def format_axes(
ax: plt.Axes,
ticklabels,
patch_size,
total_size,
block_size,
):
robust_models = [1, 2]
ax_mm_ticks = np.arange(0, total_size, block_size)
ax.set_xticks(ax_mm_ticks - 0.5)
ax.set_yticks(ax_mm_ticks - 0.5)
ax.set_xticklabels(ax_mm_ticks)
ax.set_yticklabels(ax_mm_ticks)
ax.grid(which='major', axis='both', lw=1, color='k', alpha=0.5, ls='-')
ax.set_xticks(ax_mm_ticks + block_size / 2, minor=True)
ax.set_yticks(ax_mm_ticks + block_size / 2, minor=True)
ax.set_xticklabels([])
ax.set_yticklabels([])
ax.set_xticklabels(ticklabels, minor=True, fontsize=12, weight='light')
ax.set_yticklabels(ticklabels, minor=True, fontsize=12, weight='light')
plt.setp(ax.get_yticklabels(minor=True),
rotation=90,
ha="center",
va="bottom",
rotation_mode="anchor")
# ax.set_title(f'Patch Size {patch_size}', fontsize=1, pad=10)
ax.set_xlabel('Models', fontsize=14)
ax.set_ylabel('Models', fontsize=14)
ax.xaxis.labelpad = 7
ax.yaxis.labelpad = 7
def full_extent(fig: plt.Figure, ax: plt.Axes, *extras, pad=0.01):
""" Get the full extent of an axes, including axes labels, tick labels, and
titles.
"""
# For text objects, we need to draw the figure first, otherwise the extents
# are undefined.
ax.figure.canvas.draw()
items = [ax, ax.title, ax.xaxis.label, ax.yaxis.label, *extras]
items += [*ax.get_xticklabels(), *ax.get_yticklabels()]
items += [e.xaxis.label for e in extras if hasattr(e, 'xaxis')]
items += [e.yaxis.label for e in extras if hasattr(e, 'yaxis')]
items += sum((e.get_xticklabels()
for e in extras if hasattr(e, 'get_xticklabels')), [])
items += sum((e.get_yticklabels()
for e in extras if hasattr(e, 'get_yticklabels')), [])
bbox = Bbox.union([
item.get_window_extent() for item in items
if hasattr(item, 'get_window_extent')
])
bbox = bbox.expanded(1.0 + pad, 1.0 + pad)
bbox = bbox.transformed(fig.dpi_scale_trans.inverted())
return bbox
def matrix(values: np.ndarray,
row_labels: Sequence[str] = None,
col_labels: Sequence[str] = None,
row_seps: Union[int, Collection[int]] = None,
col_seps: Union[int, Collection[int]] = None,
cmap="RdBu",
fontcolor_thresh=0.5,
norm: plt.Normalize = None,
text_len=4,
omit_leading_zero=False,
trailing_zeros=False,
grid=True,
angle_left=False,
cbar=True,
cbar_label: str = None,
ax: plt.Axes = None,
figsize: Tuple[int, int] = None,
cellsize=0.65,
title: str = None):
cmap = get_cmap(cmap)
# Create figure if necessary.
if ax is None:
if figsize is None:
# Note the extra width factor for the colorbar.
figsize = (cellsize * values.shape[1] * (1.2 if cbar else 1),
cellsize * values.shape[0])
ax = plt.figure(figsize=figsize).gca()
# Set title if applicable.
if title is not None:
ax.set_title(title)
if row_seps is not None:
values = np.insert(values, row_seps, np.nan, axis=0)
if row_labels is not None:
row_labels = np.insert(row_labels, row_seps, "")
if col_seps is not None:
values = np.insert(values, col_seps, np.nan, axis=1)
if col_labels is not None:
col_labels = np.insert(col_labels, col_seps, "")
# Plot the heatmap.
im = ax.matshow(values, cmap=cmap, norm=norm)
# Plot the text annotations showing each cell's value.
norm_values = im.norm(values)
for row, col in product(range(values.shape[0]), range(values.shape[1])):
val = values[row, col]
if not np.isnan(val):
# Find text color.
bg_color = cmap(norm_values[row, col])[:3]
luma = 0.299 * bg_color[0] + 0.587 * bg_color[
1] + 0.114 * bg_color[2]
color = "white" if luma < fontcolor_thresh else "black"
# Plot cell text.
annotation = _format_value(val, text_len, omit_leading_zero,
trailing_zeros)
ax.text(col,
row,
annotation,
ha="center",
va="center",
color=color)
# Add ticks and labels.
if col_labels is None:
ax.set_xticks([])
else:
col_labels = np.asarray(col_labels)
labeled_cols = np.where(col_labels)[0]
ax.set_xticks(labeled_cols)
ax.set_xticklabels(col_labels[labeled_cols])
if row_labels is None:
ax.set_yticks([])
else:
row_labels = np.asarray(row_labels)
labeled_rows = np.where(row_labels)[0]
ax.set_yticks(labeled_rows)
ax.set_yticklabels(row_labels[labeled_rows])
ax.tick_params(which="major", bottom=False)
plt.setp(ax.get_xticklabels(),
rotation=40,
ha="left",
rotation_mode="anchor")
# Turn off spines.
for edge, spine in ax.spines.items():
spine.set_visible(False)
# Rotate the left labels if applicable.
if angle_left:
plt.setp(ax.get_yticklabels(),
rotation=40,
ha="right",
rotation_mode="anchor")
# Create the white grid if applicable.
if grid:
# Extra ticks required to avoid glitch.
xticks = np.concatenate([[-0.56],
np.arange(values.shape[1] + 1) - 0.5,
[values.shape[1] - 0.44]])
yticks = np.concatenate([[-0.56],
np.arange(values.shape[0] + 1) - 0.5,
[values.shape[0] - 0.44]])
ax.set_xticks(xticks, minor=True)
ax.set_yticks(yticks, minor=True)
ax.grid(which="minor", color="w", linestyle='-', linewidth=3)
ax.tick_params(which="minor", bottom=False, top=False, left=False)
# Create the colorbar if applicable.
if cbar:
bar = ax.figure.colorbar(im, ax=ax)
bar.ax.set_ylabel(cbar_label)
fmt = bar.ax.yaxis.get_major_formatter()
if isinstance(fmt, FixedFormatter):
fmt.seq = [
_format_value(
eval(
re.sub(r"[a-z$\\{}]", "",
label.replace("times", "*").replace(
"^", "**"))), text_len, omit_leading_zero,
trailing_zeros) if label else "" for label in fmt.seq
]
def plot(self,
plot_these: List[np.ndarray],
ax: plt.Axes = None,
fout: str = None,
poly=True,
root: str = "",
title: str = None,
aspect: int = 3,
ticks: list = [10, 5],
grid: bool = False):
"""
Plot the lattice coordinates or lattice as an array
Arguments:
plot_these - will take up to 2 sets of lattice and crossover coordinates
ax - an axes
fout - saves the plot as a png with the name of `fout`
poly - True will plot the initial polygon vertices
root - directory to save file in, defaults to saving in current directory
title - this is the title of the plot
aspect - this is the aspect ratio of the x and y axes
"""
assert len(
plot_these) <= 4, "Max no. of plots on one axis reached, 4 or less"
if not ax:
fig, ax = plt.subplots()
if grid:
plt.grid(True)
for label in ax.get_xticklabels() + ax.get_yticklabels():
label.set_fontsize(4)
ax.xaxis.set_major_locator(MultipleLocator(ticks[0]))
ax.yaxis.set_major_locator(MultipleLocator(ticks[1]))
ax.set_xlabel("No. of nucleotides")
ax.set_ylabel("No. of strands")
point_style = itertools.cycle(["ko", "b.", "r.", "cP"])
point_size = itertools.cycle([0.5, 2.5])
for points in plot_these:
if np.shape(points)[1] not in [2, 3]: # if array
nodes = self.array_to_coords(points)
else:
nodes = points
# Lattice sites then crossover sites
ax.plot(nodes[:, 0],
nodes[:, 1],
next(point_style),
ms=next(point_size),
alpha=0.25)
if poly:
self.plotPolygon(ax, plot_these[0], coords=True)
if title:
ax.set_title(f"{title}")
plt.gca().set_aspect(aspect)
if fout:
plt.savefig(f"{root}{fout}.png", dpi=500)
if not ax:
plt.show()
def plot_f1_metrics(run_histories: List[Run],
experiment_name: str,
metric_names: Dict[str, str],
target_file_paths: Optional[List[str]] = None,
axis: plt.Axes = None,
y_lims: Tuple[float, float] = None,
mode='box',
add_legend=True,
color=None):
standalone_mode = axis is None
if color is None:
# by default we use bright orange from map 'tab20c'
color = plt.get_cmap('tab20c')(4)
runs_data = []
for run_history in run_histories:
metrics_data = []
for metric_name in metric_names.keys():
metric_data = []
for fold_history in run_history.fold_histories:
# add the metric for the last epoch
metric_data.append(fold_history.epochs[-1].metrics[metric_name])
metric_data = np.array(metric_data)
if mode == 'bar':
metric_data = np.mean(metric_data)
metrics_data.append(metric_data)
runs_data.append(metrics_data)
with plt.style.context('seaborn'):
if axis is None:
fig: plt.Figure = plt.figure()
fig.suptitle(experiment_name, fontsize=24)
axis = fig.add_subplot(111)
else:
axis.set_title(experiment_name, fontsize=22)
axis.set_ylabel('Metric Value', fontsize=22)
if y_lims is not None:
print('limits', y_lims)
axis.set_ylim(*y_lims)
num_metrics = None
num_runs = len(runs_data)
for i, metrics_data in enumerate(runs_data):
num_metrics = len(metrics_data)
xs = range(i * len(metrics_data) + i, (i + 1) * len(metrics_data) + i)
max_v = .9
min_v = .6
colors = []
for idx in range(num_metrics):
if num_metrics > 1:
norm = idx * (max_v - min_v) / (num_metrics - 1)
else:
norm = 0
fill_color = list(colorsys.rgb_to_hls(*mc.to_rgb(color)))
fill_color[1] = min_v + norm
colors.append((
color,
colorsys.hls_to_rgb(*fill_color)
))
line_styles = ['-', '-.', ':', '--']
if mode == 'box':
boxplots = axis.boxplot(
metrics_data,
meanline=True,
showmeans=True,
positions=xs,
widths=0.6,
patch_artist=True
)
for plot_idx in range(num_metrics):
dark_color = colors[plot_idx][0]
light_color = colors[plot_idx][1]
plt.setp(boxplots['boxes'][plot_idx], color=dark_color)
plt.setp(boxplots['boxes'][plot_idx], facecolor=light_color)
plt.setp(boxplots['boxes'][plot_idx], linestyle=line_styles[plot_idx])
plt.setp(boxplots['whiskers'][plot_idx * 2], color=dark_color)
plt.setp(boxplots['whiskers'][plot_idx * 2 + 1], color=dark_color)
plt.setp(boxplots['whiskers'][plot_idx * 2], linestyle=line_styles[plot_idx])
plt.setp(boxplots['whiskers'][plot_idx * 2 + 1], linestyle=line_styles[plot_idx])
plt.setp(boxplots['caps'][plot_idx * 2], color=dark_color)
plt.setp(boxplots['caps'][plot_idx * 2 + 1], color=dark_color)
plt.setp(boxplots['fliers'][plot_idx], markeredgecolor=dark_color)
plt.setp(boxplots['fliers'][plot_idx], marker='x')
plt.setp(boxplots['medians'][plot_idx], color=dark_color)
plt.setp(boxplots['means'][plot_idx], color=dark_color)
legend_styles = [boxplots['boxes'][idx] for idx in range(num_metrics)]
elif mode == 'bar':
legend_styles = []
for plot_idx in range(num_metrics):
ret = axis.bar(xs[plot_idx], metrics_data[plot_idx],
color=colors[plot_idx][1],
edgecolor=colors[plot_idx][0],
width=0.6,
linewidth=1.25,
linestyle=line_styles[plot_idx], )
legend_styles.append(ret)
tick_offset = num_metrics * 0.5 - 0.5
ticks = np.arange(start=tick_offset, stop=num_runs * num_metrics + num_runs + tick_offset,
step=num_metrics + 1.0)
axis.set_xticks(ticks)
for yticklabel in axis.get_yticklabels():
yticklabel.set_fontsize(20)
axis.set_xticklabels([r.name for r in run_histories], fontsize=20, rotation=0)
if add_legend:
axis.legend(legend_styles, metric_names.values(),
loc='lower right', fontsize=16,
facecolor="white", frameon=True,
edgecolor="black")
if standalone_mode:
fig.show()
if target_file_paths is not None:
for target_file_path in target_file_paths:
fig.savefig(target_file_path)
return legend_styles