def _get_font(self):
from matplotlib.font_manager import FontManager
from PIL import ImageFont
font_finder = FontManager()
font_path = font_finder.findfont("arial")
return ImageFont.truetype(font_path, 35)
def get_matplot_zh_font():
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"', shell=True)
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = list(mat_fonts & zh_fonts)
return available
def build_matplotlib_font_cache():
"""This is to mitigate errors on CI VMs, where you can get the message:
"Matplotlib is building the font cache" in output notebooks
"""
from matplotlib.font_manager import FontManager
FontManager()
def _plot(self, a, key, title, gx, gy, num_x, num_y):
logger.debug('Plotting: %s', key)
pp.rcParams['figure.figsize'] = (self._image_width / 300,
self._image_height / 300)
pp.title(title)
# Interpolate the data
rbf = Rbf(a['x'], a['y'], a[key], function='linear')
z = rbf(gx, gy)
z = z.reshape((num_y, num_x))
# Render the interpolated data to the plot
pp.axis('off')
# begin color mapping
if 'min' in self.thresholds.get(key, {}):
vmin = self.thresholds[key]['min']
logger.debug('Using min threshold from thresholds: %s', vmin)
else:
vmin = min(a[key])
logger.debug('Using calculated min threshold: %s', vmin)
if 'max' in self.thresholds.get(key, {}):
vmax = self.thresholds[key]['max']
logger.debug('Using max threshold from thresholds: %s', vmax)
else:
vmax = max(a[key])
logger.debug('Using calculated max threshold: %s', vmax)
norm = matplotlib.colors.Normalize(vmin=vmin, vmax=vmax, clip=True)
mapper = cm.ScalarMappable(norm=norm, cmap='RdYlBu_r')
# end color mapping
image = pp.imshow(z,
extent=(0, self._image_width, self._image_height, 0),
cmap='RdYlBu_r',
alpha=0.5,
zorder=100,
vmin=vmin,
vmax=vmax)
pp.colorbar(image)
pp.imshow(self._layout, interpolation='bicubic', zorder=1, alpha=1)
labelsize = FontManager.get_default_size() * 0.4
# begin plotting points
for idx in range(0, len(a['x'])):
if (a['x'][idx], a['y'][idx]) in self._corners:
continue
pp.plot(a['x'][idx],
a['y'][idx],
marker='o',
markeredgecolor='black',
markeredgewidth=1,
markerfacecolor=mapper.to_rgba(a[key][idx]),
markersize=6)
pp.text(a['x'][idx],
a['y'][idx] - 30,
a['ap'][idx],
fontsize=labelsize,
horizontalalignment='center')
# end plotting points
fname = '%s_%s.png' % (key, self._title)
logger.info('Writing plot to: %s', fname)
pp.savefig(fname, dpi=300)
pp.close('all')
def try_load_font_mpl(name):
try:
from matplotlib.font_manager import FontManager
except ImportError:
return
fm = FontManager()
suggestions={}
for style in [BOLD, LIGHT, LIGHT_ITALIC, ITALIC, ITALIC_BOLD, NORMAL]:
with warnings.catch_warnings():
warnings.filterwarnings("ignore")
suggestion = fm.findfont("{0}:{1}".format(name, style.lower()), fallback_to_default=True)
if get_font_family(suggestion) != name:
continue
else:
add_font_directory(os.path.dirname(suggestion))
suggestions[style] = suggestion
load_ttf_font(name, {style: os.path.splitext(os.path.basename(suggestion))[0] for style, suggestion in suggestions.items()})
def get_matplot_zh_font():
# From https://blog.csdn.net/kesalin/article/details/71214038
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh-tw -f "%{family}\n"',
shell=True)
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = list(mat_fonts & zh_fonts)
return available
def get_zh_fonts():
"""获取可用的中文字体"""
font_manger = FontManager()
mat_fonts = set(f.name for f in font_manger.ttflist)
# print(mat_fonts)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"', shell=True)
# print(output)
zh_fonts = set(f.split(',', 1)[0] for f in output.decode('utf-8').split('\n'))
available = mat_fonts & zh_fonts
# print('*' * 10, '可用的字体', '*' * 10)
for f in available:
print(f)
def _draw_mol_at(
self,
mol,
center_x,
center_y,
mol_size=None,
act_dict={},
center=False,
**kwargs,
):
img = mol
if mol_size is None:
mol_size = self.mol_size
if isinstance(mol, Chem.Mol):
img = Draw.MolToImage(mol,
mol_size,
kekulize=True,
fitImage=True,
**kwargs)
width, height = img.size
self.image.paste(
img, (int(center_x - width / 2), int(center_y - height / 2)))
txt = []
for prop, propval in act_dict.items():
if not isinstance(propval, str):
propval = "{:.2f}".format(propval)
txt.append(f"{prop}: {propval}")
if txt and self.draw is not None:
txt = "\n".join(txt)
font = None
w, h = self.draw.multiline_textsize(txt)
try:
fn = FontManager()
fontpath = fn.findfont("Droid sans")
font = ImageFont.truetype(fontpath, 18 + center * 8)
w, h = self.draw.multiline_textsize(txt, font=font)
except:
passcircle_mols
def get_matplot_zh_font():
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True)
output_list = output.decode().split('\n')
zh_fonts = set(f.split(',', 1)[0] for f in output_list)
available = list(mat_fonts & zh_fonts)
print('*' * 10, '可用的字体', '*' * 10)
for f in available:
print(f)
return available
def get_matplot_zh_font():
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
print(mat_fonts)
output = str(subprocess.check_output('fc-list :lang=zh -f "%{family}\n"', shell=True))
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
print(zh_fonts)
available = list(mat_fonts & zh_fonts)
#
print('*' * 10, '可用的字体', '*' * 10)
for f in available:
print(f)
available_font = [output.split('\n')[0].split(',', 1)[0]]
print("available_font:" + str(available_font))
return available_font
def _draw(self):
"""Create circles and slices in-memory"""
draw = ImageDraw.Draw(self.image)
self.draw = draw
all_radius = self._draw_circles(draw)
self._draw_center_mol()
self._draw_ring_mols(all_radius)
font = None
w, h = draw.textsize(self.legend)
try:
fn = FontManager()
fontpath = fn.findfont("Droid sans")
font = ImageFont.truetype(fontpath, 12 * self.size // 800)
w, h = font.getsize(self.legend)
except:
pass
draw.text(
((self.size // 2 - w) - 2, self.size - 2 * h),
self.legend,
fill="black",
font=font,
)
del draw
self.draw = None
def add_system_fonts() -> None:
try:
from matplotlib.font_manager import FontManager
except ImportError:
logger.debug('This function requires the optional matplotlib package.')
return
for entry in FontManager().ttflist:
ttf = db_key(entry.fname)
fonts[ttf] = Font(
ttf,
entry.name,
entry.style,
entry.stretch,
entry.weight,
)
def supported_chinese_font():
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True)
# print '*' * 10, '系统可用的中文字体', '*' * 10
# print output
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = mat_fonts & zh_fonts
print '*' * 10, '可用的中文字体', '*' * 10
for f in available:
print f
return available
def plot_similarity(labels, features, rotation):
"""
sentence vector evaluation
:param labels: list of sentences
:param features: matrix composed by sentence vectors
:param rotation: generally 90
:return:
"""
plt.rcParams['font.sans-serif'] = ['SimHei']
# plt.rcParams['font.serif'] = ['SimHei']
# plt.rcParams['axes.unicode_minus'] = False
print('matplotlib directory:', mpl.matplotlib_fname())
# SimHei.ttf should be in
# /home/deco/miniconda2/envs/tf17/lib/python3.6/site-packages/matplotlib/
# mpl-data/fonts/ttf
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True)
# print('Type of output:', type(output))
output = output.decode('utf-8')
# print('*' * 10, '系统可用的中文字体', '*' * 10)
# print(output)
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = mat_fonts & zh_fonts
print('*' * 10, '可用的字体', '*' * 10)
for f in available:
print(f)
corr = np.inner(features, features)
# features是多个sentence embeddings组成的矩阵，上面这个内积操作就算出了sentence两两之间的相似度
# np.inner(u, v) == u*transpose(v)
sns.set(font_scale=1.2)
sns.set_style("darkgrid", {"font.sans-serif": ['simhei', 'Arial']})
# the code above must be after sns.set()
g = sns.heatmap(
corr,
xticklabels=labels,
yticklabels=labels,
vmin=0,
# 如果值小于0，就按0处理，也就是所谓的语境相反会被当做不相关处理
vmax=1,
cmap="YlOrRd")
g.set_xticklabels(labels, rotation=rotation)
# 把labels字体进行旋转
g.set_title("Semantic Textual Similarity")
plt.show()
def load_system_fonts() -> Dict[str, fonts.FontFace]:
""" Load system fonts by the FontManager of matplotlib.
This may take a long time!
"""
font_faces = dict()
for entry in FontManager().ttflist:
ttf = fonts.cache_key(entry.fname)
font_faces[ttf] = fonts.FontFace(
ttf,
entry.name,
entry.style,
entry.stretch,
entry.weight,
)
return font_faces
def getavailablezhfont():
"""
获取系统中安装的中文字体列表
"""
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True).decode('utf-8')
# print '*' * 10, '系统可用的中文字体', '*' * 10
# print output
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = mat_fonts & zh_fonts
print('*' * 10, '可用的字体', '*' * 10)
for f in available:
print(f'{f}')
def get_matplot_zh_font():
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True)
fonts = output.decode()
fontArr = fonts.split('\n')
zh_fonts = set(f.split(',', 1)[0] for f in fontArr)
# zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = list(mat_fonts & zh_fonts)
# remove .xxxx
available = [i for i in available if not i.startswith('.')]
print('*' * 10, '可用的字体', '*' * 10)
for f in available:
print(f)
return available
def test():
# judge the availabe chinese fonts
from matplotlib.font_manager import FontManager
import subprocess
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output(
'fc-list :lang=zh -f "%{family}\n"', shell=True)
# print '*' * 10, '系统可用的中文字体', '*' * 10
print output
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = mat_fonts & zh_fonts
print '*' * 10, '可用的字体', '*' * 10
for f in available:
print f
def dispFonts():
#显示可用的中文字体，同时支持英文的
from matplotlib.font_manager import FontManager
import subprocess
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True)
output = output.decode('utf-8')
# print '*' * 10, '系统可用的中文字体', '*' * 10
# print output
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = mat_fonts & zh_fonts
print '*' * 10 + u'可用的中文字体' + '*' * 10
for f in available:
print(f)
#显示可用的中文字体，同时支持英文的
from matplotlib.font_manager import FontManager
import subprocess
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
output = subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',
shell=True)
output = output.decode('utf-8')
# print ‘*‘ * 10, ‘系统可用的中文字体‘, ‘*‘ * 10
# print output
zh_fonts = set(f.split(',', 1)[0] for f in output.split('\n'))
available = mat_fonts & zh_fonts
print('*' * 10 + u'可用的中文字体' + '*' * 10)
for f in available:
print(f)
def reset_font_manager() -> None:
global _font_manager
_font_manager = FontManager()
def plot2(regex=re.compile(pattern='-*\d+\.\d+')):
"""
lr对ddpg收敛性的影响
:return:
"""
ddpg_lr1 = None
with open(file=
r'/Users/songyunlong/Desktop/实验室/时延模型ddpg/ddpg_file/ddpg200.txt',
mode='r') as f:
#lr=1e-3
while True:
line_str = f.readline()
if not line_str:
break
line_array = np.array([float(i) for i in line_str.split(' ')])
ddpg_lr1 = line_array if ddpg_lr1 is None else np.vstack(
(ddpg_lr1, line_array))
ddpg_lr2 = None
with open(
file=
r'/Users/songyunlong/Desktop/实验室/时延模型ddpg/ddpg_file/ddpg200-1e-2.txt',
mode='r') as f:
#lr=1e-2
while True:
line_str = f.readline()
if not line_str:
break
generator = regex.finditer(string=line_str)
line_array = np.array([float(i.group(0)) for i in generator])
ddpg_lr2 = line_array if ddpg_lr2 is None else np.vstack(
(ddpg_lr2, line_array))
ddpg_lr3 = None
with open(
file=
r'/Users/songyunlong/Desktop/实验室/时延模型ddpg/ddpg_file/ddpg200-5e-3.txt',
mode='r') as f:
# lr=5e-3
while True:
line_str = f.readline()
if not line_str:
break
generator = regex.finditer(string=line_str)
line_array = np.array([float(i.group(0)) for i in generator])
ddpg_lr3 = line_array if ddpg_lr3 is None else np.vstack(
(ddpg_lr3, line_array))
from matplotlib.font_manager import FontManager
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
# print (mat_fonts)
# plt.rcParams['font.sans-serif'] = ['Arial Unicode MS']
#reward
# fig, ax = plt.subplots(ncols=1, nrows=2)
#
# x = range(1, 101)
# ax[0].plot(x_spline_new, kesi(spline(x, ddpg_lr1[:, 0]*100)(x_spline_new),1), c='r', linestyle='-', label='Lr=1e-3')
# ax[0].plot(x_spline_new, kesi(spline(x, ddpg_lr2[:, 0]*100)(x_spline_new),1), c='g', linestyle='--', label='Lr=1e-2')
# ax[0].plot(x_spline_new, kesi(spline(x, ddpg_lr3[:, 0]*100)(x_spline_new),1), c='b', linestyle=':', label='Lr=5e-3')
# ax[0].legend()
# # ax[0].set_xlabel('Training epochs')
# ax[0].set_xlabel(u'迭代训练次数', fontproperties='Noto Sans Vai')
# # ax[0].set_ylabel('Reward')
# ax[0].set_ylabel(u'奖励值', fontproperties='Noto Sans Vai')
# ax[0].grid(axis='x', linestyle='-.')
# ax[0].grid(axis='y', linestyle='-.')
# ax[0].set_title('(a)The total reward value per episode', y=-0.5)
# # fig.show()
#
# #latency
# # fig, ax = plt.subplots()
# x = range(1, 101)
# ax[1].plot(x_spline_new, kesi(spline(x, ddpg_lr1[:, -1])(x_spline_new),2), c='r', linestyle='-', label='Lr=1e-3')
# ax[1].plot(x_spline_new, kesi(spline(x, ddpg_lr2[:, -1])(x_spline_new),2), c='g', linestyle='--', label='Lr=1e-2')
# ax[1].plot(x_spline_new, kesi(spline(x, ddpg_lr3[:, -1])(x_spline_new),2), c='b', linestyle=':', label='Lr=5e-3')
# ax[1].legend()
# ax[1].set_xlabel('Training epochs')
# ax[1].set_ylabel('Average latency/s')
# ax[1].grid(axis='x', linestyle='-.')
# ax[1].grid(axis='y', linestyle='-.')
# ax[1].set_title('(b)The average latency per episode', y=-0.5)
# fig.show()
#专利绘图
plt.rcParams['font.sans-serif'] = ['Arial Unicode MS']
plt.title('搜狗')
x = range(1, 101)
# plt.plot(x_spline_new, kesi(spline(x, ddpg_lr1[:, 0] * 100)(x_spline_new), 1), c='r', linestyle='-',
# label='Lr=1e-3')
# plt.plot(x_spline_new, kesi(spline(x, ddpg_lr2[:, 0] * 100)(x_spline_new), 1), c='g', linestyle='--',
# label='Lr=1e-2')
# plt.plot(x_spline_new, kesi(spline(x, ddpg_lr3[:, 0] * 100)(x_spline_new), 1), c='b', linestyle=':',
# label='Lr=5e-3')
# plt.legend()
# # ax[0].set_xlabel('Training epochs')
# plt.xlabel(u'迭代训练次数', size=9)
# # ax[0].set_ylabel('Reward')
# plt.ylabel(u'奖励值')
# plt.grid(axis='x', linestyle='-.')
# plt.grid(axis='y', linestyle='-.')
# plt.title('(a)每条采样链的总奖励值', y=-0.16, size=10)
# plt.show()
# plt.plot(x_spline_new, kesi(spline(x, ddpg_lr1[:, -1] * 100)(x_spline_new), 1), c='r', linestyle='-',
# label='Lr=1e-3')
# plt.plot(x_spline_new, kesi(spline(x, ddpg_lr2[:, -1] * 100)(x_spline_new), 1), c='g', linestyle='--',
# label='Lr=1e-2')
# plt.plot(x_spline_new, kesi(spline(x, ddpg_lr3[:, -1] * 100)(x_spline_new), 1), c='b', linestyle=':',
# label='Lr=5e-3')
plt.plot(x_spline_new,
kesi(spline(x, ddpg_lr1[:, -1])(x_spline_new), 2),
c='r',
linestyle='-',
label='Lr=1e-3')
plt.plot(x_spline_new,
kesi(spline(x, ddpg_lr2[:, -1])(x_spline_new), 2),
c='g',
linestyle='--',
label='Lr=1e-2')
plt.plot(x_spline_new,
kesi(spline(x, ddpg_lr3[:, -1])(x_spline_new), 2),
c='b',
linestyle=':',
label='Lr=5e-3')
plt.legend()
# ax[0].set_xlabel('Training epochs')
plt.xlabel(u'迭代训练次数', size=9)
# ax[0].set_ylabel('Reward')
plt.ylabel(u'平均时延')
plt.grid(axis='x', linestyle='-.')
plt.grid(axis='y', linestyle='-.')
plt.title('(b)每条采样链的平均时延', y=-0.16, size=10)
plt.show()
def _get_font_manager():
global _font_manager
if _font_manager is None:
_font_manager = FontManager()
return _font_manager
def rebuild_system_fonts():
import matplotlib.font_manager
matplotlib.font_manager._rebuild()
global _font_manager
_font_manager = FontManager()
from matplotlib.font_manager import FontManager
import subprocess
mpl_fonts = set(f.name for f in FontManager().ttflist)
print('all font list get from matplotlib.font_manager:')
for f in sorted(mpl_fonts):
print('\t' + f)
def main():
if len(sys.argv)!=4:
print (str(len(sys.argv)))
else:
for i in range(len(sys.argv)):
print sys.argv[i]
print("read triggers list")
triggers_input=open(sys.argv[1],"r")
for line in triggers_input:
if line.count("\t")<6:
print(str(line.count("\t")))
continue
else:
tokens=line.strip().split("\t")
print(tokens[4]+":"+tokens[3])
begin_index=tokens[3].strip().split(",")[0]
print("begin index is"+begin_index)
trigger=[]
trigger.append(tokens[4])
trigger.append(begin_index)
triggers_list.append(trigger)
print("length of triggers list is "+str(len(triggers_list)))
print("read arguments list")
arguments_input=open(sys.argv[2],"r")
for line in arguments_input:
if line.count("\t")<11:
print(str(line.count("\t")))
continue
else:
tokens=line.strip().split("\t")
tokens[4]=tokens[4].replace(',','').replace('，','')
tokens[4]=tokens[4].replace('。','').replace('.','')
tokens[4]=tokens[4].replace('[','').replace(']','')
tokens[4]=tokens[4].replace('；','').replace(';','')
tokens[4]=tokens[4].replace('《','').replace('》','')
tokens[4]=tokens[4].replace('￥','').replace('$','')
tokens[4]=tokens[4].replace('、','').replace('\\','')
tokens[4]=tokens[4].strip()
print(tokens[4]+":"+tokens[11])
argument=[]
argument.append(tokens[4])
argument.append(tokens[11])
arguments_list.append(argument)
print("length of arguments list is "+str(len(arguments_list)))
print("generate event")
for i in range(len(arguments_list)):
if len(events_list)==0:
for j in range(len(triggers_list)):
if arguments_list[i][1]==triggers_list[j][1]:
print("插入第一个事件")
event=[]
event.append(arguments_list[i])
event.append(triggers_list[j])
events_list.append(event)
print("插入结果为：")
print(arguments_list[i][0]+":"+arguments_list[i][1])
print(triggers_list[j][0]+":"+triggers_list[j][1])
print(events_list[0][0][0]+":"+events_list[0][0][1])
print(events_list[0][1][0]+":"+events_list[0][1][1])
break
else:
exist_state=False
for k in range(len(events_list)):
if arguments_list[i][1]==events_list[k][0][1]:
events_list[k].append(arguments_list[i])
print("该事件已经存在于事件列表中，现向列表中添加事件元素")
print("添加的事件元素为："+arguments_list[i][0]+":"+arguments_list[i][1])
print("该事件为：")
for l in range(len(events_list[k])):
print(events_list[k][l][0]+":"+events_list[k][l][1])
exist_state=True
break
if not exist_state:
for j in range(len(triggers_list)):
if arguments_list[i][1]==triggers_list[j][1]:
print("插入新的事件如下：")
event=[]
event.append(arguments_list[i])
event.append(triggers_list[j])
events_list.append(event)
print("插入结果为：")
print(arguments_list[i][0]+":"+arguments_list[i][1])
print(triggers_list[j][0]+":"+triggers_list[j][1])
print(events_list[len(events_list)-1][0][0]+":"+events_list[len(events_list)-1][0][1])
print(events_list[len(events_list)-1][1][0]+":"+events_list[len(events_list)-1][1][1])
print("事件提取完毕")
print("\n生成词表")
for i in range(len(events_list)):
for j in range(len(events_list[i])):
if len(words_list)==0:
words_list.append(events_list[i][j][0])
print("向词表中添加第一个词,这个词是："+words_list[0])
else:
exist_state=False
for k in range(len(words_list)):
if events_list[i][j][0]==words_list[k]:
print("词语("+events_list[i][j][0]+")已经存在，不需要重复加入，这个存在的词语是:"+words_list[k])
exist_state=True
break
if not exist_state:
words_list.append(events_list[i][j][0])
print("将一个新词("+events_list[i][j][0]+")添加进词表中，这个词是:"+words_list[len(words_list)-1])
print("\n将事件触发词和事件元素替换为词典编码")
for i in range(len(events_list)):
event=[]
for j in range(len(events_list[i])):
for k in range(len(words_list)):
if words_list[k]==events_list[i][j][0]:
event.append(k)
break
events_code_list.append(event)
for l in range(len(events_list[i])):
print("词语("+events_list[i][l][0]+")替换为("+str(events_code_list[len(events_code_list)-1][l])+")")
print("\n")
print("\n生成联通图")
#dg=digraph()
dg=graph()
nodes_list=range(len(words_list))
dg.add_nodes(nodes_list)
print("将句子级的事件元素以及事件触发词串联起来")
for i in range(len(events_code_list)):
for j in range(len(events_code_list[i])):
for k in range(len(events_code_list[i])):
if not dg.has_edge((events_code_list[i][j],events_code_list[i][k])) and (j!=k):
dg.add_edge((events_code_list[i][j],events_code_list[i][k]))
print("事件"+str(i)+"中的("+words_list[events_code_list[i][j]]+")与("+words_list[events_code_list[i][k]]+")之间建立了一条边")
print("\n显示可用字体")
fm=FontManager()
mat_fonts=set(f.name for f in fm.ttflist)
output=subprocess.check_output('fc-list :lang=zh -f "%{family}\n"',shell=True)
zh_fonts=set(f.split(',',1)[0] for f in output.split('\n'))
available=mat_fonts
print '*'*10,'可用的字体','*'*10
for f in available:
print f
print("read CHINESE_WORD2VEC_FILENAME")
word2vec=open(sys.argv[3],"r")
for line in word2vec:
if line.count(" ")<5:
continue
else:
index=line.find(" ")
curWord=line[:index]
rest=line[index+1:]
tokens=rest.strip().split(" ")
numTokens=[]
for tok in tokens:
numTokens.append(float(tok))
chineseWordVecs[curWord]=numTokens
word2vec.close()
print("load CHINESE_WORD2VEC_FILENAME successfully")
print("\n提取单词的词向量")
for i in range(len(events_code_list)):
event_word2vec=[]
for j in range(len(events_code_list[i])):
event_word_word2vec=[]
words=words_list[events_code_list[i][j]].split(" ")
for l in range(len(words)):
if chineseWordVecs.has_key(words[l]):
#print("\n事件"+str(i)+"中的词语("+words[l]+")具有词向量")
#print("该词向量为:"+str(chineseWordVecs[words[l]]))
event_word_word2vec.append(chineseWordVecs[words[l]])
else:
#print("\n事件"+str(i)+"中的词语("+words[l]+")不具有词向量")
zero_list=[]
for m in range(100):
zero_list.append(0)
event_word_word2vec.append(zero_list)
event_word2vec.append(event_word_word2vec)
events_word2vec_list.append(event_word2vec)
#for i in range(len(events_code_list)):
#for j in range(len(events_code_list[i])):
#words=words_list[events_code_list[i][j]].split(" ")
#for l in range(len(words)):
#if chineseWordVecs.has_key(words[l]):
#print("\n事件"+str(i)+"中的词语("+words[l]+")具有词向量")
#print("该词向量为:"+str(chineseWordVecs[words[l]]))
#print("提取出的词向量为:"+str(events_word2vec_list[i][j][l]))
#print("提取出的词向量长度为:"+str(len(events_word2vec_list[i][j][l])))
#else:
#print("\n事件"+str(i)+"中的词语("+words[l]+")不具有词向量")
#print("提取出的词向量为:"+str(events_word2vec_list[i][j][l]))
#print("提取出的词向量长度为:"+str(len(events_word2vec_list[i][j][l])))
print("根据余弦相似度，将词组用边连接起来")
for i in range(len(events_code_list)):
for j in range(len(events_code_list)):
if i==j:
continue
else:
#print("将事件("+str(i)+")与事件("+str(j)+")做比较")
for k in range(len(events_code_list[i])):
for l in range(len(events_code_list[j])):
#print("将事件("+str(i)+")的第("+str(k)+")个词组与事件("+str(j)+")的第("+str(l)+")个词组做比较")
max_cos=0
for m in range(len(events_word2vec_list[i][k])):
for n in range(len(events_word2vec_list[j][l])):
mn=0
mm=0
nn=0
for o in range(100):
mn+=(events_word2vec_list[i][k][m][o]*events_word2vec_list[j][l][n][o])
mm+=(events_word2vec_list[i][k][m][o]*events_word2vec_list[i][k][m][o])
nn+=(events_word2vec_list[j][l][n][o]*events_word2vec_list[j][l][n][o])
if (mn==0)or(nn==0)or(mm==0):
cos=0
else:
cos=mn/(math.sqrt(mm)*math.sqrt(nn))
if cos>max_cos:
max_cos=cos
#print("词组("+words_list[events_code_list[i][k]]+")与词组("+words_list[events_code_list[j][l]]+")的余弦相似度为:"+str(max_cos))
if max_cos>0.7:
if not dg.has_edge((events_code_list[i][k],events_code_list[j][l])):
dg.add_edge((events_code_list[i][k],events_code_list[j][l]))
print("词组("+words_list[events_code_list[i][k]]+")与词组("+words_list[events_code_list[j][l]]+")的余弦相似度为("+str(max_cos)+")超过阀值，建立一条边")
pr=PRIterator(dg)
page_ranks=pr.page_rank()
print("词表长度为:"+str(len(words_list)))
print("提取page_rangk字典中的值")
for i in range(len(words_list)):
page_rank=[]
page_rank.append(i)
page_rank.append(page_ranks[i])
page_ranks_list.append(page_rank)
print(str(page_ranks_list))
print("按照PR值对词组进行排序")
for i in range(len(page_ranks_list)-1):
epoch=i+1
for j in range(len(page_ranks_list)-epoch):
page_rank=[]
if page_ranks_list[j][1]>page_ranks_list[j+1][1]:
page_rank=page_ranks_list[j]
page_ranks_list[j]=page_ranks_list[j+1]
page_ranks_list[j+1]=page_rank
print(str(page_ranks_list))
#print("提取出的篇章级事件为:")
#for i in range(len(page_ranks_list)):
#print("\n核心短语为:"+words_list[page_ranks_list[len(page_ranks_list)-i-1][0]])
#neighbors=dg.neighbors(page_ranks_list[len(page_ranks_list)-i-1][0])
#for j in range(len(neighbors)):
#print(words_list[neighbors[j]])
print("\n输出所有的连接边:")
edges=dg.edges()
for i in range(len(edges)):
print str(edges[i])
print("\n将编码替换为词组")
nodes_word=[]
for i in range(len(words_list)):
nodes_word.append(words_list[i].decode('utf8'))
edges_word=[]
for i in range(len(edges)):
tup=(words_list[edges[i][0]].decode('utf8'),words_list[edges[i][1]].decode('utf8'))
#print("\n"+tup[0]+"---"+tup[1])
edges_word.append(tup)
print("\n为节点赋予标签")
labels=dict()
for i in range(len(words_list)):
labels[i]=words_list[i]
print("\n使用networkx插件建立新的联通图")
G=nx.Graph()
G.add_nodes_from(nodes_word)
G.add_edges_from(edges_word)
#pos=nx.circular_layout(G)
#pos=nx.spectral_layout(G)
pos=nx.spring_layout(G)
#pos=nx.shell_layout(G)
#nx.draw(G,pos,node_shape='.',node_size=40)
#nx.draw_networkx_nodes(G,pos,range(len(words_list)),node_shape='.',node_size=40)
#nx.draw_networkx_labels(G,pos,labels,node_shape='.',node_size=40)
plt.subplot(2,2,1)
nx.draw_networkx(G,pos,node_shape='.',node_size=40)
# print("\n寻找权值最大的路径并绘制出来")
# dis=[]
# path=[]
# print("初始化路径矩阵")
# for i in range(len(words_list)):
# path_i=[]
# for j in range(len(words_list)):
# path_i.append(j)
# path.append(path_i)
# print("初始化邻接矩阵")
# FLOAT_MAX=0.0
# for i in range(len(words_list)):
# dis_i=[]
# for j in range(len(words_list)):
# if dg.has_edge((i,j)):
# dis_i.append(0.0-page_ranks[i]-page_ranks[j])
# print("\n节点("+words_list[i]+")与节点("+words_list[j]+")之间存在一条边，这条边的权值为:"+str(dis_i[len(dis_i)-1]))
# else:
# dis_i.append(FLOAT_MAX)
# print(
# "\n节点(" + words_list[i] + ")与节点(" + words_list[j] + ")之间不存在边，将权值矩阵此处的值初始化为:" + str(dis_i[len(dis_i) - 1]))
# dis.append(dis_i)
# print("\n开始寻找任意两个节点之间权值最大的路径")
# for temp in range(len(words_list)):
# for row in range(len(words_list)):
# for col in range(len(words_list)):
# if dg.has_edge((temp,row)) and dg.has_edge((temp,col)):
# select=dis[row][temp]+dis[temp][col]
# if dis[row][col]>select:
# dis[row][col]=select
# path[row][col]=path[row][temp]
# print("\n节点("+words_list[row]+")与节点("+words_list[col]+"之间插入了节点("+wodrs_list[temp]+")")
print("\n寻找权值最大的路径")
nodes_word=[]
exist_nodes=[]
node=""
triggers_only_list=[]
for i in range(len(triggers_list)):
triggers_only_list.append(triggers_list[i][0])
for i in range(len(page_ranks_list)):
if words_list[page_ranks_list[len(page_ranks_list)-i-1][0]] in triggers_only_list:
print(words_list[page_ranks_list[len(page_ranks_list)-i-1][0]]+":"+str(page_ranks_list[len(page_ranks_list)-i-1][1]))
for i in range(len(page_ranks_list)):
if words_list[page_ranks_list[len(page_ranks_list)-i-1][0]] in triggers_only_list:
#print(words_list[page_ranks_list[len(page_ranks_list)-i-1][0]]+":"+str(page_ranks_list[len(page_ranks_list)-i-1][1]))
node = page_ranks_list[len(page_ranks_list) - i - 1][0]
nodes_word.append(words_list[node])
exist_nodes.append(node)
print("\n加入词语(" + words_list[node] + ")")
break
while True:
neighbors = dg.neighbors(node)
finish_state=True
for i in range(len(neighbors)):
if neighbors[i] in exist_nodes:
finish_state=(finish_state and True)
else:
finish_state=(finish_state and False)
if finish_state:
break
for i in range(len(neighbors)):
if neighbors[i] in exist_nodes:
continue
else:
max_neighbor = neighbors[i]
print("\n最开始的最大邻接点是("+words_list[max_neighbor]+")")
print("这个节点的PR值为："+str(page_ranks[max_neighbor]))
break
for i in range(len(neighbors)):
if not(neighbors[i] in exist_nodes):
if (page_ranks[neighbors[i]]>page_ranks[max_neighbor]):
max_neighbor = neighbors[i]
print("\n最大邻接点更换为(" + words_list[max_neighbor] + ")")
print("这个节点的PR值为：" + str(page_ranks[max_neighbor]))
node=max_neighbor
nodes_word.append(words_list[max_neighbor])
exist_nodes.append(node)
print("加入词语(" + words_list[node] + ")")
print("\n绘制最大权重路径")
G2=nx.Graph()
edges_word=[]
for i in range(len(nodes_word)-1):
tup=(nodes_word[i].decode('utf-8'),nodes_word[i+1].decode('utf-8'))
edges_word.append(tup)
G2.add_nodes_from(nodes_word)
G2.add_edges_from(edges_word)
pos = nx.spring_layout(G2)
plt.subplot(2, 2, 4)
nx.draw_networkx(G2, pos, node_shape='.', node_size=40)
plt.savefig("graph.png")
plt.show()
def get_system_font():
fm = FontManager()
mat_fonts = set(f.name for f in fm.ttflist)
print(mat_fonts)
from math import ceil
from PIL import Image
from io import BytesIO
from bs4 import BeautifulSoup
from matplotlib import pyplot as plt
from matplotlib.font_manager import FontManager
from cjkwrap import fill
from warnings import filterwarnings
from cachetools import TTLCache
from ..aio_helper import run_async
from .utils import logger
from ..compat import cached_async
MPL_TTF_LIST = FontManager().ttflist
MPL_SANS_FONTS: Final = (
list({f.name
for f in MPL_TTF_LIST if f.name == 'WenQuanYi Micro Hei'}) +
list({f.name
for f in MPL_TTF_LIST if f.name == 'WenQuanYi Zen Hei'}) +
list({f.name
for f in MPL_TTF_LIST if f.name.startswith('Noto Sans CJK')}) +
list(
{f.name
for f in MPL_TTF_LIST if f.name.startswith('Microsoft YaHei')}) +
list({f.name
for f in MPL_TTF_LIST if f.name == 'SimHei'}) +
list({
f.name
for f in MPL_TTF_LIST if f.name in {'SimKai', 'SimSun', 'SimSun-ExtB'}
def fontsize(idx, default=FontManager.get_default_size()):
coeff = p.nodes[idx]["freq"] / p.molecules
size = default * coeff * ax_coeff
return size
def _plot(self, a, key, title, gx, gy, num_x, num_y):
if key not in a:
logger.info("Skipping {} due to insufficient data".format(key))
return
if not len(a['x']) == len(a['y']) == len(a[key]):
logger.info("Skipping {} because data has holes".format(key))
return
logger.debug('Plotting: %s', key)
pp.rcParams['figure.figsize'] = (
self._image_width / 300, self._image_height / 300
)
fig, ax = pp.subplots()
ax.set_title(title)
if 'min' in self.thresholds.get(key, {}):
vmin = self.thresholds[key]['min']
logger.debug('Using min threshold from thresholds: %s', vmin)
else:
vmin = min(a[key])
logger.debug('Using calculated min threshold: %s', vmin)
if 'max' in self.thresholds.get(key, {}):
vmax = self.thresholds[key]['max']
logger.debug('Using max threshold from thresholds: %s', vmax)
else:
vmax = max(a[key])
logger.debug('Using calculated max threshold: %s', vmax)
logger.info("{} has range [{},{}]".format(key, vmin, vmax))
# Interpolate the data only if there is something to interpolate
if vmin != vmax:
rbf = Rbf(
a['x'], a['y'], a[key], function='linear'
)
z = rbf(gx, gy)
z = z.reshape((num_y, num_x))
else:
# Uniform array with the same color everywhere
# (avoids interpolation artifacts)
z = numpy.ones((num_y, num_x))*vmin
# Render the interpolated data to the plot
ax.axis('off')
# begin color mapping
norm = matplotlib.colors.Normalize(vmin=vmin, vmax=vmax, clip=True)
mapper = cm.ScalarMappable(norm=norm, cmap=self._cmap)
# end color mapping
image = ax.imshow(
z,
extent=(0, self._image_width, self._image_height, 0),
alpha=0.5, zorder=100,
cmap=self._cmap, vmin=vmin, vmax=vmax
)
# Draw contours if requested and meaningful in this plot
if self._contours is not None and vmin != vmax:
CS = ax.contour(z, colors='k', linewidths=1, levels=self._contours,
extent=(0, self._image_width, self._image_height, 0),
alpha=0.3, zorder=150, origin='upper')
ax.clabel(CS, inline=1, fontsize=6)
cbar = fig.colorbar(image)
# Print only one ytick label when there is only one value to be shown
if vmin == vmax:
cbar.set_ticks([vmin])
# Draw floorplan itself to the lowest layer with full opacity
ax.imshow(self._layout, interpolation='bicubic', zorder=1, alpha=1)
labelsize = FontManager.get_default_size() * 0.4
if(self._showpoints):
# begin plotting points
for idx in range(0, len(a['x'])):
if (a['x'][idx], a['y'][idx]) in self._corners:
continue
ax.plot(
a['x'][idx], a['y'][idx], zorder=200,
marker='o', markeredgecolor='black', markeredgewidth=1,
markerfacecolor=mapper.to_rgba(a[key][idx]), markersize=6
)
ax.text(
a['x'][idx], a['y'][idx] - 30,
a['ap'][idx], fontsize=labelsize,
horizontalalignment='center'
)
# end plotting points
fname = '%s_%s.png' % (key, self._title)
logger.info('Writing plot to: %s', fname)
pp.savefig(fname, dpi=300)
pp.close('all')
import matplotlib
matplotlib.use('TkAgg')
from matplotlib.textpath import TextPath
from matplotlib.patches import PathPatch
from matplotlib.transforms import Affine2D, Bbox
from matplotlib.font_manager import FontManager, FontProperties
from matplotlib.colors import to_rgb
import matplotlib.pyplot as plt
from matplotlib.axes import Axes
from logomaker.src.error_handling import check, handle_errors
from logomaker.src.colors import get_rgb
import numpy as np
# Create global font manager instance. This takes a second or two
font_manager = FontManager()
# Create global list of valid font weights
VALID_FONT_WEIGHT_STRINGS = [
'ultralight', 'light', 'normal', 'regular', 'book', 'medium', 'roman',
'semibold', 'demibold', 'demi', 'bold', 'heavy', 'extra bold', 'black'
]
def list_font_names():
"""
Returns a list of valid font_name options for use in Glyph or
Logo constructors.
parameters
----------
