def test_Fit_Weibull_2P(): data = Weibull_Distribution(alpha=50, beta=2).random_samples(20, seed=5) fit = Fit_Weibull_2P(failures=data, show_probability_plot=False, print_results=False) assert fit.alpha == 47.507980313141516 assert fit.beta == 2.492960611854891 #need to add more tests.
def test_Fit_Weibull_2P(): dist = Weibull_Distribution(alpha=50, beta=2) rawdata = dist.random_samples(20, seed=5) data = make_right_censored_data(data=rawdata, threshold=dist.mean) fit = Fit_Weibull_2P(failures=data.failures, right_censored=data.right_censored, show_probability_plot=False, print_results=False) assert_allclose(fit.alpha, 45.099010886086354,rtol=rtol,atol=atol) assert_allclose(fit.beta, 2.7827531773597975,rtol=rtol,atol=atol) assert_allclose(fit.gamma, 0,rtol=rtol,atol=atol) assert_allclose(fit.AICc, 115.66971887883678,rtol=rtol,atol=atol) assert_allclose(fit.Cov_alpha_beta, 0.9178064889295382,rtol=rtol,atol=atol) assert_allclose(fit.loglik, -55.4819182629478,rtol=rtol,atol=atol) assert_allclose(fit.initial_guess[1], 2.96571536864614,rtol=rtol,atol=atol)
def test_Fit_Weibull_2P(): dist = Weibull_Distribution(alpha=50, beta=2) rawdata = dist.random_samples(20, seed=5) data = make_right_censored_data(data=rawdata, threshold=dist.mean) MLE = Fit_Weibull_2P(failures=data.failures, right_censored=data.right_censored, method='MLE', show_probability_plot=False, print_results=False) assert_allclose(MLE.alpha, 45.099010886086354, rtol=rtol, atol=atol) assert_allclose(MLE.beta, 2.7827531773597984, rtol=rtol, atol=atol) assert_allclose(MLE.gamma, 0, rtol=rtol, atol=atol) assert_allclose(MLE.AICc, 115.66971887883678, rtol=rtol, atol=atol) assert_allclose(MLE.BIC, 116.95530107300358, rtol=rtol, atol=atol) assert_allclose(MLE.loglik, -55.4819182629478, rtol=rtol, atol=atol) assert_allclose(MLE.AD, 55.60004028891652, rtol=rtol, atol=atol) assert_allclose(MLE.Cov_alpha_beta, -0.9178064889295378, rtol=rtol, atol=atol) LS = Fit_Weibull_2P(failures=data.failures, right_censored=data.right_censored, method='LS', show_probability_plot=False, print_results=False) assert_allclose(LS.alpha, 42.91333312142757, rtol=rtol, atol=atol) assert_allclose(LS.beta, 2.9657153686461033, rtol=rtol, atol=atol) assert_allclose(LS.gamma, 0, rtol=rtol, atol=atol) assert_allclose(LS.AICc, 115.93668384456019, rtol=rtol, atol=atol) assert_allclose(LS.BIC, 117.222266038727, rtol=rtol, atol=atol) assert_allclose(LS.loglik, -55.61540074580951, rtol=rtol, atol=atol) assert_allclose(LS.AD, 55.62807482958476, rtol=rtol, atol=atol) assert_allclose(LS.Cov_alpha_beta, -0.1119680481788733, rtol=rtol, atol=atol)
def get_image(repository, must_not_have_labels): issues = get_filted_issues(repository, must_not_have_labels) months = get_months(issues) hist = Counter(months) axis_y = [hist[month] for month in months] fig, ax1 = plt.subplots() ax1.set_xlabel('Meses') ax1.set_ylabel('Bugs reportados', color='tab:red') ax1.plot(months, axis_y, color='tab:red') repository_name = repository.url.replace('https://github.com/', '') plt.suptitle('Padrão de chegada de issues de\nBug do Repositório %s' % (repository_name)) wb = Fit_Weibull_2P( failures=months, show_probability_plot=False, print_results=False ) weibull = wb.distribution X = generate_X_array(dist=weibull, xvals=None, xmin=None, xmax=None) Y = ss.weibull_min.pdf(X, weibull.beta, scale=weibull.alpha, loc=weibull.gamma) count = len([i for i in X if i < months[-1]+1]) X = X[:count] Y = Y[:count] ax2 = ax1.twinx() ax2.set_ylabel('Função de densidade de probabilidade de Weibull', color='tab:blue') ax2.plot(X, Y, color='tab:blue') fig.tight_layout() img_bytes = io.BytesIO() plt.savefig(img_bytes, format='png') img_bytes.seek(0) return img_bytes
def __BIC_minimizer(common_shape_X): #lgtm [py/similar-function] ''' __BIC_minimizer is used by the minimize function to get the shape which gives the lowest overall BIC ''' BIC_tot = 0 for stress in unique_stresses_f: failure_current_stress_df = f_df[f_df['stress'] == stress] FAILURES = failure_current_stress_df['times'].values if right_censored is not None: if stress in unique_stresses_rc: right_cens_current_stress_df = rc_df[rc_df['stress'] == stress] RIGHT_CENSORED = right_cens_current_stress_df['times'].values else: RIGHT_CENSORED = None else: RIGHT_CENSORED = None weibull_fit_common_shape = Fit_Weibull_2P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False, force_beta=common_shape_X) BIC_tot += weibull_fit_common_shape.BIC return BIC_tot
def weibull_coeff(tseries): ''' This function fits a time series of wind speed data with a Weibull distribution and returns shape (k) and scale (c) parameter. These are used to calculate the Weibull PDF. Inputs tseries: array-like. distrubtion of wind speeds. Outputs k: float. shape parameter, describes the Weibull slope in a probability plot. c: float. scale parameter, describes height and width of the Weibull PDF. ''' data = list(tseries) wb = Fit_Weibull_2P(failures=data, show_probability_plot=False, print_results=False) k = wb.beta c = wb.alpha return k, c
def HistogramPLOT_all(data, month, year): #Initiate Situation = [] mon = [ 'January', 'Febuary', 'March', 'April', 'May', 'June', 'July', 'August', 'September', 'October', 'November', 'December' ] #Get just Full day data logicF = (data["isFULL"].apply(lambda x: x) == (1)) data01 = data[logicF].copy() data01.fillna(method='ffill', inplace=True) logicY = (data01["DateTime"].apply(lambda x: x.year) == (year)) data01 = data01[logicY].copy() fig = plt.figure(figsize=(24, 18), dpi=80, facecolor='w', edgecolor='r') #Plotting 12 graph xvals = np.linspace(0, 30, 1000) for i in range(month): ax = plt.subplot2grid((4, 3), (int(np.floor(i / 3)), int(i % 3))) logic = (data01["DateTime"].apply(lambda x: x.month)) == (i + 1) ws = data01['WS95'][logic] ws = ws + 0.0001 failures = [] censored = [] threshold = 30 for item in ws: if item > threshold: censored.append(threshold) else: failures.append(item) xvals = np.linspace(0, 30, 1000) print(ws.shape) if (np.sum(logic) != 0): ax.hist(ws, bins=30, normed=True) hist, edge = np.histogram(np.array(ws), bins=1000, range=(0, 30), normed=True) wb2 = Fit_Weibull_2P(failures=failures, show_probability_plot=False, print_results=False) wb3 = Fit_Weibull_3P(failures=failures, show_probability_plot=False, print_results=False) gm2 = Fit_Gamma_2P(failures=failures, show_probability_plot=False, print_results=False) gm3 = Fit_Gamma_3P(failures=failures, show_probability_plot=False, print_results=False) ln2 = Fit_Lognormal_2P(failures=failures, show_probability_plot=False, print_results=False) wbm = Fit_Weibull_Mixture(failures=failures, right_censored=censored, show_plot=False, print_results=False) wb2_pdf = Weibull_Distribution(alpha=wb2.alpha, beta=wb2.beta).PDF( xvals=xvals, show_plot=True, label='Weibull_2P') wb3_pdf = Weibull_Distribution(alpha=wb3.alpha, beta=wb3.beta, gamma=wb3.gamma).PDF( xvals=xvals, show_plot=True, label='Weibull_3P') gm2_pdf = Gamma_Distribution(alpha=gm2.alpha, beta=gm2.beta).PDF(xvals=xvals, show_plot=True, label='Gamma_2P') gm3_pdf = Gamma_Distribution(alpha=gm3.alpha, beta=gm3.beta, gamma=gm3.gamma).PDF(xvals=xvals, show_plot=True, label='Gamma_3P') ln2_pdf = Lognormal_Distribution(mu=ln2.mu, sigma=ln2.sigma).PDF( xvals=xvals, show_plot=True, label='Lognormal_2P') part1_pdf = Weibull_Distribution(alpha=wbm.alpha_1, beta=wbm.beta_1).PDF( xvals=xvals, show_plot=False) part2_pdf = Weibull_Distribution(alpha=wbm.alpha_2, beta=wbm.beta_2).PDF( xvals=xvals, show_plot=False) Mixture_PDF = part1_pdf * wbm.proportion_1 + part2_pdf * wbm.proportion_2 ax.plot(xvals, Mixture_PDF, label='Weibull_Mixture') ax.legend() ax.set_ylim(0, 0.16) ax.set_xlim(0, 30) ax.set_xticks([0, 5, 10, 15, 20, 25, 30]) ax.tick_params(axis="x", labelsize=20) ax.tick_params(axis="y", labelsize=20) ax.set_title('{}'.format(mon[i]), fontweight='bold', size=20) plt.tight_layout() plt.show()
def __init__(self, failures, failure_stress, right_censored=None, right_censored_stress=None, print_results=True, show_plot=True, common_shape_method='BIC'): # input type checking and converting to arrays in preperation for creation of dataframe if common_shape_method not in ['BIC', 'weighted_average', 'average']: raise ValueError('common_shape_method must be either BIC, weighted_average, or average. Default is BIC.') if len(failures) != len(failure_stress): raise ValueError('The length of failures does not match the length of failure_stress') if type(failures) is list: failures = np.array(failures) elif type(failures) is np.ndarray: pass else: raise ValueError('failures must be an array or list') if type(failure_stress) is list: failure_stress = np.array(failure_stress) elif type(failure_stress) is np.ndarray: pass else: raise ValueError('failure_stress must be an array or list') if right_censored is not None: if len(right_censored) != len(right_censored_stress): raise ValueError('The length of right_censored does not match the length of right_censored_stress') if type(right_censored) is list: right_censored = np.array(right_censored) elif type(right_censored) is np.ndarray: pass else: raise ValueError('right_censored must be an array or list') if type(right_censored_stress) is list: right_censored_stress = np.array(right_censored_stress) elif type(right_censored_stress) is np.ndarray: pass else: raise ValueError('right_censored_stress must be an array or list') xmin = np.floor(np.log10(min(failures))) - 1 xmax = np.ceil(np.log10(max(failures))) + 1 xvals = np.logspace(xmin, xmax, 100) if right_censored is not None: TIMES = np.hstack([failures, right_censored]) STRESS = np.hstack([failure_stress, right_censored_stress]) CENS_CODES = np.hstack([np.ones_like(failures), np.zeros_like(right_censored)]) else: TIMES = failures STRESS = failure_stress CENS_CODES = np.ones_like(failures) data = {'times': TIMES, 'stress': STRESS, 'cens_codes': CENS_CODES} df = pd.DataFrame(data, columns=['times', 'stress', 'cens_codes']) df_sorted = df.sort_values(by=['cens_codes', 'stress', 'times']) is_failure = df_sorted['cens_codes'] == 1 is_right_cens = df_sorted['cens_codes'] == 0 f_df = df_sorted[is_failure] rc_df = df_sorted[is_right_cens] unique_stresses_f = f_df.stress.unique() if right_censored is not None: unique_stresses_rc = rc_df.stress.unique() for item in unique_stresses_rc: # check that there are no unique right_censored stresses that are not also in failure stresses if item not in unique_stresses_f: raise ValueError('The right_censored_stress array contains values that are not in the failure_stress array. This is equivalent to trying to fit a distribution to only censored data and cannot be done.') weibull_fit_alpha_array = [] weibull_fit_beta_array = [] weibull_fit_alpha_array_common_shape = [] color_list = ['steelblue', 'darkorange', 'red', 'green', 'purple', 'blue', 'grey', 'deeppink', 'cyan', 'chocolate'] weights_array = [] # within this loop, each list of failures and right censored values will be unpacked for each unique stress to find the common shape parameter for stress in unique_stresses_f: failure_current_stress_df = f_df[f_df['stress'] == stress] FAILURES = failure_current_stress_df['times'].values len_f = len(FAILURES) if right_censored is not None: if stress in unique_stresses_rc: right_cens_current_stress_df = rc_df[rc_df['stress'] == stress] RIGHT_CENSORED = right_cens_current_stress_df['times'].values len_rc = len(RIGHT_CENSORED) else: RIGHT_CENSORED = None len_rc = 0 else: RIGHT_CENSORED = None len_rc = 0 weights_array.append(len_f + len_rc) weibull_fit = Fit_Weibull_2P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False) weibull_fit_alpha_array.append(weibull_fit.alpha) weibull_fit_beta_array.append(weibull_fit.beta) common_shape_guess = np.average(weibull_fit_beta_array) def __BIC_minimizer(common_shape_X): #lgtm [py/similar-function] ''' __BIC_minimizer is used by the minimize function to get the shape which gives the lowest overall BIC ''' BIC_tot = 0 for stress in unique_stresses_f: failure_current_stress_df = f_df[f_df['stress'] == stress] FAILURES = failure_current_stress_df['times'].values if right_censored is not None: if stress in unique_stresses_rc: right_cens_current_stress_df = rc_df[rc_df['stress'] == stress] RIGHT_CENSORED = right_cens_current_stress_df['times'].values else: RIGHT_CENSORED = None else: RIGHT_CENSORED = None weibull_fit_common_shape = Fit_Weibull_2P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False, force_beta=common_shape_X) BIC_tot += weibull_fit_common_shape.BIC return BIC_tot if common_shape_method == 'BIC': optimized_beta_results = minimize(__BIC_minimizer, x0=common_shape_guess, method='nelder-mead') common_shape = optimized_beta_results.x[0] elif common_shape_method == 'weighted_average': total_data = sum(weights_array) weights = np.array(weights_array) / total_data common_shape = sum(weights * np.array(weibull_fit_beta_array)) elif common_shape_method == 'average': common_shape = common_shape_guess # this was just the numerical average obtained above self.common_shape = common_shape # within this loop, each list of failures and right censored values will be unpacked for each unique stress and plotted as a probability plot as well as the CDF of the common beta plot AICc_total = 0 BIC_total = 0 AICc = True for i, stress in enumerate(unique_stresses_f): failure_current_stress_df = f_df[f_df['stress'] == stress] FAILURES = failure_current_stress_df['times'].values if right_censored is not None: if stress in unique_stresses_rc: right_cens_current_stress_df = rc_df[rc_df['stress'] == stress] RIGHT_CENSORED = right_cens_current_stress_df['times'].values else: RIGHT_CENSORED = None else: RIGHT_CENSORED = None weibull_fit_common_shape = Fit_Weibull_2P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False, force_beta=common_shape) weibull_fit_alpha_array_common_shape.append(weibull_fit_common_shape.alpha) if type(weibull_fit_common_shape.AICc) == str: AICc = False else: AICc_total += weibull_fit_common_shape.AICc BIC_total += weibull_fit_common_shape.BIC if show_plot is True: weibull_fit_common_shape.distribution.CDF(linestyle='--', color=color_list[i], xvals=xvals, plot_CI=False) # plotting of the confidence intervals has been turned off Probability_plotting.Weibull_probability_plot(failures=FAILURES, right_censored=RIGHT_CENSORED,plot_CI=False, color=color_list[i], label=str(stress)) plt.legend(title='Stress') plt.xlim(10 ** (xmin + 1), 10 ** (xmax - 1)) if common_shape_method == 'BIC': plt.title(str('ALT Weibull Probability Plot\nOptimal BIC ' + r'$\beta$ = ' + str(round(common_shape, 4)))) elif common_shape_method == 'weighted_average': plt.title(str('ALT Weibull Probability Plot\nWeighted average ' + r'$\beta$ = ' + str(round(common_shape, 4)))) elif common_shape_method == 'average': plt.title(str('ALT Weibull Probability Plot\nAverage ' + r'$\beta$ = ' + str(round(common_shape, 4)))) self.BIC_sum = np.sum(BIC_total) if AICc is True: self.AICc_sum = np.sum(AICc_total) else: self.AICc_sum = 'Insufficient Data' beta_difs = (common_shape - np.array(weibull_fit_beta_array)) / np.array(weibull_fit_beta_array) beta_differences = [] for item in beta_difs: if item > 0: beta_differences.append(str('+' + str(round(item * 100, 2)) + '%')) else: beta_differences.append(str(str(round(item * 100, 2)) + '%')) results = {'stress': unique_stresses_f, 'original alpha': weibull_fit_alpha_array, 'original beta': weibull_fit_beta_array, 'new alpha': weibull_fit_alpha_array_common_shape, 'common beta': np.ones_like(unique_stresses_f) * common_shape, 'beta change': beta_differences} results_df = pd.DataFrame(results, columns=['stress', 'original alpha', 'original beta', 'new alpha', 'common beta', 'beta change']) blankIndex = [''] * len(results_df) results_df.index = blankIndex self.results = results_df if print_results is True: pd.set_option('display.width', 200) # prevents wrapping after default 80 characters pd.set_option('display.max_columns', 9) # shows the dataframe without ... truncation print('\nALT Weibull probability plot results:') print(self.results) print('Total AICc:', self.AICc_sum) print('Total BIC:', self.BIC_sum)
def __init__(self, failures, failure_stress, right_censored=None, right_censored_stress=None, print_results=True, show_plot=True): # input type checking and converting to arrays in preperation for creation of dataframe if len(failures) != len(failure_stress): raise ValueError('The length of failures does not match the length of failure_stress') if type(failures) is list: failures = np.array(failures) elif type(failures) is np.ndarray: pass else: raise ValueError('failures must be an array or list') if type(failure_stress) is list: failure_stress = np.array(failure_stress) elif type(failure_stress) is np.ndarray: pass else: raise ValueError('failure_stress must be an array or list') if right_censored is not None: if len(right_censored) != len(right_censored_stress): raise ValueError('The length of right_censored does not match the length of right_censored_stress') if type(right_censored) is list: right_censored = np.array(right_censored) elif type(right_censored) is np.ndarray: pass else: raise ValueError('right_censored must be an array or list') if type(right_censored_stress) is list: right_censored_stress = np.array(right_censored_stress) elif type(right_censored_stress) is np.ndarray: pass else: raise ValueError('right_censored_stress must be an array or list') xmin = np.floor(np.log10(min(failures))) - 1 xmax = np.ceil(np.log10(max(failures))) + 1 xvals = np.logspace(xmin, xmax, 100) if right_censored is not None: TIMES = np.hstack([failures, right_censored]) STRESS = np.hstack([failure_stress, right_censored_stress]) CENS_CODES = np.hstack([np.ones_like(failures), np.zeros_like(right_censored)]) else: TIMES = failures STRESS = failure_stress CENS_CODES = np.ones_like(failures) data = {'times': TIMES, 'stress': STRESS, 'cens_codes': CENS_CODES} df = pd.DataFrame(data, columns=['times', 'stress', 'cens_codes']) df_sorted = df.sort_values(by=['cens_codes', 'stress', 'times']) is_failure = df_sorted['cens_codes'] == 1 is_right_cens = df_sorted['cens_codes'] == 0 f_df = df_sorted[is_failure] rc_df = df_sorted[is_right_cens] unique_stresses_f = f_df.stress.unique() if right_censored is not None: unique_stresses_rc = rc_df.stress.unique() for item in unique_stresses_rc: # check that there are no unique right_censored stresses that are not also in failure stresses if item not in unique_stresses_f: raise ValueError('The right_censored_stress array contains values that are not in the failure_stress array. This is equivalent to trying to fit a distribution to only censored data and cannot be done.') weibull_fit_alpha_array = [] weibull_fit_beta_array = [] expon_fit_lambda_array = [] color_list = ['steelblue', 'darkorange', 'red', 'green', 'purple', 'blue', 'grey', 'deeppink', 'cyan', 'chocolate'] # within this loop, each list of failures and right censored values will be unpacked for each unique stress to find the common beta parameter for stress in unique_stresses_f: failure_current_stress_df = f_df[f_df['stress'] == stress] FAILURES = failure_current_stress_df['times'].values if right_censored is not None: if stress in unique_stresses_rc: right_cens_current_stress_df = rc_df[rc_df['stress'] == stress] RIGHT_CENSORED = right_cens_current_stress_df['times'].values else: RIGHT_CENSORED = None else: RIGHT_CENSORED = None weibull_fit = Fit_Weibull_2P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False) weibull_fit_alpha_array.append(weibull_fit.alpha) weibull_fit_beta_array.append(weibull_fit.beta) # within this loop, each list of failures and right censored values will be unpacked for each unique stress and plotted as a probability plot as well as the CDF of the common beta plot AICc_total = 0 BIC_total = 0 AICc_total_weib = 0 BIC_total_weib = 0 AICc = True AICc_weib = True for i, stress in enumerate(unique_stresses_f): failure_current_stress_df = f_df[f_df['stress'] == stress] FAILURES = failure_current_stress_df['times'].values if right_censored is not None: if stress in unique_stresses_rc: right_cens_current_stress_df = rc_df[rc_df['stress'] == stress] RIGHT_CENSORED = right_cens_current_stress_df['times'].values else: RIGHT_CENSORED = None else: RIGHT_CENSORED = None expon_fit = Fit_Expon_1P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False) weib_fit = Fit_Weibull_2P(failures=FAILURES, right_censored=RIGHT_CENSORED, show_probability_plot=False, print_results=False, force_beta=np.average(weibull_fit_beta_array)) expon_fit_lambda_array.append(expon_fit.Lambda) if type(expon_fit.AICc) == str: AICc = False else: AICc_total += expon_fit.AICc if type(weib_fit.AICc) == str: AICc_weib = False else: AICc_total_weib += weib_fit.AICc BIC_total += expon_fit.BIC BIC_total_weib += weib_fit.BIC if show_plot is True: expon_fit.distribution.CDF(linestyle='--', color=color_list[i], xvals=xvals, plot_CI=False) # plotting of the confidence intervals has been turned off Probability_plotting.Weibull_probability_plot(failures=FAILURES, right_censored=RIGHT_CENSORED,plot_CI=False, color=color_list[i], label=str(stress)) plt.legend(title='Stress') plt.xlim(10 ** (xmin + 1), 10 ** (xmax - 1)) plt.title('ALT Exponential Probability Plot') self.BIC_sum = np.sum(BIC_total) self.BIC_sum_weibull = np.sum(BIC_total_weib) if AICc is True: self.AICc_sum = np.sum(AICc_total) else: self.AICc_sum = 'Insufficient Data' if AICc_weib is True: self.AICc_sum_weibull = np.sum(AICc_total_weib) else: self.AICc_sum_weibull = 'Insufficient Data' beta_difs = (1 - np.array(weibull_fit_beta_array)) / np.array(weibull_fit_beta_array) beta_differences = [] for item in beta_difs: if item > 0: beta_differences.append(str('+' + str(round(item * 100, 2)) + '%')) else: beta_differences.append(str(str(round(item * 100, 2)) + '%')) results = {'stress': unique_stresses_f, 'weibull alpha': weibull_fit_alpha_array, 'weibull beta': weibull_fit_beta_array, 'new 1/Lambda': 1 / np.array(expon_fit_lambda_array), 'common shape': np.ones_like(unique_stresses_f), 'shape change': beta_differences} results_df = pd.DataFrame(results, columns=['stress', 'weibull alpha', 'weibull beta', 'new 1/Lambda', 'common shape', 'shape change']) blankIndex = [''] * len(results_df) results_df.index = blankIndex self.results = results_df if print_results is True: pd.set_option('display.width', 200) # prevents wrapping after default 80 characters pd.set_option('display.max_columns', 9) # shows the dataframe without ... truncation print('\nALT Exponential probability plot results:') print(self.results) print('Total AICc:', self.AICc_sum) print('Total BIC:', self.BIC_sum) print('Total AICc (weibull):', self.AICc_sum_weibull) print('Total BIC (weibull):', self.BIC_sum_weibull) if self.BIC_sum > self.BIC_sum_weibull: print('WARNING: The Weibull distribution would be a more appropriate fit for this data set as it has a lower BIC (using the average method to obtain BIC) than the Exponential distribution.')
# else: # failures.append(item) x = [] file_name = "mtbf_total_failures_gpu-cpu-mem_epoch1.txt" with open(file_name) as log: for line in log: #if int(line) > threshold: # censored.append(threshold) #else: x.append(int(line)) #fit the Weibull Mixture and Weibull_2P #mixture = Fit_Weibull_Mixture(failures=failures,right_censored=censored) fig, ax = plt.subplots(figsize=(5, 2)) single = Fit_Weibull_2P( failures=x, right_censored=censored) #, show_probability_plot = False) #plot the histogram of all the data and shade the censored part white #n_bins = int(len(x)/100) #N,bins,patches = plt.hist(x, density=True, alpha=0.2, color='k', bins=n_bins, edgecolor='k') #for i in range(np.argmin(abs(np.array(bins)-threshold)),len(patches)): #this is to shade the censored part of the histogram as white # patches[i].set_facecolor('white') #extract the y_vals from the mixture and build the Mixture PDF using the proportions x = np.sort(x) xvals2 = np.linspace(0, max(x), 5211) xvals = np.arange(1, len(x) + 1) / len(x) #xvals2 = np.arange(1, len(x)+1) alpha_1 = 1197 beta_1 = 0.873
def plot_event_dist(channels_input, Tend, xlim=175, bins=50, fit='weibull', filepath=None): """Plot wait-time distributions in the ESC and EPI states. Args: n_Tcounter (np.ndarray): array of wait-times for each cell type channels_input (dict): contains Channel objects with r0 and alpha params for each reaction c1hannel Tend (float): length of simulation fit (str): (weibull, gengamma) choose what distribution to fit """ #sns.set_theme() #sns.set_style("ticks") plt.style.use('default') #sns.set_theme(style="ticks", font_scale=1.2) sns.set_context('notebook', font_scale=1.5) channels = copy.deepcopy(channels_input) n_samples_max = 15000 ######################## # FOR ESC ######################## y_esc = [item for sublist in channels['esc'].wait_times for item in sublist] # flatten list y_esc = np.array(y_esc) # subsample distribution if too many samples if n_samples_max >= len(y_esc): n_samples_max = len(y_esc)-1 nkeep = np.random.choice(len(y_esc), size=n_samples_max, replace=False) y_esc = y_esc[nkeep] #y_esc = channels['esc'].wait_times.reshape(-1) # reshape combines all simulations #y_esc = y_esc[y_esc != 0] # fix plot bug when t_wait==0 fig, ax = plt.subplots(1,2) fig.set_size_inches(15,5) sns.histplot(y_esc, bins=bins, color=sns.color_palette()[0], stat='density', ax=ax[0]) # fit x = np.linspace(0, Tend, 10000) if fit =='weibull': wbf = Fit_Weibull_2P(failures=y_esc, show_probability_plot=False, print_results=False) pdf = wbf.distribution.PDF(x,show_plot=False) else: P = gengamma.fit(y_esc) pdf = gengamma.pdf(x, *P) sns.lineplot(x=x, y=pdf, color='r', lw=3, ax=ax[0]) ax[0].set_xlim(0,xlim) ax[0].set_xlabel('Time before reaction [h]',fontsize=18) ax[0].set_ylabel('Density', fontsize=18) ax[0].set_title('Wait-time distribution for ESC to EPI differentiation', fontsize=18) # annotate ax[0].text(-0.1, 1.05, string.ascii_uppercase[1], transform=ax[0].transAxes, size=20, weight='bold') #plt.xlim(0, Tend) # overlay theoretical weibull esc_r0 = channels['esc'].r0 esc_alpha = channels['esc'].alpha esc_weibull = weibull_density(esc_alpha,esc_r0, x) sns.lineplot(x=x,y=esc_weibull, color='y', lw=3, style=True, dashes=[(3,3)], ax=ax[0]) ax[0].legend(['Simulation','Theoretical']) ######################## # FOR EPI ######################## y_epi = [item for sublist in channels['epi'].wait_times for item in sublist] # flatten y_epi = np.array(y_epi) if n_samples_max >= len(y_epi): n_samples_max = len(y_epi)-1 nkeep = np.random.choice(len(y_epi), size=n_samples_max, replace=False) y_epi = y_epi[nkeep] sns.histplot(y_epi, bins=bins, color=sns.color_palette()[4], stat='density', ax=ax[1]) # fit gamma x = np.linspace(0, Tend, 10000) if fit =='weibull': wbf = Fit_Weibull_2P(failures=y_epi, show_probability_plot=False, print_results=False) pdf = wbf.distribution.PDF(x,show_plot=False) else: P = gengamma.fit(y_epi) pdf = gengamma.pdf(x, *P) sns.lineplot(x=x, y=pdf, color='r', lw=3, ax=ax[1]) ax[1].set_xlim(0,xlim) ax[1].set(xlabel='Time before reaction [h]', ylabel='Density', title='Wait-time distribution for EPI to NPC differentiation') # overlay theoretical weibull epi_r0 = channels['epi'].r0 epi_alpha = channels['epi'].alpha epi_weibull = weibull_density(epi_alpha,epi_r0, x) sns.lineplot(x=x,y=epi_weibull, color='y', lw=3,style=True, dashes=[(3,3)], ax=ax[1]) ax[1].legend(['Simulation','Theoretical']) fig.tight_layout(h_pad=1, w_pad=0) if filepath: fig.savefig(filepath, dpi=400, bbox_inches = 'tight', facecolor='w') plt.show() print("ESC:") print(" Obtained rate is %.3f vs %.3f" % (1/np.mean(y_esc),esc_r0)) print(" Corresponding to %.1fh vs %.1fh" % (np.mean(y_esc),1/esc_r0)) print(" std is %.1fh" % (np.std(y_esc))) print("EPI:") print(" Obtained rate is %.3f vs %.3f" % (1/np.mean(y_epi),epi_r0)) print(" Corresponding to %.1fh vs %.1fh" % (np.mean(y_epi),1/epi_r0)) print(" std is %.1fh" % (np.std(y_epi))) return None