def generate_correlation_box_plots(query_pearson_spearman, subject_pearson_spearman, metric, x_label="Method"):
""" Generate box plots for correlation coefficient
query_pearson_spearman: pearson and spearman data as returned
from compute_pearson_spearman for query data
subject_pearson_spearman: pearson and spearman data as returned
from compute_pearson_spearman for subject data
metric: metric to generate plots for (choices are "pearson",
"spearman")
x_label: x axis label for the plot (default: "Method")
"""
metric_lookup = {"pearson": (4, "r"), "spearman": (6, "rho")}
try:
metric_idx, y_label = metric_lookup[metric.lower()]
except KeyError:
available_metric_desc = ", ".join(metric_lookup.keys())
error_msg = "Unknown metric: %s. Available choices are: %s" % (metric, available_metric_desc)
raise KeyError, error_msg
distributions_by_method = defaultdict(list)
for e in query_pearson_spearman:
distributions_by_method[e[2]].append(e[metric_idx])
for e in subject_pearson_spearman:
distributions_by_method[e[2]].append(e[metric_idx])
x_tick_labels, distributions = zip(*distributions_by_method.items())
generate_box_plots(
distributions, x_tick_labels=x_tick_labels, x_label=x_label, y_label=y_label, y_min=-1.0, y_max=1.0
)
def generate_alpha_diversity_boxplots(rarefaction_lines,
mapping_lines,
category,
depth=None):
rarefaction_data = parse_rarefaction(rarefaction_lines)
category_value_to_sample_ids = \
get_category_value_to_sample_ids(mapping_lines,
category)
per_sample_average_diversities = \
get_per_sample_average_diversities(rarefaction_data,
category,
depth)
per_category_value_average_diversities = \
collapse_sample_diversities_by_category_value(category_value_to_sample_ids,
per_sample_average_diversities)
x_tick_labels = []
distributions = []
for cat, avg_diversities in per_category_value_average_diversities.items():
x_tick_labels.append("%s (n=%d)" % (cat, len(avg_diversities)))
distributions.append(avg_diversities)
return generate_box_plots(distributions, x_tick_labels=x_tick_labels)
def generate_alpha_diversity_boxplots(rarefaction_lines,
mapping_lines,
category,
depth=None):
rarefaction_data = parse_rarefaction(rarefaction_lines)
category_value_to_sample_ids = \
get_category_value_to_sample_ids(mapping_lines,
category)
per_sample_average_diversities = \
get_per_sample_average_diversities(rarefaction_data,
category,
depth)
per_category_value_average_diversities = \
collapse_sample_diversities_by_category_value(category_value_to_sample_ids,
per_sample_average_diversities)
# sort the data alphabetically
sorted_per_category_value_average_diversities = \
per_category_value_average_diversities.items()
sorted_per_category_value_average_diversities.sort()
x_tick_labels = []
distributions = []
for cat, avg_diversities in sorted_per_category_value_average_diversities:
x_tick_labels.append("%s (n=%d)" % (cat, len(avg_diversities)))
distributions.append(avg_diversities)
return generate_box_plots(distributions,
x_tick_labels=x_tick_labels)
def test_generate_box_plots(self):
"""generate_box_plots() should return a valid Figure object."""
fig = generate_box_plots(self.ValidTypicalBoxData, [1, 4, 10],
["Data 1", "Data 2", "Data 3"], "Test",
"x-axis label", "y-axis label")
ax = fig.get_axes()[0]
self.assertEqual(ax.get_title(), "Test")
self.assertEqual(ax.get_xlabel(), "x-axis label")
self.assertEqual(ax.get_ylabel(), "y-axis label")
self.assertEqual(len(ax.get_xticklabels()), 3)
self.assertFloatEqual(ax.get_xticks(), [1, 4, 10])
def test_generate_box_plots(self):
"""generate_box_plots() should return a valid Figure object."""
fig = generate_box_plots(self.ValidTypicalBoxData, [1, 4, 10],
["Data 1", "Data 2", "Data 3"], "Test",
"x-axis label", "y-axis label")
ax = fig.get_axes()[0]
self.assertEqual(ax.get_title(), "Test")
self.assertEqual(ax.get_xlabel(), "x-axis label")
self.assertEqual(ax.get_ylabel(), "y-axis label")
self.assertEqual(len(ax.get_xticklabels()), 3)
self.assertFloatEqual(ax.get_xticks(), [1, 4, 10])
def generate_prf_box_plots(query_prf, subject_prf, metric, x_label="Method"):
""" Generate box plots for precision, recall, or f-measure
query_prf: precision, recall, and f-measure values as returned
from compute_prfs for query data
subject_prf: precision, recall, and f-measure values as returned
from compute_prfs for subject data
metric: metric to generate plots for (choices are "precision",
"recall", and "f-measure")
x_label: x axis label for the plot (default: "Method")
"""
metric_lookup = {
"precision": (4, "Precision"),
"p": (4, "Precision"),
"recall": (5, "Recall"),
"r": (5, "Recall"),
"f-measure": (6, "F-measure"),
"f": (6, "F-measure"),
}
try:
metric_idx, y_label = metric_lookup[metric.lower()]
except KeyError:
available_metric_desc = ", ".join(metric_lookup.keys())
error_msg = "Unknown metric: %s. Available choices are: %s" % (metric, available_metric_desc)
raise KeyError, error_msg
distributions_by_method = defaultdict(list)
for e in subject_prf:
distributions_by_method[e[2]].append(e[metric_idx])
for e in query_prf:
distributions_by_method[e[2]].append(e[metric_idx])
x_tick_labels, distributions = zip(*distributions_by_method.items())
generate_box_plots(
distributions, x_tick_labels=x_tick_labels, x_label=x_label, y_label=y_label, y_min=0.0, y_max=1.0
)
def main():
option_parser, opts, args = parse_command_line_parameters(**script_info)
# Create the output dir if it doesn't already exist.
try:
create_dir(opts.output_dir)
except:
option_parser.error("Could not create or access output directory "
"specified with the -o option.")
# Parse the distance matrix and mapping file.
try:
dist_matrix_header, dist_matrix = parse_distmat(
open(opts.distance_matrix_fp, 'U'))
except:
option_parser.error(
"This does not look like a valid distance matrix "
"file. Please supply a valid distance matrix file using the -d "
"option.")
try:
mapping, mapping_header, mapping_comments = parse_mapping_file(
open(opts.mapping_fp, 'U'))
except QiimeParseError:
option_parser.error(
"This does not look like a valid metadata mapping "
"file. Please supply a valid mapping file using the -m option.")
fields = opts.fields
fields = map(strip, fields.split(','))
fields = [field.strip('"').strip("'") for field in fields]
if fields is None:
option_parser.error("You must provide at least one field using the -f "
"option.")
# Make sure each field is in the mapping file.
for field in fields:
if field not in mapping_header:
option_parser.error(
"The field '%s' is not in the provided "
"mapping file. Please supply correct fields (using the -f "
"option) corresponding to fields in the mapping file." % field)
# Make sure the y_min and y_max options make sense, as they can be either
# 'auto' or a number.
y_min = opts.y_min
y_max = opts.y_max
try:
y_min = float(y_min)
except ValueError:
if y_min == 'auto':
y_min = None
else:
option_parser.error("The --y_min option must be either a number "
"or 'auto'.")
try:
y_max = float(y_max)
except ValueError:
if y_max == 'auto':
y_max = None
else:
option_parser.error("The --y_max option must be either a number "
"or 'auto'.")
# Generate the various boxplots, depending on what the user wanted
# suppressed. Add them all to one encompassing plot.
for field in fields:
plot_data = []
plot_labels = []
if not opts.suppress_all_within:
plot_data.append(
get_all_grouped_distances(dist_matrix_header,
dist_matrix,
mapping_header,
mapping,
field,
within=True))
plot_labels.append("All within %s" % field)
if not opts.suppress_all_between:
plot_data.append(
get_all_grouped_distances(dist_matrix_header,
dist_matrix,
mapping_header,
mapping,
field,
within=False))
plot_labels.append("All between %s" % field)
if not opts.suppress_individual_within:
within_dists = get_grouped_distances(dist_matrix_header,
dist_matrix,
mapping_header,
mapping,
field,
within=True)
for grouping in within_dists:
plot_data.append(grouping[2])
plot_labels.append("%s vs. %s" % (grouping[0], grouping[1]))
if not opts.suppress_individual_between:
between_dists = get_grouped_distances(dist_matrix_header,
dist_matrix,
mapping_header,
mapping,
field,
within=False)
for grouping in between_dists:
plot_data.append(grouping[2])
plot_labels.append("%s vs. %s" % (grouping[0], grouping[1]))
# We now have our data and labels ready, so plot them!
assert (len(plot_data) == len(plot_labels)), "The number " +\
"of boxplot labels does not match the number of " +\
"boxplots."
if plot_data:
if opts.sort:
# Sort our plot data in order of increasing median.
sorted_data = []
for label, distribution in zip(plot_labels, plot_data):
sorted_data.append(
(label, distribution, median(distribution)))
sorted_data.sort(key=itemgetter(2))
plot_labels = []
plot_data = []
for label, distribution, median_value in sorted_data:
plot_labels.append(label)
plot_data.append(distribution)
width = opts.width
height = opts.height
if width is None:
width = len(plot_data) * opts.box_width + 2
if width <= 0 or height <= 0:
option_parser.error("The specified width and height of the "
"image must be greater than zero.")
plot_figure = generate_box_plots(
plot_data,
x_tick_labels=plot_labels,
title="%s Distances" % field,
x_label="Grouping",
y_label="Distance",
x_tick_labels_orientation='vertical',
y_min=y_min,
y_max=y_max,
whisker_length=opts.whisker_length,
box_width=opts.box_width,
box_color=opts.box_color,
figure_width=width,
figure_height=height)
output_plot_fp = join(opts.output_dir,
"%s_Distances.%s" % (field, opts.imagetype))
plot_figure.savefig(output_plot_fp,
format=opts.imagetype,
transparent=opts.transparent)
else:
option_parser.error("You have chosen to suppress all plots. At "
"least one type of plot must be unsuppressed.")
if not opts.suppress_significance_tests:
sig_tests_f = open(join(opts.output_dir, "%s_Stats.xls" % field),
'w')
sig_tests_results = all_pairs_t_test(
plot_labels,
plot_data,
tail_type=opts.tail_type,
num_permutations=opts.num_permutations)
sig_tests_f.write(sig_tests_results)
sig_tests_f.close()
if opts.save_raw_data:
# Write the raw plot data into a tab-delimited file.
assert (len(plot_labels) == len(plot_data))
raw_data_fp = join(opts.output_dir, "%s_Distances.xls" % field)
raw_data_f = open(raw_data_fp, 'w')
for label, data in zip(plot_labels, plot_data):
raw_data_f.write(label.replace(" ", "_") + "\t")
raw_data_f.write("\t".join(map(str, data)))
raw_data_f.write("\n")
raw_data_f.close()
def main():
option_parser, opts, args = parse_command_line_parameters(**script_info)
# Create the output dir if it doesn't already exist.
try:
create_dir(opts.output_dir)
except:
option_parser.error("Could not create or access output directory "
"specified with the -o option.")
# Parse the distance matrix and mapping file.
try:
dist_matrix_header, dist_matrix = parse_distmat(
open(opts.distance_matrix_fp, 'U'))
except:
option_parser.error("This does not look like a valid distance matrix "
"file. Please supply a valid distance matrix file using the -d "
"option.")
try:
mapping, mapping_header, mapping_comments = parse_mapping_file(
open(opts.mapping_fp, 'U'))
except QiimeParseError:
option_parser.error("This does not look like a valid metadata mapping "
"file. Please supply a valid mapping file using the -m option.")
fields = opts.fields
fields = map(strip, fields.split(','))
fields = [field.strip('"').strip("'") for field in fields]
if fields is None:
option_parser.error("You must provide at least one field using the -f "
"option.")
# Make sure each field is in the mapping file.
for field in fields:
if field not in mapping_header:
option_parser.error("The field '%s' is not in the provided "
"mapping file. Please supply correct fields (using the -f "
"option) corresponding to fields in the mapping file."
% field)
# Make sure the y_min and y_max options make sense, as they can be either
# 'auto' or a number.
y_min = opts.y_min
y_max = opts.y_max
try:
y_min = float(y_min)
except ValueError:
if y_min == 'auto':
y_min = None
else:
option_parser.error("The --y_min option must be either a number "
"or 'auto'.")
try:
y_max = float(y_max)
except ValueError:
if y_max == 'auto':
y_max = None
else:
option_parser.error("The --y_max option must be either a number "
"or 'auto'.")
# Generate the various boxplots, depending on what the user wanted
# suppressed. Add them all to one encompassing plot.
for field in fields:
plot_data = []
plot_labels = []
if not opts.suppress_all_within:
plot_data.append(get_all_grouped_distances(dist_matrix_header,
dist_matrix, mapping_header, mapping, field, within=True))
plot_labels.append("All within %s" % field)
if not opts.suppress_all_between:
plot_data.append(get_all_grouped_distances(dist_matrix_header,
dist_matrix, mapping_header, mapping, field, within=False))
plot_labels.append("All between %s" % field)
if not opts.suppress_individual_within:
within_dists = get_grouped_distances(dist_matrix_header,
dist_matrix, mapping_header, mapping, field, within=True)
for grouping in within_dists:
plot_data.append(grouping[2])
plot_labels.append("%s vs. %s" % (grouping[0], grouping[1]))
if not opts.suppress_individual_between:
between_dists = get_grouped_distances(dist_matrix_header,
dist_matrix, mapping_header, mapping, field, within=False)
for grouping in between_dists:
plot_data.append(grouping[2])
plot_labels.append("%s vs. %s" % (grouping[0], grouping[1]))
# We now have our data and labels ready, so plot them!
assert (len(plot_data) == len(plot_labels)), "The number " +\
"of boxplot labels does not match the number of " +\
"boxplots."
if plot_data:
if opts.sort:
# Sort our plot data in order of increasing median.
sorted_data = []
for label, distribution in zip(plot_labels, plot_data):
sorted_data.append((label, distribution,
median(distribution)))
sorted_data.sort(key=itemgetter(2))
plot_labels = []
plot_data = []
for label, distribution, median_value in sorted_data:
plot_labels.append(label)
plot_data.append(distribution)
width = opts.width
height = opts.height
if width is None:
width = len(plot_data) * opts.box_width + 2
if width <= 0 or height <= 0:
option_parser.error("The specified width and height of the "
"image must be greater than zero.")
plot_figure = generate_box_plots(plot_data,
x_tick_labels=plot_labels, title="%s Distances" % field,
x_label="Grouping", y_label="Distance",
x_tick_labels_orientation='vertical', y_min=y_min,
y_max=y_max, whisker_length=opts.whisker_length,
box_width=opts.box_width, box_color=opts.box_color,
figure_width=width, figure_height=height)
output_plot_fp = join(opts.output_dir, "%s_Distances.%s"
% (field, opts.imagetype))
plot_figure.savefig(output_plot_fp, format=opts.imagetype,
transparent=opts.transparent)
else:
option_parser.error("You have chosen to suppress all plots. At "
"least one type of plot must be unsuppressed.")
if not opts.suppress_significance_tests:
sig_tests_f = open(join(
opts.output_dir, "%s_Stats.xls" % field), 'w')
sig_tests_results = all_pairs_t_test(plot_labels, plot_data,
tail_type=opts.tail_type,
num_permutations=opts.num_permutations)
sig_tests_f.write(sig_tests_results)
sig_tests_f.close()
if opts.save_raw_data:
# Write the raw plot data into a tab-delimited file.
assert(len(plot_labels) == len(plot_data))
raw_data_fp = join(opts.output_dir, "%s_Distances.xls"
% field)
raw_data_f = open(raw_data_fp, 'w')
for label, data in zip(plot_labels, plot_data):
raw_data_f.write(label.replace(" ", "_") + "\t")
raw_data_f.write("\t".join(map(str, data)))
raw_data_f.write("\n")
raw_data_f.close()
