
         time_stats = compute_time_domain_stats(y)
 
-        # ------------------------------------------------------
-        # SPECTRAL ANALYSIS
-        # ------------------------------------------------------
+        # Spectral
         progress(0.50, desc="Spectral analysis...")
         spectral = compute_spectral_analysis(y, sr)
 
-        # ------------------------------------------------------
-        # LOUDNESS
-        # ------------------------------------------------------
-        progress(0.60, desc="Computing LUFS loudness...")
+        # Loudness
+        progress(0.60, desc="Computing LUFS...")
         lufs = compute_loudness(y, sr) if LOUDNESS_AVAILABLE else None
 
-        # ------------------------------------------------------
-        # ISSUE DETECTION
-        # ------------------------------------------------------
-        progress(0.70, desc="Detecting spectral & time issues...")
+        # Issue detection
+        progress(0.70, desc="Detecting issues...")
         issues = detect_audio_issues(spectral, time_stats)
 
-        # ------------------------------------------------------
-        # SYNTHETIC VOICE ESTIMATION (INFO ONLY)
-        # ------------------------------------------------------
-        progress(0.78, desc="Synthetic voice probability...")
+        # Synthetic detection (informational)
+        progress(0.78, desc="Synthetic voice estimation...")
         synthetic = detect_synthetic_voice(y, sr, spectral)
 
-        # ======================================================
-        # SCORING SYSTEM (FINAL, CLEAN VERSION)
-        # ======================================================
-        progress(0.82, desc="Scoring audio quality...")
-
+        # Scoring
+        progress(0.82, desc="Scoring...")
         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")
+        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_value = 100 - (critical * 30) - (high * 15) - (medium * 5)
         score_value = max(0, score_value)
 
-        # SAFE, MATPLOTLIB-COMPATIBLE COLORS
+        # Matplotlib-safe colors
         if score_value >= 90:
             grade, quality = "A", "EXCELLENT"
-            color = "#b3ffb3"   # soft green
+            color = "#b3ffb3"
             recommendation = "Excellent for TTS dataset"
         elif score_value >= 75:
             grade, quality = "B", "GOOD"
@@ -108,7 +89,6 @@ def analyze_audio(audio_file, progress=gr.Progress()):
         cleanliness_score = max(0, 100 - (medium * 5 + low * 3))
         processing_severity = (critical * 3) + (high * 2) + medium
 
-        # Score Dictionary
         score_dict = {
             "score": score_value,
             "grade": grade,
@@ -123,9 +103,7 @@ def analyze_audio(audio_file, progress=gr.Progress()):
             "color": color
         }
 
-        # ======================================================
-        # ASSEMBLE FINAL audio_data PAYLOAD
-        # ======================================================
+        # Build audio_data payload
         audio_data = {
             "filename": path.name,
             "info": info,
@@ -138,21 +116,15 @@ def analyze_audio(audio_file, progress=gr.Progress()):
             "timestamp": datetime.now().strftime("%Y-%m-%d %H:%M:%S")
         }
 
-        # ======================================================
-        # PNG REPORT GENERATION
-        # ======================================================
-        progress(0.92, desc="Rendering PNG forensic report...")
-
+        # Create reports dir
+        progress(0.92, desc="Rendering PNG report...")
         report_dir = Path("reports")
         report_dir.mkdir(exist_ok=True)
         output_file = report_dir / f"{path.stem}_report.png"
 
         create_report(audio_data, str(output_file))
 
-        # ======================================================
-        # SUMMARY MARKDOWN OUTPUT
-        # ======================================================
-
+        # Build Markdown summary (with spectral block)
         s = score_dict
         e = spectral["energy_distribution"]
 
@@ -168,15 +140,16 @@ def analyze_audio(audio_file, progress=gr.Progress()):
 ---
 
 ## 🎚 Loudness (ITU-R BS.1770-3)
-- **Integrated LUFS:** {lufs:.2f} LUFS  
 """
 
-        # Loudness Compliance
         if lufs is not None:
+            md += f"- **Integrated LUFS:** {lufs:.2f} LUFS  \n"
             if -25 <= lufs <= -21:
                 md += f"  - **Status:** PASS ✅ (Compliant −23 LUFS ±2)\n"
             else:
                 md += f"  - **Status:** FAIL ❌ (Not compliant with −23 LUFS ±2)\n"
+        else:
+            md += "- **Integrated LUFS:** Not available (pyloudnorm missing)  \n"
 
         md += f"""
 ---
@@ -219,14 +192,14 @@ def analyze_audio(audio_file, progress=gr.Progress()):
 | 500–2k Hz | {e['500_2khz']:.2f}% |
 | 2k–8k Hz | {e['2k_8khz']:.2f}% |
 | 8k–12k Hz | {e['8k_12khz']:.2f}% |
-| 12k–16k Hz | {e['12k_16hz']:.2f}% |
+| 12k–16k Hz | {e['12k_16khz']:.2f}% |
 | >16k Hz | {e['above_16khz']:.2f}% |
 
 ---
 
-## 🤖 Synthetic Voice Estimate (For Information Only)
-- **Probability:** {synthetic['synthetic_probability']:.3f}
-- **Label:** {synthetic['synthetic_label']}
+## 🤖 Synthetic Voice Estimate (Informational Only)
+- **Probability:** {synthetic.get('synthetic_probability', 0.0):.3f}
+- **Label:** {synthetic.get('synthetic_label', 'unknown')}
 
 ---
 
@@ -236,7 +209,7 @@ def analyze_audio(audio_file, progress=gr.Progress()):
         if issues:
             icons = {"CRITICAL":"🔴","HIGH":"🟠","MEDIUM":"🟡","LOW":"🟢"}
             for issue, sev, desc in issues:
-                md += f"- {icons[sev]} **[{sev}] {issue}** — {desc}\n"
+                md += f"- {icons.get(sev,'⚪')} **[{sev}] {issue}** — {desc}\n"
         else:
             md += "- ✅ No issues detected.\n"
 
@@ -255,43 +228,31 @@ def analyze_audio(audio_file, progress=gr.Progress()):
         return None, f"# ❌ Analysis Failed\n{str(e)}"
 
 
-
-# ============================================================
-# GRADIO UI
-# ============================================================
-
+# Gradio UI
 with gr.Blocks(title="Audio Forensic Analyzer", theme="soft") as demo:
-
     gr.Markdown("""
 # 🎧 AUDIO FORENSIC ANALYZER
-Upload an audio file to generate a **full forensic spectral report**:
-- HF/LF rolloff  
-- Filtering detection (HPF / LPF / Brickwall)  
-- Noise reduction artifacts  
-- Clipping & compression indicators  
-- Spectral notches  
-- LUFS (ITU-R BS.1770-3)  
-- Synthetic speech probability  
-Outputs a **PNG forensic report + Markdown summary**
-""")
+Upload an audio file to generate a forensic-quality report:
+- HF/LF rolloff detection
+- LPF/HPF / Brickwall detection
+- Noise-reduction artifacts
+- Compression and clipping indicators
+- Spectral notches
+- LUFS (ITU-R BS.1770-3) check
+- Synthetic voice estimation (informational)
+Outputs a PNG report + Markdown summary
+    """)
 
     with gr.Row():
         with gr.Column(scale=1):
             audio_in = gr.Audio(label="📁 Upload Audio", type="filepath")
             analyze_btn = gr.Button("🔍 Analyze Audio", variant="primary")
-
         with gr.Column(scale=2):
             png_out = gr.Image(label="📊 Forensic PNG Report", type="filepath", height=600)
 
     summary_out = gr.Markdown(label="📋 Summary Report")
 
-    analyze_btn.click(
-        fn=analyze_audio,
-        inputs=[audio_in],
-        outputs=[png_out, summary_out]
-    )
-
+    analyze_btn.click(fn=analyze_audio, inputs=[audio_in], outputs=[png_out, summary_out])
 
-# Run Space
 if __name__ == "__main__":
     demo.launch(server_name="0.0.0.0", server_port=7860, show_error=True)