#!/usr/bin/env python3 # -*- coding: utf-8 -*- """ Main script to generate all figures for: Dynamic Order Dispersion and Volatility Persistence in a Simple Limit Order Book Model This is the turnkey script that runs all simulations and generates all figures. """ import numpy as np from random import choices import pandas as pd import matplotlib # Shutdown interactive plotting matplotlib.use('Agg') # Use the 'Agg' backend for non-GUI rendering import matplotlib.pyplot as plt from scipy.stats import kurtosis, norm from statsmodels.tsa.arima.model import ARIMA from time import time import csv import os # Import everything we need from LOBModel from LOBModel import ( modelIteration, fastautocorr, fastautocorr1, fastxcorr, archAdjust, getSummedBids, getSummedAsks ) # Import plotting function from make_plots4 from make_plots4 import plot_four_panel, process_data_file # ============================================================================ # PARAMETER DEFINITIONS # ============================================================================ def get_base_parameters(): """Base case parameters from Table 1""" return { 'nAgents': 2000, 'Tinit': 10000, 'Tmax': 200000, 'Lmin': 10, 'Lmax': 500, 'pf': 1000., 'deltaP': 0.025, 'sigmae': 0.0001, 'kMax': 0.1, 'adaptiveK': True, 'multipleFundamentals': True, 'simpleDemands': True, 'maxTrade': 1, 'tau': 25, 'sigmaF': 0.10, 'sigmaM': 0.00, 'sigmaN': 0.90, 'portfolioAdj': False, 'deltaT': 50, 'lam': 1.00, 'maxHold': 50., 'beta': 3., 'volLag': 150, 'agentsPerPeriod': 30, 'rhoBar': 0.999, 'orderSigma': 0.0, 'trimVol': False, 'probMarketOrder': 0.05 } # ============================================================================ # SIMULATION RUNNERS # ============================================================================ def run_simulation(run_name, params, seed=4242): """Run a single simulation""" print(f"\n{'='*70}") print(f"Running: {run_name}") print(f"{'='*70}") np.random.seed(seed) t0 = time() results = modelIteration(**params) t1 = time() print(f"Completed in {(t1-t0)/60:.2f} minutes") return results def save_simulation_csv(filename, results, params, seed): """Save simulation results to CSV""" (price, ret, rret, rRV, rRV2, totalV, rPrice, rSpread, rbidDepth, raskDepth, portDev, rportDev, holdings, dholdings, orders, rportDev_non_nan, Xsmooth_centered_non_nan, Xsmooth_trailing_non_nan, holdingsDiff, holdingsDiffABS, holdingsDist, holdingsDistABS, rVol, rOrders, marketBook, agentList, forecastSet, rdholdingsts, bidsdf, rbidSlope, asksdf, raskSlope, wealthByType, rpf, logPriceFund, rCRV, rtotalOrdersOnBook, orderFlow, pft) = results with open(filename, 'w', newline='') as f: writer = csv.writer(f) writer.writerow(["price", "ret", "rret", "rRV", "rRV2", "rCRV", "autocorrelation sum", "rvol", "rPrice", "spread", "totalOrders", "bid depth", "bidslope", "ask depth", "askslope", "tau", "sigmaE", "seed", "fundamental"]) for i in range(len(rret)): writer.writerow([ rPrice[i], 0, rret[i], rRV[i], rRV2[i], rCRV[i], 0.0, rVol[i], rPrice[i], rSpread[i], rtotalOrdersOnBook[i], rbidDepth[i], rbidSlope[i], raskDepth[i], raskSlope[i], params['tau'], params['sigmae'], seed, pft[i] ]) print(f"Saved: {filename}") # ============================================================================ # PLOTTING FUNCTIONS (using existing code) # ============================================================================ # vol distribution facts def create_vol_facts_plot(df, output_file, title="", start=500): start = 0 rret = df['rret'].values rRV = df['rRV'].values ret = rret[start:] vol = rRV[start:] svol = np.sqrt(vol) lvol = np.log(vol) fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(8, 4)) # Panel 1 : Normalized returns normret = ret/svol nbins = 50 n, bins, patches = ax[0].hist(normret, nbins, density=True,facecolor='green', alpha=0.5) ax[0].grid() mu = np.mean(normret) sigma = np.std(normret) xmin, xmax = ax[0].get_xlim() # Generate normal pdf from scipy.stats x = np.linspace(xmin, xmax, 100) p = norm.pdf(x, mu, sigma) ax[0].plot(x, p, 'k', linewidth=2) # y = plt.mlab.normpdf(bins,mu,sigma) # ax[0, 1].plot(bins, y, 'r') ax[0].set_title('Standardized returns, kurtosis = '+str(round(kurtosis(normret, fisher=False, bias=False),1))) # panel 2 : log vol histogram nbins = 50 n, bins, patches = ax[1].hist(lvol, nbins, density=True,facecolor='green', alpha=0.5) ax[1].grid() mu = np.mean(lvol) sigma = np.std(lvol) xmin, xmax = ax[1].get_xlim() # Generate normal pdf from scipy.stats x = np.linspace(xmin, xmax, 100) p = norm.pdf(x, mu, sigma) ax[1].plot(x, p, 'k', linewidth=2) # y = plt.mlab.normpdf(bins,mu,sigma) # ax[0, 1].plot(bins, y, 'r') ax[1].set_title('Log(RV), kurtosis = '+str(round(kurtosis(lvol, fisher=False, bias=False),1))) plt.tight_layout() plt.savefig(output_file+".png", dpi=300, bbox_inches='tight') plt.savefig(output_file+".pdf", dpi=300, bbox_inches='tight') print(f"Saved: {output_file}") plt.close() def create_long_memory_plotold(rRV, output_file, title="", start=500): """Create long memory variance scaling plot""" fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(10, 10)) if title: fig.suptitle(title, fontsize=14, fontweight='bold') start = 500 rv = rRV[start:] log_rv = np.log(rv + 1e-10) # Variance scaling max_agg = min(75, len(log_rv) // 10) aggregations = np.arange(1, max_agg) variances = np.zeros(len(aggregations)) for i, K in enumerate(aggregations): n_blocks = len(log_rv) // K block_sums = np.array([np.sum(log_rv[j*K:(j+1)*K]) for j in range(n_blocks)]) variances[i] = np.var(block_sums) log_agg = np.log(aggregations) log_var = np.log(variances) slope = np.polyfit(log_agg, log_var, 1)[0] ax1.loglog(aggregations, variances, 'o-', label=f'Slope={slope:.2f}') ax1.set_xlabel('Aggregation (Days)') ax1.set_ylabel('var(log(RV))_h') ax1.set_title('Variance scaling') ax1.grid(True, which="both", ls="-", alpha=0.2) ax1.legend() # Autocorrelations max_lag = min(70, len(log_rv) // 3) acf = fastautocorr(log_rv, max_lag) lags = np.arange(max_lag + 1) ax2.plot(lags[1:], acf[1:], 'o-') ax2.axhline(0, color='k', linestyle='-', linewidth=0.5) ax2.set_xlabel('Lag') ax2.set_ylabel('Autocorrelation') ax2.set_title('Log(RV) autocorrelations') ax2.grid(True) plt.tight_layout() plt.savefig(output_file, dpi=300, bbox_inches='tight') print(f"Saved: {output_file}") plt.close() # Find non overlapping 1 to k horizon variances of vol estimates # Part of long memory scaling (see equation 16 in paper) def varhorizon(vol,k): svol = np.cumsum(vol) varh = [] varn = [] for j in range(1,k): dvol = np.diff( svol[::j]) varh.append(np.var(dvol)) varn.append(j) # make sure return numpy arrays. I think this routine can data pandas dataframes as args, but not sure return np.array(varh), np.array(varn) def ols(y,x): y = y-np.mean(y) x = x-np.mean(x) vx = np.var(x) cvxy = np.mean( x * y) beta = cvxy/vx return(beta) def create_long_memory_plot(rRV, output_file, title="", start=500): from random import choices Delta = 1 # Start = 25000 rRV = rRV[start::1] # Use log(rv) as is standard vol = np.log(rRV) # generate bootstrap volatility series voln = np.array(choices(vol,k=len(vol)+1)) # Range for ACF estimations timeRange = 75 timeRangehalf = 25 varRange = 75 xindex = range(1,varRange+1) xindex2 = range(1,timeRange+1) # first order autocorrelation acf1 = fastautocorr(vol,timeRange)[1:] # theoretical ACF for an AR(1) proces acf1th = acf1[1]**range(1,timeRange+1) # Find summed variances for possible scaling varh, varn = varhorizon(vol,varRange) # Do the same for bootstrap scrambled sample varhn, varnn = varhorizon(voln,varRange) # Find slope of log/log regression (this is equal 2H, H=self similarity) # Also, d = H-(1/2) beta = ols(np.log(varh),np.log(varn)) betastr = str(round(beta,2)) leg1 = 'Base case: Slope ='+betastr # Same for bootstrap betarw = ols(np.log(varhn),np.log(varnn)) betastr = str(round(betarw,2)) leg2 = 'IID Bootstrap: Slope ='+betastr # Find H and d, and gamma (see equations 17-19) H = 0.5*beta d = H - 0.5 gamma = 2-2*H # gamma = 0.41 print("H,d,gamma = ",H,d,gamma) # Line up long memory ACF so that it crosses data ACF at timeRangeHalf timeRangehalf = 30 A = acf1[timeRangehalf-1]*timeRangehalf**(gamma) # Long memory ACF (see equation 18) theoryACF = A * varn**(-gamma) # Now get ARMA(1,1) comparison arma = ARIMA(vol, order=(1,0,1)).fit() print(arma.summary()) theta = arma.params[2] rho = arma.params[1] armaA = (1+rho*theta)*(rho+theta)/( 1+2*rho*theta + theta**2) acfARMA = armaA * rho**(varn) fig, ax = plt.subplots(nrows=2, ncols=1, figsize=(6, 6)) fig.subplots_adjust(hspace=0.6, wspace=0.4) # fig.suptitle('Returns for tau = '+str(tau)+", simgae = "+str(sigmae)) ax[0].set_xscale('log') ax[0].set_yscale('log') ax[0].plot(varn,varh) ax[0].plot(varnn,varhn) ax[0].grid() ax[0].set_title('Variance scaling') ax[0].set_xlabel('Aggregation (Days)') ax[0].set_ylabel('$var(\log(RV)_h)$') ax[0].legend([leg1,leg2]) ax[1].plot(xindex2,acf1) ax[1].plot(xindex2,acf1th) ax[1].grid() ax[1].set_title("Log(RV) autocorrelations") # ax[1].set_xlabel('Lag(j)') ax[1].set_ylabel('Autocorrelation') ax[1].set_xlabel('Lag') ax[1].plot(varn,theoryACF) leg1 = 'Base case' ax[1].legend([leg1,'AR(1)','Long memory']) ax[1].plot(varn,acfARMA) ax[1].legend([leg1,'AR(1)','Long memory','ARMA(1,1)']) plt.tight_layout() plt.savefig(output_file+".png", dpi=300, bbox_inches='tight') plt.savefig(output_file+".pdf", dpi=300, bbox_inches='tight') print(f"Saved: {output_file}") plt.close() def create_orderbook_dynamics_plot(rret, rSpread, rPrice, rbidSlope, raskSlope, rbidDepth, raskDepth, output_file, title="", start=500): """Create 4-panel order book dynamics plot""" # fig, axes = plt.subplots(4, 1, figsize=(12, 10), sharex=True) fig, axes = plt.subplots(4,1,figsize=(6, 6)) fig.subplots_adjust(hspace=0.4, wspace=0.4) """ if title: fig.suptitle(title, fontsize=14, fontweight='bold') """ returns = rret[start:] spread = rSpread[start:] spread = 100*spread/rPrice slope = 0.5 * (np.abs(rbidSlope[start:]) + np.abs(raskSlope[start:])) depth = rbidDepth[start:] + raskDepth[start:] days = np.arange(len(returns)) axes[0].plot(days, returns) axes[0].set_title("Returns") axes[0].set_ylabel('') axes[0].grid(True) axes[1].plot(days, spread) axes[1].set_title("Bid/Ask Spread") axes[1].set_ylabel('% of price') axes[1].grid(True) axes[2].plot(days, slope) axes[2].set_title("Order Book Slope") axes[2].grid(True) axes[3].plot(days, depth) axes[3].set_title("Depth") axes[3].set_ylabel('Shares') axes[3].set_xlabel('Period') axes[3].grid(True) plt.tight_layout() plt.savefig(output_file+".png", dpi=300, bbox_inches='tight') plt.savefig(output_file+".pdf", dpi=300, bbox_inches='tight') print(f"Saved: {output_file}") plt.close() def create_cross_correlation_plot(rret, rRV, rCRV, rSpread, rbidSlope, raskSlope, rbidDepth, raskDepth, rVol,volatilityXC, output_file, title="", max_lag=50, start=500): """Create cross-correlation plot""" abs_ret = np.abs(rret[start:]) rv = rRV[start:] rv = np.sqrt(rv) crv = rCRV[start:] spread = rSpread[start:] slope = 0.5 * (np.abs(rbidSlope[start:]) + np.abs(raskSlope[start:])) depth = rbidDepth[start:] + raskDepth[start:] volume = rVol[start:] print(rv[0:10]) print(spread[0:10]) """ xcorr_spread = fastxcorr(abs_ret, spread, max_lag) xcorr_slope = fastxcorr(abs_ret, slope, max_lag) xcorr_depth = fastxcorr(abs_ret, depth, max_lag) xcorr_volume = fastxcorr(abs_ret, volume, max_lag) """ T = len(rv) T2 = int(T/2) """ rv = rv[T2:] crv = crv[T2:] spread = spread[T2:] slope = slope[T2:] depth = depth[T2:] volume = volume[T2:] """ xcorr_spread = fastxcorr(spread,rv, max_lag) xcorr_slope = fastxcorr(slope,rv, max_lag) xcorr_depth = fastxcorr(depth,rv, max_lag) xcorr_volume = fastxcorr(rv,volume, max_lag) xcorr_slopevol = fastxcorr(slope,volume,max_lag) xcorr_slopecvar = fastxcorr(slope,crv,max_lag) fig, ax = plt.subplots(figsize=(6, 6)) fig.subplots_adjust(hspace=0.4, wspace=0.4) lags = np.arange(-max_lag, max_lag + 1) if(volatilityXC): ax.plot(lags, xcorr_slope, lw=2, label='slope(t),vol(t+j)', alpha=0.7) ax.plot(lags, xcorr_spread, lw=2, label='spread(t),vol(t+j)', alpha=0.7) ax.plot(lags, xcorr_depth, lw=2, label='depth(t),vol(t+j)', alpha=0.7) # ax.plot(lags, xcorr_volume, lw=2, label='|r(t)| vs Volume', alpha=0.7) else : # vikyne cross correlations ax.plot(lags, xcorr_volume, lw=2, label='vol(t),volume(t+j)', alpha=0.7) ax.plot(lags, xcorr_slopevol, lw=2, label='slope(t),volume(t+j)', alpha=0.7) ax.plot(lags, xcorr_slopecvar, lw=2, label='slope(t),cvar(t+j)', alpha=0.7) ax.axhline(0, color='k', linestyle='-', linewidth=0.5) ax.axvline(0, color='k', linestyle='--', linewidth=0.5) ax.set_xlabel('Cross Correlation Lag (j)') ax.set_ylabel('Correlation') # No title for paper """ if title: ax.set_title(title) """ ax.legend() ax.grid(True, alpha=0.3) plt.tight_layout() plt.savefig(output_file+".png", dpi=300, bbox_inches='tight') plt.savefig(output_file+".pdf", dpi=300, bbox_inches='tight') print(f"Saved: {output_file}") plt.close() # ============================================================================ # MAIN EXECUTION # ============================================================================ def main(): """Main execution: run all experiments and generate all figures""" print("\n" + "="*70) print("LIMIT ORDER BOOK MODEL - FIGURE GENERATION") print("="*70) os.makedirs("figures", exist_ok=True) seed = 4242 seed = 14 experiments = {} # ======================================================================== # RUN SIMULATIONS # ======================================================================== print("\n" + "="*70) print("RUNNING SIMULATIONS") print("="*70) # Base case params_base = get_base_parameters() experiments['base'] = run_simulation("Base Case", params_base, seed) save_simulation_csv("dataOutputFile_base.csv", experiments['base'], params_base, seed) # Fixed shading params_fixed = get_base_parameters() params_fixed['adaptiveK'] = False experiments['fixed'] = run_simulation("Fixed Shading", params_fixed, seed) save_simulation_csv("dataOutputFile_fixed.csv", experiments['fixed'], params_fixed, seed) # Homogeneous fundamental params_homFund = get_base_parameters() params_homFund['multipleFundamentals'] = False experiments['homFund'] = run_simulation("Homogeneous Fundamentals", params_homFund, seed) save_simulation_csv("dataOutputFile_homFund.csv", experiments['homFund'], params_homFund, seed) # Additional experiments (optional - comment out to save time) if True : # Set to False to skip these params_rc = get_base_parameters() params_rc['tau'] = 75 experiments['reduced_cancel'] = run_simulation("Reduced Cancelations", params_rc, seed) save_simulation_csv("dataOutputFile_reduced_cancel.csv", experiments['reduced_cancel'], params_rc, seed) params_tv = get_base_parameters() params_tv['trimVol'] = True experiments['trimmed'] = run_simulation("Trimmed Volatility", params_tv, seed) save_simulation_csv("dataOutputFile_trimmed.csv", experiments['trimmed'], params_tv, seed) params_noise = get_base_parameters() params_noise['sigmae'] = 0.001 experiments['noise'] = run_simulation("Increased Noise", params_noise, seed) save_simulation_csv("dataOutputFile_noise.csv", experiments['noise'], params_noise, seed) # Long base run for vol scaling, double standard run params_baseLong = get_base_parameters() params_baseLong['Tmax'] = 400000 experiments['baseLong'] = run_simulation("Base Long", params_baseLong, seed) save_simulation_csv("dataOutputFile_baseLong.csv", experiments['baseLong'], params_baseLong, seed) # Long base run for vol scaling, double standard run, heterogeneous vol forecasts params_baseHetVol = get_base_parameters() params_baseHetVol['Tmax'] = 400000 params_baseHetVol['volLag'] = 0 # set for heterogeneous vol forecast experiments['baseHetVol'] = run_simulation("Base VolHet", params_baseHetVol, seed) save_simulation_csv("dataOutputFile_baseHetVol.csv", experiments['baseHetVol'], params_baseHetVol, seed) # ======================================================================== # GENERATE FIGURES # ======================================================================== print("\n" + "="*70) print("GENERATING FIGURES") print("="*70) # Unpack base results (price, ret, rret, rRV, rRV2, totalV, rPrice, rSpread, rbidDepth, raskDepth, portDev, rportDev, holdings, dholdings, orders, rportDev_non_nan, Xsmooth_centered_non_nan, Xsmooth_trailing_non_nan, holdingsDiff, holdingsDiffABS, holdingsDist, holdingsDistABS, rVol, rOrders, marketBook, agentList, forecastSet, rdholdingsts, bidsdf, rbidSlope, asksdf, raskSlope, wealthByType, rpf, logPriceFund, rCRV, rtotalOrdersOnBook, orderFlow, pft) = experiments['base'] # Figure 4: Base case print("\nGenerating Figure 4...") df_base = process_data_file("dataOutputFile_base.csv") plot_four_panel(df_base, params_base['tau'], params_base['sigmae'], "figures/Figure4_base_case") # Figure 6: Fixed shading print("\nGenerating Figure 6...") df_fixed = process_data_file("dataOutputFile_fixed.csv") plot_four_panel(df_fixed, params_fixed['tau'], params_fixed['sigmae'], "figures/Figure6_fixed_shading") # Figure 7: Homogeneous fundamentals print("\nGenerating Figure 7...") df_homFund = process_data_file("dataOutputFile_homFund.csv") plot_four_panel(df_homFund, params_homFund['tau'], params_homFund['sigmae'], "figures/Figure7_homFund") # Figure 10: Vol distribution facts print("\nGenerating Figure 10...") df_baseLong = process_data_file("dataOutputFile_baseLong.csv") create_vol_facts_plot(df_baseLong, "figures/Figure10_vol_facts", 'Figure 10: Realized volatility distribution facts', start=500) # Figure 11: Long memory print("\nGenerating Figure 11...") df_baseLong = process_data_file("dataOutputFile_baseLong.csv") rRVLong = df_baseLong['rRV'].values create_long_memory_plot(rRVLong, "figures/Figure11_long_memory", "Figure 11: Long memory tests", start=0) # Figure 13: Long memory, het volatility forecasts print("\nGenerating Figure 13...") df_baseHetVol = process_data_file("dataOutputFile_baseHetVol.csv") rRVLongHetVol = df_baseHetVol['rRV'].values create_long_memory_plot(rRVLongHetVol, "figures/Figure13_long_memoryHetVol", "Figure 13: Long memory tests (Heterogeneous memory)", start=0) # Figure 20: Order book dynamics print("\nGenerating Figure 20...") df_base = process_data_file("dataOutputFile_base.csv", Start=0) rret = df_base['rret'].values rRV = df_base['rRV'].values rCRV = df_base['rCRV'].values rVol = df_base['rvol'].values rSpread = df_base['spread'].values rPrice = df_base['rPrice'].values rbidSlope = df_base['bidslope'].values raskSlope = df_base['askslope'].values rbidDepth = df_base['bid depth'].values raskDepth = df_base['ask depth'].values create_orderbook_dynamics_plot(rret, rSpread, rPrice, rbidSlope, raskSlope, rbidDepth, raskDepth, "figures/Figure20_orderbook_dynamics", "Figure 20: Order Book Dynamics",start=0) # Figure 21: Cross-correlations print("\nGenerating Figure 21...") volatilityXC = True create_cross_correlation_plot(rret, rRV, rCRV, rSpread, rbidSlope, raskSlope, rbidDepth, raskDepth, rVol,volatilityXC, "figures/Figure21_cross_correlations", "Figure 21: Volatility Cross-Correlations",start=0) # Figure 22: Cross-correlations print("\nGenerating Figure 22...") volatilityXC = False create_cross_correlation_plot(rret, rRV, rCRV, rSpread, rbidSlope, raskSlope, rbidDepth, raskDepth, rVol,volatilityXC, "figures/Figure22_cross_correlations", "Figure 22: Slope Cross-Correlations",start=0) # Figure 23: Order book dynamics, fixed shading) print("\nGenerating Figure 23...") df_fixed = process_data_file("dataOutputFile_fixed.csv",Start=0) rretFix = df_fixed['rret'].values rRVFix = df_fixed['rRV'].values rCRVFix = df_fixed['rCRV'].values rVolFix = df_fixed['rvol'].values rSpreadFix = df_fixed['spread'].values rPriceFix = df_fixed['rPrice'].values rbidSlopeFix = df_fixed['bidslope'].values raskSlopeFix = df_fixed['askslope'].values rbidDepthFix = df_fixed['bid depth'].values raskDepthFix = df_fixed['ask depth'].values create_orderbook_dynamics_plot(rretFix, rSpreadFix, rPriceFix, rbidSlopeFix, raskSlopeFix, rbidDepthFix, raskDepthFix, "figures/Figure23_orderbook_dynamics", "Figure 23: Order Book Dynamics (static shading)",start=0) # Figure 24: Cross-correlations (fixed shading) print("\nGenerating Figure 24...") volatilityXCFix = True create_cross_correlation_plot(rretFix, rRVFix, rCRVFix, rSpreadFix, rbidSlopeFix, raskSlopeFix, rbidDepthFix, raskDepthFix, rVolFix,volatilityXCFix, "figures/Figure24_cross_correlations", "Figure 24: Volatility Cross-Correlations (static shading)",start=0) # Additional figures if experiments were run if 'reduced_cancel' in experiments: print("\nGenerating Figure 14...") df_rc = process_data_file("dataOutputFile_reduced_cancel.csv") plot_four_panel(df_rc, params_rc['tau'], params_rc['sigmae'], "figures/Figure14_reduced_cancelations") if 'trimmed' in experiments: print("\nGenerating Figure 15...") df_tv = process_data_file("dataOutputFile_trimmed.csv") plot_four_panel(df_tv, params_tv['tau'], params_tv['sigmae'], "figures/Figure15_trimmed_volatility") if 'noise' in experiments: print("\nGenerating Figure 16...") df_noise = process_data_file("dataOutputFile_noise.csv") plot_four_panel(df_noise, params_noise['tau'], params_noise['sigmae'], "figures/Figure16_increased_noise") # ======================================================================== # SUMMARY # ======================================================================== print("\n" + "="*70) print("COMPLETE!") print("="*70) print("\nGenerated figures are in the 'figures/' directory") print("Generated CSV files are in the current directory") if __name__ == '__main__': main()