def construct_ax(figsize=(10, 10), despine=True):
"""
Construct a new pyplot fig, ax object pair and despine.
Parameters
----------
figsize : tuple of floats
figure size for pyplot.plot.subplots call
despine : bool (default True)
whether or not to remove axes and spines
Returns
-------
fig,ax : tuple
fig, ax pair generated by pyplot.subplots
"""
# Check sanity of figsize
try:
if len(figsize) != 2:
raise TypeError
for v in figsize:
check.float_sanity(v, min_allowed=0)
except (TypeError, ValueError):
err = "figsize should be a length 2 tuple of floats. See documentation\n"
err += "on matplotlib.pyplot.subplots"
raise ValueError(err)
fig, ax = plt.subplots(figsize=figsize)
if despine:
ax = despine_ax(ax)
return fig, ax
def _parse_add_sizemap(data_column, vmin, vmax, size_min, size_max, df,
df_name):
"""
Do standard checking and format selection for adding sizemap, either for
nodes or edges.
Parameters
----------
data_column : str
name of data column
vmin, vmax : float, float
minimum and maximum data values
size_min, size_max, float, float
minimum and maximum size values
df : pandas DataFrame
data frame to look for column in
df_name : str
name of dataframe for errors
Returns
-------
"_gpm" : str
indicates this should be interpreted in a special way by gpmap
data_column : str
vmin, vmax : float, float
size_min, size_max : float, float
"""
# Make sure the data column exists
try:
v = df.loc[:, data_column]
except KeyError:
err = f"GenotypePhenotypeMap {df_name} dataframe does not have column\n"
err += f"'{data_column}'\n"
raise KeyError(err)
# Deal with minimum and maximum for color map
if vmin is None:
vmin = np.nanmin(v)
if vmax is None:
vmax = np.nanmax(v)
# Check sanity of inputs
check.float_sanity(vmin)
check.float_sanity(vmax)
check.float_sanity(size_min)
check.float_sanity(size_max)
# Make sure vmin <= than vmax
if vmin > vmax:
err = f"vmax ({vmax}) must be >= than vmin ({vmin})\n"
raise ValueError(err)
# make sure size_min is <= size_max
if size_min > size_max:
err = f"size_max ({size_max}) must be >= size_min ({size_min})\n"
raise ValueError(err)
# make sure size_min and size_max are both >= 0
if size_min < 0 or size_max < 0:
err = f"size_max ({size_max}) and size_min ({size_min}) must be >= 0\n"
raise ValueError(err)
return "_gpm", data_column, vmin, vmax, size_min, size_max
def test_float_sanity():
"""
Test float sanity checker.
"""
with pytest.raises(ValueError):
check.int_sanity("stupid")
with pytest.raises(ValueError):
check.float_sanity([])
# should not throw error
check.float_sanity(11234521)
check.float_sanity(1.0)
check.float_sanity(-1)
with pytest.raises(ValueError):
check.float_sanity(0,min_allowed=0)
check.float_sanity(100,min_allowed=0)
with pytest.raises(ValueError):
check.float_sanity(10,max_allowed=0)
check.float_sanity(10,max_allowed=100)
def _parse_add_cmap(data_column, cmap, vmin, vmax, df, df_name):
"""
Do standard checking and format selection for adding sizemap, either for
nodes or edges.
Parameters
----------
data_column : str
name of data column
cmap : str or matplotlib.colors.Colormap
color map to apply
vmin, vmax : float, float
minimum and maximum data values
df : pandas DataFrame
data frame to look for column in
df_name : str
name of dataframe for errors
Returns
-------
"_gpm" : str
indicates this should be interpreted in a special way by gpmap
data_column : str
cmap : matplotlib.colors.Colormap
vmin, vmax : float, float
"""
# Make sure the data column exists
try:
v = df.loc[:, data_column]
except KeyError:
err = f"GenotypePhenotypeMap {df_name} dataframe does not have column\n"
err += f"'{data_column}'\n"
raise KeyError(err)
# Make sure the cmap is recognized
if not isinstance(cmap, matplotlib.colors.Colormap):
try:
cmap = matplotlib.cm.get_cmap(cmap)
except ValueError as msg:
col = str(msg).split(",")
err = f"cmap '{cmap}' not recognized. This should either be a\n"
err += "matplotlib Colormap object or the name of one of the\n"
err += "colormaps in matplotlib. Available colormaps on your\n"
err += "system are:\n"
col[0] = col[0].split(" ")[-1]
for c in col:
err += f" {c.strip()}\n"
err += "\n"
raise ValueError(err)
# Deal with minimum and maximum for color map
if vmin is None:
vmin = np.nanmin(v)
if vmax is None:
vmax = np.nanmax(v)
check.float_sanity(vmin)
check.float_sanity(vmax)
# Make sure vmin <= than vmax
if vmin > vmax:
err = f"vmax ({vmax}) must be >= than vmin ({vmin})\n"
raise ValueError(err)
return "_gpm", data_column, cmap, vmin, vmax
def flattened(G, scale=1, vertical=False):
"""
Get flattened positions for a genotype-phenotype graph.
Parameters
----------
G : GenotypePhenotypeGraph object
A genotype-phenotype objects
scale : float (default=1)
density of the nodes. Must be > 0.
vertical : bool (default=False)
position nodes top-to-bottom (vertical) rather than left-to-right
Returns
-------
positions: dict
positions of all nodes in network (i.e. {index: [x,y]})
"""
# Make sure this looks like a GenotypePhenotypeGraph that has a genotype
# phenotype map loaded.
try:
if G.gpm is None:
raise AttributeError
except AttributeError:
err = "G must be a GenotypePhenotypeGraph object with a loaded \n"
err += "genotype map. (see the add_gpm method)."
raise ValueError(err)
# Make sure the scale is sane
check.float_sanity(scale, min_allowed=0.0)
# Get the binary genotypes from the GenotypePhenotypeGraph
# Set level of nodes and begin calc offset on the fly
graph = G
offsets = {}
positions = {}
for n in range(len(list(G.nodes()))):
node = graph.nodes[n]
# Calculate the level of each node
level = node["binary"].count("1")
if level in offsets:
offsets[level] += 1
else:
offsets[level] = 1
positions[n] = [level]
# Center the offsets on 0
for key, val in offsets.items():
offsets[key] = list(np.arange(val) - (val - 1) / 2.0)
# Offset positions
if vertical:
for n in range(len(list(G.nodes()))):
pos = offsets[positions[n][0]].pop(0)
scaled = scale * pos
positions[n].insert(0, scaled)
positions[n][-1] *= -1
else:
for n in range(len(list(G.nodes()))):
pos = offsets[positions[n][0]].pop(0)
scaled = scale * pos
positions[n].append(scaled)
return positions