#!/usr/bin/env python3
from seed_common import run_seed

# Generate 500 distinct prophecy module descriptions (forecasting modules)

def gen_prophecy(n: int):
    domains = [
        "market volatility", "equity returns", "crypto price movement", "energy demand",
        "load forecasting", "weather extremes", "hurricane landfall", "wildfire risk",
        "epidemiology spread", "hospital admissions", "supply chain delays", "port congestion",
        "shipping times", "inventory replenishment", "customer churn", "lifetime value",
        "ad conversion", "marketing lift", "sales pipeline", "cash flow", "credit risk",
        "fraud detection", "cyber intrusion", "phishing campaigns", "patch urgency",
        "geopolitical risk", "election turnout", "policy passage", "legislative gridlock",
        "sports outcomes", "injury likelihood", "travel demand", "hotel occupancy",
        "traffic flow", "accident risk", "air quality index", "earthquake aftershocks",
        "product demand", "feature adoption", "release risk", "incident rate", "SLA breach",
        "compute capacity", "GPU queue time", "training time", "inference latency",
    ]
    methods = [
        "Bayesian structural time series", "state-space model", "seasonal ARIMA",
        "prophet-style additive model", "Gaussian Process regression", "quantile regression",
        "gradient boosted trees", "random forest", "temporal convolution network",
        "transformer encoder", "graph temporal network", "survival analysis",
        "copula-based simulation", "Monte Carlo scenario generation", "change point detection",
        "extreme value theory", "hierarchical modeling", "spatio-temporal kriging",
    ]
    horizons = [
        "nowcasting", "short-term (hours)", "near-term (days)", "mid-term (weeks)", "long-term (months)",
    ]
    constraints = [
        "with uncertainty quantification", "with prediction intervals", "with conformal bands",
        "with causal adjustments", "with holidays and events", "with regime shifts",
        "with exogenous regressors", "with missing data imputation",
    ]

    seen = set()
    i = 0
    for d in domains:
        for m in methods:
            phrase = f"Prophecy: {m} for {d}"
            if phrase in seen:
                continue
            seen.add(phrase)
            yield phrase
            i += 1
            if i >= n:
                return
    if i < n:
        for d in domains:
            for h in horizons:
                for c in constraints:
                    phrase = f"Prophecy: {h} {d} {c}"
                    if phrase in seen:
                        continue
                    seen.add(phrase)
                    yield phrase
                    i += 1
                    if i >= n:
                        return

if __name__ == "__main__":
    run_seed("prophecy", gen_prophecy)