def _write_sounding_figure(figure_object, title_string, output_file_name):
"""Writes sounding figure to file.
:param figure_object: Figure handle (instance of
`matplotlib.figure.Figure`).
:param title_string: Figure title.
:param output_file_name: Path to output file (figure will be saved as image
here).
"""
print('Saving figure to: "{0:s}"...'.format(output_file_name))
figure_object.savefig(output_file_name,
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
imagemagick_utils.resize_image(input_file_name=output_file_name,
output_file_name=output_file_name,
output_size_pixels=SOUNDING_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(input_file_name=output_file_name,
output_file_name=output_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=output_file_name,
text_string=title_string,
x_offset_from_center_px=0,
y_offset_from_top_px=0)
imagemagick_utils.trim_whitespace(input_file_name=output_file_name,
output_file_name=output_file_name,
border_width_pixels=10)
def _run(input_dir_name, output_dir_name):
"""Creates figure showing overall model evaluation.
This is effectively the main method.
:param input_dir_name: See documentation at top of file.
:param output_dir_name: Same.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
panel_file_names = [
'{0:s}/{1:s}'.format(input_dir_name, p)
for p in PATHLESS_INPUT_FILE_NAMES
]
resized_panel_file_names = [
'{0:s}/{1:s}'.format(output_dir_name, p)
for p in PATHLESS_INPUT_FILE_NAMES
]
letter_label = None
for i in range(len(panel_file_names)):
print('Resizing panel and saving to: "{0:s}"...'.format(
resized_panel_file_names[i]))
imagemagick_utils.trim_whitespace(
input_file_name=panel_file_names[i],
output_file_name=resized_panel_file_names[i])
if letter_label is None:
letter_label = 'a'
else:
letter_label = chr(ord(letter_label) + 1)
_overlay_text(image_file_name=resized_panel_file_names[i],
x_offset_from_left_px=0,
y_offset_from_top_px=TITLE_FONT_SIZE,
text_string='({0:s})'.format(letter_label))
imagemagick_utils.resize_image(
input_file_name=resized_panel_file_names[i],
output_file_name=resized_panel_file_names[i],
output_size_pixels=PANEL_SIZE_PX)
concat_figure_file_name = '{0:s}/overall_evaluation.jpg'.format(
output_dir_name)
print(
'Concatenating panels to: "{0:s}"...'.format(concat_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=resized_panel_file_names,
output_file_name=concat_figure_file_name,
num_panel_rows=NUM_PANEL_ROWS,
num_panel_columns=NUM_PANEL_COLUMNS)
imagemagick_utils.resize_image(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _run(input_dir_name, output_dir_name):
"""Creates figure showing model performance as a function of hyperparams.
This is effectively the main method.
:param input_dir_name: See documentation at top of file.
:param output_dir_name: Same.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
num_scores = len(SCORE_NAMES)
num_dense_layer_counts = len(DENSE_LAYER_COUNTS)
for j in range(num_scores):
panel_file_names = [
'{0:s}/num-dense-layers={1:d}_{2:s}_grid.jpg'.format(
input_dir_name, d, SCORE_NAMES[j]) for d in DENSE_LAYER_COUNTS
]
resized_panel_file_names = [
'{0:s}/num-dense-layers={1:d}_{2:s}_grid.jpg'.format(
output_dir_name, d, SCORE_NAMES[j]) for d in DENSE_LAYER_COUNTS
]
for i in range(num_dense_layer_counts):
imagemagick_utils.trim_whitespace(
input_file_name=panel_file_names[i],
output_file_name=resized_panel_file_names[i])
imagemagick_utils.resize_image(
input_file_name=resized_panel_file_names[i],
output_file_name=resized_panel_file_names[i],
output_size_pixels=PANEL_SIZE_PX)
concat_figure_file_name = '{0:s}/{1:s}_grid.jpg'.format(
output_dir_name, SCORE_NAMES[j])
print('Concatenating panels to: "{0:s}"...'.format(
concat_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=resized_panel_file_names,
output_file_name=concat_figure_file_name,
num_panel_rows=NUM_PANEL_ROWS,
num_panel_columns=NUM_PANEL_COLUMNS)
imagemagick_utils.resize_image(
input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(
input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name)
def _run():
"""Plots dilation of WPC fronts.
This is effectively the main method.
"""
print 'Reading data from: "{0:s}"...'.format(FRONT_LINE_FILE_NAME)
front_line_table = fronts_io.read_polylines_from_file(FRONT_LINE_FILE_NAME)
print 'Reading data from: "{0:s}"...'.format(FRONTAL_GRID_FILE_NAME)
frontal_grid_table = fronts_io.read_narr_grids_from_file(
FRONTAL_GRID_FILE_NAME)
num_grid_rows, num_grid_columns = nwp_model_utils.get_grid_dimensions(
model_name=nwp_model_utils.NARR_MODEL_NAME)
ternary_front_matrix = ml_utils.front_table_to_images(
frontal_grid_table=frontal_grid_table,
num_rows_per_image=num_grid_rows,
num_columns_per_image=num_grid_columns)
_plot_fronts(front_line_table=front_line_table,
ternary_front_matrix=ternary_front_matrix,
title_string='Observed fronts before dilation',
annotation_string='(a)',
output_file_name=BEFORE_FILE_NAME)
ternary_front_matrix = ml_utils.dilate_ternary_target_images(
target_matrix=ternary_front_matrix,
dilation_distance_metres=DILATION_DISTANCE_METRES,
verbose=False)
_plot_fronts(front_line_table=front_line_table,
ternary_front_matrix=ternary_front_matrix,
title_string='Observed fronts after dilation',
annotation_string='(b)',
output_file_name=AFTER_FILE_NAME)
print 'Concatenating figures to: "{0:s}"...'.format(CONCAT_FILE_NAME)
imagemagick_utils.concatenate_images(
input_file_names=[BEFORE_FILE_NAME, AFTER_FILE_NAME],
output_file_name=CONCAT_FILE_NAME,
num_panel_rows=1,
num_panel_columns=2)
imagemagick_utils.resize_image(input_file_name=CONCAT_FILE_NAME,
output_file_name=CONCAT_FILE_NAME,
output_size_pixels=CONCAT_SIZE_PIXELS)
def _run():
"""Concatenates figures showing dilation procedure and double penalty.
This is effectively the main method.
"""
dilation_image_object = Image.open(DILATION_FILE_NAME)
double_penalty_image_object = Image.open(DOUBLE_PENALTY_FILE_NAME)
new_dilation_width_px = double_penalty_image_object.size[0]
width_ratio = (float(new_dilation_width_px) /
dilation_image_object.size[0])
new_dilation_height_px = int(
numpy.round(dilation_image_object.size[1] * width_ratio))
small_dilation_file_name = DILATION_FILE_NAME.replace(
'.jpg', '_resized.jpg')
print 'Resizing dilation figure to: "{0:s}"...'.format(
small_dilation_file_name)
command_string = (
'/usr/bin/convert "{0:s}" -resize {1:d}x{2:d} "{3:s}"').format(
DILATION_FILE_NAME, new_dilation_width_px, new_dilation_height_px,
small_dilation_file_name)
os.system(command_string)
print('Concatenating dilation and double-penalty figures to: "{0:s}"...'
).format(CONCAT_FILE_NAME)
imagemagick_utils.concatenate_images(
input_file_names=[small_dilation_file_name, DOUBLE_PENALTY_FILE_NAME],
output_file_name=CONCAT_FILE_NAME,
num_panel_rows=2,
num_panel_columns=1)
imagemagick_utils.trim_whitespace(input_file_name=CONCAT_FILE_NAME,
output_file_name=CONCAT_FILE_NAME)
imagemagick_utils.resize_image(input_file_name=CONCAT_FILE_NAME,
output_file_name=CONCAT_FILE_NAME,
output_size_pixels=CONCAT_SIZE_PIXELS)
os.remove(small_dilation_file_name)
def _run():
"""Plots determinization of front probabilities.
This is effectively the main method.
"""
prediction_dict = ml_utils.read_gridded_predictions(PREDICTION_FILE_NAME)
class_probability_matrix = prediction_dict[ml_utils.PROBABILITY_MATRIX_KEY]
for this_id in front_utils.VALID_INTEGER_IDS:
if this_id == front_utils.NO_FRONT_INTEGER_ID:
class_probability_matrix[..., this_id][numpy.isnan(
class_probability_matrix[..., this_id])] = 1.
else:
class_probability_matrix[..., this_id][numpy.isnan(
class_probability_matrix[..., this_id])] = 0.
_plot_predictions(output_file_name=BEFORE_FILE_NAME,
title_string='Probabilities',
annotation_string='(a)',
class_probability_matrix=class_probability_matrix)
predicted_label_matrix = object_eval.determinize_probabilities(
class_probability_matrix=class_probability_matrix,
binarization_threshold=BINARIZATION_THRESHOLD)
_plot_predictions(output_file_name=AFTER_FILE_NAME,
title_string='Deterministic predictions',
annotation_string='(b)',
predicted_label_matrix=predicted_label_matrix)
print 'Concatenating figures to: "{0:s}"...'.format(CONCAT_FILE_NAME)
imagemagick_utils.concatenate_images(
input_file_names=[BEFORE_FILE_NAME, AFTER_FILE_NAME],
output_file_name=CONCAT_FILE_NAME,
num_panel_rows=2,
num_panel_columns=1)
imagemagick_utils.resize_image(input_file_name=CONCAT_FILE_NAME,
output_file_name=CONCAT_FILE_NAME,
output_size_pixels=CONCAT_SIZE_PIXELS)
def _run(evaluation_dir_name, num_months_per_chunk, num_hours_per_chunk,
confidence_level, output_dir_name):
"""Plots temporally subset model evaluation.
This is effectively the main method.
:param evaluation_dir_name: See documentation at top of file.
:param num_months_per_chunk: Same.
:param num_hours_per_chunk: Same.
:param confidence_level: Same.
:param output_dir_name: Same.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
panel_file_names = _plot_by_month(
evaluation_dir_name=evaluation_dir_name,
num_months_per_chunk=num_months_per_chunk,
confidence_level=confidence_level,
output_dir_name=output_dir_name)
panel_file_names += _plot_by_hour(evaluation_dir_name=evaluation_dir_name,
num_hours_per_chunk=num_hours_per_chunk,
confidence_level=confidence_level,
output_dir_name=output_dir_name)
concat_file_name = '{0:s}/temporally_subset_eval.jpg'.format(
output_dir_name)
print('Concatenating panels to: "{0:s}"...'.format(concat_file_name))
imagemagick_utils.concatenate_images(input_file_names=panel_file_names,
output_file_name=concat_file_name,
num_panel_rows=2,
num_panel_columns=2)
imagemagick_utils.resize_image(input_file_name=concat_file_name,
output_file_name=concat_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _write_radar_figures(figure_objects, field_names, composite_name,
concat_title_string, output_dir_name):
"""Writes radar figures to file.
F = number of radar fields
:param figure_objects: length-F list of figure handles (each an instance of
`matplotlib.figure.Figure`).
:param field_names: length-F list of field names.
:param composite_name: Name of composite (e.g., "after" or "difference").
:param concat_title_string: Title for concatenated figure.
:param output_dir_name: Name of output directory (figures will be saved
here).
:return: concat_figure_file_name: Path to image file with concatenated
figure.
"""
num_fields = len(field_names)
panel_file_names = [None] * num_fields
for k in range(num_fields):
panel_file_names[k] = '{0:s}/{1:s}_{2:s}.jpg'.format(
output_dir_name, field_names[k].replace('_', '-'), composite_name)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[k]))
figure_objects[k].savefig(panel_file_names[k],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_objects[k])
concat_figure_file_name = '{0:s}/radar_{1:s}.jpg'.format(
output_dir_name, composite_name)
print(
'Concatenating panels to: "{0:s}"...'.format(concat_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=concat_figure_file_name,
num_panel_rows=1,
num_panel_columns=num_fields,
border_width_pixels=50)
imagemagick_utils.resize_image(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=concat_figure_file_name,
text_string=concat_title_string,
x_offset_from_center_px=0,
y_offset_from_top_px=0)
imagemagick_utils.trim_whitespace(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
border_width_pixels=10)
return concat_figure_file_name
def plot_many_soundings(list_of_metpy_dictionaries,
title_strings,
num_panel_rows,
output_file_name,
temp_directory_name=None,
option_dict=None):
"""Creates paneled figure with many soundings.
N = number of soundings to plot
:param list_of_metpy_dictionaries: length-N list of dictionaries. Each
dictionary must satisfy the input format for `sounding_dict_for_metpy`
in `plot_sounding`.
:param title_strings: length-N list of titles.
:param num_panel_rows: Number of rows in paneled figure.
:param output_file_name: Path to output (image) file.
:param temp_directory_name: Name of temporary directory. Each panel will be
stored here, then deleted after the panels have been concatenated into
the final image. If `temp_directory_name is None`, will use the default
temp directory on the local machine.
:param option_dict: See doc for `plot_sounding`.
"""
error_checking.assert_is_numpy_array(numpy.array(title_strings),
num_dimensions=1)
num_soundings = len(title_strings)
error_checking.assert_is_list(list_of_metpy_dictionaries)
error_checking.assert_is_geq(len(list_of_metpy_dictionaries),
num_soundings)
error_checking.assert_is_leq(len(list_of_metpy_dictionaries),
num_soundings)
error_checking.assert_is_integer(num_panel_rows)
error_checking.assert_is_geq(num_panel_rows, 1)
error_checking.assert_is_leq(num_panel_rows, num_soundings)
file_system_utils.mkdir_recursive_if_necessary(file_name=output_file_name)
if temp_directory_name is not None:
file_system_utils.mkdir_recursive_if_necessary(
directory_name=temp_directory_name)
temp_file_names = [None] * num_soundings
num_panel_columns = int(numpy.ceil(float(num_soundings) / num_panel_rows))
for i in range(num_panel_rows):
for j in range(num_panel_columns):
this_sounding_index = i * num_panel_columns + j
if this_sounding_index >= num_soundings:
break
plot_sounding(sounding_dict_for_metpy=list_of_metpy_dictionaries[
this_sounding_index],
title_string=title_strings[this_sounding_index],
option_dict=option_dict)
temp_file_names[this_sounding_index] = '{0:s}.jpg'.format(
tempfile.NamedTemporaryFile(dir=temp_directory_name,
delete=False).name)
print('Saving sounding to: "{0:s}"...'.format(
temp_file_names[this_sounding_index]))
pyplot.savefig(temp_file_names[this_sounding_index],
dpi=DOTS_PER_INCH)
pyplot.close()
imagemagick_utils.trim_whitespace(
input_file_name=temp_file_names[this_sounding_index],
output_file_name=temp_file_names[this_sounding_index],
border_width_pixels=SINGLE_IMAGE_BORDER_WIDTH_PX)
imagemagick_utils.resize_image(
input_file_name=temp_file_names[this_sounding_index],
output_file_name=temp_file_names[this_sounding_index],
output_size_pixels=SINGLE_IMAGE_SIZE_PX)
print('Concatenating panels into one figure: "{0:s}"...'.format(
output_file_name))
imagemagick_utils.concatenate_images(
input_file_names=temp_file_names,
output_file_name=output_file_name,
num_panel_rows=num_panel_rows,
num_panel_columns=num_panel_columns,
border_width_pixels=PANELED_IMAGE_BORDER_WIDTH_PX)
for i in range(num_soundings):
os.remove(temp_file_names[i])
def _run():
"""Plots weird WPC fronts, along with theta_w and wind barbs from NARR.
This is effectively the main method.
"""
narr_latitude_matrix_deg, narr_longitude_matrix_deg = (
nwp_model_utils.get_latlng_grid_point_matrices(
model_name=nwp_model_utils.NARR_MODEL_NAME))
narr_rotation_cos_matrix, narr_rotation_sin_matrix = (
nwp_model_utils.get_wind_rotation_angles(
latitudes_deg=narr_latitude_matrix_deg,
longitudes_deg=narr_longitude_matrix_deg,
model_name=nwp_model_utils.NARR_MODEL_NAME))
num_panels = len(VALID_TIME_STRINGS)
panel_file_names = [''] * num_panels
this_annotation_string = None
for i in range(num_panels):
this_time_unix_sec = time_conversion.string_to_unix_sec(
VALID_TIME_STRINGS[i], DEFAULT_TIME_FORMAT)
this_title_string = time_conversion.unix_sec_to_string(
this_time_unix_sec, NICE_TIME_FORMAT)
if PRESSURE_LEVELS_MB[i] == 1000:
this_title_string += ' at 1000 mb'
else:
this_title_string += ' at surface'
if this_annotation_string is None:
this_annotation_string = '(a)'
else:
this_orig_character = this_annotation_string[1]
this_new_character = chr(ord(this_orig_character) + 1)
this_annotation_string = this_annotation_string.replace(
this_orig_character, this_new_character)
this_file_name = _plot_one_time(
valid_time_string=VALID_TIME_STRINGS[i],
pressure_level_mb=PRESSURE_LEVELS_MB[i],
title_string=this_title_string,
annotation_string=this_annotation_string,
narr_rotation_cos_matrix=narr_rotation_cos_matrix,
narr_rotation_sin_matrix=narr_rotation_sin_matrix)
panel_file_names[i] = this_file_name
print '\n'
concat_file_name = '{0:s}/weird_fronts.jpg'.format(OUTPUT_DIR_NAME)
print 'Concatenating figures to: "{0:s}"...'.format(concat_file_name)
num_panels = len(panel_file_names)
num_panel_columns = int(numpy.ceil(float(num_panels) / 2))
imagemagick_utils.concatenate_images(input_file_names=panel_file_names,
output_file_name=concat_file_name,
num_panel_rows=NUM_PANEL_ROWS,
num_panel_columns=num_panel_columns)
imagemagick_utils.trim_whitespace(input_file_name=concat_file_name,
output_file_name=concat_file_name)
imagemagick_utils.resize_image(input_file_name=concat_file_name,
output_file_name=concat_file_name,
output_size_pixels=CONCAT_SIZE_PIXELS)
def _run(valid_time_string, smoothing_radius_pixels, front_percentile,
num_closing_iters, pressure_level_mb, top_narr_directory_name,
narr_mask_file_name, output_dir_name):
"""Plots NFA (numerical frontal analysis) procedure.
This is effectively the main method.
:param valid_time_string: See documentation at top of file.
:param smoothing_radius_pixels: Same.
:param front_percentile: Same.
:param num_closing_iters: Same.
:param pressure_level_mb: Same.
:param top_narr_directory_name: Same.
:param narr_mask_file_name: Same.
:param output_dir_name: Same.
"""
cutoff_radius_pixels = 4 * smoothing_radius_pixels
valid_time_unix_sec = time_conversion.string_to_unix_sec(
valid_time_string, INPUT_TIME_FORMAT)
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
if narr_mask_file_name == '':
num_grid_rows, num_grid_columns = nwp_model_utils.get_grid_dimensions(
model_name=nwp_model_utils.NARR_MODEL_NAME)
narr_mask_matrix = numpy.full(
(num_grid_rows, num_grid_columns), 1, dtype=int)
else:
print 'Reading mask from: "{0:s}"...\n'.format(narr_mask_file_name)
narr_mask_matrix = ml_utils.read_narr_mask(narr_mask_file_name)
wet_bulb_theta_file_name = processed_narr_io.find_file_for_one_time(
top_directory_name=top_narr_directory_name,
field_name=processed_narr_io.WET_BULB_THETA_NAME,
pressure_level_mb=pressure_level_mb,
valid_time_unix_sec=valid_time_unix_sec)
print 'Reading data from: "{0:s}"...'.format(wet_bulb_theta_file_name)
wet_bulb_theta_matrix_kelvins = processed_narr_io.read_fields_from_file(
wet_bulb_theta_file_name)[0][0, ...]
wet_bulb_theta_matrix_kelvins = general_utils.fill_nans(
wet_bulb_theta_matrix_kelvins)
u_wind_file_name = processed_narr_io.find_file_for_one_time(
top_directory_name=top_narr_directory_name,
field_name=processed_narr_io.U_WIND_GRID_RELATIVE_NAME,
pressure_level_mb=pressure_level_mb,
valid_time_unix_sec=valid_time_unix_sec)
print 'Reading data from: "{0:s}"...'.format(u_wind_file_name)
u_wind_matrix_m_s01 = processed_narr_io.read_fields_from_file(
u_wind_file_name)[0][0, ...]
u_wind_matrix_m_s01 = general_utils.fill_nans(u_wind_matrix_m_s01)
v_wind_file_name = processed_narr_io.find_file_for_one_time(
top_directory_name=top_narr_directory_name,
field_name=processed_narr_io.V_WIND_GRID_RELATIVE_NAME,
pressure_level_mb=pressure_level_mb,
valid_time_unix_sec=valid_time_unix_sec)
print 'Reading data from: "{0:s}"...'.format(v_wind_file_name)
v_wind_matrix_m_s01 = processed_narr_io.read_fields_from_file(
v_wind_file_name)[0][0, ...]
v_wind_matrix_m_s01 = general_utils.fill_nans(v_wind_matrix_m_s01)
unsmoothed_narr_file_name = '{0:s}/unsmoothed_narr_fields.jpg'.format(
output_dir_name)
_plot_narr_fields(
wet_bulb_theta_matrix_kelvins=wet_bulb_theta_matrix_kelvins,
u_wind_matrix_m_s01=u_wind_matrix_m_s01,
v_wind_matrix_m_s01=v_wind_matrix_m_s01,
title_string='Predictors before smoothing', annotation_string='(a)',
output_file_name=unsmoothed_narr_file_name)
wet_bulb_theta_matrix_kelvins = nfa.gaussian_smooth_2d_field(
field_matrix=wet_bulb_theta_matrix_kelvins,
standard_deviation_pixels=smoothing_radius_pixels,
cutoff_radius_pixels=cutoff_radius_pixels)
u_wind_matrix_m_s01 = nfa.gaussian_smooth_2d_field(
field_matrix=u_wind_matrix_m_s01,
standard_deviation_pixels=smoothing_radius_pixels,
cutoff_radius_pixels=cutoff_radius_pixels)
v_wind_matrix_m_s01 = nfa.gaussian_smooth_2d_field(
field_matrix=v_wind_matrix_m_s01,
standard_deviation_pixels=smoothing_radius_pixels,
cutoff_radius_pixels=cutoff_radius_pixels)
smoothed_narr_file_name = '{0:s}/smoothed_narr_fields.jpg'.format(
output_dir_name)
_plot_narr_fields(
wet_bulb_theta_matrix_kelvins=wet_bulb_theta_matrix_kelvins,
u_wind_matrix_m_s01=u_wind_matrix_m_s01,
v_wind_matrix_m_s01=v_wind_matrix_m_s01,
title_string='Predictors after smoothing', annotation_string='(b)',
output_file_name=smoothed_narr_file_name)
x_spacing_metres, y_spacing_metres = nwp_model_utils.get_xy_grid_spacing(
model_name=nwp_model_utils.NARR_MODEL_NAME)
tfp_matrix_kelvins_m02 = nfa.get_thermal_front_param(
thermal_field_matrix_kelvins=wet_bulb_theta_matrix_kelvins,
x_spacing_metres=x_spacing_metres, y_spacing_metres=y_spacing_metres)
tfp_matrix_kelvins_m02[narr_mask_matrix == 0] = 0.
tfp_file_name = '{0:s}/tfp.jpg'.format(output_dir_name)
tfp_title_string = (
r'Thermal front parameter ($\times$ 10$^{-10}$ K m$^{-2}$)')
_plot_tfp(tfp_matrix_kelvins_m02=tfp_matrix_kelvins_m02,
title_string=tfp_title_string, annotation_string='(c)',
output_file_name=tfp_file_name)
proj_velocity_matrix_m_s01 = nfa.project_wind_to_thermal_gradient(
u_matrix_grid_relative_m_s01=u_wind_matrix_m_s01,
v_matrix_grid_relative_m_s01=v_wind_matrix_m_s01,
thermal_field_matrix_kelvins=wet_bulb_theta_matrix_kelvins,
x_spacing_metres=x_spacing_metres, y_spacing_metres=y_spacing_metres)
locating_var_matrix_m01_s01 = nfa.get_locating_variable(
tfp_matrix_kelvins_m02=tfp_matrix_kelvins_m02,
projected_velocity_matrix_m_s01=proj_velocity_matrix_m_s01)
locating_var_file_name = '{0:s}/locating_variable.jpg'.format(
output_dir_name)
locating_var_title_string = (
r'Locating variable ($\times$ 10$^{-9}$ K m$^{-1}$ s$^{-1}$)')
_plot_locating_variable(
locating_var_matrix_m01_s01=locating_var_matrix_m01_s01,
title_string=locating_var_title_string, annotation_string='(d)',
output_file_name=locating_var_file_name)
predicted_label_matrix = nfa.get_front_types(
locating_var_matrix_m01_s01=locating_var_matrix_m01_s01,
warm_front_percentile=front_percentile,
cold_front_percentile=front_percentile)
unclosed_fronts_file_name = '{0:s}/unclosed_fronts.jpg'.format(
output_dir_name)
_plot_front_types(
predicted_label_matrix=predicted_label_matrix,
title_string='Frontal regions before closing', annotation_string='(e)',
output_file_name=unclosed_fronts_file_name)
predicted_label_matrix = front_utils.close_frontal_image(
ternary_image_matrix=predicted_label_matrix,
num_iterations=num_closing_iters)
closed_fronts_file_name = '{0:s}/closed_fronts.jpg'.format(output_dir_name)
_plot_front_types(
predicted_label_matrix=predicted_label_matrix,
title_string='Frontal regions after closing', annotation_string='(f)',
output_file_name=closed_fronts_file_name)
concat_file_name = '{0:s}/nfa_procedure.jpg'.format(output_dir_name)
print 'Concatenating figures to: "{0:s}"...'.format(concat_file_name)
panel_file_names = [
unsmoothed_narr_file_name, smoothed_narr_file_name, tfp_file_name,
locating_var_file_name, unclosed_fronts_file_name,
closed_fronts_file_name
]
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=concat_file_name, num_panel_rows=3,
num_panel_columns=2)
imagemagick_utils.resize_image(
input_file_name=concat_file_name, output_file_name=concat_file_name,
output_size_pixels=CONCAT_SIZE_PIXELS)
def _plot_one_composite(gradcam_file_name, monte_carlo_file_name,
composite_name_abbrev, composite_name_verbose,
colour_map_object, min_colour_value, max_colour_value,
num_contours, smoothing_radius_grid_cells,
monte_carlo_max_fdr, output_dir_name):
"""Plots class-activation map for one composite.
:param gradcam_file_name: Path to input file (will be read by
`gradcam.read_file`).
:param monte_carlo_file_name: Path to Monte Carlo file (will be read by
`_read_monte_carlo_file`).
:param composite_name_abbrev: Abbrev composite name (will be used in file
names).
:param composite_name_verbose: Verbose composite name (will be used in
figure title).
:param colour_map_object: See documentation at top of file.
:param min_colour_value: Minimum value in colour bar (may be NaN).
:param max_colour_value: Max value in colour bar (may be NaN).
:param num_contours: See documentation at top of file.
:param smoothing_radius_grid_cells: Same.
:param monte_carlo_max_fdr: Same.
:param output_dir_name: Name of output directory (figures will be saved
here).
:return: main_figure_file_name: Path to main image file created by this
method.
:return: min_colour_value: Same as input but cannot be None.
:return: max_colour_value: Same as input but cannot be None.
"""
(mean_radar_matrix, mean_class_activn_matrix, significance_matrix,
model_metadata_dict) = _read_one_composite(
gradcam_file_name=gradcam_file_name,
smoothing_radius_grid_cells=smoothing_radius_grid_cells,
monte_carlo_file_name=monte_carlo_file_name,
monte_carlo_max_fdr=monte_carlo_max_fdr)
if numpy.isnan(min_colour_value) or numpy.isnan(max_colour_value):
min_colour_value_log10 = numpy.log10(
numpy.percentile(mean_class_activn_matrix, 1.))
max_colour_value_log10 = numpy.log10(
numpy.percentile(mean_class_activn_matrix, 99.))
min_colour_value_log10 = max([min_colour_value_log10, -2.])
max_colour_value_log10 = max([max_colour_value_log10, -1.])
min_colour_value_log10 = min([min_colour_value_log10, 1.])
max_colour_value_log10 = min([max_colour_value_log10, 2.])
min_colour_value = 10**min_colour_value_log10
max_colour_value = 10**max_colour_value_log10
else:
min_colour_value_log10 = numpy.log10(min_colour_value)
max_colour_value_log10 = numpy.log10(max_colour_value)
contour_interval_log10 = (
(max_colour_value_log10 - min_colour_value_log10) / (num_contours - 1))
mean_activn_matrix_log10 = numpy.log10(mean_class_activn_matrix)
training_option_dict = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY]
field_names = training_option_dict[trainval_io.RADAR_FIELDS_KEY]
num_fields = mean_radar_matrix.shape[-1]
num_heights = mean_radar_matrix.shape[-2]
handle_dict = plot_examples.plot_one_example(
list_of_predictor_matrices=[mean_radar_matrix],
model_metadata_dict=model_metadata_dict,
pmm_flag=True,
allow_whitespace=True,
plot_panel_names=True,
panel_name_font_size=PANEL_NAME_FONT_SIZE,
add_titles=False,
label_colour_bars=True,
colour_bar_length=COLOUR_BAR_LENGTH,
colour_bar_font_size=COLOUR_BAR_FONT_SIZE,
num_panel_rows=num_heights)
figure_objects = handle_dict[plot_examples.RADAR_FIGURES_KEY]
axes_object_matrices = handle_dict[plot_examples.RADAR_AXES_KEY]
for k in range(num_fields):
cam_plotting.plot_many_2d_grids(
class_activation_matrix_3d=numpy.flip(
mean_activn_matrix_log10[0, ...], axis=0),
axes_object_matrix=axes_object_matrices[k],
colour_map_object=colour_map_object,
min_contour_level=min_colour_value_log10,
max_contour_level=max_colour_value_log10,
contour_interval=contour_interval_log10)
significance_plotting.plot_many_2d_grids_without_coords(
significance_matrix=numpy.flip(significance_matrix[0, ...],
axis=0),
axes_object_matrix=axes_object_matrices[k])
panel_file_names = [None] * num_fields
for k in range(num_fields):
panel_file_names[k] = '{0:s}/{1:s}_{2:s}.jpg'.format(
output_dir_name, composite_name_abbrev,
field_names[k].replace('_', '-'))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[k]))
figure_objects[k].savefig(panel_file_names[k],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_objects[k])
main_figure_file_name = '{0:s}/{1:s}_gradcam.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Concatenating panels to: "{0:s}"...'.format(main_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=main_figure_file_name,
num_panel_rows=1,
num_panel_columns=num_fields,
border_width_pixels=50)
imagemagick_utils.resize_image(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=main_figure_file_name,
x_offset_from_center_px=0,
y_offset_from_top_px=0,
text_string=composite_name_verbose)
imagemagick_utils.trim_whitespace(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
border_width_pixels=10)
return main_figure_file_name, min_colour_value, max_colour_value
def _run():
"""Plots conversion of gridded probabilities into objects.
This is effectively the main method.
"""
prediction_dict = ml_utils.read_gridded_predictions(PREDICTION_FILE_NAME)
class_probability_matrix = prediction_dict[ml_utils.PROBABILITY_MATRIX_KEY]
for this_id in front_utils.VALID_INTEGER_IDS:
if this_id == front_utils.NO_FRONT_INTEGER_ID:
class_probability_matrix[..., this_id][numpy.isnan(
class_probability_matrix[..., this_id])] = 1.
else:
class_probability_matrix[..., this_id][numpy.isnan(
class_probability_matrix[..., this_id])] = 0.
predicted_label_matrix = object_eval.determinize_probabilities(
class_probability_matrix=class_probability_matrix,
binarization_threshold=BINARIZATION_THRESHOLD)
_plot_predictions(predicted_label_matrix=predicted_label_matrix,
title_string='All frontal regions',
annotation_string='(b)',
output_file_name=ALL_REGIONS_FILE_NAME)
valid_time_unix_sec = time_conversion.string_to_unix_sec(
VALID_TIME_STRING, TIME_FORMAT)
valid_times_unix_sec = numpy.array([valid_time_unix_sec], dtype=int)
predicted_region_table = object_eval.images_to_regions(
predicted_label_matrix=predicted_label_matrix,
image_times_unix_sec=valid_times_unix_sec)
num_grid_rows, num_grid_columns = nwp_model_utils.get_grid_dimensions(
model_name=nwp_model_utils.NARR_MODEL_NAME)
grid_spacing_metres = nwp_model_utils.get_xy_grid_spacing(
model_name=nwp_model_utils.NARR_MODEL_NAME)[0]
predicted_region_table = object_eval.discard_regions_with_small_area(
predicted_region_table=predicted_region_table,
x_grid_spacing_metres=grid_spacing_metres,
y_grid_spacing_metres=grid_spacing_metres,
min_area_metres2=MIN_REGION_AREA_METRES2)
predicted_label_matrix = object_eval.regions_to_images(
predicted_region_table=predicted_region_table,
num_grid_rows=num_grid_rows,
num_grid_columns=num_grid_columns)
_plot_predictions(predicted_label_matrix=predicted_label_matrix,
title_string='Large frontal regions',
annotation_string='(c)',
output_file_name=LARGE_REGIONS_FILE_NAME)
predicted_region_table = object_eval.skeletonize_frontal_regions(
predicted_region_table=predicted_region_table,
num_grid_rows=num_grid_rows,
num_grid_columns=num_grid_columns)
predicted_label_matrix = object_eval.regions_to_images(
predicted_region_table=predicted_region_table,
num_grid_rows=num_grid_rows,
num_grid_columns=num_grid_columns)
_plot_predictions(predicted_label_matrix=predicted_label_matrix,
title_string='All skeleton lines',
annotation_string='(d)',
output_file_name=ALL_SKELETONS_FILE_NAME)
predicted_region_table = object_eval.find_main_skeletons(
predicted_region_table=predicted_region_table,
image_times_unix_sec=valid_times_unix_sec,
num_grid_rows=num_grid_rows,
num_grid_columns=num_grid_columns,
x_grid_spacing_metres=grid_spacing_metres,
y_grid_spacing_metres=grid_spacing_metres,
min_endpoint_length_metres=MIN_ENDPOINT_LENGTH_METRES)
predicted_label_matrix = object_eval.regions_to_images(
predicted_region_table=predicted_region_table,
num_grid_rows=num_grid_rows,
num_grid_columns=num_grid_columns)
_plot_predictions(predicted_label_matrix=predicted_label_matrix,
title_string='Main skeleton lines',
annotation_string='(e)',
output_file_name=MAIN_SKELETONS_FILE_NAME)
print 'Concatenating figures to: "{0:s}"...'.format(CONCAT_FILE_NAME)
panel_file_names = [
PROBABILITY_FILE_NAME, ALL_REGIONS_FILE_NAME, LARGE_REGIONS_FILE_NAME,
ALL_SKELETONS_FILE_NAME, MAIN_SKELETONS_FILE_NAME
]
imagemagick_utils.concatenate_images(input_file_names=panel_file_names,
output_file_name=CONCAT_FILE_NAME,
num_panel_rows=2,
num_panel_columns=3)
imagemagick_utils.resize_image(input_file_name=CONCAT_FILE_NAME,
output_file_name=CONCAT_FILE_NAME,
output_size_pixels=CONCAT_SIZE_PIXELS)
def _plot_histogram_one_target(target_values, target_name, num_bins,
letter_label, output_dir_name):
"""Plots histogram for one target variable.
:param target_values: 1-D numpy array of values.
:param target_name: Name of target variable.
:param num_bins: Number of bins in histogram.
:param letter_label: Letter label (will be used to label panel).
:param output_dir_name: Name of output directory. Figure will be saved
here.
:return: output_file_name: Path to output file.
"""
min_value = (numpy.min(target_values)
if target_name == SHORTWAVE_NET_FLUX_NAME else 0.)
max_value = numpy.max(target_values)
num_examples_by_bin = histograms.create_histogram(
input_values=target_values,
num_bins=num_bins,
min_value=min_value,
max_value=max_value)[1]
frequency_by_bin = (num_examples_by_bin.astype(float) /
numpy.sum(num_examples_by_bin))
bin_edges = numpy.linspace(min_value, max_value, num=num_bins + 1)
bin_centers = numpy.array(
[numpy.mean(bin_edges[[k, k + 1]]) for k in range(num_bins)])
x_tick_coords = 0.5 + numpy.linspace(
0, num_bins - 1, num=num_bins, dtype=float)
if target_name == example_utils.SHORTWAVE_HEATING_RATE_NAME:
x_tick_labels = ['{0:.1f}'.format(c) for c in bin_centers]
else:
x_tick_labels = [
'{0:d}'.format(int(numpy.round(c))) for c in bin_centers
]
x_tick_labels = [
x_tick_labels[k] if numpy.mod(k, 3) == 0 else ' '
for k in range(num_bins)
]
figure_object, axes_object = pyplot.subplots(
1, 1, figsize=(FIGURE_WIDTH_INCHES, FIGURE_HEIGHT_INCHES))
axes_object.bar(x=x_tick_coords,
height=frequency_by_bin,
width=1.,
color=FACE_COLOUR,
edgecolor=EDGE_COLOUR,
linewidth=EDGE_WIDTH)
axes_object.set_xlim([x_tick_coords[0] - 0.5, x_tick_coords[-1] + 0.5])
axes_object.set_xticks(x_tick_coords)
axes_object.set_xticklabels(x_tick_labels, rotation=90.)
axes_object.set_ylabel('Frequency')
axes_object.set_xlabel(TARGET_NAME_TO_VERBOSE[target_name])
plotting_utils.label_axes(axes_object=axes_object,
label_string='({0:s})'.format(letter_label))
output_file_name = '{0:s}/histogram_{1:s}.jpg'.format(
output_dir_name, target_name.replace('_', '-'))
print('Saving figure to: "{0:s}"...'.format(output_file_name))
figure_object.savefig(output_file_name,
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
imagemagick_utils.trim_whitespace(input_file_name=output_file_name,
output_file_name=output_file_name)
imagemagick_utils.resize_image(input_file_name=output_file_name,
output_file_name=output_file_name,
output_size_pixels=int(2.5e6))
return output_file_name
def _run(input_dir_name, site_names_for_reliability, site_name_for_histogram,
output_dir_name):
"""Creates figure showing overall model evaluation.
This is effectively the main method.
:param input_dir_name: See documentation at top of file.
:param site_names_for_reliability: Same.
:param site_name_for_histogram: Same.
:param output_dir_name: Same.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
pathless_input_file_names = [
'shortwave-surface-down-flux-w-m02_attributes_{0:s}.jpg'.format(
site_name_for_histogram),
'shortwave-toa-up-flux-w-m02_attributes_{0:s}.jpg'.format(
site_name_for_histogram),
'net-shortwave-flux-w-m02_attributes_{0:s}.jpg'.format(
site_name_for_histogram),
'shortwave-heating-rate-k-day01_bias_profile.jpg',
'shortwave-heating-rate-k-day01_mean-absolute-error_profile.jpg',
'shortwave-heating-rate-k-day01_mae-skill-score_profile.jpg'
]
num_sites_for_relia = len(site_names_for_reliability)
for j in range(num_sites_for_relia):
this_file_name = (
'shortwave-heating-rate-k-day01_reliability_{0:s}.jpg').format(
site_names_for_reliability[j])
pathless_input_file_names.append(this_file_name)
panel_file_names = [
'{0:s}/{1:s}'.format(input_dir_name, p)
for p in pathless_input_file_names
]
resized_panel_file_names = [
'{0:s}/{1:s}'.format(output_dir_name, p)
for p in pathless_input_file_names
]
letter_label = None
for i in range(len(panel_file_names)):
print('Resizing panel and saving to: "{0:s}"...'.format(
resized_panel_file_names[i]))
imagemagick_utils.trim_whitespace(
input_file_name=panel_file_names[i],
output_file_name=resized_panel_file_names[i])
if letter_label is None:
letter_label = 'a'
else:
letter_label = chr(ord(letter_label) + 1)
_overlay_text(image_file_name=resized_panel_file_names[i],
x_offset_from_left_px=0,
y_offset_from_top_px=TITLE_FONT_SIZE,
text_string='({0:s})'.format(letter_label))
imagemagick_utils.resize_image(
input_file_name=resized_panel_file_names[i],
output_file_name=resized_panel_file_names[i],
output_size_pixels=PANEL_SIZE_PX)
concat_figure_file_name = '{0:s}/evaluation_by_site.jpg'.format(
output_dir_name)
print(
'Concatenating panels to: "{0:s}"...'.format(concat_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=resized_panel_file_names,
output_file_name=concat_figure_file_name,
num_panel_rows=NUM_PANEL_ROWS,
num_panel_columns=NUM_PANEL_COLUMNS)
imagemagick_utils.resize_image(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _run(saliency_file_names, monte_carlo_file_names, composite_names,
colour_map_name, max_colour_values, half_num_contours,
smoothing_radius_grid_cells, output_dir_name):
"""Makes figure with sanity checks for saliency maps.
This is effectively the main method.
:param saliency_file_names: See documentation at top of file.
:param monte_carlo_file_names: Same.
:param composite_names: Same.
:param colour_map_name: Same.
:param max_colour_values: Same.
:param half_num_contours: Same.
:param smoothing_radius_grid_cells: Same.
:param output_dir_name: Same.
"""
# Process input args.
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
if smoothing_radius_grid_cells <= 0:
smoothing_radius_grid_cells = None
colour_map_object = pyplot.cm.get_cmap(colour_map_name)
error_checking.assert_is_geq(half_num_contours, 5)
num_composites = len(saliency_file_names)
expected_dim = numpy.array([num_composites], dtype=int)
error_checking.assert_is_numpy_array(numpy.array(composite_names),
exact_dimensions=expected_dim)
error_checking.assert_is_numpy_array(numpy.array(monte_carlo_file_names),
exact_dimensions=expected_dim)
monte_carlo_file_names = [
None if f in NONE_STRINGS else f for f in monte_carlo_file_names
]
max_colour_values[max_colour_values <= 0] = numpy.nan
error_checking.assert_is_numpy_array(max_colour_values,
exact_dimensions=expected_dim)
composite_names_abbrev = [
n.replace('_', '-').lower() for n in composite_names
]
composite_names_verbose = [
'({0:s}) {1:s}'.format(chr(ord('a') + i),
composite_names[i].replace('_', ' '))
for i in range(num_composites)
]
panel_file_names = [None] * num_composites
for i in range(num_composites):
panel_file_names[i], max_colour_values[i] = _plot_one_composite(
saliency_file_name=saliency_file_names[i],
monte_carlo_file_name=monte_carlo_file_names[i],
composite_name_abbrev=composite_names_abbrev[i],
composite_name_verbose=composite_names_verbose[i],
colour_map_object=colour_map_object,
max_colour_value=max_colour_values[i],
half_num_contours=half_num_contours,
smoothing_radius_grid_cells=smoothing_radius_grid_cells,
output_dir_name=output_dir_name)
_add_colour_bar(figure_file_name=panel_file_names[i],
colour_map_object=colour_map_object,
max_colour_value=max_colour_values[i],
temporary_dir_name=output_dir_name)
print('\n')
figure_file_name = '{0:s}/saliency_concat.jpg'.format(output_dir_name)
print('Concatenating panels to: "{0:s}"...'.format(figure_file_name))
num_panel_rows = int(numpy.ceil(numpy.sqrt(num_composites)))
num_panel_columns = int(numpy.floor(
float(num_composites) / num_panel_rows))
imagemagick_utils.concatenate_images(input_file_names=panel_file_names,
output_file_name=figure_file_name,
border_width_pixels=100,
num_panel_rows=num_panel_rows,
num_panel_columns=num_panel_columns)
imagemagick_utils.trim_whitespace(input_file_name=figure_file_name,
output_file_name=figure_file_name,
border_width_pixels=10)
imagemagick_utils.resize_image(input_file_name=figure_file_name,
output_file_name=figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _plot_sounding_saliency(
saliency_matrix, colour_map_object, max_colour_value, sounding_matrix,
saliency_dict, model_metadata_dict, output_dir_name, pmm_flag,
example_index=None):
"""Plots saliency for sounding.
H = number of sounding heights
F = number of sounding fields
If plotting a composite rather than one example, `full_storm_id_string` and
`storm_time_unix_sec` can be None.
:param saliency_matrix: H-by-F numpy array of saliency values.
:param colour_map_object: See documentation at top of file.
:param max_colour_value: Max value in colour scheme for saliency.
:param sounding_matrix: H-by-F numpy array of actual sounding values.
:param saliency_dict: Dictionary returned from
`saliency_maps.read_standard_file` or `saliency_maps.read_pmm_file`.
:param model_metadata_dict: Dictionary returned by
`cnn.read_model_metadata`.
:param output_dir_name: Path to output directory. Figure will be saved
here.
:param pmm_flag: Boolean flag. If True, will plot composite rather than one
example.
:param example_index: [used only if `pmm_flag == False`]
Will plot the [i]th example, where i = `example_index`.
"""
if pmm_flag:
example_index = 0
training_option_dict = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY]
sounding_field_names = training_option_dict[trainval_io.SOUNDING_FIELDS_KEY]
sounding_heights_m_agl = training_option_dict[
trainval_io.SOUNDING_HEIGHTS_KEY]
sounding_matrix = numpy.expand_dims(sounding_matrix, axis=0)
if saliency_maps.SOUNDING_PRESSURES_KEY in saliency_dict:
pressure_matrix_pascals = numpy.expand_dims(
saliency_dict[saliency_maps.SOUNDING_PRESSURES_KEY], axis=-1
)
pressure_matrix_pascals = pressure_matrix_pascals[[example_index], ...]
sounding_matrix = numpy.concatenate(
(sounding_matrix, pressure_matrix_pascals), axis=-1
)
sounding_dict_for_metpy = dl_utils.soundings_to_metpy_dictionaries(
sounding_matrix=sounding_matrix,
field_names=sounding_field_names + [soundings.PRESSURE_NAME]
)[0]
else:
sounding_dict_for_metpy = dl_utils.soundings_to_metpy_dictionaries(
sounding_matrix=sounding_matrix, field_names=sounding_field_names,
height_levels_m_agl=sounding_heights_m_agl,
storm_elevations_m_asl=numpy.array([0.])
)[0]
if pmm_flag:
full_storm_id_string = None
storm_time_unix_sec = None
title_string = 'PMM composite'
else:
full_storm_id_string = saliency_dict[saliency_maps.FULL_IDS_KEY][
example_index]
storm_time_unix_sec = saliency_dict[saliency_maps.STORM_TIMES_KEY][
example_index]
title_string = 'Storm "{0:s}" at {1:s}'.format(
full_storm_id_string,
time_conversion.unix_sec_to_string(
storm_time_unix_sec, plot_input_examples.TIME_FORMAT)
)
sounding_plotting.plot_sounding(
sounding_dict_for_metpy=sounding_dict_for_metpy,
title_string=title_string)
left_panel_file_name = plot_input_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=False, pmm_flag=pmm_flag,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec,
radar_field_name='sounding-actual')
print('Saving figure to file: "{0:s}"...'.format(left_panel_file_name))
pyplot.savefig(left_panel_file_name, dpi=FIGURE_RESOLUTION_DPI)
pyplot.close()
imagemagick_utils.trim_whitespace(
input_file_name=left_panel_file_name,
output_file_name=left_panel_file_name)
saliency_plotting.plot_saliency_for_sounding(
saliency_matrix=saliency_matrix,
sounding_field_names=sounding_field_names,
pressure_levels_mb=sounding_dict_for_metpy[
soundings.PRESSURE_COLUMN_METPY],
colour_map_object=colour_map_object,
max_absolute_colour_value=max_colour_value)
right_panel_file_name = plot_input_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=False, pmm_flag=pmm_flag,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec,
radar_field_name='sounding-saliency')
print('Saving figure to file: "{0:s}"...'.format(right_panel_file_name))
pyplot.savefig(right_panel_file_name, dpi=FIGURE_RESOLUTION_DPI)
pyplot.close()
imagemagick_utils.trim_whitespace(
input_file_name=right_panel_file_name,
output_file_name=right_panel_file_name)
concat_file_name = plot_input_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=True, pmm_flag=pmm_flag,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec)
print('Concatenating figures to: "{0:s}"...\n'.format(concat_file_name))
imagemagick_utils.concatenate_images(
input_file_names=[left_panel_file_name, right_panel_file_name],
output_file_name=concat_file_name, num_panel_rows=1,
num_panel_columns=2)
imagemagick_utils.resize_image(
input_file_name=concat_file_name, output_file_name=concat_file_name,
output_size_pixels=SOUNDING_IMAGE_SIZE_PX)
os.remove(left_panel_file_name)
os.remove(right_panel_file_name)
def _run(input_dir_name, heights_m_agl, output_dir_name):
"""Creates figure showing overall model evaluation.
This is effectively the main method.
:param input_dir_name: See documentation at top of file.
:param heights_m_agl: Same.
:param output_dir_name: Same.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
num_heights = len(heights_m_agl)
pathless_input_file_names = []
for j in range(num_heights):
first_file_name = (
'shortwave-heating-rate-k-day01_{0:05d}metres_scores_without_units'
'.jpg').format(heights_m_agl[j])
second_file_name = (
'shortwave-heating-rate-k-day01_{0:05d}metres_scores_with_units.jpg'
).format(heights_m_agl[j])
pathless_input_file_names += [first_file_name, second_file_name]
pathless_input_file_names += [
'net-shortwave-flux-w-m02_scores_without_units.jpg',
'net-shortwave-flux-w-m02_scores_with_units.jpg'
]
panel_file_names = [
'{0:s}/{1:s}'.format(input_dir_name, p)
for p in pathless_input_file_names
]
resized_panel_file_names = [
'{0:s}/{1:s}'.format(output_dir_name, p)
for p in pathless_input_file_names
]
letter_label = None
for i in range(len(panel_file_names)):
print('Resizing panel and saving to: "{0:s}"...'.format(
resized_panel_file_names[i]))
imagemagick_utils.trim_whitespace(
input_file_name=panel_file_names[i],
output_file_name=resized_panel_file_names[i])
if letter_label is None:
letter_label = 'a'
else:
letter_label = chr(ord(letter_label) + 1)
image_matrix = Image.open(resized_panel_file_names[i])
_, figure_height_px = image_matrix.size
_overlay_text(image_file_name=resized_panel_file_names[i],
x_offset_from_left_px=0,
y_offset_from_top_px=figure_height_px - 50,
text_string='({0:s})'.format(letter_label))
imagemagick_utils.resize_image(
input_file_name=resized_panel_file_names[i],
output_file_name=resized_panel_file_names[i],
output_size_pixels=PANEL_SIZE_PX)
concat_figure_file_name = '{0:s}/evaluation_by_time.jpg'.format(
output_dir_name)
print(
'Concatenating panels to: "{0:s}"...'.format(concat_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=resized_panel_file_names,
output_file_name=concat_figure_file_name,
num_panel_rows=num_heights + 1,
num_panel_columns=NUM_PANEL_COLUMNS)
imagemagick_utils.resize_image(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _plot_one_example(
radar_matrix, sounding_matrix, sounding_pressures_pascals,
full_storm_id_string, storm_time_unix_sec, model_metadata_dict,
output_dir_name):
"""Plots predictors for one example.
M = number of rows in radar grid
N = number of columns in radar grid
H_r = number of heights in radar grid
F_r = number of radar fields
H_s = number of sounding heights
F_s = number of sounding fields
:param radar_matrix: numpy array (1 x M x N x H_r x F_r) of radar values.
:param sounding_matrix: numpy array (1 x H_s x F_s) of sounding values.
:param sounding_pressures_pascals: numpy array (length H_s) of sounding
pressures.
:param full_storm_id_string: Full storm ID.
:param storm_time_unix_sec: Valid time.
:param model_metadata_dict: Dictionary returned by
`cnn.read_model_metadata`.
:param output_dir_name: Name of output directory (figures will be saved
here).
:return: radar_figure_file_name: Path to radar figure created by this
method.
"""
training_option_dict = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY]
radar_field_names = training_option_dict[trainval_io.RADAR_FIELDS_KEY]
num_radar_fields = radar_matrix.shape[-1]
num_radar_heights = radar_matrix.shape[-2]
handle_dict = plot_examples.plot_one_example(
list_of_predictor_matrices=[radar_matrix, sounding_matrix],
model_metadata_dict=model_metadata_dict,
pmm_flag=False, example_index=0, plot_sounding=True,
sounding_pressures_pascals=sounding_pressures_pascals,
allow_whitespace=True, plot_panel_names=True,
panel_name_font_size=PANEL_NAME_FONT_SIZE,
add_titles=False, label_colour_bars=True,
colour_bar_length=COLOUR_BAR_LENGTH,
colour_bar_font_size=COLOUR_BAR_FONT_SIZE,
num_panel_rows=num_radar_heights
)
sounding_file_name = plot_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=True, pmm_flag=False,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec
)
print('Saving figure to: "{0:s}"...'.format(sounding_file_name))
sounding_figure_object = handle_dict[plot_examples.SOUNDING_FIGURE_KEY]
sounding_figure_object.savefig(
sounding_file_name, dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0, bbox_inches='tight'
)
pyplot.close(sounding_figure_object)
figure_objects = handle_dict[plot_examples.RADAR_FIGURES_KEY]
panel_file_names = [None] * num_radar_fields
for k in range(num_radar_fields):
panel_file_names[k] = plot_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=False, pmm_flag=False,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec,
radar_field_name=radar_field_names[k]
)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[k]))
figure_objects[k].savefig(
panel_file_names[k], dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0, bbox_inches='tight'
)
pyplot.close(figure_objects[k])
radar_figure_file_name = plot_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=False, pmm_flag=False,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec
)
print('Concatenating panels to: "{0:s}"...'.format(radar_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=radar_figure_file_name,
num_panel_rows=1, num_panel_columns=num_radar_fields,
border_width_pixels=50
)
imagemagick_utils.resize_image(
input_file_name=radar_figure_file_name,
output_file_name=radar_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX
)
imagemagick_utils.trim_whitespace(
input_file_name=radar_figure_file_name,
output_file_name=radar_figure_file_name,
border_width_pixels=10
)
for this_file_name in panel_file_names:
os.remove(this_file_name)
return radar_figure_file_name
def _run(gradcam_file_names, monte_carlo_file_names, composite_names,
colour_map_name, min_colour_values, max_colour_values, num_contours,
smoothing_radius_grid_cells, monte_carlo_max_fdr, output_dir_name):
"""Makes figure with gradient-weighted class-activation maps (Grad-CAM).
This is effectively the main method.
:param gradcam_file_names: See documentation at top of file.
:param monte_carlo_file_names: Same.
:param composite_names: Same.
:param colour_map_name: Same.
:param min_colour_values: Same.
:param max_colour_values: Same.
:param num_contours: Same.
:param smoothing_radius_grid_cells: Same.
:param output_dir_name: Same.
"""
if smoothing_radius_grid_cells <= 0:
smoothing_radius_grid_cells = None
if monte_carlo_max_fdr <= 0:
monte_carlo_max_fdr = None
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
colour_map_object = pyplot.cm.get_cmap(colour_map_name)
error_checking.assert_is_geq(num_contours, 10)
num_composites = len(gradcam_file_names)
expected_dim = numpy.array([num_composites], dtype=int)
error_checking.assert_is_numpy_array(numpy.array(composite_names),
exact_dimensions=expected_dim)
error_checking.assert_is_numpy_array(numpy.array(monte_carlo_file_names),
exact_dimensions=expected_dim)
monte_carlo_file_names = [
None if f in NONE_STRINGS else f for f in monte_carlo_file_names
]
nan_indices = numpy.where(
numpy.logical_or(max_colour_values < 0, min_colour_values < 0))[0]
min_colour_values[nan_indices] = numpy.nan
max_colour_values[nan_indices] = numpy.nan
error_checking.assert_is_numpy_array(min_colour_values,
exact_dimensions=expected_dim)
error_checking.assert_is_numpy_array(max_colour_values,
exact_dimensions=expected_dim)
assert not numpy.any(max_colour_values <= min_colour_values)
composite_names_abbrev = [
n.replace('_', '-').lower() for n in composite_names
]
composite_names_verbose = [
'({0:s}) {1:s}'.format(chr(ord('a') + i),
composite_names[i].replace('_', ' '))
for i in range(num_composites)
]
panel_file_names = [None] * num_composites
for i in range(num_composites):
(panel_file_names[i], min_colour_values[i],
max_colour_values[i]) = _plot_one_composite(
gradcam_file_name=gradcam_file_names[i],
monte_carlo_file_name=monte_carlo_file_names[i],
composite_name_abbrev=composite_names_abbrev[i],
composite_name_verbose=composite_names_verbose[i],
colour_map_object=colour_map_object,
min_colour_value=min_colour_values[i],
max_colour_value=max_colour_values[i],
num_contours=num_contours,
smoothing_radius_grid_cells=smoothing_radius_grid_cells,
monte_carlo_max_fdr=monte_carlo_max_fdr,
output_dir_name=output_dir_name)
_add_colour_bar(figure_file_name=panel_file_names[i],
colour_map_object=colour_map_object,
min_colour_value=min_colour_values[i],
max_colour_value=max_colour_values[i],
temporary_dir_name=output_dir_name)
print('\n')
figure_file_name = '{0:s}/gradcam_concat.jpg'.format(output_dir_name)
print('Concatenating panels to: "{0:s}"...'.format(figure_file_name))
num_panel_rows = int(numpy.ceil(numpy.sqrt(num_composites)))
num_panel_columns = int(numpy.floor(
float(num_composites) / num_panel_rows))
imagemagick_utils.concatenate_images(input_file_names=panel_file_names,
output_file_name=figure_file_name,
border_width_pixels=25,
num_panel_rows=num_panel_rows,
num_panel_columns=num_panel_columns)
imagemagick_utils.trim_whitespace(input_file_name=figure_file_name,
output_file_name=figure_file_name,
border_width_pixels=10)
imagemagick_utils.resize_image(input_file_name=figure_file_name,
output_file_name=figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _plot_one_composite(saliency_file_name, monte_carlo_file_name,
composite_name_abbrev, composite_name_verbose,
colour_map_object, max_colour_value, half_num_contours,
smoothing_radius_grid_cells, output_dir_name):
"""Plots saliency map for one composite.
:param saliency_file_name: Path to saliency file (will be read by
`saliency.read_file`).
:param monte_carlo_file_name: Path to Monte Carlo file (will be read by
`_read_monte_carlo_file`).
:param composite_name_abbrev: Abbrev composite name (will be used in file
names).
:param composite_name_verbose: Verbose composite name (will be used in
figure title).
:param colour_map_object: See documentation at top of file.
:param max_colour_value: Same.
:param half_num_contours: Same.
:param smoothing_radius_grid_cells: Same.
:param output_dir_name: Name of output directory (figures will be saved
here).
:return: main_figure_file_name: Path to main image file created by this
method.
"""
(mean_radar_matrices, mean_saliency_matrices, significance_matrices,
model_metadata_dict) = _read_one_composite(
saliency_file_name=saliency_file_name,
smoothing_radius_grid_cells=smoothing_radius_grid_cells,
monte_carlo_file_name=monte_carlo_file_name)
refl_heights_m_agl = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY][
trainval_io.RADAR_HEIGHTS_KEY]
num_refl_heights = len(refl_heights_m_agl)
handle_dict = plot_examples.plot_one_example(
list_of_predictor_matrices=mean_radar_matrices,
model_metadata_dict=model_metadata_dict,
pmm_flag=True,
plot_sounding=False,
allow_whitespace=True,
plot_panel_names=True,
panel_name_font_size=PANEL_NAME_FONT_SIZE,
add_titles=False,
label_colour_bars=True,
colour_bar_length=COLOUR_BAR_LENGTH,
colour_bar_font_size=COLOUR_BAR_FONT_SIZE,
num_panel_rows=num_refl_heights)
axes_object_matrices = handle_dict[plot_examples.RADAR_AXES_KEY]
this_saliency_matrix = numpy.flip(mean_saliency_matrices[0][0, ..., 0],
axis=0)
saliency_plotting.plot_many_2d_grids_with_contours(
saliency_matrix_3d=this_saliency_matrix,
axes_object_matrix=axes_object_matrices[0],
colour_map_object=colour_map_object,
max_absolute_contour_level=max_colour_value,
contour_interval=max_colour_value / half_num_contours,
row_major=True)
this_sig_matrix = numpy.flip(significance_matrices[0][0, ..., 0], axis=0)
significance_plotting.plot_many_2d_grids_without_coords(
significance_matrix=this_sig_matrix,
axes_object_matrix=axes_object_matrices[0],
marker_size=2,
row_major=True)
this_saliency_matrix = numpy.flip(mean_saliency_matrices[1][0, ...],
axis=0)
saliency_plotting.plot_many_2d_grids_with_contours(
saliency_matrix_3d=this_saliency_matrix,
axes_object_matrix=axes_object_matrices[1],
colour_map_object=colour_map_object,
max_absolute_contour_level=max_colour_value,
contour_interval=max_colour_value / half_num_contours,
row_major=False)
this_sig_matrix = numpy.flip(significance_matrices[1][0, ...], axis=0)
significance_plotting.plot_many_2d_grids_without_coords(
significance_matrix=this_sig_matrix,
axes_object_matrix=axes_object_matrices[1],
marker_size=2,
row_major=False)
refl_figure_object = handle_dict[plot_examples.RADAR_FIGURES_KEY][0]
refl_figure_file_name = '{0:s}/{1:s}_reflectivity.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Saving figure to: "{0:s}"...'.format(refl_figure_file_name))
refl_figure_object.savefig(refl_figure_file_name,
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(refl_figure_object)
shear_figure_object = handle_dict[plot_examples.RADAR_FIGURES_KEY][1]
shear_figure_file_name = '{0:s}/{1:s}_shear.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Saving figure to: "{0:s}"...'.format(shear_figure_file_name))
shear_figure_object.savefig(shear_figure_file_name,
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(shear_figure_object)
main_figure_file_name = '{0:s}/{1:s}_saliency.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Concatenating panels to: "{0:s}"...'.format(main_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=[refl_figure_file_name, shear_figure_file_name],
output_file_name=main_figure_file_name,
num_panel_rows=1,
num_panel_columns=2,
border_width_pixels=50,
extra_args_string='-gravity south')
imagemagick_utils.resize_image(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=main_figure_file_name,
x_offset_from_center_px=0,
y_offset_from_top_px=0,
text_string=composite_name_verbose)
imagemagick_utils.trim_whitespace(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
border_width_pixels=10)
return main_figure_file_name
def _plot_composite(composite_file_name, composite_name_abbrev,
composite_name_verbose, output_dir_name):
"""Plots one composite.
:param composite_file_name: Path to input file. Will be read by
`_read_composite`.
:param composite_name_abbrev: Abbreviated name for composite. Will be used
in names of output files.
:param composite_name_verbose: Verbose name for composite. Will be used as
figure title.
:param output_dir_name: Path to output directory. Figures will be saved
here.
:return: radar_figure_file_name: Path to file with radar figure for this
composite.
:return: sounding_figure_file_name: Path to file with sounding figure for
this composite.
"""
mean_predictor_matrices, model_metadata_dict, mean_sounding_pressures_pa = (
_read_composite(composite_file_name))
radar_field_names = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY][
trainval_io.RADAR_FIELDS_KEY]
radar_heights_m_agl = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY][
trainval_io.RADAR_HEIGHTS_KEY]
num_radar_fields = len(radar_field_names)
num_radar_heights = len(radar_heights_m_agl)
handle_dict = plot_examples.plot_one_example(
list_of_predictor_matrices=mean_predictor_matrices,
model_metadata_dict=model_metadata_dict,
pmm_flag=True,
plot_sounding=True,
sounding_pressures_pascals=mean_sounding_pressures_pa,
allow_whitespace=True,
plot_panel_names=True,
panel_name_font_size=PANEL_NAME_FONT_SIZE,
add_titles=False,
label_colour_bars=True,
colour_bar_length=COLOUR_BAR_LENGTH,
colour_bar_font_size=COLOUR_BAR_FONT_SIZE,
sounding_font_size=SOUNDING_FONT_SIZE,
num_panel_rows=num_radar_heights)
sounding_figure_file_name = '{0:s}/{1:s}_sounding.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Saving figure to: "{0:s}"...'.format(sounding_figure_file_name))
sounding_figure_object = handle_dict[plot_examples.SOUNDING_FIGURE_KEY]
sounding_figure_object.savefig(sounding_figure_file_name,
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(sounding_figure_object)
imagemagick_utils.resize_image(input_file_name=sounding_figure_file_name,
output_file_name=sounding_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(
input_file_name=sounding_figure_file_name,
output_file_name=sounding_figure_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=sounding_figure_file_name,
x_offset_from_center_px=0,
y_offset_from_top_px=0,
text_string=composite_name_verbose)
imagemagick_utils.trim_whitespace(
input_file_name=sounding_figure_file_name,
output_file_name=sounding_figure_file_name,
border_width_pixels=10)
radar_figure_objects = handle_dict[plot_examples.RADAR_FIGURES_KEY]
panel_file_names = [None] * num_radar_fields
for j in range(num_radar_fields):
panel_file_names[j] = '{0:s}/{1:s}_{2:s}.jpg'.format(
output_dir_name, composite_name_abbrev,
radar_field_names[j].replace('_', '-'))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[j]))
radar_figure_objects[j].savefig(panel_file_names[j],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(radar_figure_objects[j])
radar_figure_file_name = '{0:s}/{1:s}_radar.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Concatenating panels to: "{0:s}"...'.format(radar_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=radar_figure_file_name,
num_panel_rows=1,
num_panel_columns=num_radar_fields,
border_width_pixels=50)
imagemagick_utils.resize_image(input_file_name=radar_figure_file_name,
output_file_name=radar_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(input_file_name=radar_figure_file_name,
output_file_name=radar_figure_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=radar_figure_file_name,
x_offset_from_center_px=0,
y_offset_from_top_px=0,
text_string=composite_name_verbose)
imagemagick_utils.trim_whitespace(input_file_name=radar_figure_file_name,
output_file_name=radar_figure_file_name,
border_width_pixels=10)
return radar_figure_file_name, sounding_figure_file_name
def _run(input_file_name, output_dir_name):
"""Plots mean example created by average_examples.py.
This is effectively the main method.
:param input_file_name: See documentation at top of file.
:param output_dir_name: Same.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
pickle_file_handle = open(input_file_name, 'rb')
mean_example_dict = pickle.load(pickle_file_handle)
pickle_file_handle.close()
# Plot predictors with liquid-water content (LWC).
figure_object = profile_plotting.plot_predictors(
example_dict=mean_example_dict,
example_index=0,
plot_ice=False,
use_log_scale=True)[0]
panel_file_names = ['foo'] * 3
panel_file_names[0] = '{0:s}/predictors_with_lwc.jpg'.format(
output_dir_name)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[0]))
figure_object.savefig(panel_file_names[0],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot predictors with ice-water content (IWC).
figure_object = profile_plotting.plot_predictors(
example_dict=mean_example_dict,
example_index=0,
plot_ice=True,
use_log_scale=True)[0]
panel_file_names[1] = '{0:s}/predictors_with_iwc.jpg'.format(
output_dir_name)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[1]))
figure_object.savefig(panel_file_names[1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot targets.
figure_object = profile_plotting.plot_targets(
example_dict=mean_example_dict, example_index=0, use_log_scale=True)[0]
panel_file_names[2] = '{0:s}/targets.jpg'.format(output_dir_name)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[2]))
figure_object.savefig(panel_file_names[2],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Concatenate panels.
concat_figure_file_name = '{0:s}/predictors_and_targets.jpg'.format(
output_dir_name)
print(
'Concatenating panels to: "{0:s}"...'.format(concat_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=concat_figure_file_name,
num_panel_rows=2,
num_panel_columns=2,
border_width_pixels=50)
imagemagick_utils.resize_image(input_file_name=concat_figure_file_name,
output_file_name=concat_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _run(evaluation_dir_name, smoothing_radius_grid_cells,
score_colour_map_name, num_ex_colour_map_name, max_colour_percentile,
output_dir_name):
"""Plots spatially subset model evaluation.
This is effectively the main method.
:param evaluation_dir_name: See documentation at top of file.
:param smoothing_radius_grid_cells: Same.
:param score_colour_map_name: Same.
:param num_ex_colour_map_name: Same.
:param max_colour_percentile: Same.
:param output_dir_name: Same.
"""
if smoothing_radius_grid_cells <= 0:
smoothing_radius_grid_cells = None
score_colour_map_object = pyplot.get_cmap(score_colour_map_name)
num_ex_colour_map_object = pyplot.get_cmap(num_ex_colour_map_name)
error_checking.assert_is_geq(max_colour_percentile, 90.)
error_checking.assert_is_leq(max_colour_percentile, 100.)
grid_metafile_name = grids.find_equidistant_metafile(
directory_name=evaluation_dir_name, raise_error_if_missing=True)
print('Reading grid metadata from: "{0:s}"...'.format(grid_metafile_name))
grid_metadata_dict = grids.read_equidistant_metafile(grid_metafile_name)
print(SEPARATOR_STRING)
num_grid_rows = len(grid_metadata_dict[grids.Y_COORDS_KEY])
num_grid_columns = len(grid_metadata_dict[grids.X_COORDS_KEY])
auc_matrix = numpy.full((num_grid_rows, num_grid_columns), numpy.nan)
csi_matrix = numpy.full((num_grid_rows, num_grid_columns), numpy.nan)
pod_matrix = numpy.full((num_grid_rows, num_grid_columns), numpy.nan)
far_matrix = numpy.full((num_grid_rows, num_grid_columns), numpy.nan)
num_examples_matrix = numpy.full((num_grid_rows, num_grid_columns),
0,
dtype=int)
num_positive_examples_matrix = numpy.full(
(num_grid_rows, num_grid_columns), 0, dtype=int)
for i in range(num_grid_rows):
for j in range(num_grid_columns):
this_eval_file_name = model_eval.find_file(
directory_name=evaluation_dir_name,
grid_row=i,
grid_column=j,
raise_error_if_missing=False)
if not os.path.isfile(this_eval_file_name):
warning_string = (
'Cannot find file (this may or may not be a problem). '
'Expected at: "{0:s}"').format(this_eval_file_name)
warnings.warn(warning_string)
continue
print('Reading data from: "{0:s}"...'.format(this_eval_file_name))
this_evaluation_dict = model_eval.read_evaluation(
this_eval_file_name)
num_examples_matrix[i, j] = len(
this_evaluation_dict[model_eval.OBSERVED_LABELS_KEY])
num_positive_examples_matrix[i, j] = numpy.sum(
this_evaluation_dict[model_eval.OBSERVED_LABELS_KEY])
this_evaluation_table = this_evaluation_dict[
model_eval.EVALUATION_TABLE_KEY]
auc_matrix[i, j] = numpy.nanmean(
this_evaluation_table[model_eval.AUC_KEY].values)
csi_matrix[i, j] = numpy.nanmean(
this_evaluation_table[model_eval.CSI_KEY].values)
pod_matrix[i, j] = numpy.nanmean(
this_evaluation_table[model_eval.POD_KEY].values)
far_matrix[i, j] = 1. - numpy.nanmean(
this_evaluation_table[model_eval.SUCCESS_RATIO_KEY].values)
print(SEPARATOR_STRING)
auc_matrix[num_positive_examples_matrix == 0] = numpy.nan
csi_matrix[num_positive_examples_matrix == 0] = numpy.nan
pod_matrix[num_positive_examples_matrix == 0] = numpy.nan
far_matrix[num_positive_examples_matrix == 0] = numpy.nan
if smoothing_radius_grid_cells is not None:
print((
'Applying Gaussian smoother with e-folding radius of {0:.1f} grid '
'cells...').format(smoothing_radius_grid_cells))
orig_num_examples_matrix = num_examples_matrix + 0
num_examples_matrix = general_utils.apply_gaussian_filter(
input_matrix=num_examples_matrix.astype(float),
e_folding_radius_grid_cells=smoothing_radius_grid_cells)
num_examples_matrix = numpy.round(num_examples_matrix).astype(int)
num_examples_matrix[orig_num_examples_matrix == 0] = 0 # HACK
num_positive_examples_matrix = general_utils.apply_gaussian_filter(
input_matrix=num_positive_examples_matrix.astype(float),
e_folding_radius_grid_cells=smoothing_radius_grid_cells)
num_positive_examples_matrix = (
numpy.round(num_positive_examples_matrix).astype(int))
num_positive_examples_matrix[num_examples_matrix == 0] = 0
auc_matrix = general_utils.apply_gaussian_filter(
input_matrix=ge_utils.fill_nans(auc_matrix),
e_folding_radius_grid_cells=smoothing_radius_grid_cells)
csi_matrix = general_utils.apply_gaussian_filter(
input_matrix=ge_utils.fill_nans(csi_matrix),
e_folding_radius_grid_cells=smoothing_radius_grid_cells)
pod_matrix = general_utils.apply_gaussian_filter(
input_matrix=ge_utils.fill_nans(pod_matrix),
e_folding_radius_grid_cells=smoothing_radius_grid_cells)
far_matrix = general_utils.apply_gaussian_filter(
input_matrix=ge_utils.fill_nans(far_matrix),
e_folding_radius_grid_cells=smoothing_radius_grid_cells)
auc_matrix[num_positive_examples_matrix == 0] = numpy.nan
csi_matrix[num_positive_examples_matrix == 0] = numpy.nan
pod_matrix[num_positive_examples_matrix == 0] = numpy.nan
far_matrix[num_positive_examples_matrix == 0] = numpy.nan
panel_file_names = []
file_system_utils.mkdir_recursive_if_necessary(
directory_name=output_dir_name)
# Plot number of examples.
this_data_matrix = numpy.maximum(numpy.log10(num_examples_matrix), 0.)
this_data_matrix[this_data_matrix == 0] = numpy.nan
max_colour_value = numpy.nanpercentile(this_data_matrix,
max_colour_percentile)
figure_object, axes_object = _plot_one_value(
data_matrix=this_data_matrix,
grid_metadata_dict=grid_metadata_dict,
colour_map_object=num_ex_colour_map_object,
min_colour_value=0.,
max_colour_value=max_colour_value,
plot_cbar_min_arrow=False,
plot_cbar_max_arrow=True,
log_scale=True)
axes_object.set_title(r'Number of examples')
plotting_utils.label_axes(axes_object=axes_object, label_string='(a)')
panel_file_names.append('{0:s}/num_examples.jpg'.format(output_dir_name))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(panel_file_names[-1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot number of positive examples.
this_data_matrix = num_positive_examples_matrix.astype(float)
this_data_matrix[this_data_matrix == 0] = numpy.nan
max_colour_value = numpy.nanpercentile(this_data_matrix,
max_colour_percentile)
min_colour_value = numpy.nanpercentile(this_data_matrix,
100. - max_colour_percentile)
figure_object, axes_object = _plot_one_value(
data_matrix=this_data_matrix,
grid_metadata_dict=grid_metadata_dict,
colour_map_object=num_ex_colour_map_object,
min_colour_value=min_colour_value,
max_colour_value=max_colour_value,
plot_cbar_min_arrow=True,
plot_cbar_max_arrow=True)
axes_object.set_title('Number of tornadic examples')
plotting_utils.label_axes(axes_object=axes_object, label_string='(b)')
panel_file_names.append(
'{0:s}/num_positive_examples.jpg'.format(output_dir_name))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(panel_file_names[-1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot AUC.
max_colour_value = numpy.nanpercentile(auc_matrix, max_colour_percentile)
min_colour_value = numpy.maximum(
numpy.nanpercentile(auc_matrix, 100. - max_colour_percentile), 0.5)
figure_object, axes_object = _plot_one_value(
data_matrix=auc_matrix,
grid_metadata_dict=grid_metadata_dict,
colour_map_object=score_colour_map_object,
min_colour_value=min_colour_value,
max_colour_value=max_colour_value,
plot_cbar_min_arrow=True,
plot_cbar_max_arrow=max_colour_value < 1.)
axes_object.set_title('AUC (area under ROC curve)')
plotting_utils.label_axes(axes_object=axes_object, label_string='(c)')
panel_file_names.append('{0:s}/auc.jpg'.format(output_dir_name))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(panel_file_names[-1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot CSI.
max_colour_value = numpy.nanpercentile(csi_matrix, max_colour_percentile)
min_colour_value = numpy.nanpercentile(csi_matrix,
100. - max_colour_percentile)
figure_object, axes_object = _plot_one_value(
data_matrix=csi_matrix,
grid_metadata_dict=grid_metadata_dict,
colour_map_object=score_colour_map_object,
min_colour_value=min_colour_value,
max_colour_value=max_colour_value,
plot_cbar_min_arrow=min_colour_value > 0.,
plot_cbar_max_arrow=max_colour_value < 1.)
axes_object.set_title('CSI (critical success index)')
plotting_utils.label_axes(axes_object=axes_object, label_string='(d)')
panel_file_names.append('{0:s}/csi.jpg'.format(output_dir_name))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(panel_file_names[-1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot POD.
max_colour_value = numpy.nanpercentile(pod_matrix, max_colour_percentile)
min_colour_value = numpy.nanpercentile(pod_matrix,
100. - max_colour_percentile)
figure_object, axes_object = _plot_one_value(
data_matrix=pod_matrix,
grid_metadata_dict=grid_metadata_dict,
colour_map_object=score_colour_map_object,
min_colour_value=min_colour_value,
max_colour_value=max_colour_value,
plot_cbar_min_arrow=min_colour_value > 0.,
plot_cbar_max_arrow=max_colour_value < 1.)
axes_object.set_title('POD (probability of detection)')
plotting_utils.label_axes(axes_object=axes_object, label_string='(e)')
panel_file_names.append('{0:s}/pod.jpg'.format(output_dir_name))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(panel_file_names[-1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Plot FAR.
max_colour_value = numpy.nanpercentile(far_matrix, max_colour_percentile)
min_colour_value = numpy.nanpercentile(far_matrix,
100. - max_colour_percentile)
figure_object, axes_object = _plot_one_value(
data_matrix=far_matrix,
grid_metadata_dict=grid_metadata_dict,
colour_map_object=score_colour_map_object,
min_colour_value=min_colour_value,
max_colour_value=max_colour_value,
plot_cbar_min_arrow=min_colour_value > 0.,
plot_cbar_max_arrow=max_colour_value < 1.)
axes_object.set_title('FAR (false-alarm ratio)')
plotting_utils.label_axes(axes_object=axes_object, label_string='(f)')
panel_file_names.append('{0:s}/far.jpg'.format(output_dir_name))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(panel_file_names[-1],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_object)
# Concatenate panels.
concat_file_name = '{0:s}/spatially_subset_evaluation.jpg'.format(
output_dir_name)
print('Concatenating panels to: "{0:s}"...'.format(concat_file_name))
imagemagick_utils.concatenate_images(input_file_names=panel_file_names,
output_file_name=concat_file_name,
num_panel_rows=NUM_PANEL_ROWS,
num_panel_columns=NUM_PANEL_COLUMNS)
imagemagick_utils.resize_image(input_file_name=concat_file_name,
output_file_name=concat_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _run():
"""Makes event-attribution schematics for 2019 tornado-prediction paper.
This is effectively the main method.
"""
file_system_utils.mkdir_recursive_if_necessary(
directory_name=OUTPUT_DIR_NAME)
# Interpolation with merger.
figure_object, axes_object = _plot_interp_two_times(
storm_object_table=_get_data_for_interp_with_merger()[0],
tornado_table=_get_data_for_interp_with_merger()[1],
legend_font_size=SMALL_LEGEND_FONT_SIZE, legend_position_string='upper right'
)
axes_object.set_title('Interpolation with merger')
this_file_name = '{0:s}/interp_with_merger_standalone.jpg'.format(
OUTPUT_DIR_NAME)
print('Saving figure to: "{0:s}"...'.format(this_file_name))
figure_object.savefig(
this_file_name, dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
plotting_utils.label_axes(axes_object=axes_object, label_string='(a)')
panel_file_names = ['{0:s}/interp_with_merger.jpg'.format(OUTPUT_DIR_NAME)]
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(
panel_file_names[-1], dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
pyplot.close(figure_object)
# Interpolation with split.
figure_object, axes_object = _plot_interp_two_times(
storm_object_table=_get_data_for_interp_with_split()[0],
tornado_table=_get_data_for_interp_with_split()[1],
legend_font_size=DEFAULT_FONT_SIZE,
legend_position_string='upper left'
)
axes_object.set_title('Interpolation with split')
this_file_name = '{0:s}/interp_with_split_standalone.jpg'.format(
OUTPUT_DIR_NAME)
print('Saving figure to: "{0:s}"...'.format(this_file_name))
figure_object.savefig(
this_file_name, dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
plotting_utils.label_axes(axes_object=axes_object, label_string='(b)')
panel_file_names.append(
'{0:s}/interp_with_split.jpg'.format(OUTPUT_DIR_NAME)
)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(
panel_file_names[-1], dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
pyplot.close(figure_object)
# Simple successors.
figure_object, axes_object = _plot_attribution_one_track(
storm_object_table=_get_track_for_simple_succ(),
plot_legend=True, plot_x_ticks=True,
legend_font_size=SMALL_LEGEND_FONT_SIZE, legend_location='lower right'
)
this_file_name = '{0:s}/simple_successors_standalone.jpg'.format(
OUTPUT_DIR_NAME)
print('Saving figure to: "{0:s}"...'.format(this_file_name))
figure_object.savefig(
this_file_name, dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
plotting_utils.label_axes(axes_object=axes_object, label_string='(c)')
axes_object.set_title('Linking to simple successors')
panel_file_names.append(
'{0:s}/simple_successors.jpg'.format(OUTPUT_DIR_NAME)
)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(
panel_file_names[-1], dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
pyplot.close(figure_object)
# Simple predecessors, example 1.
figure_object, axes_object = _plot_attribution_one_track(
storm_object_table=_get_track1_for_simple_pred(),
plot_legend=True, plot_x_ticks=False,
legend_font_size=DEFAULT_FONT_SIZE, legend_location=(0.28, 0.1)
)
axes_object.set_title('Simple predecessors, example 1')
this_file_name = '{0:s}/simple_predecessors_track1_standalone.jpg'.format(
OUTPUT_DIR_NAME)
print('Saving figure to: "{0:s}"...'.format(this_file_name))
figure_object.savefig(
this_file_name, dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
plotting_utils.label_axes(axes_object=axes_object, label_string='(d)')
axes_object.set_title('Linking to simple predecessors, example 1')
panel_file_names.append(
'{0:s}/simple_predecessors_track1.jpg'.format(OUTPUT_DIR_NAME)
)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(
panel_file_names[-1], dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
pyplot.close(figure_object)
# Simple predecessors, example 2.
figure_object, axes_object = _plot_attribution_one_track(
storm_object_table=_get_track2_for_simple_pred(),
plot_legend=False, plot_x_ticks=False
)
axes_object.set_title('Simple predecessors, example 2')
this_file_name = '{0:s}/simple_predecessors_track2_standalone.jpg'.format(
OUTPUT_DIR_NAME)
print('Saving figure to: "{0:s}"...'.format(this_file_name))
figure_object.savefig(
this_file_name, dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
plotting_utils.label_axes(axes_object=axes_object, label_string='(e)')
axes_object.set_title('Linking to simple predecessors, example 2')
panel_file_names.append(
'{0:s}/simple_predecessors_track2.jpg'.format(OUTPUT_DIR_NAME)
)
print('Saving figure to: "{0:s}"...'.format(panel_file_names[-1]))
figure_object.savefig(
panel_file_names[-1], dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight'
)
pyplot.close(figure_object)
# Concatenate all panels into one figure.
concat_file_name = '{0:s}/attribution_schemas.jpg'.format(OUTPUT_DIR_NAME)
print('Concatenating panels to: "{0:s}"...'.format(concat_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names, output_file_name=concat_file_name,
num_panel_rows=2, num_panel_columns=3)
imagemagick_utils.resize_image(
input_file_name=concat_file_name, output_file_name=concat_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
def _run():
"""Plots results of CNN experiment.
This is effectively the main method.
:raises: ValueError: if any Unix command fails.
"""
num_predictor_combos = len(UNIQUE_PREDICTOR_COMBO_STRINGS)
num_image_sizes = len(UNIQUE_HALF_IMAGE_SIZES)
num_dropout_fractions = len(UNIQUE_DROPOUT_FRACTIONS)
gerrity_score_matrix = numpy.full(
(num_predictor_combos, num_image_sizes, num_dropout_fractions),
numpy.nan)
peirce_score_matrix = gerrity_score_matrix + 0.
hss_matrix = gerrity_score_matrix + 0.
accuracy_matrix = gerrity_score_matrix + 0.
for i in range(num_predictor_combos):
for j in range(num_image_sizes):
for k in range(num_dropout_fractions):
this_num_predictors = len(
UNIQUE_PREDICTOR_COMBO_STRINGS[i].split())
this_eval_file_name = (
'{0:s}/{1:s}_init-num-filters={2:d}_half-image-size-px='
'{3:d}_num-conv-layer-sets={4:d}_dropout={5:.2f}/validation'
'/model_evaluation.p'
).format(
TOP_EXPERIMENT_DIR_NAME,
UNIQUE_PREDICTOR_COMBO_STRINGS[i].replace(
'_', '-').replace(' ', '_'),
this_num_predictors * 8,
UNIQUE_HALF_IMAGE_SIZES[j],
2 + int(UNIQUE_HALF_IMAGE_SIZES[j] > 8),
UNIQUE_DROPOUT_FRACTIONS[k]
)
if not os.path.isfile(this_eval_file_name):
warning_string = (
'POTENTIAL PROBLEM. Cannot find file expected at: '
'"{0:s}"'
).format(this_eval_file_name)
warnings.warn(warning_string)
continue
print 'Reading data from: "{0:s}"...'.format(
this_eval_file_name)
this_evaluation_dict = eval_utils.read_evaluation_results(
this_eval_file_name)
gerrity_score_matrix[i, j, k] = this_evaluation_dict[
eval_utils.GERRITY_SCORE_KEY]
peirce_score_matrix[i, j, k] = this_evaluation_dict[
eval_utils.PEIRCE_SCORE_KEY]
hss_matrix[i, j, k] = this_evaluation_dict[
eval_utils.HEIDKE_SCORE_KEY]
accuracy_matrix[i, j, k] = this_evaluation_dict[
eval_utils.ACCURACY_KEY]
print SEPARATOR_STRING
panel_file_names = numpy.full((4, num_dropout_fractions), '', dtype=object)
for k in range(num_dropout_fractions):
if k == 0:
this_y_axis_text_colour = BLACK_COLOUR + 0.
else:
this_y_axis_text_colour = WHITE_COLOUR + 0.
this_title_string = 'Gerrity score; dropout = {0:.2f}'.format(
UNIQUE_DROPOUT_FRACTIONS[k])
panel_file_names[0, k] = (
'{0:s}/gerrity_score_dropout-fraction={1:.2f}.jpg'
).format(TOP_EXPERIMENT_DIR_NAME, UNIQUE_DROPOUT_FRACTIONS[k])
print gerrity_score_matrix[..., k]
_plot_scores_as_grid(
score_matrix=gerrity_score_matrix[..., k],
colour_map_object=COLOUR_MAP_OBJECT,
min_colour_value=numpy.nanpercentile(
gerrity_score_matrix, MIN_COLOUR_PERCENTILE),
max_colour_value=numpy.nanpercentile(
gerrity_score_matrix, MAX_COLOUR_PERCENTILE),
x_tick_labels=UNIQUE_IMAGE_SIZE_STRINGS,
x_axis_label=IMAGE_SIZE_AXIS_LABEL,
x_axis_text_colour=WHITE_COLOUR,
y_tick_labels=UNIQUE_PREDICTOR_ABBREV_STRINGS,
y_axis_label='',
y_axis_text_colour=this_y_axis_text_colour,
title_string=this_title_string,
output_file_name=panel_file_names[0, k])
this_title_string = 'Peirce score; dropout = {0:.2f}'.format(
UNIQUE_DROPOUT_FRACTIONS[k])
panel_file_names[1, k] = (
'{0:s}/peirce_score_dropout-fraction={1:.2f}.jpg'
).format(TOP_EXPERIMENT_DIR_NAME, UNIQUE_DROPOUT_FRACTIONS[k])
_plot_scores_as_grid(
score_matrix=peirce_score_matrix[..., k],
colour_map_object=COLOUR_MAP_OBJECT,
min_colour_value=numpy.nanpercentile(
peirce_score_matrix, MIN_COLOUR_PERCENTILE),
max_colour_value=numpy.nanpercentile(
peirce_score_matrix, MAX_COLOUR_PERCENTILE),
x_tick_labels=UNIQUE_IMAGE_SIZE_STRINGS,
x_axis_label=IMAGE_SIZE_AXIS_LABEL,
x_axis_text_colour=WHITE_COLOUR,
y_tick_labels=UNIQUE_PREDICTOR_ABBREV_STRINGS,
y_axis_label='',
y_axis_text_colour=this_y_axis_text_colour,
title_string=this_title_string,
output_file_name=panel_file_names[1, k])
this_title_string = 'Heidke skill score; dropout = {0:.2f}'.format(
UNIQUE_DROPOUT_FRACTIONS[k])
panel_file_names[2, k] = (
'{0:s}/hss_dropout-fraction={1:.2f}.jpg'
).format(TOP_EXPERIMENT_DIR_NAME, UNIQUE_DROPOUT_FRACTIONS[k])
_plot_scores_as_grid(
score_matrix=hss_matrix[..., k],
colour_map_object=COLOUR_MAP_OBJECT,
min_colour_value=numpy.nanpercentile(
hss_matrix, MIN_COLOUR_PERCENTILE),
max_colour_value=numpy.nanpercentile(
hss_matrix, MAX_COLOUR_PERCENTILE),
x_tick_labels=UNIQUE_IMAGE_SIZE_STRINGS,
x_axis_label=IMAGE_SIZE_AXIS_LABEL,
x_axis_text_colour=WHITE_COLOUR,
y_tick_labels=UNIQUE_PREDICTOR_ABBREV_STRINGS,
y_axis_label='',
y_axis_text_colour=this_y_axis_text_colour,
title_string=this_title_string,
output_file_name=panel_file_names[2, k])
this_title_string = 'Accuracy; dropout = {0:.2f}'.format(
UNIQUE_DROPOUT_FRACTIONS[k])
panel_file_names[3, k] = (
'{0:s}/accuracy_dropout-fraction={1:.2f}.jpg'
).format(TOP_EXPERIMENT_DIR_NAME, UNIQUE_DROPOUT_FRACTIONS[k])
_plot_scores_as_grid(
score_matrix=accuracy_matrix[..., k],
colour_map_object=COLOUR_MAP_OBJECT,
min_colour_value=numpy.nanpercentile(
accuracy_matrix, MIN_COLOUR_PERCENTILE),
max_colour_value=numpy.nanpercentile(
accuracy_matrix, MAX_COLOUR_PERCENTILE),
x_tick_labels=UNIQUE_IMAGE_SIZE_STRINGS,
x_axis_label=IMAGE_SIZE_AXIS_LABEL,
x_axis_text_colour=BLACK_COLOUR,
y_tick_labels=UNIQUE_PREDICTOR_ABBREV_STRINGS,
y_axis_label='',
y_axis_text_colour=this_y_axis_text_colour,
title_string=this_title_string,
output_file_name=panel_file_names[3, k])
for m in range(4):
second_image_object = Image.open(panel_file_names[m, 1])
command_string = '"{0:s}" "{1:s}" -resize {2:d}x{3:d}\! "{1:s}"'.format(
imagemagick_utils.DEFAULT_CONVERT_EXE_NAME, panel_file_names[m, 0],
second_image_object.size[0], second_image_object.size[1])
print 'Resizing image: "{0:s}"...'.format(panel_file_names[m, 0])
exit_code = os.system(command_string)
if exit_code != 0:
raise ValueError('\nUnix command failed (log messages shown '
'above should explain why).')
concat_file_name = '{0:s}/validation.jpg'.format(TOP_EXPERIMENT_DIR_NAME)
print 'Concatenating panels to: "{0:s}"...'.format(concat_file_name)
imagemagick_utils.concatenate_images(
input_file_names=numpy.ravel(panel_file_names).tolist(),
output_file_name=concat_file_name, num_panel_rows=4,
num_panel_columns=num_dropout_fractions)
imagemagick_utils.resize_image(
input_file_name=concat_file_name, output_file_name=concat_file_name,
output_size_pixels=FIGURE_SIZE_PIXELS)
def _run():
"""Plots example of double penalty.
This is effectively the main method.
"""
print 'Reading data from: "{0:s}"...'.format(INPUT_FILE_NAME)
actual_grid_point_table = fronts_io.read_narr_grids_from_file(
INPUT_FILE_NAME)
predicted_grid_point_table = copy.deepcopy(actual_grid_point_table)
predicted_grid_point_table[
front_utils.WARM_FRONT_COLUMN_INDICES_COLUMN] += 1
predicted_grid_point_table[
front_utils.COLD_FRONT_COLUMN_INDICES_COLUMN] += 1
predicted_grid_point_table[front_utils.WARM_FRONT_ROW_INDICES_COLUMN] += 1
predicted_grid_point_table[front_utils.COLD_FRONT_ROW_INDICES_COLUMN] += 1
num_grid_rows, num_grid_columns = nwp_model_utils.get_grid_dimensions(
model_name=nwp_model_utils.NARR_MODEL_NAME)
actual_binary_matrix = ml_utils.front_table_to_images(
frontal_grid_table=actual_grid_point_table,
num_rows_per_image=num_grid_rows,
num_columns_per_image=num_grid_columns)
actual_binary_matrix = ml_utils.binarize_front_images(actual_binary_matrix)
predicted_binary_matrix = ml_utils.front_table_to_images(
frontal_grid_table=predicted_grid_point_table,
num_rows_per_image=num_grid_rows,
num_columns_per_image=num_grid_columns)
predicted_binary_matrix = ml_utils.binarize_front_images(
predicted_binary_matrix)
_plot_fronts(actual_binary_matrix=actual_binary_matrix,
predicted_binary_matrix=predicted_binary_matrix,
title_string='Observed and predicted front\nwithout dilation',
annotation_string='(c)',
output_file_name=NO_DILATION_FILE_NAME)
actual_binary_matrix = ml_utils.dilate_binary_target_images(
target_matrix=actual_binary_matrix,
dilation_distance_metres=DILATION_DISTANCE_METRES,
verbose=False)
predicted_binary_matrix = ml_utils.dilate_binary_target_images(
target_matrix=predicted_binary_matrix,
dilation_distance_metres=DILATION_DISTANCE_METRES,
verbose=False)
_plot_fronts(actual_binary_matrix=actual_binary_matrix,
predicted_binary_matrix=predicted_binary_matrix,
title_string='Observed and predicted front\nwith dilation',
annotation_string='(d)',
output_file_name=WITH_DILATION_FILE_NAME)
print 'Concatenating figures to: "{0:s}"...'.format(CONCAT_FILE_NAME)
imagemagick_utils.concatenate_images(
input_file_names=[NO_DILATION_FILE_NAME, WITH_DILATION_FILE_NAME],
output_file_name=CONCAT_FILE_NAME,
num_panel_rows=1,
num_panel_columns=2)
imagemagick_utils.resize_image(input_file_name=CONCAT_FILE_NAME,
output_file_name=CONCAT_FILE_NAME,
output_size_pixels=CONCAT_SIZE_PIXELS)
def _plot_sounding_saliency(
saliency_matrix, colour_map_object, max_colour_value,
sounding_figure_object, sounding_axes_object,
sounding_pressures_pascals, saliency_dict, model_metadata_dict,
add_title, output_dir_name, pmm_flag, example_index=None):
"""Plots saliency for sounding.
H = number of sounding heights
F = number of sounding fields
:param saliency_matrix: H-by-F numpy array of saliency values.
:param colour_map_object: See documentation at top of file.
:param max_colour_value: Same.
:param sounding_figure_object: Figure handle (instance of
`matplotlib.figure.Figure`) for sounding itself.
:param sounding_axes_object: Axes handle (instance of
`matplotlib.axes._subplots.AxesSubplot`) for sounding itself.
:param sounding_pressures_pascals: length-H numpy array of sounding
pressures.
:param saliency_dict: Dictionary returned by `saliency_maps.read_file`.
:param model_metadata_dict: Dictionary returned by
`cnn.read_model_metadata`.
:param add_title: Boolean flag.
:param output_dir_name: Name of output directory. Figures will be saved
here.
:param pmm_flag: Boolean flag. If True, plotting PMM composite rather than
one example.
:param example_index: [used only if `pmm_flag == False`]
Plotting the [i]th example, where i = `example_index`.
"""
if max_colour_value is None:
max_colour_value = numpy.percentile(
numpy.absolute(saliency_matrix), MAX_COLOUR_PERCENTILE
)
training_option_dict = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY]
sounding_field_names = training_option_dict[trainval_io.SOUNDING_FIELDS_KEY]
if pmm_flag:
full_storm_id_string = None
storm_time_unix_sec = None
else:
full_storm_id_string = saliency_dict[saliency_maps.FULL_STORM_IDS_KEY][
example_index]
storm_time_unix_sec = saliency_dict[saliency_maps.STORM_TIMES_KEY][
example_index]
if add_title:
title_string = 'Max absolute saliency = {0:.2e}'.format(
max_colour_value)
sounding_axes_object.set_title(title_string)
left_panel_file_name = plot_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=False, pmm_flag=pmm_flag,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec,
radar_field_name='sounding-actual')
print('Saving figure to file: "{0:s}"...'.format(left_panel_file_name))
sounding_figure_object.savefig(
left_panel_file_name, dpi=FIGURE_RESOLUTION_DPI, pad_inches=0,
bbox_inches='tight')
pyplot.close(sounding_figure_object)
saliency_plotting.plot_saliency_for_sounding(
saliency_matrix=saliency_matrix,
sounding_field_names=sounding_field_names,
pressure_levels_mb=PASCALS_TO_MB * sounding_pressures_pascals,
colour_map_object=colour_map_object,
max_absolute_colour_value=max_colour_value)
right_panel_file_name = plot_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=False, pmm_flag=pmm_flag,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec,
radar_field_name='sounding-saliency')
print('Saving figure to file: "{0:s}"...'.format(right_panel_file_name))
pyplot.savefig(right_panel_file_name, dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0, bbox_inches='tight')
pyplot.close()
concat_file_name = plot_examples.metadata_to_file_name(
output_dir_name=output_dir_name, is_sounding=True, pmm_flag=pmm_flag,
full_storm_id_string=full_storm_id_string,
storm_time_unix_sec=storm_time_unix_sec)
print('Concatenating figures to: "{0:s}"...\n'.format(concat_file_name))
imagemagick_utils.concatenate_images(
input_file_names=[left_panel_file_name, right_panel_file_name],
output_file_name=concat_file_name,
num_panel_rows=1, num_panel_columns=2
)
imagemagick_utils.resize_image(
input_file_name=concat_file_name, output_file_name=concat_file_name,
output_size_pixels=SOUNDING_IMAGE_SIZE_PX)
os.remove(left_panel_file_name)
os.remove(right_panel_file_name)
def _plot_one_composite(saliency_file_name, composite_name_abbrev,
composite_name_verbose, colour_map_object,
max_colour_value, half_num_contours,
smoothing_radius_grid_cells, output_dir_name):
"""Plots saliency map for one composite.
:param saliency_file_name: Path to input file (will be read by
`saliency.read_file`).
:param composite_name_abbrev: Abbrev composite name (will be used in file
names).
:param composite_name_verbose: Verbose composite name (will be used in
figure title).
:param colour_map_object: See documentation at top of file.
:param max_colour_value: Same.
:param half_num_contours: Same.
:param smoothing_radius_grid_cells: Same.
:param output_dir_name: Name of output directory (figures will be saved
here).
:return: main_figure_file_name: Path to main image file created by this
method.
:return: max_colour_value: See input doc.
"""
mean_radar_matrix, mean_saliency_matrix, model_metadata_dict = (
_read_one_composite(
saliency_file_name=saliency_file_name,
smoothing_radius_grid_cells=smoothing_radius_grid_cells))
if numpy.isnan(max_colour_value):
max_colour_value = numpy.percentile(mean_saliency_matrix,
MAX_COLOUR_PERCENTILE)
training_option_dict = model_metadata_dict[cnn.TRAINING_OPTION_DICT_KEY]
field_names = training_option_dict[trainval_io.RADAR_FIELDS_KEY]
num_fields = mean_radar_matrix.shape[-1]
num_heights = mean_radar_matrix.shape[-2]
handle_dict = plot_examples.plot_one_example(
list_of_predictor_matrices=[mean_radar_matrix],
model_metadata_dict=model_metadata_dict,
pmm_flag=True,
allow_whitespace=True,
plot_panel_names=True,
panel_name_font_size=PANEL_NAME_FONT_SIZE,
add_titles=False,
label_colour_bars=True,
colour_bar_length=COLOUR_BAR_LENGTH,
colour_bar_font_size=COLOUR_BAR_FONT_SIZE,
num_panel_rows=num_heights)
figure_objects = handle_dict[plot_examples.RADAR_FIGURES_KEY]
axes_object_matrices = handle_dict[plot_examples.RADAR_AXES_KEY]
for k in range(num_fields):
this_saliency_matrix = mean_saliency_matrix[0, ..., k]
saliency_plotting.plot_many_2d_grids_with_contours(
saliency_matrix_3d=numpy.flip(this_saliency_matrix, axis=0),
axes_object_matrix=axes_object_matrices[k],
colour_map_object=colour_map_object,
max_absolute_contour_level=max_colour_value,
contour_interval=max_colour_value / half_num_contours)
panel_file_names = [None] * num_fields
for k in range(num_fields):
panel_file_names[k] = '{0:s}/{1:s}_{2:s}.jpg'.format(
output_dir_name, composite_name_abbrev,
field_names[k].replace('_', '-'))
print('Saving figure to: "{0:s}"...'.format(panel_file_names[k]))
figure_objects[k].savefig(panel_file_names[k],
dpi=FIGURE_RESOLUTION_DPI,
pad_inches=0,
bbox_inches='tight')
pyplot.close(figure_objects[k])
main_figure_file_name = '{0:s}/{1:s}_saliency.jpg'.format(
output_dir_name, composite_name_abbrev)
print('Concatenating panels to: "{0:s}"...'.format(main_figure_file_name))
imagemagick_utils.concatenate_images(
input_file_names=panel_file_names,
output_file_name=main_figure_file_name,
num_panel_rows=1,
num_panel_columns=num_fields,
border_width_pixels=50)
imagemagick_utils.resize_image(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
output_size_pixels=CONCAT_FIGURE_SIZE_PX)
imagemagick_utils.trim_whitespace(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
border_width_pixels=TITLE_FONT_SIZE + 25)
_overlay_text(image_file_name=main_figure_file_name,
x_offset_from_center_px=0,
y_offset_from_top_px=0,
text_string=composite_name_verbose)
imagemagick_utils.trim_whitespace(input_file_name=main_figure_file_name,
output_file_name=main_figure_file_name,
border_width_pixels=10)
return main_figure_file_name, max_colour_value
def _plot_sounding_saliency(
sounding_matrix, sounding_saliency_matrix, sounding_field_names,
saliency_metadata_dict, colour_map_object, max_colour_value_by_example,
output_dir_name):
"""Plots soundings along with their saliency maps.
E = number of examples
H = number of sounding heights
F = number of sounding fields
:param sounding_matrix: E-by-H-by-F numpy array of sounding values.
:param sounding_saliency_matrix: E-by-H-by-F numpy array of corresponding
saliency values.
:param sounding_field_names: length-F list of field names.
:param saliency_metadata_dict: Dictionary returned by
`saliency_maps.read_standard_file`.
:param colour_map_object: See doc for `_plot_2d3d_radar_saliency`.
:param max_colour_value_by_example: Same.
:param output_dir_name: Name of output directory (figures will be saved
here).
"""
# TODO(thunderhoser): Generalize this method to deal with
# probability-matched means.
num_examples = sounding_matrix.shape[0]
try:
metpy_dict_by_example = dl_utils.soundings_to_metpy_dictionaries(
sounding_matrix=sounding_matrix, field_names=sounding_field_names)
except:
sounding_pressure_matrix_pa = numpy.expand_dims(
saliency_metadata_dict[saliency_maps.SOUNDING_PRESSURES_KEY],
axis=-1)
this_sounding_matrix = numpy.concatenate(
(sounding_matrix, sounding_pressure_matrix_pa), axis=-1)
metpy_dict_by_example = dl_utils.soundings_to_metpy_dictionaries(
sounding_matrix=this_sounding_matrix,
field_names=sounding_field_names + [soundings.PRESSURE_NAME])
for i in range(num_examples):
this_storm_id = saliency_metadata_dict[
saliency_maps.STORM_IDS_KEY][i]
this_storm_time_string = time_conversion.unix_sec_to_string(
saliency_metadata_dict[saliency_maps.STORM_TIMES_KEY][i],
TIME_FORMAT)
this_title_string = 'Storm "{0:s}" at {1:s}'.format(
this_storm_id, this_storm_time_string)
sounding_plotting.plot_sounding(
sounding_dict_for_metpy=metpy_dict_by_example[i],
title_string=this_title_string)
this_left_file_name = (
'{0:s}/{1:s}_{2:s}_sounding-actual.jpg'
).format(
output_dir_name, this_storm_id.replace('_', '-'),
this_storm_time_string
)
print 'Saving figure to file: "{0:s}"...'.format(
this_left_file_name)
pyplot.savefig(this_left_file_name, dpi=FIGURE_RESOLUTION_DPI)
pyplot.close()
imagemagick_utils.trim_whitespace(
input_file_name=this_left_file_name,
output_file_name=this_left_file_name)
saliency_plotting.plot_saliency_for_sounding(
saliency_matrix=sounding_saliency_matrix[i, ...],
sounding_field_names=sounding_field_names,
pressure_levels_mb=metpy_dict_by_example[i][
soundings.PRESSURE_COLUMN_METPY],
colour_map_object=colour_map_object,
max_absolute_colour_value=max_colour_value_by_example[i])
this_right_file_name = (
'{0:s}/{1:s}_{2:s}_sounding-saliency.jpg'
).format(
output_dir_name, this_storm_id.replace('_', '-'),
this_storm_time_string
)
print 'Saving figure to file: "{0:s}"...'.format(
this_right_file_name)
pyplot.savefig(this_right_file_name, dpi=FIGURE_RESOLUTION_DPI)
pyplot.close()
imagemagick_utils.trim_whitespace(
input_file_name=this_right_file_name,
output_file_name=this_right_file_name)
this_file_name = (
'{0:s}/saliency_{1:s}_{2:s}_sounding.jpg'
).format(
output_dir_name, this_storm_id.replace('_', '-'),
this_storm_time_string
)
print 'Concatenating panels into file: "{0:s}"...\n'.format(
this_file_name)
imagemagick_utils.concatenate_images(
input_file_names=[this_left_file_name, this_right_file_name],
output_file_name=this_file_name, num_panel_rows=1,
num_panel_columns=2)
imagemagick_utils.resize_image(
input_file_name=this_file_name,
output_file_name=this_file_name,
output_size_pixels=SOUNDING_IMAGE_SIZE_PX)
os.remove(this_left_file_name)
os.remove(this_right_file_name)