def plot_tune_acc(results):
'''Plot the model's tuning set accuracy over time'''
metric_list = results['tune_acc']
metric_name = 'Tuning Set Accuracy'
plot = plot_metric(metric_name, metric_list)
plot = plot + ggtitle(
'{} vs epochs'.format(metric_name)) + make_baseline_geom(results)
return plot
def plot_metric(metric_name, metric_list):
'''This function is the generic plotting function for plotting a given metric
Arguments
---------
metric_name: string
The human understandable name of the metric being plotted
metric_list: list
A list containing the various values of the metric over the epochs
Returns
-------
plot: plotnine.Plot
A plot showing the change in the metric over the epochs
'''
data_df = create_data_df(metric_name, metric_list)
plot = ggplot(data_df, aes(x='epochs', y=metric_name)) + geom_line() +\
ggtitle('{} vs epochs'.format(metric_name))
return plot
def histogram(gray_img, mask=None, bins=256, color='red', title=None):
"""Plot a histogram using ggplot.
Inputs:
gray_img = grayscale image to analyze
mask = binary mask made from selected contours
bins = number of classes to divide spectrum into
color = color of the line drawn
title = custom title for the plot gets drawn if title is not None
:param gray_img: numpy.ndarray
:param mask: numpy.ndarray
:param bins: int
:param color: str
:param title: str
:return fig_hist: ggplot
"""
params.device += 1
debug = params.debug
# Apply mask if one is supplied
if mask is not None:
# apply plant shaped mask to image
params.debug = None
mask1 = binary_threshold(mask, 0, 255, 'light')
mask1 = (mask1 / 255)
masked = np.multiply(gray_img, mask1)
else:
masked = gray_img
params.debug = debug
if gray_img.dtype == 'uint16':
maxval = 65536
else:
maxval = 256
# Store histogram data
hist_gray_data, hist_bins = np.histogram(masked, bins, (1, maxval))
# make hist percentage for plotting
pixels = cv2.countNonZero(masked)
hist_percent = (hist_gray_data / float(pixels)) * 100
hist_x = hist_percent
bin_labels = np.arange(0, bins)
dataset = pd.DataFrame({
'Grayscale pixel intensity': bin_labels,
'Proportion of pixels (%)': hist_x
})
if title is None:
fig_hist = (ggplot(data=dataset,
mapping=aes(x='Grayscale pixel intensity',
y='Proportion of pixels (%)')) +
geom_line(color=color) +
scale_x_continuous(breaks=list(range(0, bins, 25))))
elif title is not None:
fig_hist = (ggplot(data=dataset,
mapping=aes(x='Grayscale pixel intensity',
y='Proportion of pixels (%)')) +
geom_line(color=color) +
scale_x_continuous(breaks=list(range(0, bins, 25))) +
labels.ggtitle(title))
if params.debug is not None:
if params.debug == "print":
fig_hist.save(
os.path.join(params.debug_outdir,
str(params.device) + '_hist.png'))
if params.debug == "plot":
print(fig_hist)
return fig_hist
df_hpi = df_hpi.assign(sidx=lambda x: x.groupby(cn_gg).idx.diff(1)).dropna().reset_index(None,True)
df_hpi = df_hpi.assign(year=lambda x: x.date.dt.year, sidx=lambda x: np.sign(x.sidx).astype(int))
df_hpi_sidx = df_hpi.pivot_table(index=cn_gg+['year'],columns='sidx',values='idx',aggfunc='count')
df_hpi_sidx = df_hpi_sidx.fillna(0).astype(int).reset_index().melt(cn_gg+['year'],None,None,'n')
df_hpi_sidx = df_hpi_sidx.assign(hpi=lambda x: x.hpi.map(di_hpi),tt=lambda x: x.tt.map(di_tt))
# Plot the number of negative months
tmp = df_hpi_sidx.query('sidx == -1 & city.isin(@cities)', engine='python').drop(columns='sidx')
gg_hpi_sidx = (ggplot(tmp, aes(x='year',y='n',color='tt')) +
geom_point() + geom_line() + theme_bw() +
scale_y_continuous(breaks=list(range(1,13,1))) +
scale_x_continuous(breaks=list(range(2005,2021,1))) +
facet_grid('city~hpi') +
theme(axis_title_x=element_blank(),axis_text_x=element_text(angle=90)) +
scale_color_discrete(name='Housing type') +
ggtitle('Index adjusted for seasonality') +
labs(x='Year',y='# of negative months'))
gg_save('gg_hpi_sidx.png', dir_figures, gg_hpi_sidx, 8, 8)
###########################################
# --- (2) QUADRANT STRATEGY (MONTHLY) --- #
# (i) Calculate for stock
mm_stock = pd.concat([df_other[cn_ticker],df_reit[cn_ticker]])
mm_stock = get_delta(mm_stock,'price','ticker', 1).dropna()
mm_stock = mm_stock.assign(sidx=lambda x: np.sign(x.mm).astype(int))
mm_stock = mm_stock.drop(columns=['mm','price'])#.rename(columns={'price':'idx'})
ticker12 = mm_stock[mm_stock.date>=dmin].ticker.value_counts().reset_index().query('ticker>=12')['index']
mm_stock = mm_stock[mm_stock.ticker.isin(ticker12)].reset_index(None, True)
# (ii) Get "any" negative change CREA/Teranet/Housing type
mm_hpi_s = df_hpi.drop(columns=['year','idx']).pivot_table('sidx',['date','city'],['tt','hpi'])
def plot_bar(data,nuclstr,column='value',factor=None,ymin=None,ymax=None,stat='identity',dpi=300,features=None,feature_types=['all'],add_features=[],funcgroups=None,shading_modes=['charge_functional'],usd=False,right_overhang_fix=None,debug=False,startnumber=1,cropseq=(0,None),aspect_ratio=None,reverse_seq=False,double_seq=False,transparent=True,fill_params=None,bar_position='stack',title=None):
"""
A wrapper function to make a plot of data with bars along the sequnce
input should be a dataframe with resid, segid column and 'value'
This one is inspired by seqplot/seqplot/pdb_plot.py
"""
segid=data['segid'].values[0]
if title is None:
title="Segid: %s, Type: %s"%(segid,nuclstr.components[segid]['type'])
seq=Seq(str(nuclstr.seqs[segid]['fullseq']),generic_protein \
if nuclstr.components[segid]['entity'] is 'DNA' or 'histone' or 'protein' else generic_dna)
msar=MultipleSeqAlignment([SeqRecord(seq=seq,id=nuclstr.components[segid]['type']+':'+segid,\
name=nuclstr.components[segid]['type']+':'+segid)])
if(reverse_seq):
logger.info("Experimental feature will reverse the sequence")
msar[0].seq=msar[0].seq[::-1]
if double_seq:
msar.add_sequence('reverse',str(msar[0].seq[::-1]))
msar=msar[:,cropseq[0]:cropseq[1]]
# print("Seq to plot:",msar)
#We need to get starting residue, currently for DNA chains only cifseq gets it correctly
resid_start=nuclstr.seqs[segid]['resid_start']
logger.debug("Starting resid",resid_start)
overhang=nuclstr.seqs[segid]['overhangL']
datafixed=data.copy()
datafixed.loc[:,'resid']=datafixed.loc[:,'resid']-resid_start+overhang+1-cropseq[0]
sl=len(msar[0].seq)
# fn=shade.seqfeat2shadefeat(msar,feature_types=feature_types,force_feature_pos='bottom',debug=debug)
if features is None:
fn=nuclstr.shading_features[segid]
else:
fn=features
fn2=[]
for i in fn:
if (i['style'] in feature_types) or ('all' in feature_types) :
fn2.append(i)
fn2.extend(add_features)
if usd:
ruler='top'
else:
ruler=None
shaded=ipyshade.shadedmsa4plot(msar,features=fn2,shading_modes=shading_modes,debug=debug,startnumber=startnumber,setends=[startnumber-2,sl+startnumber+2],funcgroups=funcgroups,ruler=ruler,density=200)
#If sl%10=10 se will have a ruler number hanging beyond the sequence image, and we need to correct for that.
if right_overhang_fix is None:
if sl%10==0:
if sl<100:
rof= 0.1
else:
rof=0.5
else:
rof=0
else:
rof=right_overhang_fix
if (not aspect_ratio is None ):
ar=aspect_ratio
else:
ar=0.2*100./sl
# print(datafixed)
plot=(ggplot(data=datafixed,mapping=aes(x='resid', y=column))
# + geom_point(size=0.1)
# +geom_bar(stat='identity',width=0.5,mapping=aes(fill=factor))
+ scale_x_continuous(limits=(0.5,sl+0.5+rof),expand=(0,0.2),name='',breaks=[])
# + scale_y_continuous(breaks=[0,0.5,1.0])
+ theme_light()+theme(aspect_ratio=ar,dpi=dpi,plot_margin=0,text=element_text(size=6), legend_key_size=5 ,legend_position='bottom',legend_direction='horizontal'))
#+ facet_wrap('~ segid',dir='v') +guides(color=guide_legend(ncol=10))
if factor is None:
plot=plot+geom_bar(stat=stat,width=0.5)
else:
plot=plot+geom_bar(stat=stat,width=0.5,mapping=aes(fill=factor),position=bar_position)
if fill_params is not None:
plot=plot+scale_fill_manual(**fill_params)
if not usd:
if (ymax is not None) :
plot=plot+scale_y_continuous(limits=(None,ymax))
else:
if (ymin is not None) :
plot=plot+scale_y_continuous(limits=(ymin,None))
if ymax is None:
ymax=data[column].max()
if ymin is None:
ymin=data[column].min()
# print(ymax)
plot = plot + geom_seq_x(seqimg=shaded.img,\
xlim=(1,sl+rof),ylim=(ymin,ymax),usd=usd,aspect_ratio=ar,transparent=transparent)+ggtitle(title)
return plot
def plot_line(data,nuclstr,columns=['value'],ymin=None,ymax=None,dpi=300,features=None,feature_types=['all'],add_features=[],funcgroups=None,shading_modes=['charge_functional'],right_overhang_fix=None,debug=False,startnumber=1,cropseq=(0,None),aspect_ratio=None,reverse_seq=False,transparent=True,xshift=0):
"""
A wrapper function to make a plot of data with bars along the sequnce
input should be a dataframe with resid, segid column and 'value'
This one is inspired by seqplot/seqplot/pdb_plot.py
funcgroup example fg="\\funcgroup{xxx}{CT}{White}{Green}{upper}{up} \\funcgroup{xxx}{GA}{White}{Blue}{upper}{up}"
"""
if isinstance(columns,str):
columns=[columns]
segid=data['segid'].values[0]
title="Segid: %s, Type: %s"%(segid,nuclstr.components[segid]['type'])
seq=Seq(str(nuclstr.seqs[segid]['fullseq']),generic_protein \
if nuclstr.components[segid]['entity'] is 'DNA' or 'histone' or 'protein' else generic_dna)
msar=MultipleSeqAlignment([SeqRecord(seq=seq,id=nuclstr.components[segid]['type']+':'+segid,\
name=nuclstr.components[segid]['type']+':'+segid)])
if(reverse_seq):
logger.info("Experimental feature will reverse the sequence")
msar[0].seq=msar[0].seq[::-1]
msar=msar[:,cropseq[0]:cropseq[1]]
# print("Seq to plot:",msar)
#We need to get starting residue, currently for DNA chains only cifseq gets it correctly
resid_start=nuclstr.seqs[segid]['resid_start']
logger.debug("Starting resid %d"%int(resid_start))
overhang=nuclstr.seqs[segid]['overhangL']
datafixed=data.copy()
datafixed.loc[:,'resid']=datafixed.loc[:,'resid']-resid_start+overhang+1-cropseq[0]+xshift
# print(datafixed)
sl=len(msar[0].seq)
# fn=shade.seqfeat2shadefeat(msar,feature_types=feature_types,force_feature_pos='bottom',debug=debug)
if features is None:
fn=nuclstr.shading_features[segid]
else:
fn=features
fn2=[]
for i in fn:
if (i['style'] in feature_types) or ('all' in feature_types) :
fn2.append(i)
fn2.extend(add_features)
shaded=ipyshade.shadedmsa4plot(msar,features=fn2,shading_modes=shading_modes,debug=debug,startnumber=startnumber,setends=[startnumber-2,sl+startnumber+2],funcgroups=funcgroups,density=200)
#If sl%10=10 se will have a ruler number hanging beyond the sequence image, and we need to correct for that.
if right_overhang_fix is None:
if sl%10==0:
if sl<100:
rof= 0.1
else:
rof=0.5
else:
rof=0
else:
rof=right_overhang_fix
if (not aspect_ratio is None ):
ar=aspect_ratio
else:
ar=0.15*100./sl
md=pd.melt(datafixed,id_vars=['segid','resid'],value_vars=columns)
# print(md)
# print(md)
# print(md['variable'])
plot=(ggplot(data=md,mapping=aes(x='resid', y='value'))
+ geom_point(aes(color='variable'),size=0.1)+geom_line(aes(color='variable'),stat='identity')
+ scale_x_continuous(limits=(0.5,sl+0.5+rof),expand=(0,0.2),name='',breaks=[])
# + scale_y_continuous()
+ theme_light()+theme(aspect_ratio=ar,dpi=dpi,plot_margin=0)) #+ facet_wrap('~ segid',dir='v')
if ymax is not None:
plot=plot+scale_y_continuous(limits=(None,ymax))
if ymin is None:
ymin=md['value'].min()
if ymax is None:
ymax=md['value'].max()
plot = plot + geom_seq_x(seqimg=shaded.img,\
xlim=(1,sl+rof),ylim=(ymin,ymax),aspect_ratio=ar,transparent=transparent)+ggtitle(title)
return plot
def histogram(img,
mask=None,
bins=100,
lower_bound=None,
upper_bound=None,
title=None,
hist_data=False):
"""Plot histograms of each input image channel
Inputs:
img = an RGB or grayscale image to analyze
mask = binary mask, calculate histogram from masked area only (default=None)
bins = divide the data into n evenly spaced bins (default=100)
lower_bound = the lower bound of the bins (x-axis min value) (default=None)
upper_bound = the upper bound of the bins (x-axis max value) (default=None)
title = a custom title for the plot (default=None)
hist_data = return the frequency distribution data if True (default=False)
Returns:
fig_hist = histogram figure
hist_df = dataframe with histogram data, with columns "pixel intensity" and "proportion of pixels (%)"
:param img: numpy.ndarray
:param mask: numpy.ndarray
:param bins: int
:param lower_bound: int
:param upper_bound: int
:param title: str
:param hist_data: bool
:return fig_hist: plotnine.ggplot.ggplot
:return hist_df: pandas.core.frame.DataFrame
"""
if not isinstance(img, np.ndarray):
fatal_error("Only image of type numpy.ndarray is supported input!")
if len(img.shape) < 2:
fatal_error("Input image should be at least a 2d array!")
if mask is not None:
masked = img[np.where(mask > 0)]
img_min, img_max = np.nanmin(masked), np.nanmax(masked)
else:
img_min, img_max = np.nanmin(img), np.nanmax(img)
# for lower / upper bound, if given, use the given value, otherwise, use the min / max of the image
lower_bound = lower_bound if lower_bound is not None else img_min
upper_bound = upper_bound if upper_bound is not None else img_max
if len(img.shape) > 2:
if img.shape[2] == 3:
b_names = ['blue', 'green', 'red']
else:
b_names = [str(i) for i in range(img.shape[2])]
if len(img.shape) == 2:
bin_labels, hist_percent, hist_ = _hist_gray(img,
bins=bins,
lower_bound=lower_bound,
upper_bound=upper_bound,
mask=mask)
hist_df = pd.DataFrame({
'pixel intensity':
bin_labels,
'proportion of pixels (%)':
hist_percent,
'hist_count':
hist_,
'color channel': ['0' for _ in range(len(hist_percent))]
})
else:
# Assumption: RGB image
# Initialize dataframe column arrays
px_int = np.array([])
prop = np.array([])
hist_count = np.array([])
channel = []
for (b, b_name) in enumerate(b_names):
bin_labels, hist_percent, hist_ = _hist_gray(
img[:, :, b],
bins=bins,
lower_bound=lower_bound,
upper_bound=upper_bound,
mask=mask)
# Append histogram data for each channel
px_int = np.append(px_int, bin_labels)
prop = np.append(prop, hist_percent)
hist_count = np.append(hist_count, hist_)
channel = channel + [b_name for _ in range(len(hist_percent))]
# Create dataframe
hist_df = pd.DataFrame({
'pixel intensity': px_int,
'proportion of pixels (%)': prop,
'hist_count': hist_count,
'color channel': channel
})
fig_hist = (ggplot(data=hist_df,
mapping=aes(x='pixel intensity',
y='proportion of pixels (%)',
color='color channel')) + geom_line())
if title is not None:
fig_hist = fig_hist + labels.ggtitle(title)
if len(img.shape) > 2 and img.shape[2] == 3:
fig_hist = fig_hist + scale_color_manual(['blue', 'green', 'red'])
# Plot or print the histogram
_debug(visual=fig_hist,
filename=os.path.join(params.debug_outdir,
str(params.device) + '_hist.png'))
if hist_data is True:
return fig_hist, hist_df
return fig_hist