def hard_coded_analysis():
branch_length = 5.0
sequence_length = 1000
nsequences = 1000
estimate_triple_list = []
column_headers = ('most.info', 'less.info', 'least.info')
for i in range(nsequences):
# sample sequence changes at three levels of informativeness
sequence_changes = sample_sequence_changes(
branch_length, sequence_length)
# get a distance estimate for each level of informativeness
estimate_triple = sample_distance(*sequence_changes)
estimate_triple_list.append(estimate_triple)
print RUtil.get_table_string(estimate_triple_list, column_headers)
def get_table_string_and_scripts_from_logs(start_stop_pairs, log_paths,
nsamples):
"""
This is for analysis of remote execution.
"""
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_stop_pair, log_path in zip(start_stop_pairs, log_paths):
start_pos, stop_pos = start_stop_pair
sequence_length = stop_pos - start_pos + 1
means, variations, covs = read_log(log_path, nsamples)
midpoint = (start_pos + stop_pos) / 2.0
row = [sequence_length, midpoint]
for values in means, variations, covs:
corr_info = mcmc.Correlation()
corr_info.analyze(values)
hpd_low, hpd_high = mcmc.get_hpd_interval(0.95, values)
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = get_ggplot2_scripts(nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_response_content(fs):
f_info = ctmcmi.get_mutual_info_known_distn
# define the R table headers
headers = ['log.probability.ratio', 'mutual.information']
# make the array
arr = []
for x in np.linspace(fs.x_min, fs.x_max, 101):
row = [x]
proc = evozoo.AlternatingHypercube_d_1(3)
X = np.array([x])
distn = proc.get_distn(X)
Q = proc.get_rate_matrix(X)
info = f_info(Q, distn, fs.t)
row.append(info)
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts_from_logs(
start_stop_pairs, log_paths, nsamples):
"""
This is for analysis of remote execution.
"""
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_stop_pair, log_path in zip(
start_stop_pairs, log_paths):
start_pos, stop_pos = start_stop_pair
sequence_length = stop_pos - start_pos + 1
means, variations, covs = read_log(log_path, nsamples)
midpoint = (start_pos + stop_pos) / 2.0
row = [sequence_length, midpoint]
for values in means, variations, covs:
corr_info = mcmc.Correlation()
corr_info.analyze(values)
hpd_low, hpd_high = mcmc.get_hpd_interval(0.95, values)
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = get_ggplot2_scripts(nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_table_string_and_scripts_par(start_stop_pairs, nsamples):
"""
Local command-line multi-process only.
"""
# define the pool of processes corresponding to the number of cores
mypool = Pool(processes=4)
# do the multiprocessing
start_stop_n_triples = [(a, b, nsamples) for a, b in start_stop_pairs]
post_pairs_list = mypool.map(forked_function, start_stop_n_triples)
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_stop_pair, post_pairs in zip(start_stop_pairs, post_pairs_list):
start_pos, stop_pos = start_stop_pair
sequence_length = stop_pos - start_pos + 1
midpoint = (start_pos + stop_pos) / 2.0
row = [sequence_length, midpoint]
for corr_info, hpd_interval in post_pairs:
hpd_low, hpd_high = hpd_interval
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = get_ggplot2_scripts(nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_table_string_and_scripts(start_stop_pairs, nsamples):
"""
Command-line only.
"""
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_pos, stop_pos in start_stop_pairs:
sequence_length = stop_pos - start_pos + 1
means, variations, covs = get_value_lists(
start_pos, stop_pos, nsamples)
midpoint = (start_pos + stop_pos) / 2.0
row = [sequence_length, midpoint]
for values in means, variations, covs:
corr_info = mcmc.Correlation()
corr_info.analyze(values)
hpd_low, hpd_high = mcmc.get_hpd_interval(0.95, values)
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = get_ggplot2_scripts(nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_response_content(fs):
# validate and store user input
if fs.x_max <= fs.x_min:
raise ValueError('check the min and max logs')
f_info = divtime.get_fisher_info_known_distn_fast
# define the R table headers
headers = ['log.probability.ratio', 'fisher.information']
# make the array
arr = []
for x in np.linspace(fs.x_min, fs.x_max, 101):
row = [x]
proc = evozoo.DistinguishedCornerPairHypercube_d_1(3)
X = np.array([x])
distn = proc.get_distn(X)
Q = proc.get_rate_matrix(X)
info = f_info(Q, distn, fs.t)
row.append(info)
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_latex_documentbody(fs):
"""
This is obsolete because I am now using pure R output.
The latex documentbody should have a bunch of tikz pieces in it.
Each tikz piece should have been generated from R.
"""
Q_mut, Q_sels = get_qmut_qsels(fs)
# compute the statistics
ER_ratios, NSR_ratios, ER_NSR_ratios = get_statistic_ratios(Q_mut, Q_sels)
M = zip(*(ER_ratios, NSR_ratios, ER_NSR_ratios))
column_headers = ('ER.ratio', 'NSR.ratio', 'ER.times.NSR.ratio')
table_string = RUtil.get_table_string(M, column_headers)
nsels = len(Q_sels)
# define the R scripts
scripts = []
for name in column_headers:
scripts.append(get_r_tikz_script(nsels, name))
# get the tikz codes from R, for each histogram
retcode, r_out, r_err, tikz_code_list = RUtil.run_plotter_multiple_scripts(
table_string, scripts, 'tikz', width=3, height=2)
if retcode:
raise RUtil.RError(r_err)
#
# show some timings
print 'R did not fail, but here is its stderr:'
print r_err
#
# write the latex code
out = StringIO()
#print >> out, '\\pagestyle{empty}'
for tikz_code in tikz_code_list:
print >> out, tikz_code
# return the latex code, consisting mainly of a bunch of tikz plots
return out.getvalue()
def get_latex_documentbody(fs):
"""
This is obsolete because I am now using pure R output.
The latex documentbody should have a bunch of tikz pieces in it.
Each tikz piece should have been generated from R.
"""
Q_mut, Q_sels = get_qmut_qsels(fs)
# compute the statistics
ER_ratios, NSR_ratios, ER_NSR_ratios = get_statistic_ratios(Q_mut, Q_sels)
M = zip(*(ER_ratios, NSR_ratios, ER_NSR_ratios))
column_headers = ('ER.ratio', 'NSR.ratio', 'ER.times.NSR.ratio')
table_string = RUtil.get_table_string(M, column_headers)
nsels = len(Q_sels)
# define the R scripts
scripts = []
for name in column_headers:
scripts.append(get_r_tikz_script(nsels, name))
# get the tikz codes from R, for each histogram
retcode, r_out, r_err, tikz_code_list = RUtil.run_plotter_multiple_scripts(
table_string, scripts, 'tikz',
width=3, height=2)
if retcode:
raise RUtil.RError(r_err)
#
# show some timings
print 'R did not fail, but here is its stderr:'
print r_err
#
# write the latex code
out = StringIO()
#print >> out, '\\pagestyle{empty}'
for tikz_code in tikz_code_list:
print >> out, tikz_code
# return the latex code, consisting mainly of a bunch of tikz plots
return out.getvalue()
def get_response_content(fs):
M, R = get_input_matrices(fs)
# create the R table string and scripts
headers = [
't',
'mi.true.mut',
'mi.true.mutsel',
'mi.analog.mut',
'mi.analog.mutsel']
npoints = 100
t_low = 0.0
t_high = 5.0
t_incr = (t_high - t_low) / (npoints - 1)
t_values = [t_low + t_incr*i for i in range(npoints)]
# get the data for the R table
arr = []
for t in t_values:
mi_mut = ctmcmi.get_mutual_information(M, t)
mi_mutsel = ctmcmi.get_mutual_information(R, t)
mi_analog_mut = ctmcmi.get_ll_ratio_wrong(M, t)
mi_analog_mutsel = ctmcmi.get_ll_ratio_wrong(R, t)
row = [t, mi_mut, mi_mutsel, mi_analog_mut, mi_analog_mutsel]
arr.append(row)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts(stop_positions, nsamples):
"""
Command-line only.
"""
start_position = 1
# build the array for the R table
data_arr = []
for stop_position in stop_positions:
sequence_length = stop_position - start_position + 1
means, variations, covs = get_value_lists(start_position,
stop_position, nsamples)
row = [sequence_length]
for values in means, variations, covs:
corr_info = mcmc.Correlation()
corr_info.analyze(values)
hpd_low, hpd_high = mcmc.get_hpd_interval(0.95, values)
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
sequence_lengths = [x - start_position + 1 for x in stop_positions]
scripts = get_ggplot2_scripts(sequence_lengths)
# return the table string and scripts
return table_string, scripts
def get_response_content(fs):
# precompute some transition matrices
P_drift_selection = pgmsinglesite.create_drift_selection_transition_matrix(
fs.npop, fs.selection_ratio)
MatrixUtil.assert_transition_matrix(P_drift_selection)
P_mutation = pgmsinglesite.create_mutation_transition_matrix(
fs.npop, fs.mutation_ab, fs.mutation_ba)
MatrixUtil.assert_transition_matrix(P_mutation)
# define the R table headers
headers = ['generation', 'number.of.mutants']
# compute the path samples
P = np.dot(P_drift_selection, P_mutation)
mypath = PathSampler.sample_endpoint_conditioned_path(
fs.nmutants_initial, fs.nmutants_final, fs.ngenerations, P)
arr = [[i, nmutants] for i, nmutants in enumerate(mypath)]
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# validate and store user input
if fs.x_max <= fs.x_min:
raise ValueError('check the min and max logs')
f_info = divtime.get_fisher_info_known_distn_fast
# define the R table headers
headers = ['log.probability.ratio', 'fisher.information']
# make the array
arr = []
for x in np.linspace(fs.x_min, fs.x_max, 101):
row = [x]
proc = evozoo.DistinguishedCornerPairHypercube_d_1(3)
X = np.array([x])
distn = proc.get_distn(X)
Q = proc.get_rate_matrix(X)
info = f_info(Q, distn, fs.t)
row.append(info)
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
f_info = ctmcmi.get_mutual_info_known_distn
# define the R table headers
headers = ['log.probability.ratio', 'mutual.information']
# make the array
arr = []
for x in np.linspace(fs.x_min, fs.x_max, 101):
row = [x]
proc = evozoo.AlternatingHypercube_d_1(3)
X = np.array([x])
distn = proc.get_distn(X)
Q = proc.get_rate_matrix(X)
info = f_info(Q, distn, fs.t)
row.append(info)
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts(start_stop_pairs, nsamples):
"""
Command-line only.
"""
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_pos, stop_pos in start_stop_pairs:
sequence_length = stop_pos - start_pos + 1
means, variations, covs = get_value_lists(start_pos, stop_pos,
nsamples)
midpoint = (start_pos + stop_pos) / 2.0
row = [sequence_length, midpoint]
for values in means, variations, covs:
corr_info = mcmc.Correlation()
corr_info.analyze(values)
hpd_low, hpd_high = mcmc.get_hpd_interval(0.95, values)
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = get_ggplot2_scripts(nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_response_content(fs):
M, R = get_input_matrices(fs)
# create the R table string and scripts
headers = [
't', 'mi.true.mut', 'mi.true.mutsel', 'mi.analog.mut',
'mi.analog.mutsel'
]
npoints = 100
t_low = 0.0
t_high = 5.0
t_incr = (t_high - t_low) / (npoints - 1)
t_values = [t_low + t_incr * i for i in range(npoints)]
# get the data for the R table
arr = []
for t in t_values:
mi_mut = ctmcmi.get_mutual_information(M, t)
mi_mutsel = ctmcmi.get_mutual_information(R, t)
mi_analog_mut = ctmcmi.get_ll_ratio_wrong(M, t)
mi_analog_mutsel = ctmcmi.get_ll_ratio_wrong(R, t)
row = [t, mi_mut, mi_mutsel, mi_analog_mut, mi_analog_mutsel]
arr.append(row)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts_par(start_stop_pairs, nsamples):
"""
Local command-line multi-process only.
"""
# define the pool of processes corresponding to the number of cores
mypool = Pool(processes=4)
# do the multiprocessing
start_stop_n_triples = [(a, b, nsamples) for a, b in start_stop_pairs]
post_pairs_list = mypool.map(forked_function, start_stop_n_triples)
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_stop_pair, post_pairs in zip(start_stop_pairs, post_pairs_list):
start_pos, stop_pos = start_stop_pair
sequence_length = stop_pos - start_pos + 1
midpoint = (start_pos + stop_pos) / 2.0
row = [sequence_length, midpoint]
for corr_info, hpd_interval in post_pairs:
hpd_low, hpd_high = hpd_interval
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = get_ggplot2_scripts(nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_table_string_and_scripts(stop_positions, nsamples):
"""
Command-line only.
"""
start_position = 1
# build the array for the R table
data_arr = []
for stop_position in stop_positions:
sequence_length = stop_position - start_position + 1
means, variations, covs = get_value_lists(
start_position, stop_position, nsamples)
row = [sequence_length]
for values in means, variations, covs:
corr_info = mcmc.Correlation()
corr_info.analyze(values)
hpd_low, hpd_high = mcmc.get_hpd_interval(0.95, values)
row.extend([hpd_low, corr_info.mean, hpd_high])
data_arr.append(row)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
sequence_lengths = [x - start_position + 1 for x in stop_positions]
scripts = get_ggplot2_scripts(sequence_lengths)
# return the table string and scripts
return table_string, scripts
def get_response_content(fs):
# precompute some transition matrices
P_drift_selection = pgmsinglesite.create_drift_selection_transition_matrix(
fs.npop, fs.selection_ratio)
MatrixUtil.assert_transition_matrix(P_drift_selection)
P_mutation = pgmsinglesite.create_mutation_transition_matrix(
fs.npop, fs.mutation_ab, fs.mutation_ba)
MatrixUtil.assert_transition_matrix(P_mutation)
# define the R table headers
headers = ['generation', 'number.of.mutants']
# compute the path samples
P = np.dot(P_drift_selection, P_mutation)
mypath = PathSampler.sample_endpoint_conditioned_path(
fs.nmutants_initial, fs.nmutants_final, fs.ngenerations, P)
arr = [[i, nmutants] for i, nmutants in enumerate(mypath)]
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts(fs):
nstates = fs.nresidues**fs.nsites
if nstates > 256:
raise ValueError('the mutation rate matrix is too big')
# get the mutation matrix
Q_mut = mrate.get_sparse_sequence_rate_matrix(fs.nresidues, fs.nsites)
# sample a bunch of mutation-selection rate matrices
Q_sels = []
for selection_index in range(fs.nselections):
# sample the selection parameters
if fs.low_var:
v = 0.2
elif fs.medium_var:
v = 1
elif fs.high_var:
v = 5.0
elif fs.really_high_var:
v = 25.0
s = math.sqrt(v)
if fs.neg_skew:
sels = [-random.expovariate(1 / s) for i in range(nstates)]
elif fs.no_skew:
sels = [random.gauss(0, s) for i in range(nstates)]
elif fs.pos_skew:
sels = [random.expovariate(1 / s) for i in range(nstates)]
# define the mutation-selection rate matrix using Halpern-Bruno
Q = np.zeros_like(Q_mut)
for i in range(nstates):
for j in range(nstates):
if i != j:
tau = math.exp(-(sels[j] - sels[i]))
coeff = math.log(tau) / (1 - 1 / tau)
Q[i, j] = Q_mut[i, j] * coeff
for i in range(nstates):
Q[i, i] = -np.sum(Q[i])
Q_sels.append(Q)
# define the time points
incr = (fs.t_high - fs.t_low) / (fs.ntimes - 1)
times = [fs.t_low + i * incr for i in range(fs.ntimes)]
# compute the statistics
nsels = len(Q_sels)
pairs = [get_time_point_summary(Q_mut, Q_sels, t) for t in times]
mi_sign_lists, time_stats = zip(*pairs)
ncrossing_list = []
# look at how the signs change over time for each selection sample
for signs in zip(*mi_sign_lists):
count = 0
for sign_a, sign_b in iterutils.pairwise(signs):
if sign_a != sign_b:
count += 1
ncrossing_list.append(count)
# get the R scripts
scripts = [
get_r_band_script(nsels, time_stats),
get_r_prop_script(nsels, time_stats),
get_r_cross_script(ncrossing_list)
]
table_string = RUtil.get_table_string(time_stats, g_time_stats_headers)
return table_string, scripts
def get_table_string_and_scripts(fs):
nstates = fs.nresidues ** fs.nsites
if nstates > 256:
raise ValueError('the mutation rate matrix is too big')
# get the mutation matrix
Q_mut = mrate.get_sparse_sequence_rate_matrix(fs.nresidues, fs.nsites)
# sample a bunch of mutation-selection rate matrices
Q_sels = []
for selection_index in range(fs.nselections):
# sample the selection parameters
if fs.low_var:
v = 0.2
elif fs.medium_var:
v = 1
elif fs.high_var:
v = 5.0
elif fs.really_high_var:
v = 25.0
s = math.sqrt(v)
if fs.neg_skew:
sels = [-random.expovariate(1/s) for i in range(nstates)]
elif fs.no_skew:
sels = [random.gauss(0, s) for i in range(nstates)]
elif fs.pos_skew:
sels = [random.expovariate(1/s) for i in range(nstates)]
# define the mutation-selection rate matrix using Halpern-Bruno
Q = np.zeros_like(Q_mut)
for i in range(nstates):
for j in range(nstates):
if i != j:
tau = math.exp(-(sels[j] - sels[i]))
coeff = math.log(tau) / (1 - 1/tau)
Q[i, j] = Q_mut[i, j] * coeff
for i in range(nstates):
Q[i, i] = -np.sum(Q[i])
Q_sels.append(Q)
# define the time points
incr = (fs.t_high - fs.t_low) / (fs.ntimes - 1)
times = [fs.t_low + i*incr for i in range(fs.ntimes)]
# compute the statistics
nsels = len(Q_sels)
pairs = [get_time_point_summary(Q_mut, Q_sels, t) for t in times]
mi_sign_lists, time_stats = zip(*pairs)
ncrossing_list = []
# look at how the signs change over time for each selection sample
for signs in zip(*mi_sign_lists):
count = 0
for sign_a, sign_b in iterutils.pairwise(signs):
if sign_a != sign_b:
count += 1
ncrossing_list.append(count)
# get the R scripts
scripts = [
get_r_band_script(nsels, time_stats),
get_r_prop_script(nsels, time_stats),
get_r_cross_script(ncrossing_list)]
table_string = RUtil.get_table_string(time_stats, g_time_stats_headers)
return table_string, scripts
def get_response_content(fs):
f_info = divtime.get_fisher_info_known_distn_fast
requested_triples = []
for triple in g_process_triples:
name, desc, zoo_obj = triple
if getattr(fs, name):
requested_triples.append(triple)
if not requested_triples:
raise ValueError('nothing to plot')
# define the R table headers
r_names = [a.replace('_', '.') for a, b, c in requested_triples]
headers = ['t'] + r_names
# Spend a lot of time doing the optimizations
# to construct the points for the R table.
arr = []
for t in cbreaker.throttled(progrid.gen_binary(fs.start_time,
fs.stop_time),
nseconds=5,
ncount=200):
row = [t]
for python_name, desc, zoo_class in requested_triples:
zoo_obj = zoo_class(fs.d)
df = zoo_obj.get_df()
opt_dep = OptDep(zoo_obj, t, f_info)
if df:
X0 = np.random.randn(df)
xopt = scipy.optimize.fmin(opt_dep,
X0,
maxiter=10000,
maxfun=10000)
# I would like to use scipy.optimize.minimize
# except that this requires a newer version of
# scipy than is packaged for ubuntu right now.
# fmin_bfgs seems to have problems sometimes
# either hanging or maxiter=10K is too big.
"""
xopt = scipy.optimize.fmin_bfgs(opt_dep, X0,
gtol=1e-8, maxiter=10000)
"""
else:
xopt = np.array([])
info_value = -opt_dep(xopt)
row.append(info_value)
arr.append(row)
arr.sort()
npoints = len(arr)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# legend labels
label_a = 'N=%d mu=%f' % (fs.nstates_a, fs.mu_a)
label_b = 'N=%d mu=%f' % (fs.nstates_b, fs.mu_b)
arr, headers = make_table(fs)
# compute the max value
ymax = math.log(max(fs.nstates_a, fs.nstates_b))
nfifths = int(math.floor(ymax * 5.0)) + 1
ylim = RUtil.mk_call_str('c', 0, 0.2 * nfifths)
# write the R script body
out = StringIO()
print >> out, RUtil.mk_call_str(
'plot',
'my.table$t',
'my.table$alpha',
type='"n"',
ylim=ylim,
xlab='"time"',
ylab='"information"',
main='"comparison of an information criterion for two processes"',
)
# draw some horizontal lines
for i in range(nfifths+1):
print >> out, RUtil.mk_call_str(
'abline',
h=0.2*i,
col='"lightgray"',
lty='"dotted"')
colors = ('darkblue', 'darkred')
for c, header in zip(colors, headers[1:]):
print >> out, RUtil.mk_call_str(
'lines',
'my.table$t',
'my.table$%s' % header,
col='"%s"' % c,
)
legend_names = (label_a, label_b)
legend_name_str = 'c(' + ', '.join('"%s"' % s for s in legend_names) + ')'
legend_col_str = 'c(' + ', '.join('"%s"' % s for s in colors) + ')'
legend_lty_str = 'c(' + ', '.join('1' for s in colors) + ')'
print >> out, RUtil.mk_call_str(
'legend',
'"%s"' % fs.legend_placement,
legend_name_str,
col=legend_col_str,
lty=legend_lty_str,
)
script_body = out.getvalue()
# create the R plot image
table_string = RUtil.get_table_string(arr, headers)
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script_body, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# precompute some transition matrices
P_drift_selection = pgmsinglesite.create_drift_selection_transition_matrix(
fs.npop, fs.selection_ratio)
MatrixUtil.assert_transition_matrix(P_drift_selection)
P_mutation = pgmsinglesite.create_mutation_transition_matrix(
fs.npop, fs.mutation_ab, fs.mutation_ba)
MatrixUtil.assert_transition_matrix(P_mutation)
# define the R table headers
headers = [
'generation',
'number.of.mutants',
'probability',
'log.prob',
]
# compute the transition matrix
P = np.dot(P_drift_selection, P_mutation)
# Compute the endpoint conditional probabilities for various states
# along the unobserved path.
nstates = fs.npop + 1
M = np.zeros((nstates, fs.ngenerations))
M[fs.nmutants_initial, 0] = 1.0
M[fs.nmutants_final, fs.ngenerations-1] = 1.0
for i in range(fs.ngenerations-2):
A_exponent = i + 1
B_exponent = fs.ngenerations - 1 - A_exponent
A = np.linalg.matrix_power(P, A_exponent)
B = np.linalg.matrix_power(P, B_exponent)
weights = np.zeros(nstates)
for k in range(nstates):
weights[k] = A[fs.nmutants_initial, k] * B[k, fs.nmutants_final]
weights /= np.sum(weights)
for k, p in enumerate(weights):
M[k, i+1] = p
arr = []
for g in range(fs.ngenerations):
for k in range(nstates):
p = M[k, g]
if p:
logp = math.log(p)
else:
logp = float('-inf')
row = [g, k, p, logp]
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts(fs):
"""
The latex documentbody should have a bunch of tikz pieces in it.
Each tikz piece should have been generated from R.
"""
nstates = fs.nresidues ** fs.nsites
if nstates > 256:
raise ValueError("the mutation rate matrix is too big")
# get the mutation matrix
Q_mut = mrate.get_sparse_sequence_rate_matrix(fs.nresidues, fs.nsites)
# sample a bunch of mutation-selection rate matrices
Q_sels = []
for selection_index in range(fs.nselections):
# sample the selection parameters
if fs.low_var:
v = 0.2
elif fs.medium_var:
v = 1
elif fs.high_var:
v = 5.0
elif fs.really_high_var:
v = 25.0
s = math.sqrt(v)
if fs.neg_skew:
sels = [-random.expovariate(1 / s) for i in range(nstates)]
elif fs.no_skew:
sels = [random.gauss(0, s) for i in range(nstates)]
elif fs.pos_skew:
sels = [random.expovariate(1 / s) for i in range(nstates)]
# define the mutation-selection rate matrix using Halpern-Bruno
Q = np.zeros_like(Q_mut)
for i in range(nstates):
for j in range(nstates):
if i != j:
tau = math.exp(-(sels[j] - sels[i]))
coeff = math.log(tau) / (1 - 1 / tau)
Q[i, j] = Q_mut[i, j] * coeff
for i in range(nstates):
Q[i, i] = -np.sum(Q[i])
Q_sels.append(Q)
# define the time points
incr = (fs.t_high - fs.t_low) / (fs.ntimes - 1)
times = [fs.t_low + i * incr for i in range(fs.ntimes)]
# compute the statistics
nsels = len(Q_sels)
time_stats = [get_time_point_summary(Q_mut, Q_sels, t) for t in times]
# get the R scripts
scripts = [
# get_r_tikz_mi_plot(nsels, time_stats),
get_r_tikz_corr_plot(nsels, time_stats),
get_r_tikz_prop_plot(nsels, time_stats),
get_r_tikz_info_plot(nsels, time_stats),
]
table_string = RUtil.get_table_string(time_stats, g_time_stats_headers)
return table_string, scripts
def get_response_content(fs):
# legend labels
label_a = 'N=%d mu=%f' % (fs.nstates_a, fs.mu_a)
label_b = 'N=%d mu=%f' % (fs.nstates_b, fs.mu_b)
arr, headers = make_table(fs)
# compute the max value
ymax = math.log(max(fs.nstates_a, fs.nstates_b))
nfifths = int(math.floor(ymax * 5.0)) + 1
ylim = RUtil.mk_call_str('c', 0, 0.2 * nfifths)
# write the R script body
out = StringIO()
print >> out, RUtil.mk_call_str(
'plot',
'my.table$t',
'my.table$alpha',
type='"n"',
ylim=ylim,
xlab='"time"',
ylab='"information"',
main='"comparison of an information criterion for two processes"',
)
# draw some horizontal lines
for i in range(nfifths + 1):
print >> out, RUtil.mk_call_str('abline',
h=0.2 * i,
col='"lightgray"',
lty='"dotted"')
colors = ('darkblue', 'darkred')
for c, header in zip(colors, headers[1:]):
print >> out, RUtil.mk_call_str(
'lines',
'my.table$t',
'my.table$%s' % header,
col='"%s"' % c,
)
legend_names = (label_a, label_b)
legend_name_str = 'c(' + ', '.join('"%s"' % s for s in legend_names) + ')'
legend_col_str = 'c(' + ', '.join('"%s"' % s for s in colors) + ')'
legend_lty_str = 'c(' + ', '.join('1' for s in colors) + ')'
print >> out, RUtil.mk_call_str(
'legend',
'"%s"' % fs.legend_placement,
legend_name_str,
col=legend_col_str,
lty=legend_lty_str,
)
script_body = out.getvalue()
# create the R plot image
table_string = RUtil.get_table_string(arr, headers)
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script_body, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
distn_modes = [x for x in g_ordered_modes if x in fs.distribution]
if not distn_modes:
raise ValueError('no distribution mode was specified')
colors = [g_mode_to_color[m] for m in distn_modes]
arr, headers = make_table(fs, distn_modes)
distn_headers = headers[1:]
# Get the largest value in the array,
# skipping the first column.
arrmax = np.max(arr[:, 1:])
# write the R script body
out = StringIO()
ylim = RUtil.mk_call_str('c', 0, arrmax + 0.1)
sel_str = {
BALANCED: 'balanced',
HALPERN_BRUNO: 'Halpern-Bruno',
}[fs.selection]
print >> out, RUtil.mk_call_str(
'plot',
'my.table$t',
'my.table$%s' % distn_headers[0],
type='"n"',
ylim=ylim,
xlab='""',
ylab='"relaxation time"',
main='"Effect of selection (%s) on relaxation time for %d states"' %
(sel_str, fs.nstates),
)
for c, header in zip(colors, distn_headers):
print >> out, RUtil.mk_call_str(
'lines',
'my.table$t',
'my.table$%s' % header,
col='"%s"' % c,
)
mode_names = [s.replace('_', ' ') for s in distn_modes]
legend_name_str = 'c(' + ', '.join('"%s"' % s for s in mode_names) + ')'
legend_col_str = 'c(' + ', '.join('"%s"' % s for s in colors) + ')'
legend_lty_str = 'c(' + ', '.join(['1'] * len(distn_modes)) + ')'
print >> out, RUtil.mk_call_str(
'legend',
'"%s"' % fs.legend_placement,
legend_name_str,
col=legend_col_str,
lty=legend_lty_str,
)
script_body = out.getvalue()
# create the R plot image
table_string = RUtil.get_table_string(arr, headers)
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script_body, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
f_info = divtime.get_fisher_info_known_distn_fast
requested_triples = []
for triple in g_process_triples:
name, desc, zoo_obj = triple
if getattr(fs, name):
requested_triples.append(triple)
if not requested_triples:
raise ValueError('nothing to plot')
# define the R table headers
r_names = [a.replace('_', '.') for a, b, c in requested_triples]
headers = ['t'] + r_names
# Spend a lot of time doing the optimizations
# to construct the points for the R table.
arr = []
for t in cbreaker.throttled(
progrid.gen_binary(fs.start_time, fs.stop_time),
nseconds=5, ncount=200):
row = [t]
for python_name, desc, zoo_class in requested_triples:
zoo_obj = zoo_class(fs.d)
df = zoo_obj.get_df()
opt_dep = OptDep(zoo_obj, t, f_info)
if df:
X0 = np.random.randn(df)
xopt = scipy.optimize.fmin(
opt_dep, X0, maxiter=10000, maxfun=10000)
# I would like to use scipy.optimize.minimize
# except that this requires a newer version of
# scipy than is packaged for ubuntu right now.
# fmin_bfgs seems to have problems sometimes
# either hanging or maxiter=10K is too big.
"""
xopt = scipy.optimize.fmin_bfgs(opt_dep, X0,
gtol=1e-8, maxiter=10000)
"""
else:
xopt = np.array([])
info_value = -opt_dep(xopt)
row.append(info_value)
arr.append(row)
arr.sort()
npoints = len(arr)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
distn_modes = [x for x in g_ordered_modes if x in fs.distribution]
if not distn_modes:
raise ValueError('no distribution mode was specified')
colors = [g_mode_to_color[m] for m in distn_modes]
arr, headers = make_table(fs, distn_modes)
distn_headers = headers[1:]
# Get the largest value in the array,
# skipping the first column.
arrmax = np.max(arr[:,1:])
# write the R script body
out = StringIO()
ylim = RUtil.mk_call_str('c', 0, arrmax + 0.1)
sel_str = {
BALANCED : 'balanced',
HALPERN_BRUNO : 'Halpern-Bruno',
}[fs.selection]
print >> out, RUtil.mk_call_str(
'plot',
'my.table$t',
'my.table$%s' % distn_headers[0],
type='"n"',
ylim=ylim,
xlab='""',
ylab='"relaxation time"',
main='"Effect of selection (%s) on relaxation time for %d states"' % (sel_str, fs.nstates),
)
for c, header in zip(colors, distn_headers):
print >> out, RUtil.mk_call_str(
'lines',
'my.table$t',
'my.table$%s' % header,
col='"%s"' % c,
)
mode_names = [s.replace('_', ' ') for s in distn_modes]
legend_name_str = 'c(' + ', '.join('"%s"' % s for s in mode_names) + ')'
legend_col_str = 'c(' + ', '.join('"%s"' % s for s in colors) + ')'
legend_lty_str = 'c(' + ', '.join(['1']*len(distn_modes)) + ')'
print >> out, RUtil.mk_call_str(
'legend',
'"%s"' % fs.legend_placement,
legend_name_str,
col=legend_col_str,
lty=legend_lty_str,
)
script_body = out.getvalue()
# create the R plot image
table_string = RUtil.get_table_string(arr, headers)
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script_body, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_string_and_scripts(fs):
"""
The latex documentbody should have a bunch of tikz pieces in it.
Each tikz piece should have been generated from R.
"""
nstates = fs.nresidues ** fs.nsites
if nstates > 256:
raise ValueError('the mutation rate matrix is too big')
# get the mutation matrix
Q_mut = mrate.get_sparse_sequence_rate_matrix(fs.nresidues, fs.nsites)
# sample a bunch of mutation-selection rate matrices
Q_sels = []
for selection_index in range(fs.nselections):
# sample the selection parameters
if fs.low_var:
v = 0.2
elif fs.medium_var:
v = 1
elif fs.high_var:
v = 5.0
elif fs.really_high_var:
v = 25.0
s = math.sqrt(v)
if fs.neg_skew:
sels = [-random.expovariate(1/s) for i in range(nstates)]
elif fs.no_skew:
sels = [random.gauss(0, s) for i in range(nstates)]
elif fs.pos_skew:
sels = [random.expovariate(1/s) for i in range(nstates)]
# define the mutation-selection rate matrix using Halpern-Bruno
Q = np.zeros_like(Q_mut)
for i in range(nstates):
for j in range(nstates):
if i != j:
tau = math.exp(-(sels[j] - sels[i]))
coeff = math.log(tau) / (1 - 1/tau)
Q[i, j] = Q_mut[i, j] * coeff
for i in range(nstates):
Q[i, i] = -np.sum(Q[i])
Q_sels.append(Q)
# define the time points
incr = (fs.t_high - fs.t_low) / (fs.ntimes - 1)
times = [fs.t_low + i*incr for i in range(fs.ntimes)]
# compute the statistics
nsels = len(Q_sels)
time_stats = [get_time_point_summary(Q_mut, Q_sels, t) for t in times]
# get the R scripts
scripts = [
#get_r_tikz_mi_plot(nsels, time_stats),
get_r_tikz_corr_plot(nsels, time_stats),
get_r_tikz_prop_plot(nsels, time_stats),
get_r_tikz_info_plot(nsels, time_stats)]
table_string = RUtil.get_table_string(time_stats, g_time_stats_headers)
return table_string, scripts
def get_response_content(fs):
f_info = ctmcmi.get_mutual_info_known_distn
requested_triples = []
for triple in g_process_triples:
name, desc, zoo_obj = triple
if getattr(fs, name):
requested_triples.append(triple)
if not requested_triples:
raise ValueError('nothing to plot')
# define the R table headers
headers = ['t']
if fs.log4:
headers.append('log.4')
if fs.log3:
headers.append('log.3')
r_names = [a.replace('_', '.') for a, b, c in requested_triples]
headers.extend(r_names)
# Spend a lot of time doing the optimizations
# to construct the points for the R table.
times = np.linspace(fs.start_time, fs.stop_time, 101)
arr = []
for t in times:
row = [t]
if fs.log4:
row.append(math.log(4))
if fs.log3:
row.append(math.log(3))
for python_name, desc, zoo_obj in requested_triples:
X = np.array([])
info_value = f_info(
zoo_obj.get_rate_matrix(X),
zoo_obj.get_distn(X),
t)
row.append(info_value)
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
Q_mut, Q_sels = get_qmut_qsels(fs)
# compute the statistics
ER_ratios, NSR_ratios, ER_NSR_ratios = get_statistic_ratios(Q_mut, Q_sels)
M = zip(*(ER_ratios, NSR_ratios, ER_NSR_ratios))
column_headers = ('ER.ratio', 'NSR.ratio', 'ER.times.NSR.ratio')
table_string = RUtil.get_table_string(M, column_headers)
nsels = len(Q_sels)
# get the R script
comboscript = get_r_comboscript(nsels, column_headers)
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, comboscript, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
Q_mut, Q_sels = get_qmut_qsels(fs)
# compute the statistics
ER_ratios, NSR_ratios, ER_NSR_ratios = get_statistic_ratios(Q_mut, Q_sels)
M = zip(*(ER_ratios, NSR_ratios, ER_NSR_ratios))
column_headers = ('ER.ratio', 'NSR.ratio', 'ER.times.NSR.ratio')
table_string = RUtil.get_table_string(M, column_headers)
nsels = len(Q_sels)
# get the R script
comboscript = get_r_comboscript(nsels, column_headers)
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, comboscript, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
f_info = ctmcmi.get_mutual_info_known_distn
requested_triples = []
for triple in g_process_triples:
name, desc, zoo_obj = triple
if getattr(fs, name):
requested_triples.append(triple)
if not requested_triples:
raise ValueError('nothing to plot')
# define the R table headers
headers = ['t']
if fs.log4:
headers.append('log.4')
if fs.log3:
headers.append('log.3')
r_names = [a.replace('_', '.') for a, b, c in requested_triples]
headers.extend(r_names)
# Spend a lot of time doing the optimizations
# to construct the points for the R table.
times = np.linspace(fs.start_time, fs.stop_time, 101)
arr = []
for t in times:
row = [t]
if fs.log4:
row.append(math.log(4))
if fs.log3:
row.append(math.log(3))
for python_name, desc, zoo_obj in requested_triples:
X = np.array([])
info_value = f_info(zoo_obj.get_rate_matrix(X),
zoo_obj.get_distn(X), t)
row.append(info_value)
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# create the R table string and scripts
headers = [
'entropy',
'analog']
distributions = []
nstates = 4
npoints = 5000
arr = []
best_pair = None
for i in range(npoints):
weights = [random.expovariate(1) for j in range(nstates)]
total = sum(weights)
distn = [x / total for x in weights]
entropy = -sum(p * math.log(p) for p in distn)
sum_squares = sum(p*p for p in distn)
sum_cubes = sum(p*p*p for p in distn)
analog = math.log(sum_squares / sum_cubes)
row = [entropy, analog]
arr.append(row)
dist = (entropy - 1.0)**2 + (analog - 0.4)**2
if (best_pair is None) or (dist < best_pair[0]):
best_pair = (dist, distn)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
out = StringIO()
title = ', '.join(str(x) for x in best_pair[1])
print >> out, RUtil.mk_call_str(
'plot',
'my.table$entropy',
'my.table$analog',
pch='20',
main='"%s"' % title)
script = out.getvalue()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
distn_modes = [x for x in g_ordered_modes if x in fs.distribution]
if not distn_modes:
raise ValueError("no distribution mode was specified")
colors = [g_mode_to_color[m] for m in distn_modes]
arr, headers = make_table(fs, distn_modes)
distn_headers = headers[1:]
# Get the largest value in the array,
# skipping the first column.
arrmax = np.max(arr[:, 1:])
# write the R script body
out = StringIO()
ylim = RUtil.mk_call_str("c", 0, arrmax + 0.1)
sel_str = {BALANCED: "f=1/2", HALPERN_BRUNO: "Halpern-Bruno"}[fs.selection]
print >> out, RUtil.mk_call_str(
"plot",
"my.table$t",
"my.table$%s" % distn_headers[0],
type='"n"',
ylim=ylim,
xlab='"time"',
ylab='"expected log L-ratio"',
main='"Effect of selection (%s) on log L-ratio for %d states"' % (sel_str, fs.nstates),
)
for c, header in zip(colors, distn_headers):
print >> out, RUtil.mk_call_str("lines", "my.table$t", "my.table$%s" % header, col='"%s"' % c)
mode_names = [s.replace("_", " ") for s in distn_modes]
legend_name_str = "c(" + ", ".join('"%s"' % s for s in mode_names) + ")"
legend_col_str = "c(" + ", ".join('"%s"' % s for s in colors) + ")"
legend_lty_str = "c(" + ", ".join(["1"] * len(distn_modes)) + ")"
print >> out, RUtil.mk_call_str(
"legend", '"%s"' % fs.legend_placement, legend_name_str, col=legend_col_str, lty=legend_lty_str
)
script_body = out.getvalue()
# create the R plot image
table_string = RUtil.get_table_string(arr, headers)
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(table_string, script_body, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# unpack some options
npoints = fs.npoints
stddev = fs.stddev
# define the data rows and the headers
if fs.add_labels:
headers = ('x', 'y', 'label')
data_rows = list(SpiralSampler.gen_labeled_points(npoints, stddev))
else:
headers = ('x', 'y')
data_rows = list(SpiralSampler.gen_points(npoints, stddev))
# begin the response
if fs.raw:
lines = []
for data_row in data_rows:
line = '\t'.join(str(x) for x in data_row)
lines.append(line)
response_text = '\n'.join(lines)
elif fs.table:
response_text = RUtil.get_table_string(data_rows, headers)
# return the response
return response_text
def get_response_content(fs):
# create the R table string and scripts
headers = ['entropy', 'analog']
distributions = []
nstates = 4
npoints = 5000
arr = []
best_pair = None
for i in range(npoints):
weights = [random.expovariate(1) for j in range(nstates)]
total = sum(weights)
distn = [x / total for x in weights]
entropy = -sum(p * math.log(p) for p in distn)
sum_squares = sum(p * p for p in distn)
sum_cubes = sum(p * p * p for p in distn)
analog = math.log(sum_squares / sum_cubes)
row = [entropy, analog]
arr.append(row)
dist = (entropy - 1.0)**2 + (analog - 0.4)**2
if (best_pair is None) or (dist < best_pair[0]):
best_pair = (dist, distn)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
out = StringIO()
title = ', '.join(str(x) for x in best_pair[1])
print >> out, RUtil.mk_call_str('plot',
'my.table$entropy',
'my.table$analog',
pch='20',
main='"%s"' % title)
script = out.getvalue()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# unpack some options
npoints = fs.npoints
stddev = fs.stddev
# define the data rows and the headers
if fs.add_labels:
headers = ('x', 'y', 'label')
data_rows = list(SpiralSampler.gen_labeled_points(npoints, stddev))
else:
headers = ('x', 'y')
data_rows = list(SpiralSampler.gen_points(npoints, stddev))
# begin the response
if fs.raw:
lines = []
for data_row in data_rows:
line = '\t'.join(str(x) for x in data_row)
lines.append(line)
response_text = '\n'.join(lines)
elif fs.table:
response_text = RUtil.get_table_string(data_rows, headers)
# return the response
return response_text
def get_table_string_and_scripts(start_stop_pairs, nsamples, header_seq_pairs):
"""
Command-line only.
"""
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_pos, stop_pos in start_stop_pairs:
sequence_length = stop_pos - start_pos + 1
midpoint = (start_pos + stop_pos) / 2.0
arr = get_loganalysis_array(
start_pos, stop_pos, nsamples, header_seq_pairs)
mean_low = arr[1][4]
mean_mean = arr[1][1]
mean_high = arr[1][5]
var_low = arr[2][4]
var_mean = arr[2][1]
var_high = arr[2][5]
cov_low = arr[3][4]
cov_mean = arr[3][1]
cov_high = arr[3][5]
row = [
sequence_length, midpoint,
mean_low, mean_mean, mean_high,
var_low, var_mean, var_high,
cov_low, cov_mean, cov_high]
data_arr.append(row)
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table string
table_string = RUtil.get_table_string(data_arr, g_headers)
# get the scripts
scripts = beasttut.get_ggplot2_scripts(
nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, scripts
def get_response_content(fs):
# create the R table string and scripts
headers = [
'z',
'c.neg2.0',
'c.neg0.5',
'c.0.5',
'c.2.0',
#'c.a',
#'c.b',
#'c.c',
#'c.d',
]
#C = numpy.array([-0.5, -0.2, 0.2, 0.5], dtype=float)
#C = numpy.array([-1.0, -0.4, 0.4, 1.0], dtype=float)
C = numpy.array([-2.0, -0.5, 0.5, 2.0], dtype=float)
Z = numpy.linspace(-5, 5, 101)
# get the data for the R table
arr = []
for z in Z:
row = [z]
for c in C:
rate = 1.0 / kimrecessive.denom_piecewise(c, z*numpy.sign(c))
row.append(rate)
arr.append(row)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# create the R table string and scripts
headers = [
'z',
'c.neg2.0',
'c.neg0.5',
'c.0.5',
'c.2.0',
#'c.a',
#'c.b',
#'c.c',
#'c.d',
]
#C = numpy.array([-0.5, -0.2, 0.2, 0.5], dtype=float)
#C = numpy.array([-1.0, -0.4, 0.4, 1.0], dtype=float)
C = numpy.array([-2.0, -0.5, 0.5, 2.0], dtype=float)
Z = numpy.linspace(-5, 5, 101)
# get the data for the R table
arr = []
for z in Z:
row = [z]
for c in C:
rate = 1.0 / kimrecessive.denom_piecewise(c, z * numpy.sign(c))
row.append(rate)
arr.append(row)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_strings_and_scripts(
xmldata, alignment_id, start_stop_pairs,
nsamples):
"""
Command-line only.
@param xmldata: xml data already adjusted for nsamples and log filename
@param alignment_id: xml element id
@param start_stop_pairs: alignment interval bounds
@param nsamples: an extra parameter for script generation
@return: short table string, long table string, scripts (for short table)
"""
# init the array for the full R table
full_data_arr = []
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_pos, stop_pos in start_stop_pairs:
sequence_length = stop_pos - start_pos + 1
midpoint = (start_pos + stop_pos) / 2.0
interval_xml_data = beast.set_alignment_interval(
xmldata, alignment_id, start_pos, stop_pos)
row_labels, col_labels, arr = get_loganalysis_labeled_array(
interval_xml_data)
stat_name_to_row_index = dict(
(x, i) for i, x in enumerate(row_labels))
summary_name_to_col_index = dict(
(x, i) for i, x in enumerate(col_labels))
# define row indices of interest
mean_row_index = stat_name_to_row_index['meanRate']
var_row_index = stat_name_to_row_index['coefficientOfVariation']
cov_row_index = stat_name_to_row_index['covariance']
# define column indices of interest
mean_col_index = summary_name_to_col_index['mean']
low_col_index = summary_name_to_col_index['hpdLower']
high_col_index = summary_name_to_col_index['hpdUpper']
row = [
sequence_length,
midpoint,
arr[mean_row_index][low_col_index],
arr[mean_row_index][mean_col_index],
arr[mean_row_index][high_col_index],
arr[var_row_index][low_col_index],
arr[var_row_index][mean_col_index],
arr[var_row_index][high_col_index],
arr[cov_row_index][low_col_index],
arr[cov_row_index][mean_col_index],
arr[cov_row_index][high_col_index],
]
data_arr.append(row)
# add rows to the full data array
for row_index, row_label in enumerate(row_labels):
for col_index, col_label in enumerate(col_labels):
row = [
sequence_length,
midpoint,
'"' + row_label + '"',
'"' + col_label + '"',
arr[row_index][col_index],
]
full_data_arr.append(row)
# add entries to some utility arrays
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table strings
table_string = RUtil.get_table_string(data_arr, g_headers)
full_table_string = RUtil.get_table_string(
full_data_arr,
[
'sequence.length',
'midpoint',
'statistic.name',
'posterior.analysis',
'value'
],
force_float=False,
)
# get the scripts
scripts = beasttut.get_ggplot2_scripts(
nsamples, sequence_lengths, midpoints)
# return the table string and scripts
return table_string, full_table_string, scripts
def get_response_content(fs):
M = get_input_matrix(fs)
nstates = len(M)
nsites = fs.nsites
if nstates ** nsites > 16:
raise ValueError('the site dependent rate matrix is too big')
# precompute some stuff
M_site_indep = get_site_independent_process(M, nsites)
v = mrate.R_to_distn(M)
v_site_indep = get_site_independent_distn(v, nsites)
if fs.info_fis:
f_info = divtime.get_fisher_info_known_distn_fast
elif fs.info_mut:
f_info = ctmcmi.get_mutual_info_known_distn
else:
raise ValueError('no info type specified')
f_selection = mrate.to_gtr_hb_known_energies
# Spend a lot of time doing the optimizations
# to construct the points for the R table.
arr = []
for t in cbreaker.throttled(
progrid.gen_binary(fs.start_time, fs.stop_time),
nseconds=4, ncount=100):
row = [t]
# get the site-dependent mutation selection balance information
if fs.dep_balance:
dep_balance = OptDep(
M_site_indep, v_site_indep, t, f_info, f_selection)
X0 = np.random.randn(nstates ** nsites - 1)
xopt = scipy.optimize.fmin(dep_balance, X0)
max_dep_balance_info = -dep_balance(xopt)
row.append(max_dep_balance_info)
# for debug
Q_bal, v_bal = dep_balance.get_process(xopt)
print 'dependent balance:'
print max_dep_balance_info
print v_bal
print Q_bal
print
# get the site-independent mutation selection balance information
if fs.indep_balance:
indep_balance = OptIndep(
M, v, nsites, t, f_info, f_selection)
X0 = np.random.randn(nstates-1)
xopt = scipy.optimize.fmin(indep_balance, X0)
max_indep_balance_info = -indep_balance(xopt)
row.append(max_indep_balance_info)
# for debug
Q_bal, v_bal = indep_balance.get_process(xopt)
print 'independent balance:'
print max_indep_balance_info
print v_bal
print Q_bal
print
# get the site-independent mutation process information
if fs.indep_mutation:
indep_mut_info = f_info(M_site_indep, v_site_indep, t)
row.append(indep_mut_info)
# add the data row to the table
arr.append(row)
arr.sort()
npoints = len(arr)
# create the R table string and scripts
headers = ['t']
if fs.dep_balance:
headers.append('max.site.dep.balance')
if fs.indep_balance:
headers.append('max.site.indep.balance')
if fs.indep_mutation:
headers.append('site.indep.mutation')
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def do_hard_coded_analysis_b(tree, tree_remark):
"""
Do a hardcoded analysis of tree reconstruction methods.
Make R files of ordered reconstruction losses.
@param tree: a tree object
@param tree_remark: a string that is a comment about the tree
"""
# define an arbitrary order for the names of the leaves of the tree
ordered_names = list(node.name for node in tree.gen_tips())
# use some replicates
reconstruction_count = 100
# Make R files for reconstruction results from sequences
# of some number of nucleotides in length.
sequence_length = 2000
# define the tree reconstruction methods to be used
sims = [
Simulation(Clustering.NeighborJoiningDMS(), 'nj', 'neighbor joining'),
Simulation(Clustering.StoneSpectralSignDMS(), 'nj', 'spectral sign')
]
# set tree reconstruction parameters
for sim in sims:
sim.set_original_tree(tree)
# initialize the distance matrix sampler
sampler = DMSampler.InfiniteAllelesSampler(tree, ordered_names,
sequence_length)
sampler.set_inf_replacement(20.0)
sampler.set_zero_replacement(0.0)
# start the progress bar
pbar = Progress.Bar(1.0)
# sample some distance matrices
distance_matrix_start_time = time.time()
distance_matrices = []
for result in sampler.gen_samples_or_none():
# if we got a result then update the distance matrix list
if result:
sequence_list, D = result
distance_matrices.append(D)
# Update the progressbar regardless of whether or not
# the proposal was accepted.
remaining_acceptances = reconstruction_count - len(distance_matrices)
numerator = sampler.get_completed_proposals()
denominator = numerator + sampler.get_remaining_proposals(
remaining_acceptances)
dms_fraction = float(numerator) / float(denominator)
dms_total = 1.0 / (1 + len(sims))
pbar.update(dms_fraction * dms_total)
# if we have enough samples then break the loop
if not remaining_acceptances:
break
distance_matrix_seconds = time.time() - distance_matrix_start_time
# reconstruct trees using various methods
reconstruction_seconds = []
for i, sim in enumerate(sims):
reconstruction_start_time = time.time()
print 'reconstructing', len(distance_matrices), 'trees'
print 'using', sim.description
sim.run(distance_matrices, ordered_names)
pbar.update(float(i + 2) / float(1 + len(sims)))
reconstruction_seconds.append(time.time() - reconstruction_start_time)
# stop the progress bar
pbar.finish()
# consider the neighbor joining and the spectral sign results
nj_sim, ss_sim = sims
# extract the simulation data
label_list_pairs = [
('nj.unweighted', nj_sim.get_normalized_error_counts()),
('ss.unweighted', ss_sim.get_normalized_error_counts()),
('nj.weighted', nj_sim.get_normalized_loss_values()),
('ss.weighted', ss_sim.get_normalized_loss_values())
]
labels, transposed_table = zip(*label_list_pairs)
table = zip(*transposed_table)
table_string = RUtil.get_table_string(table, labels)
# write the table
filename = 'out3.table'
with open(filename, 'w') as fout:
print >> fout, '# tree source:', tree_remark
print >> fout, '# number of taxa:', len(ordered_names)
print >> fout, '# sampled distance matrices:', len(distance_matrices)
print >> fout, '# sampling seconds elapsed:', distance_matrix_seconds
print >> fout, '# sites per sequence:', sequence_length
for sim, seconds in zip(sims, reconstruction_seconds):
msg_a = '# seconds elapsed for tree reconstruction using '
msg_b = sim.description + ': ' + str(seconds)
print >> fout, msg_a + msg_b
print >> fout, table_string
print 'wrote', filename
def get_response_content(fs):
M = get_input_matrix(fs)
nstates = len(M)
nsites = fs.nsites
if nstates**nsites > 16:
raise ValueError('the site dependent rate matrix is too big')
# precompute some stuff
M_site_indep = get_site_independent_process(M, nsites)
v = mrate.R_to_distn(M)
v_site_indep = get_site_independent_distn(v, nsites)
if fs.info_fis:
f_info = divtime.get_fisher_info_known_distn_fast
elif fs.info_mut:
f_info = ctmcmi.get_mutual_info_known_distn
else:
raise ValueError('no info type specified')
f_selection = mrate.to_gtr_hb_known_energies
# Spend a lot of time doing the optimizations
# to construct the points for the R table.
arr = []
for t in cbreaker.throttled(progrid.gen_binary(fs.start_time,
fs.stop_time),
nseconds=4,
ncount=100):
row = [t]
# get the site-dependent mutation selection balance information
if fs.dep_balance:
dep_balance = OptDep(M_site_indep, v_site_indep, t, f_info,
f_selection)
X0 = np.random.randn(nstates**nsites - 1)
xopt = scipy.optimize.fmin(dep_balance, X0)
max_dep_balance_info = -dep_balance(xopt)
row.append(max_dep_balance_info)
# for debug
Q_bal, v_bal = dep_balance.get_process(xopt)
print 'dependent balance:'
print max_dep_balance_info
print v_bal
print Q_bal
print
# get the site-independent mutation selection balance information
if fs.indep_balance:
indep_balance = OptIndep(M, v, nsites, t, f_info, f_selection)
X0 = np.random.randn(nstates - 1)
xopt = scipy.optimize.fmin(indep_balance, X0)
max_indep_balance_info = -indep_balance(xopt)
row.append(max_indep_balance_info)
# for debug
Q_bal, v_bal = indep_balance.get_process(xopt)
print 'independent balance:'
print max_indep_balance_info
print v_bal
print Q_bal
print
# get the site-independent mutation process information
if fs.indep_mutation:
indep_mut_info = f_info(M_site_indep, v_site_indep, t)
row.append(indep_mut_info)
# add the data row to the table
arr.append(row)
arr.sort()
npoints = len(arr)
# create the R table string and scripts
headers = ['t']
if fs.dep_balance:
headers.append('max.site.dep.balance')
if fs.indep_balance:
headers.append('max.site.indep.balance')
if fs.indep_mutation:
headers.append('site.indep.mutation')
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# transform the arguments according to the diffusion approximation
mutation_ab = (fs.pop * fs.mutation_ab) / fs.pop_gran
mutation_ba = (fs.pop * fs.mutation_ba) / fs.pop_gran
if mutation_ab > 1 or mutation_ba > 1:
raise Exception(
'the mutation probability is not small enough '
'for the diffusion approximation to be meaningful')
selection_ratio = 1 + (fs.pop * fs.additive_selection) / fs.pop_gran
npop = fs.pop_gran
ngenerations = fs.ngenerations
nmutants_initial = int(fs.initial_freq * fs.pop_gran)
nmutants_final = int(fs.final_freq * fs.pop_gran)
# precompute some transition matrices
P_drift_selection = pgmsinglesite.create_drift_selection_transition_matrix(
npop, selection_ratio)
MatrixUtil.assert_transition_matrix(P_drift_selection)
P_mutation = pgmsinglesite.create_mutation_transition_matrix(
npop, mutation_ab, mutation_ba)
MatrixUtil.assert_transition_matrix(P_mutation)
# define the R table headers
headers = [
'generation',
'allele.frequency',
'probability',
'log.density',
]
# compute the transition matrix
P = np.dot(P_drift_selection, P_mutation)
# Compute the endpoint conditional probabilities for various states
# along the unobserved path.
nstates = npop + 1
M = np.zeros((nstates, ngenerations))
M[nmutants_initial, 0] = 1.0
M[nmutants_final, ngenerations-1] = 1.0
for i in range(ngenerations-2):
A_exponent = i + 1
B_exponent = ngenerations - 1 - A_exponent
A = np.linalg.matrix_power(P, A_exponent)
B = np.linalg.matrix_power(P, B_exponent)
weights = np.zeros(nstates)
for k in range(nstates):
weights[k] = A[nmutants_initial, k] * B[k, nmutants_final]
weights /= np.sum(weights)
for k, p in enumerate(weights):
M[k, i+1] = p
arr = []
for g in range(ngenerations):
for k in range(nstates):
p = M[k, g]
allele_frequency = k / float(npop)
# Finer gridding needs larger scaling for the density
# because each interval has a smaller support.
density = p * nstates
if density:
log_density = math.log(density)
else:
log_density = float('-inf')
row = [g, allele_frequency, p, log_density]
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot(nstates)
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_response_content(fs):
# transform the arguments according to the diffusion approximation
mutation_ab = (fs.pop * fs.mutation_ab) / fs.pop_gran
mutation_ba = (fs.pop * fs.mutation_ba) / fs.pop_gran
if mutation_ab > 1 or mutation_ba > 1:
raise Exception('the mutation probability is not small enough '
'for the diffusion approximation to be meaningful')
selection_ratio = 1 + (fs.pop * fs.additive_selection) / fs.pop_gran
npop = fs.pop_gran
ngenerations = fs.ngenerations
nmutants_initial = int(fs.initial_freq * fs.pop_gran)
nmutants_final = int(fs.final_freq * fs.pop_gran)
# precompute some transition matrices
P_drift_selection = pgmsinglesite.create_drift_selection_transition_matrix(
npop, selection_ratio)
MatrixUtil.assert_transition_matrix(P_drift_selection)
P_mutation = pgmsinglesite.create_mutation_transition_matrix(
npop, mutation_ab, mutation_ba)
MatrixUtil.assert_transition_matrix(P_mutation)
# define the R table headers
headers = [
'generation',
'allele.frequency',
'probability',
'log.density',
]
# compute the transition matrix
P = np.dot(P_drift_selection, P_mutation)
# Compute the endpoint conditional probabilities for various states
# along the unobserved path.
nstates = npop + 1
M = np.zeros((nstates, ngenerations))
M[nmutants_initial, 0] = 1.0
M[nmutants_final, ngenerations - 1] = 1.0
for i in range(ngenerations - 2):
A_exponent = i + 1
B_exponent = ngenerations - 1 - A_exponent
A = np.linalg.matrix_power(P, A_exponent)
B = np.linalg.matrix_power(P, B_exponent)
weights = np.zeros(nstates)
for k in range(nstates):
weights[k] = A[nmutants_initial, k] * B[k, nmutants_final]
weights /= np.sum(weights)
for k, p in enumerate(weights):
M[k, i + 1] = p
arr = []
for g in range(ngenerations):
for k in range(nstates):
p = M[k, g]
allele_frequency = k / float(npop)
# Finer gridding needs larger scaling for the density
# because each interval has a smaller support.
density = p * nstates
if density:
log_density = math.log(density)
else:
log_density = float('-inf')
row = [g, allele_frequency, p, log_density]
arr.append(row)
# create the R table string and scripts
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot(nstates)
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
def get_table_strings_and_scripts(xmldata, alignment_id, start_stop_pairs,
nsamples):
"""
Command-line only.
@param xmldata: xml data already adjusted for nsamples and log filename
@param alignment_id: xml element id
@param start_stop_pairs: alignment interval bounds
@param nsamples: an extra parameter for script generation
@return: short table string, long table string, scripts (for short table)
"""
# init the array for the full R table
full_data_arr = []
# build the array for the R table
data_arr = []
sequence_lengths = []
midpoints = []
for start_pos, stop_pos in start_stop_pairs:
sequence_length = stop_pos - start_pos + 1
midpoint = (start_pos + stop_pos) / 2.0
interval_xml_data = beast.set_alignment_interval(
xmldata, alignment_id, start_pos, stop_pos)
row_labels, col_labels, arr = get_loganalysis_labeled_array(
interval_xml_data)
stat_name_to_row_index = dict((x, i) for i, x in enumerate(row_labels))
summary_name_to_col_index = dict(
(x, i) for i, x in enumerate(col_labels))
# define row indices of interest
mean_row_index = stat_name_to_row_index['meanRate']
var_row_index = stat_name_to_row_index['coefficientOfVariation']
cov_row_index = stat_name_to_row_index['covariance']
# define column indices of interest
mean_col_index = summary_name_to_col_index['mean']
low_col_index = summary_name_to_col_index['hpdLower']
high_col_index = summary_name_to_col_index['hpdUpper']
row = [
sequence_length,
midpoint,
arr[mean_row_index][low_col_index],
arr[mean_row_index][mean_col_index],
arr[mean_row_index][high_col_index],
arr[var_row_index][low_col_index],
arr[var_row_index][mean_col_index],
arr[var_row_index][high_col_index],
arr[cov_row_index][low_col_index],
arr[cov_row_index][mean_col_index],
arr[cov_row_index][high_col_index],
]
data_arr.append(row)
# add rows to the full data array
for row_index, row_label in enumerate(row_labels):
for col_index, col_label in enumerate(col_labels):
row = [
sequence_length,
midpoint,
'"' + row_label + '"',
'"' + col_label + '"',
arr[row_index][col_index],
]
full_data_arr.append(row)
# add entries to some utility arrays
sequence_lengths.append(sequence_length)
midpoints.append(midpoint)
# build the table strings
table_string = RUtil.get_table_string(data_arr, g_headers)
full_table_string = RUtil.get_table_string(
full_data_arr,
[
'sequence.length', 'midpoint', 'statistic.name',
'posterior.analysis', 'value'
],
force_float=False,
)
# get the scripts
scripts = beasttut.get_ggplot2_scripts(nsamples, sequence_lengths,
midpoints)
# return the table string and scripts
return table_string, full_table_string, scripts
def do_hard_coded_analysis_b(tree, tree_remark):
"""
Do a hardcoded analysis of tree reconstruction methods.
Make R files of ordered reconstruction losses.
@param tree: a tree object
@param tree_remark: a string that is a comment about the tree
"""
# define an arbitrary order for the names of the leaves of the tree
ordered_names = list(node.name for node in tree.gen_tips())
# use some replicates
reconstruction_count = 100
# Make R files for reconstruction results from sequences
# of some number of nucleotides in length.
sequence_length = 2000
# define the tree reconstruction methods to be used
sims = [
Simulation(Clustering.NeighborJoiningDMS(), "nj", "neighbor joining"),
Simulation(Clustering.StoneSpectralSignDMS(), "nj", "spectral sign"),
]
# set tree reconstruction parameters
for sim in sims:
sim.set_original_tree(tree)
# initialize the distance matrix sampler
sampler = DMSampler.InfiniteAllelesSampler(tree, ordered_names, sequence_length)
sampler.set_inf_replacement(20.0)
sampler.set_zero_replacement(0.0)
# start the progress bar
pbar = Progress.Bar(1.0)
# sample some distance matrices
distance_matrix_start_time = time.time()
distance_matrices = []
for result in sampler.gen_samples_or_none():
# if we got a result then update the distance matrix list
if result:
sequence_list, D = result
distance_matrices.append(D)
# Update the progressbar regardless of whether or not
# the proposal was accepted.
remaining_acceptances = reconstruction_count - len(distance_matrices)
numerator = sampler.get_completed_proposals()
denominator = numerator + sampler.get_remaining_proposals(remaining_acceptances)
dms_fraction = float(numerator) / float(denominator)
dms_total = 1.0 / (1 + len(sims))
pbar.update(dms_fraction * dms_total)
# if we have enough samples then break the loop
if not remaining_acceptances:
break
distance_matrix_seconds = time.time() - distance_matrix_start_time
# reconstruct trees using various methods
reconstruction_seconds = []
for i, sim in enumerate(sims):
reconstruction_start_time = time.time()
print "reconstructing", len(distance_matrices), "trees"
print "using", sim.description
sim.run(distance_matrices, ordered_names)
pbar.update(float(i + 2) / float(1 + len(sims)))
reconstruction_seconds.append(time.time() - reconstruction_start_time)
# stop the progress bar
pbar.finish()
# consider the neighbor joining and the spectral sign results
nj_sim, ss_sim = sims
# extract the simulation data
label_list_pairs = [
("nj.unweighted", nj_sim.get_normalized_error_counts()),
("ss.unweighted", ss_sim.get_normalized_error_counts()),
("nj.weighted", nj_sim.get_normalized_loss_values()),
("ss.weighted", ss_sim.get_normalized_loss_values()),
]
labels, transposed_table = zip(*label_list_pairs)
table = zip(*transposed_table)
table_string = RUtil.get_table_string(table, labels)
# write the table
filename = "out3.table"
with open(filename, "w") as fout:
print >> fout, "# tree source:", tree_remark
print >> fout, "# number of taxa:", len(ordered_names)
print >> fout, "# sampled distance matrices:", len(distance_matrices)
print >> fout, "# sampling seconds elapsed:", distance_matrix_seconds
print >> fout, "# sites per sequence:", sequence_length
for sim, seconds in zip(sims, reconstruction_seconds):
msg_a = "# seconds elapsed for tree reconstruction using "
msg_b = sim.description + ": " + str(seconds)
print >> fout, msg_a + msg_b
print >> fout, table_string
print "wrote", filename
def get_response_content(fs):
M = get_input_matrix(fs)
# create the R table string and scripts
headers = ['t']
if fs.show_entropy:
headers.append('ub.entropy')
headers.extend([
'ub.jc.spectral',
'ub.f81.spectral',
'mutual.information',
'lb.2.state.spectral',
'lb.2.state',
'lb.f81',
])
npoints = 100
t_low = fs.start_time
t_high = fs.stop_time
t_incr = (t_high - t_low) / (npoints - 1)
t_values = [t_low + t_incr*i for i in range(npoints)]
# define some extra stuff
v = mrate.R_to_distn(M)
entropy = -np.dot(v, np.log(v))
n = len(M)
gap = sorted(abs(x) for x in np.linalg.eigvals(M))[1]
print 'stationary distn:', v
print 'entropy:', entropy
print 'spectral gap:', gap
M_slow_jc = gap * (1.0 / n) * (np.ones((n,n)) - n*np.eye(n))
M_slow_f81 = gap * np.outer(np.ones(n), v)
M_slow_f81 -= np.diag(np.sum(M_slow_f81, axis=1))
M_f81 = msimpl.get_fast_f81(M)
M_2state = msimpl.get_fast_two_state_autobarrier(M)
M_2state_spectral = -gap * M_2state / np.trace(M_2state)
# get the data for the R table
arr = []
for u in t_values:
# experiment with log time
#t = math.exp(u)
t = u
mi_slow_jc = ctmcmi.get_mutual_information(M_slow_jc, t)
mi_slow_f81 = ctmcmi.get_mutual_information(M_slow_f81, t)
mi_mut = ctmcmi.get_mutual_information(M, t)
mi_2state_spectral = ctmcmi.get_mutual_information(M_2state_spectral, t)
mi_f81 = ctmcmi.get_mutual_information(M_f81, t)
mi_2state = ctmcmi.get_mutual_information(M_2state, t)
row = [u]
if fs.show_entropy:
row.append(entropy)
row.extend([mi_slow_jc, mi_slow_f81,
mi_mut, mi_2state_spectral, mi_2state, mi_f81])
arr.append(row)
# get the R table
table_string = RUtil.get_table_string(arr, headers)
# get the R script
script = get_ggplot()
# create the R plot image
device_name = Form.g_imageformat_to_r_function[fs.imageformat]
retcode, r_out, r_err, image_data = RUtil.run_plotter(
table_string, script, device_name)
if retcode:
raise RUtil.RError(r_err)
return image_data
