def draw_venn_RQ3(issues_dict, target):
imp.reload(venn)
leven_text = "LEV ({})".format(len(issues_dict["LEVEN_Norm"]))
dep_text = "DEP ({})".format(len(issues_dict["DEP_Norm"]))
bleu_text = "BLEU ({})".format(len(issues_dict["BLEU_Norm"]))
sbert_text = "SBERT ({})".format(len(issues_dict["SBERT_Norm"]))
dfa_text = "DFA ({})".format(len(issues_dict["DFA_Norm"]))
reg_text = "REG ({})".format(len(issues_dict["REG_Norm"]))
hyb_text = "HYB ({})".format(len(issues_dict["HYB_Norm"]))
replace = {"LEVEN_Norm":leven_text,
"DEP_Norm": dep_text,
"BLEU_Norm": bleu_text,
"SBERT_Norm": sbert_text,
"DFA_Norm": dfa_text,
"REG_Norm": reg_text,
"HYB_Norm": hyb_text}
new_issues_dict = {replace[k]: set(map(lambda x: str(x), v)) for k, v in issues_dict.items()}
cmap = matplotlib.cm.get_cmap("viridis_r", 5)
color_list = []
for i in range(cmap.N):
rgb = cmap(i)[:3]
color_list.append(matplotlib.colors.rgb2hex(rgb))
color_list[0] = "#E68C8D"
if target == "DFA":
_ = new_issues_dict.pop(reg_text)
_ = new_issues_dict.pop(hyb_text)
labels = venn.get_labels([new_issues_dict[leven_text], new_issues_dict[dep_text], new_issues_dict[bleu_text], \
new_issues_dict[sbert_text], new_issues_dict[dfa_text]], fill=['number'])
fig, ax = venn.venn5(labels, names=[leven_text, dep_text, bleu_text, sbert_text, dfa_text]
, colors=color_list
)
elif target == "REG":
_ = new_issues_dict.pop(dfa_text)
_ = new_issues_dict.pop(hyb_text)
labels = venn.get_labels([new_issues_dict[leven_text], new_issues_dict[dep_text], new_issues_dict[bleu_text], \
new_issues_dict[sbert_text], new_issues_dict[reg_text]], fill=['number'])
fig, ax = venn.venn5(labels, names=[leven_text, dep_text, bleu_text, sbert_text, reg_text]
, colors=color_list
)
elif target == "HYB":
_ = new_issues_dict.pop(dfa_text)
_ = new_issues_dict.pop(reg_text)
labels = venn.get_labels([new_issues_dict[leven_text], new_issues_dict[dep_text], new_issues_dict[bleu_text], \
new_issues_dict[sbert_text], new_issues_dict[hyb_text]], fill=['number'])
fig, ax = venn.venn5(labels, names=[leven_text, dep_text, bleu_text, sbert_text, hyb_text]
, colors=color_list
)
def venn(inp, names, title="venn", folder=''):
"""
Plots a venn diagram using the pyvenn package
Args:
-----
inp: list[set()] of sets of values (e.g. [(1,2,3,4),(2,3),(1,3,4,5)])
names: list[str] of the name of each leaf
title: str the plot title
folder: str of location where to save the plot, won't save if empty
"""
labels = pyvenn.get_labels(inp, fill=['number', 'logic'])
if len(inp) == 2:
fig, ax = pyvenn.venn2(labels, names=names)
elif len(inp) == 3:
fig, ax = pyvenn.venn3(labels, names=names)
elif len(inp) == 4:
fig, ax = pyvenn.venn4(labels, names=names)
elif len(inp) == 5:
fig, ax = pyvenn.venn5(labels, names=names)
elif len(inp) == 6:
fig, ax = pyvenn.venn6(labels, names=names)
else:
raise ValueError('need to be between 2 to 6')
ax.set_title(title)
if folder:
fig.savefig(folder + title + '_venn.pdf')
fig.show()
plt.pause(0.1)
def generate_venn_diagram(nb_dim, files, names, sep, label):
"""Cette fonction fait l'appel du fichier venn.py (qui doit être présent dans le même dossier que generate_venn.py)
pour générer les diagrammes de Venn.
:param: nombre de dimensions présent dans le diagramme de Venn (1 < nb_dim <= 6)
:param: files: liste des chemins d'accès aux fichiers contenant les résultats des modèles
:param: names: liste des noms des modèles (doit être en adéquation avec la liste des modèles files)
:param: sep: séparateur présent dans filename séparant les id-instance des label
:label: label que l'on souhaite extraire
"""
if nb_dim < 2 or nb_dim > 6:
print(
'Le diagramme de Venn ne peut être généré que pour un nombre de dimensions supérieur à un et inférieur ou égal à 6.'
)
else:
i = 0
models = []
while i < nb_dim:
models.append(
get_id_instances_title(files[i], sep=sep, label=label))
i += 1
labels = venn.get_labels(models)
if nb_dim == 2:
fig, ax = venn.venn2(labels, names=names)
plt.show()
elif nb_dim == 3:
fig, ax = venn.venn3(labels, names=names)
plt.show()
elif nb_dim == 4:
fig, ax = venn.venn4(labels, names=names)
plt.show()
elif nb_dim == 5:
fig, ax = venn.venn5(labels, names=names)
plt.show()
elif nb_dim == 6:
fig, ax = venn.venn6(labels, names=names)
plt.show()
else:
print(
'Le diagramme de Venn ne peut être généré que pour un nombre de dimensions supérieur à un et inférieur ou égal à 6.'
)
def uniq_share_cal(self, infile):
for line in infile:
s = line.rstrip()
orths = list(s.split('\t'))
ids = []
for x in xrange(len(orths)):
item = orths[x].split('|')[0]
if item in self.ids:
ids.append(item)
target = set(ids)
for key in self.data:
if set(key) == target:
self.data[key] += 1
labels = {}
for key in self.data:
key2 = ''
for x in xrange(len(self.ids)):
if self.ids[x] in key:
key2 += '1'
else:
key2 += '0'
labels[key2] = self.data[key]
if len(self.ids) == 3:
fig, ax = venn.venn3(labels, names=self.ids)
fig.savefig('gene_family_3taxa.pdf')
elif len(self.ids) == 4:
fig, ax = venn.venn4(labels, names=self.ids)
fig.savefig('gene_family_4taxa.pdf')
elif len(self.ids) == 5:
fig, ax = venn.venn5(labels, names=self.ids)
fig.savefig('gene_family_5taxa.pdf')
elif len(self.ids) == 6:
fig, ax = venn.venn6(labels, names=self.ids)
fig.savefig('gene_family_6taxa.pdf')
n34 = len(intersection(tp[ids[2]], tp[ids[3]]))
n35 = len(intersection(tp[ids[2]], tp[ids[4]]))
n45 = len(intersection(tp[ids[3]], tp[ids[4]]))
area5 = len(tp[ids[4]])
area4 = len(tp[ids[3]])
area3 = len(tp[ids[2]])
area2 = len(tp[ids[1]])
area1 = len(tp[ids[0]])
venn_diag_list = [area1, area2, area3, area4, area5, n12, n13, n14, n15,n23, n24, n25,
n34, n35, n45, n123, n124, n125, n134, n135, n145, n234, n235, n245, n345, n1234,
n1235, n1245, n1345, n2345, n12345]
print('Overlap list needed for VennDiagram in R')
print(venn_diag_list)
out1.write('\n\n## Overlap list needed for VennDiagram in R:\n\n%s' % venn_diag_list)
if plot.lower() == 't' or plot.lower() == 'true':
if n_comparing == 5:
labels = venn.get_labels([tp[ids[0]], tp[ids[1]], tp[ids[2]], tp[ids[3]], tp[ids[4]]], fill = ['number'])
fig, ax = venn.venn5(labels, names = ids)
elif n_comparing == 4:
labels = venn.get_labels([tp[ids[0]], tp[ids[1]], tp[ids[2]], tp[ids[3]]], fill = ['number'])
fig, ax = venn.venn4(labels, names = ids)
elif n_comparing == 3:
labels = venn.get_labels([tp[ids[0]], tp[ids[1]], tp[ids[2]]], fill = ['number'])
fig, ax = venn.venn3(labels, names = ids)
elif n_comparing == 2:
labels = venn.get_labels([tp[ids[0]], tp[ids[1]]], fill = ['number'])
fig, ax = venn.venn2(labels, names = ids)
filename = save+'_pred_compared.pdf'
fig.savefig(filename)
labels = venn.get_labels(
[range(10), range(5, 15),
range(3, 8), range(8, 17)],
fill=['number', 'logic'])
fig, ax = venn.venn4(labels, names=['list 1', 'list 2', 'list 3', 'list 4'])
fig.savefig('venn4.png', bbox_inches='tight')
plt.close()
labels = venn.get_labels(
[range(10),
range(5, 15),
range(3, 8),
range(8, 17),
range(10, 20)],
fill=['number', 'logic'])
fig, ax = venn.venn5(labels,
names=['list 1', 'list 2', 'list 3', 'list 4', 'list 5'])
fig.savefig('venn5.png', bbox_inches='tight')
plt.close()
labels = venn.get_labels([
range(10),
range(5, 15),
range(3, 8),
range(8, 17),
range(10, 20),
range(13, 25)
],
fill=['number', 'logic'])
fig, ax = venn.venn6(
labels, names=['list 1', 'list 2', 'list 3', 'list 4', 'list 5', 'list 6'])
fig.savefig('venn6.png', bbox_inches='tight')
'Human': "Vertebrate",
'Mouse': "Vertebrate",
'Worm': "Non-vertebrate",
'Fly': "Non-vertebrate",
'Yeast': "Non-vertebrate",
}
targets = defaultdict(list)
targets_v = defaultdict(list)
for index, row in species_genes.iterrows():
if row[0] == "Rat":
continue
targets[row[0]].append(row[1])
targets_v[is_vertebrate[row[0]]].append(row[1])
list_sp = []
list_targets = []
for sp, t in targets.items():
list_sp.append(sp)
list_targets.append(set(t))
labels = venn.get_labels(list_targets, fill=['number'])
fig1, ax = venn.venn5(labels, names=list_sp)
fig1.savefig('species.svg', bbox_inches='tight')
plt.figure(figsize=(4, 4))
venn2([set(targets_v["Vertebrate"]),
set(targets_v["Non-vertebrate"])],
set_labels=("Vertebrate", "Non-vertebrate"))
plt.savefig('species_vertebrate_venn.svg')
n25 = len(intersection(comp[ids[1]], comp[ids[4]]))
n34 = len(intersection(comp[ids[2]], comp[ids[3]]))
n35 = len(intersection(comp[ids[2]], comp[ids[4]]))
n45 = len(intersection(comp[ids[3]], comp[ids[4]]))
area5 = len(comp[ids[4]])
area4 = len(comp[ids[3]])
area3 = len(comp[ids[2]])
area2 = len(comp[ids[1]])
area1 = len(comp[ids[0]])
venn_diag_list = [area1, area2, area3, area4, area5, n12, n13, n14, n15,n23, n24, n25,
n34, n35, n45, n123, n124, n125, n134, n135, n145, n234, n235, n245, n345, n1234,
n1235, n1245, n1345, n2345, n12345]
print('Overlap list needed for VennDiagram in R')
print(venn_diag_list)
#out1.write('\n\n## Overlap list needed for VennDiagram in R:\n\n%s' % venn_diag_list)
if n_comparing == 5:
labels = venn.get_labels([comp[ids[0]], comp[ids[1]], comp[ids[2]], comp[ids[3]], comp[ids[4]]], fill = ['number'])
fig, ax = venn.venn5(labels, names = ids)
elif n_comparing == 4:
labels = venn.get_labels([comp[ids[0]], comp[ids[1]], comp[ids[2]], comp[ids[3]]], fill = ['number'])
fig, ax = venn.venn4(labels, names = ids)
elif n_comparing == 3:
labels = venn.get_labels([comp[ids[0]], comp[ids[1]], comp[ids[2]]], fill = ['number'])
fig, ax = venn.venn3(labels, names = ids)
elif n_comparing == 2:
labels = venn.get_labels([comp[ids[0]], comp[ids[1]]], fill = ['number'])
fig, ax = venn.venn2(labels, names = ids)
filename = save+'_pred_compared.pdf'
fig.savefig(filename)
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import venn
labels = venn.get_labels([range(10), range(5, 15)], fill=['number', 'logic'])
fig, ax = venn.venn2(labels, names=['list 1', 'list 2'])
fig.savefig('venn2.png', bbox_inches='tight')
plt.close()
labels = venn.get_labels([range(10), range(5, 15), range(3, 8)], fill=['number', 'logic'])
fig, ax = venn.venn3(labels, names=['list 1', 'list 2', 'list 3'])
fig.savefig('venn3.png', bbox_inches='tight')
plt.close()
labels = venn.get_labels([range(10), range(5, 15), range(3, 8), range(8, 17)], fill=['number', 'logic'])
fig, ax = venn.venn4(labels, names=['list 1', 'list 2', 'list 3', 'list 4'])
fig.savefig('venn4.png', bbox_inches='tight')
plt.close()
labels = venn.get_labels([range(10), range(5, 15), range(3, 8), range(8, 17), range(10, 20)], fill=['number', 'logic'])
fig, ax = venn.venn5(labels, names=['list 1', 'list 2', 'list 3', 'list 4', 'list 5'])
fig.savefig('venn5.png', bbox_inches='tight')
plt.close()
labels = venn.get_labels([range(10), range(5, 15), range(3, 8), range(8, 17), range(10, 20), range(13, 25)], fill=['number', 'logic'])
fig, ax = venn.venn6(labels, names=['list 1', 'list 2', 'list 3', 'list 4', 'list 5', 'list 6'])
fig.savefig('venn6.png', bbox_inches='tight')
plt.close()
elif "full_model" in results_file:
all_labels.append("Full model")
for r_label in results:
# print unique TP to this result
tps = set(results[r_label]["all"]) & set(gold_pairs["all"])
for r_label2 in results:
if r_label2 == r_label:
continue
tps = tps - set(results[r_label2]["all"])
print()
print("unique to {}:".format(r_label))
print(len(tps))
print(tps)
print()
print(all_labels)
# print(results)
labels = venn.get_labels(all_pairs, fill=["number"])
if len(all_pairs) == 2:
fig, ax = venn.venn2(labels, names=all_labels)
elif len(all_pairs) == 3:
fig, ax = venn.venn3(labels, names=all_labels)
elif len(all_pairs) == 4:
fig, ax = venn.venn4(labels, names=all_labels)
elif len(all_pairs) == 5:
fig, ax = venn.venn5(labels, names=all_labels)
elif len(all_pairs) == 6:
fig, ax = venn.venn6(labels, names=all_labels)
fig.savefig("{}.png".format("_".join(all_labels)), bbox_inches="tight")
plt.close()