import os, requests, time
MV = os.environ["MAVERA_API_KEY"]
MV_BASE = "https://app.mavera.io/api/v1"
MV_H = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}
VIDEO_VARIANTS = [
{"path": "ads/product_demo_trending_sound.mp4", "label": "Trending Sound", "audio": "Espresso remix — high energy, beat drops at cuts"},
{"path": "ads/product_demo_original_music.mp4", "label": "Original Music", "audio": "Custom lo-fi track — chill, ambient, brand-composed"},
{"path": "ads/product_demo_voiceover.mp4", "label": "Voiceover Only", "audio": "Founder narration — direct, educational, no background music"},
{"path": "ads/product_demo_asmr.mp4", "label": "ASMR / Product Sounds", "audio": "Product interaction sounds — tapping, pouring, unboxing"},
]
# 1. Upload all variants and run analysis
analyses = []
for variant in VIDEO_VARIANTS:
upload = requests.post(f"{MV_BASE}/assets",
headers={"Authorization": f"Bearer {MV}"},
files={"file": (variant["label"] + ".mp4", open(variant["path"], "rb"), "video/mp4")},
).json()
analysis = requests.post(f"{MV_BASE}/video-analysis", headers=MV_H, json={
"asset_id": upload["id"],
"analysis_types": ["emotional_arc", "mood_congruence", "pacing", "hook_score", "audio_impact"],
"metadata": {"label": variant["label"], "audio_description": variant["audio"]},
}).json()
analyses.append({"id": analysis["id"], "label": variant["label"], "audio": variant["audio"]})
time.sleep(0.5)
# 2. Poll all analyses
results = []
for a in analyses:
for _ in range(30):
time.sleep(3)
status = requests.get(f"{MV_BASE}/video-analysis/{a['id']}", headers=MV_H).json()
if status.get("status") == "completed":
break
r = status.get("results", {})
results.append({
"label": a["label"],
"audio": a["audio"],
"emotional_intensity": r.get("emotional_arc", {}).get("intensity_avg", 0),
"peak_emotion": r.get("emotional_arc", {}).get("peak_emotion", "N/A"),
"peak_timestamp": r.get("emotional_arc", {}).get("peak_timestamp", 0),
"mood_congruence": r.get("mood_congruence", {}).get("score", 0),
"pacing_score": r.get("pacing", {}).get("score", 0),
"hook_score": r.get("hook_score", {}).get("score", 0),
"audio_energy": r.get("audio_impact", {}).get("energy", 0),
})
# 3. Comparative report
results.sort(key=lambda x: -x["emotional_intensity"])
print("SOUND/MUSIC IMPACT ANALYSIS — Same Visual, Different Audio")
print("=" * 70)
print(f"{'Variant':<22} {'Emotion':<10} {'Mood Fit':<10} {'Pacing':<10} {'Hook':<8} {'Audio E'}")
print("-" * 70)
for r in results:
print(f"{r['label']:<22} {r['emotional_intensity']:.1f}/10 {r['mood_congruence']:.1f}/10 {r['pacing_score']:.1f}/10 {r['hook_score']}/100 {r['audio_energy']:.1f}")
# 4. Recommendation
best = results[0]
worst = results[-1]
delta = best["emotional_intensity"] - worst["emotional_intensity"]
print(f"\nWINNER: {best['label']}")
print(f" Emotional intensity: {best['emotional_intensity']:.1f}/10 (peak: {best['peak_emotion']} at {best['peak_timestamp']}s)")
print(f" Mood congruence: {best['mood_congruence']:.1f}/10")
print(f" vs worst ({worst['label']}): +{delta:.1f} emotional intensity")
print(f"\nRECOMMENDATION: Use '{best['label']}' audio treatment for this visual.")
if best["mood_congruence"] < 6:
print(f" ⚠ Mood congruence is low ({best['mood_congruence']:.1f}). Audio energy may not match visual tone — test with audience.")
