File size: 11,659 Bytes
66f749a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 | """
Simplified simulation that runs without Ray (compatible with Celery)
Uses Gemini directly to simulate agent reactions
"""
import os
import logging
import random
from typing import Dict, Any, List, Optional
from pathlib import Path
# Load .env file from project root
from dotenv import load_dotenv
project_root = Path(__file__).parent.parent
load_dotenv(project_root / ".env")
from google import genai
logger = logging.getLogger(__name__)
def get_gemini_client():
"""Get Gemini client"""
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
# Try loading from parent .env again
load_dotenv(Path(__file__).parent.parent / ".env")
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
raise ValueError("GEMINI_API_KEY not set in .env file")
return genai.Client(api_key=api_key)
def generate_agent_profiles(num_agents: int, demographic_filter: Optional[Dict] = None) -> List[Dict]:
"""Generate diverse agent profiles"""
ages = list(range(18, 70))
genders = ["Male", "Female", "Non-binary"]
locations = ["Urban", "Suburban", "Rural"]
education_levels = ["High School", "Bachelor's", "Master's", "PhD"]
income_levels = ["Low", "Middle", "Upper-Middle", "High"]
values = [
"Family-oriented", "Career-focused", "Environmentalist",
"Traditional", "Progressive", "Religious", "Secular",
"Health-conscious", "Tech-savvy", "Minimalist"
]
profiles = []
for i in range(num_agents):
profile = {
"agent_id": f"agent_{i+1:04d}",
"age": random.choice(ages),
"gender": random.choice(genders),
"location": random.choice(locations),
"education": random.choice(education_levels),
"income": random.choice(income_levels),
"values": random.sample(values, k=random.randint(2, 4))
}
# Apply demographic filter if provided
if demographic_filter:
if "age_min" in demographic_filter:
profile["age"] = max(profile["age"], demographic_filter["age_min"])
if "age_max" in demographic_filter:
profile["age"] = min(profile["age"], demographic_filter["age_max"])
if "gender" in demographic_filter:
profile["gender"] = demographic_filter["gender"]
profiles.append(profile)
return profiles
def simulate_agent_reaction(client, ad_content: str, profile: Dict) -> Dict:
"""Simulate a single agent's reaction using Gemini"""
prompt = f"""You are simulating how a person would react to an advertisement.
PERSON PROFILE:
- Age: {profile['age']}
- Gender: {profile['gender']}
- Location: {profile['location']} area
- Education: {profile['education']}
- Income Level: {profile['income']}
- Core Values: {', '.join(profile['values'])}
ADVERTISEMENT CONTENT:
{ad_content[:2000]} # Limit context size
Based on this person's profile, determine their reaction to this advertisement.
Respond with EXACTLY this JSON format (no markdown, no extra text):
{{"opinion": "POSITIVE|NEUTRAL|NEGATIVE", "reasoning": "brief 1-2 sentence explanation", "would_share": true|false, "controversy_flag": true|false}}"""
try:
response = client.models.generate_content(
model="gemini-3-flash-preview",
contents=prompt,
config={
"max_output_tokens": 150,
"temperature": 0.7
}
)
text = response.text.strip()
# Parse JSON response
import json
# Clean up the response
if text.startswith("```"):
text = text.split("```")[1]
if text.startswith("json"):
text = text[4:]
text = text.strip()
result = json.loads(text)
return {
"agent_id": profile["agent_id"],
"profile": profile,
"opinion": result.get("opinion", "NEUTRAL"),
"reasoning": result.get("reasoning", ""),
"would_share": result.get("would_share", False),
"controversy_flag": result.get("controversy_flag", False)
}
except Exception as e:
logger.warning(f"Failed to simulate agent {profile['agent_id']}: {e}")
# Return a neutral default
return {
"agent_id": profile["agent_id"],
"profile": profile,
"opinion": random.choice(["POSITIVE", "NEUTRAL", "NEGATIVE"]),
"reasoning": "Simulation fallback",
"would_share": random.random() > 0.7,
"controversy_flag": False
}
def detect_controversies(reactions: List[Dict]) -> List[Dict]:
"""Detect controversial patterns in reactions"""
flags = []
# Group by demographics
groups = {
'age': {},
'gender': {},
'location': {},
'values': {}
}
for reaction in reactions:
profile = reaction.get('profile', {})
opinion = reaction.get('opinion')
if not opinion:
continue
# Age groups
age = profile.get('age', 30)
age_bracket = f"{(age // 10) * 10}s"
if age_bracket not in groups['age']:
groups['age'][age_bracket] = []
groups['age'][age_bracket].append(reaction)
# Gender
gender = profile.get('gender', 'Unknown')
if gender not in groups['gender']:
groups['gender'][gender] = []
groups['gender'][gender].append(reaction)
# Location
location = profile.get('location', 'Unknown')
if location not in groups['location']:
groups['location'][location] = []
groups['location'][location].append(reaction)
# Values
for value in profile.get('values', []):
if value not in groups['values']:
groups['values'][value] = []
groups['values'][value].append(reaction)
# Check each group for high negativity
for group_type, group_data in groups.items():
for group_name, group_reactions in group_data.items():
if len(group_reactions) < 3:
continue
negative_count = sum(1 for r in group_reactions if r.get('opinion') == 'NEGATIVE')
total = len(group_reactions)
negative_rate = negative_count / total
if negative_rate > 0.5:
if negative_rate > 0.8:
severity = "CRITICAL"
elif negative_rate > 0.7:
severity = "HIGH"
elif negative_rate > 0.6:
severity = "MEDIUM"
else:
severity = "LOW"
sample_reactions = [
{
"agent_id": r.get('agent_id'),
"reasoning": r.get('reasoning', '')[:100]
}
for r in group_reactions if r.get('opinion') == 'NEGATIVE'
][:3]
flags.append({
"flag_type": f"{group_type.upper()}_BACKLASH",
"severity": severity,
"description": f"{int(negative_rate * 100)}% of {group_type}={group_name} reacted negatively",
"affected_demographics": {group_type: group_name},
"sample_agent_reactions": sample_reactions
})
# Sort by severity
severity_order = {"CRITICAL": 0, "HIGH": 1, "MEDIUM": 2, "LOW": 3}
flags.sort(key=lambda x: severity_order.get(x['severity'], 4))
return flags[:10]
def run_simulation_simple(
experiment_id: str,
ad_content: str,
demographic_filter: Optional[Dict[str, Any]] = None,
num_agents: int = 10,
simulation_days: int = 5,
redis_client = None
) -> Dict[str, Any]:
"""
Run a simplified simulation without Ray
This version is compatible with Celery and uses Gemini directly
"""
logger.info(f"Starting simplified simulation {experiment_id} with {num_agents} agents")
try:
client = get_gemini_client()
# Generate profiles
profiles = generate_agent_profiles(num_agents, demographic_filter)
logger.info(f"Generated {len(profiles)} agent profiles")
# Update progress
if redis_client:
import json
redis_client.setex(f"sim:{experiment_id}:status", 60, json.dumps({
"progress": 10, "current_day": 0, "active_agents": 0
}))
# Simulate each agent's reaction
reactions = []
for i, profile in enumerate(profiles):
reaction = simulate_agent_reaction(client, ad_content, profile)
reactions.append(reaction)
# Update progress
if redis_client and i % 5 == 0:
progress = 10 + int((i / len(profiles)) * 80)
import json
redis_client.setex(f"sim:{experiment_id}:status", 60, json.dumps({
"progress": progress,
"current_day": min(i // (len(profiles) // simulation_days + 1) + 1, simulation_days),
"active_agents": i + 1
}))
logger.info(f"Simulated {len(reactions)} agent reactions")
# Calculate results
opinions = [r.get('opinion') for r in reactions if r.get('opinion')]
sentiment_counts = {
"positive": sum(1 for o in opinions if o == 'POSITIVE'),
"neutral": sum(1 for o in opinions if o == 'NEUTRAL'),
"negative": sum(1 for o in opinions if o == 'NEGATIVE')
}
total = len(opinions) or 1
strong_reactions = sentiment_counts['positive'] + sentiment_counts['negative']
engagement_score = (strong_reactions / total) * 100
# Detect controversies
risk_flags = detect_controversies(reactions)
# Prepare agent logs
agent_logs = [
{
"agent_id": r['agent_id'],
"event_type": "AD_REACTION",
"event_data": {
"opinion": r.get('opinion'),
"reasoning": r.get('reasoning'),
"would_share": r.get('would_share')
}
}
for r in reactions[:100] # Limit logs
]
# Final progress update
if redis_client:
import json
redis_client.setex(f"sim:{experiment_id}:status", 60, json.dumps({
"progress": 100, "current_day": simulation_days, "active_agents": len(reactions)
}))
logger.info(f"Simulation complete. Engagement score: {engagement_score:.1f}")
return {
"engagement_score": round(engagement_score, 2),
"sentiment_breakdown": sentiment_counts,
"total_agents": len(reactions),
"responding_agents": len(opinions),
"risk_flags": risk_flags,
"agent_logs": agent_logs
}
except Exception as e:
logger.error(f"Simulation failed: {e}")
raise
|