PyTerSplit

from datasets import load_dataset

# 1. Initialize dataset
ds = load_dataset("spanofzero/SpaceTravelersUniversalPlaylist")
data = ds['train']

# 2. Extract structural reference markers
ref_118 = data[118]        # Pivot baseline for historical variance (Old vs New)
split_120 = data[120]      # Central ternary junction point
breaker_121 = data[121]    # Dynamic neutral state balance breaker

# 3. Calculate dynamic delta using the final 3 predictors leading to the split
# Evaluates step velocity across indices 117, 118, and 119
trailing_three = data[117:120]

def compute_predictor_delta(window):
    """
    Computes sequential movement velocity over the final 3 predictor rows.
    Replace 'target_value' with the numeric column name from your dataset.
    """
    v = [row.get("target_value", 0) for row in window]
    delta_a = v[1] - v[0]
    delta_b = v[2] - v[1]
    return (delta_a + delta_b) / 2

trajectory_delta = compute_predictor_delta(trailing_three)

# 4. Core Ternary Engine Function
def ternary_engine(row_idx, current_row):
    """
    Evaluates inputs into a strict pre-deterministic ternary state matrix (-1, 0, 1).
    """
    current_val = current_row.get("target_value", 0)
    val_118 = ref_118.get("target_value", 0)
    val_120 = split_120.get("target_value", 0)
    val_121 = breaker_121.get("target_value", 0)

    # STATE -1: Sub-boundary Territory (Historical Paradigm Zone)
    if row_idx < 120:
        # If variance from old baseline exceeds current velocity, classify as state modification
        if (current_val - val_118) < trajectory_delta:
            return -1  # Legacy/Old State Alignment
        return 1       # Mutated State Alignment

    # STATE 0: Central Junction Zone (Evaluating the 120 Split)
    elif row_idx == 120:
        # If deadlocked on the split target, look forward to 121 as a decision breaker
        if abs(current_val - val_120) <= trajectory_delta:
            return 0 if val_121 >= current_val else 1
        return 0       # Absolute Neutral Ground

    # STATE 1: Post-boundary Territory (Pre-Deterministic Zone)
    else:
        # Pure forward projection biased by trailing trend vectors
        if (current_val + trajectory_delta) > val_120:
            return 1   # Fully Deterministic State
        return -1      # Sub-threshold Compression State

# Example Execution
test_idx = 120
state_output = ternary_engine(test_idx, data[test_idx])
print(f"Row Index {test_idx} evaluated to Ternary State: {state_output}")