def __init__(self, color_to_words, default_color):
from wordcloud import get_single_color_func
self.color_func_to_words = [(get_single_color_func(color), set(words))
for (color,
words) in color_to_words.items()]
self.default_color_func = get_single_color_func(default_color)
def __init__(self, color_to_words, default_color):
self.color_func_to_words = [
(wordcloud.get_single_color_func(color), set(words))
for (color, words) in color_to_words.items()
]
self.default_color_func = wordcloud.get_single_color_func(
default_color)
def test_single_color_func_grey():
# grey is special as it's a corner case
random = Random(42)
red_function = get_single_color_func('darkgrey')
assert_equal(red_function(random_state=random), 'rgb(181, 181, 181)')
assert_equal(red_function(random_state=random), 'rgb(56, 56, 56)')
def parse_args(arguments):
# prog = 'python wordcloud_cli.py'
parser = make_parser()
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
args = vars(args)
with args.pop('text') as f:
text = f.read()
if args['stopwords']:
with args.pop('stopwords') as f:
args['stopwords'] = set(map(lambda l: l.strip(), f.readlines()))
if args['mask']:
mask = args.pop('mask')
args['mask'] = np.array(Image.open(mask))
color_func = wc.random_color_func
colormask = args.pop('colormask')
color = args.pop('color')
if colormask:
image = np.array(Image.open(colormask))
color_func = wc.ImageColorGenerator(image)
if color:
color_func = wc.get_single_color_func(color)
args['color_func'] = color_func
imagefile = args.pop('imagefile')
return args, text, imagefile
def test_single_color_func_grey():
# grey is special as it's a corner case
random = Random(42)
red_function = get_single_color_func('darkgrey')
assert red_function(random_state=random) == 'rgb(181, 181, 181)'
assert red_function(random_state=random) == 'rgb(56, 56, 56)'
def parse_args(arguments):
# prog = 'python wordcloud_cli.py'
parser = make_parser()
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
args = vars(args)
with args.pop('text') as f:
text = f.read()
if args['stopwords']:
with args.pop('stopwords') as f:
args['stopwords'] = set(map(lambda l: l.strip(), f.readlines()))
if args['mask']:
mask = args.pop('mask')
args['mask'] = np.array(Image.open(mask))
color_func = wc.random_color_func
colormask = args.pop('colormask')
color = args.pop('color')
if colormask:
image = np.array(Image.open(colormask))
color_func = wc.ImageColorGenerator(image)
if color:
color_func = wc.get_single_color_func(color)
args['color_func'] = color_func
imagefile = args.pop('imagefile')
return args, text, imagefile
def parse_args(arguments):
prog = 'python wordcloud_cli.py'
description = ('A simple command line interface for wordcloud module.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument('--text', metavar='file', type=argparse.FileType(), default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument('--stopwords', metavar='file', type=argparse.FileType(),
help='specify file of stopwords (containing one word per line) to remove from the given text after parsing')
parser.add_argument('--imagefile', metavar='file', type=argparse.FileType('w'), default='-',
help='file the completed PNG image should be written to (default: stdout)')
parser.add_argument('--fontfile', metavar='path', dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument('--mask', metavar='file', type=argparse.FileType(),
help='mask to use for the image form')
parser.add_argument('--colormask', metavar='file', type=argparse.FileType(),
help='color mask to use for image coloring')
parser.add_argument('--relative_scaling', type=float, default=0,
metavar='rs', help=' scaling of words by frequency (0 - 1)')
parser.add_argument('--margin', type=int, default=2,
metavar='width', help='spacing to leave around words')
parser.add_argument('--width', type=int, default=400,
metavar='width', help='define output image width')
parser.add_argument('--height', type=int, default=200,
metavar='height', help='define output image height')
parser.add_argument('--color', metavar='color',
help='use given color as coloring for the image - accepts any value from PIL.ImageColor.getcolor')
parser.add_argument('--background', metavar='color', default='black', type=str, dest='background_color',
help='use given color as background color for the image - accepts any value from PIL.ImageColor.getcolor')
parser.add_argument('--no_collocations', action='store_true',
help='do not add collocations (bigrams) to word cloud (default: add unigrams and bigrams)')
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
with args.text:
args.text = args.text.read()
if args.stopwords:
with args.stopwords:
args.stopwords = set(map(str.strip, args.stopwords.readlines()))
if args.mask:
args.mask = np.array(Image.open(args.mask))
color_func = wc.random_color_func
if args.colormask:
image = np.array(Image.open(args.colormask))
color_func = wc.ImageColorGenerator(image)
if args.color:
color_func = wc.get_single_color_func(args.color)
args.collocations = not args.no_collocations
args.color_func = color_func
return args
def test_single_color_func():
# test single color function for different color formats
random = Random(42)
red_function = get_single_color_func('red')
assert_equal(red_function(random_state=random), 'rgb(181, 0, 0)')
hex_function = get_single_color_func('#00b4d2')
assert_equal(hex_function(random_state=random), 'rgb(0, 48, 56)')
rgb_function = get_single_color_func('rgb(0,255,0)')
assert_equal(rgb_function(random_state=random), 'rgb(0, 107, 0)')
rgb_perc_fun = get_single_color_func('rgb(80%,60%,40%)')
assert_equal(rgb_perc_fun(random_state=random), 'rgb(97, 72, 48)')
hsl_function = get_single_color_func('hsl(0,100%,50%)')
assert_equal(hsl_function(random_state=random), 'rgb(201, 0, 0)')
def test_single_color_func():
# test single color function for different color formats
random = Random(42)
red_function = get_single_color_func('red')
assert red_function(random_state=random) == 'rgb(181, 0, 0)'
hex_function = get_single_color_func('#00b4d2')
assert hex_function(random_state=random) == 'rgb(0, 48, 56)'
rgb_function = get_single_color_func('rgb(0,255,0)')
assert rgb_function(random_state=random) == 'rgb(0, 107, 0)'
rgb_perc_fun = get_single_color_func('rgb(80%,60%,40%)')
assert rgb_perc_fun(random_state=random) == 'rgb(97, 72, 48)'
hsl_function = get_single_color_func('hsl(0,100%,50%)')
assert hsl_function(random_state=random) == 'rgb(201, 0, 0)'
def run(self):
matplotlib.rcParams['figure.figsize'] = (10.0, 5.0)
file1 = open("D:\\PycharmProject\\spider\\comment\\allcomment.txt", 'r')
xt = file1.read()
pattern = re.compile(r'[\u4e00-\u9fa5]+')
filedata = re.findall(pattern, xt)
finaldata = ''.join(filedata)
file1.close()
#读取文件
file2 = open("D:\\PycharmProject\\spider\\comment\\allcomment_worked.txt", 'w')
file2.write(finaldata)
clear = jieba.cut(finaldata)
cleared = pd.DataFrame({'clear': list(clear)})
#print(clear)
stopwords = pd.read_csv("chineseStopWords.txt", index_col=False, quoting=3, sep="\t", names=['stopword'], encoding='GBK')
#添加额外停止词
stopwords = self.add_stopword(self.str_list,stopwords)
# print(str(stopwords))
cleared = cleared[~cleared.clear.isin(stopwords.stopword)]
#清洗数据
count_words=cleared.groupby(by=['clear'])['clear'].agg({"num": numpy.size})
count_words=count_words.reset_index().sort_values(by=["num"], ascending=False)
#词云展示
# RGB_coloring =numpy.array(Image.open(path.join("RGB.png")))
# image_colors = ImageColorGenerator(RGB_coloring)
wordcloud=WordCloud(color_func=get_single_color_func('red'),font_path="simhei.ttf",background_color="white",max_font_size=300,height=800,width=1000,random_state=None) #指定字体类型、字体大小和字体颜色
word_frequence = {x[0]:x[1] for x in count_words.head(400).values}
wordcloud=wordcloud.fit_words(word_frequence)
plt.imshow(wordcloud,interpolation="bilinear")
# plt.imshow(wordcloud.recolor(color_func=get_single_color_func('white')),interpolation="bilinear")
plt.axis("off")
plt.savefig("D:\\PycharmProject\\spider\\comment\\wordcloud.jpg")
plt.close()
#词频统计
file3 = open("D:\\PycharmProject\\spider\\comment\\word_counting.txt",'w')
file3.write("词频统计结果：\n")
file3.close()
file3 = open("D:\\PycharmProject\\spider\\comment\\word_counting.txt",'a')
count = 0
name_list = []
num_list = []
for y in count_words.head(40).values:
count = count + 1
file3.write(str(count) + str(y[0])+":"+str(y[1])+"\n")
name_list.append(str(y[0]))
num_list.append(int(y[1]))
file3.close()
#词频统计条形图
plt.barh(range(len(num_list)), num_list,tick_label = name_list)
# 设置刻度字体大小
plt.xticks(fontsize=10)
plt.yticks(fontsize=7)
plt.savefig("D:\\PycharmProject\\spider\\comment\\counting.jpg")
def draw_wordcloud_single(fname,name,color,mtx):
wordcloud=WordCloud(
prefer_horizontal=1.0,color_func=get_single_color_func(color),
max_font_size=50, relative_scaling=0.3,font_path='font\\verdanab.ttf',
width=600,background_color=None,mode='RGBA'
).generate_from_frequencies(mtx)
plt.figure()
plt.imshow(wordcloud)
plt.axis("off")
plt.savefig('png\\'+fname+'_wcloud.png')
plt.close()
def parse_args(arguments):
prog = 'python wordcloud_cli.py'
description = ('A simple command line interface for wordcloud module.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument('--text', metavar='file', type=argparse.FileType(), default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument('--stopwords', metavar='file', type=argparse.FileType(),
help='specify file of stopwords (containing one word per line) to remove from the given text after parsing')
parser.add_argument('--imagefile', metavar='file', type=argparse.FileType('w'), default='-',
help='file the completed PNG image should be written to (default: stdout)')
parser.add_argument('--fontfile', metavar='path', dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument('--mask', metavar='file', type=argparse.FileType(),
help='mask to use for the image form')
parser.add_argument('--colormask', metavar='file', type=argparse.FileType(),
help='color mask to use for image coloring')
parser.add_argument('--relative_scaling', type=float, default=0,
metavar='rs', help=' scaling of words by frequency (0 - 1)')
parser.add_argument('--margin', type=int, default=2,
metavar='width', help='spacing to leave around words')
parser.add_argument('--width', type=int, default=400,
metavar='width', help='define output image width')
parser.add_argument('--height', type=int, default=200,
metavar='height', help='define output image height')
parser.add_argument('--color', metavar='color',
help='use given color as coloring for the image - accepts any value from PIL.ImageColor.getcolor')
parser.add_argument('--background', metavar='color', default='black', type=str, dest='background_color',
help='use given color as background color for the image - accepts any value from PIL.ImageColor.getcolor')
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
with args.text:
args.text = args.text.read()
if args.stopwords:
with args.stopwords:
args.stopwords = set(map(str.strip, args.stopwords.readlines()))
if args.mask:
args.mask = np.array(Image.open(args.mask))
color_func = wc.random_color_func
if args.colormask:
image = np.array(Image.open(args.colormask))
color_func = wc.ImageColorGenerator(image)
if args.color:
color_func = wc.get_single_color_func(args.color)
args.color_func = color_func
return args
def plotly_wordcloud(token_count_dic: dict) -> plotly.graph_objects.Scatter:
"""Create a world cloud trace for plotly.
Args:
token_count_dic (dictionary): Dictionary of token to its count
Returns:
trace (plotly.graph_objects.Scatter)
"""
wc = WordCloud(color_func=get_single_color_func("deepskyblue"), max_words=100)
wc.generate_from_frequencies(token_count_dic)
word_list = []
rel_freq_list = []
freq_list = []
fontsize_list = []
position_list = []
orientation_list = []
color_list = []
for (word, rel_freq), fontsize, position, orientation, color in wc.layout_:
word_list.append(word)
rel_freq_list.append(rel_freq)
freq_list.append(token_count_dic[word])
fontsize_list.append(fontsize)
position_list.append(position)
orientation_list.append(orientation)
color_list.append(color)
# get the positions
x = []
y = []
for i in position_list:
x.append(i[0])
y.append(i[1])
# get the relative occurrence frequencies
new_freq_list = []
for i in rel_freq_list:
i_tmp = round(i*100, 4)
i_tmp = i_tmp if i_tmp > 1 else 1 # Plotly textfont.size in go.Scatter throws exception for values below 1.
new_freq_list.append(i_tmp)
try:
trace = go.Scatter(
x=x,
y=y,
textfont=dict(size=new_freq_list, color=color_list),
hoverinfo="text",
hovertext=["{0}: {1}".format(w, f) for w, f in zip(word_list, freq_list)],
mode="text",
text=word_list,
)
return trace
except Exception as E:
logger.error(f'While creating the word cloud, plotly.go returned the following error \
\n{E}\nfor relative frequencies: {rel_freq_list}\nthat were mapped to {new_freq_list}')
def gen_word_cloud(words_path):
text, path = words_path
# Generate a word cloud image
wordcloud = WordCloud().generate(text)
# take relative word frequencies into account, lower max_font_size
wordcloud = WordCloud(width=250, height=125, prefer_horizontal=1, max_font_size=60, max_words=20,
min_font_size=20, relative_scaling=1, background_color=None, mode="RGBA",
color_func=wc.get_single_color_func("#242426"), stopwords=['bar', "the", "place"]+list(stops)).generate(text)
plt.figure(figsize=(2.5,1.25), dpi=100)
plt.imshow(wordcloud)
plt.axis("off")
plt.savefig('/home/carlson/web/'+path, transparent=True, bbox_inches='tight')
def __init__(self, colour_words_dict, default):
self.colour_words_dict = [
(get_single_color_func(colour), set(words))
for (colour, words) in colour_words_dict.items()]
self.default = get_single_color_func(default)
def __init__(self, cdict, default_color):
self.cdict = cdict
self.color_func_to_words = [(get_single_color_func(color), words)
for (color, words) in self.cdict.items()]
self.default_color_func = get_single_color_func(default_color)
def __init__(self, color_to_words, default_color):
self.color_func_to_words = [(get_single_color_func(color), set(words))
for (color, words) in color_to_words.items()]
self.default_color_func = get_single_color_func(default_color)
def __init__(self, positive, neutral, negative):
self.positive_func = get_single_color_func(positive)
self.neutral_func = get_single_color_func(neutral)
self.negative_func = get_single_color_func(negative)
def parse_args(arguments):
# prog = 'python wordcloud_cli.py'
description = ('A simple command line interface for wordcloud module.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument(
'--text', metavar='file', type=FileType(), default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument(
'--stopwords', metavar='file', type=FileType(),
help='specify file of stopwords (containing one word per line)'
' to remove from the given text after parsing')
parser.add_argument(
'--wordlist', metavar='file', type=FileType(), default=None,
help='specify a list of words and frequencies to build the word cloud')
parser.add_argument(
'--colormode', metavar='colormode', type=str, default='RGB',
help='specify the color mode (for transparent background use RGBA'
'and background color that is either transparent, or equal to "None")')
parser.add_argument(
'--imagefile', metavar='file', type=argparse.FileType('wb'),
default='-',
help='file the completed PNG image should be written to'
' (default: stdout)')
parser.add_argument(
'--fontfile', metavar='path', dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument(
'--mask', metavar='file', type=argparse.FileType('rb'),
help='mask to use for the image form')
parser.add_argument(
'--colormask', metavar='file', type=argparse.FileType('rb'),
help='color mask to use for image coloring')
parser.add_argument(
'--relative_scaling', type=float, default=0,
metavar='rs', help=' scaling of words by frequency (0 - 1)')
parser.add_argument(
'--margin', type=int, default=2,
metavar='width', help='spacing to leave around words')
parser.add_argument(
'--width', type=int, default=400,
metavar='width', help='define output image width')
parser.add_argument(
'--height', type=int, default=200,
metavar='height', help='define output image height')
parser.add_argument(
'--color', metavar='color',
help='use given color as coloring for the image -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--background', metavar='color', default='black', type=str,
dest='background_color',
help='use given color as background color for the image -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--no_collocations', action='store_true',
help='do not add collocations (bigrams) to word cloud '
'(default: add unigrams and bigrams)')
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
if not args.wordlist is None:
wordlist = []
with args.wordlist as wordlist_file:
for line in wordlist_file:
line = line.strip()
if(not len(line)): continue
items = line.split(';')
wordlist.append((items[0], int(items[1])))
args.wordlist = dict(wordlist)
else:
with args.text:
args.text = args.text.read()
if args.background_color == 'None' or args.background_color == 'none':
args.background_color = "#ffffff00"
if args.stopwords:
with args.stopwords:
args.stopwords = set(map(str.strip, args.stopwords.readlines()))
if args.mask:
args.mask = np.array(Image.open(args.mask))
color_func = wc.random_color_func
if args.colormask:
image = np.array(Image.open(args.colormask))
color_func = wc.ImageColorGenerator(image)
if args.color:
color_func = wc.get_single_color_func(args.color)
args.collocations = not args.no_collocations
args.color_func = color_func
return args
'''
设置背景
scipy.misc imread()：返回的是 numpy.ndarray 也即 numpy 下的多维数组对象
'''
backgroud_Image = plt.imread('/Users/joshuazk/Downloads/back_image.jpg')
wc = WordCloud(
background_color='white', # 设置背景颜色
mask=backgroud_Image, # 设置背景图片
font_path='wryh.ttf', # 若是有中文的话，这句代码必须添加，不然会出现方框，不出现汉字
max_words=100, # 设置最大显示词数
stopwords=STOPWORDS, # 设置停用词
max_font_size=100, # 设置字体最大值
# min_font_size=20,
random_state=10 # 设置有多少种随机生成状态，即有多少种配色方案
# color_func = get_single_color_func('blue')
)
# wc.generate_from_text(text)
wc.generate_from_frequencies(da['job_name'].value_counts())
img_colors = ImageColorGenerator(backgroud_Image)
wc.recolor(color_func=img_colors)
plt.imshow(wc.recolor(color_func=get_single_color_func('green')),
interpolation="bilinear")
plt.axis('off')
plt.show()
d = path.dirname(__file__)
wc.to_file(path.join(d, "jobspider.jpg"))
# words = da['job_name']
def word_cloud_view(request):
# if Generate Word Cloud is submitted
if 'word_cloud_submit' in request.POST:
word_cloud_dict_updated_1 = {}
word_cloud_dict_updated_2 = {}
word_cloud_dict_3 = {}
# dictionary containing common tokens and frequencies to both country 1 and country 2
for key_1 in word_cloud_dict_1:
for key_2 in word_cloud_dict_2:
if key_1 == key_2:
word_cloud_dict_3[key_1] = word_cloud_dict_1[
key_1] + word_cloud_dict_2[key_2]
# dictionary containing tokens and frequencies exclusive to country 1
for key_1 in word_cloud_dict_1:
if key_1 not in word_cloud_dict_3:
word_cloud_dict_updated_1[key_1] = word_cloud_dict_1[key_1]
# dictionary containing tokens and frequencies exclusive to country 2
for key_2 in word_cloud_dict_2:
if key_2 not in word_cloud_dict_3:
word_cloud_dict_updated_2[key_2] = word_cloud_dict_2[key_2]
plt.clf()
# word cloud for country 1 terms colored in dark blue
color_func1 = get_single_color_func('darkblue')
# word cloud for country 1 terms colored in dark red
color_func2 = get_single_color_func('darkred')
# word cloud for common terms colored in dark green
color_func3 = get_single_color_func('darkgreen')
# generate word cloud for country 1 terms
wordcloud_1 = WordCloud(
background_color='#DCDCDC',
color_func=color_func1).generate_from_frequencies(
word_cloud_dict_updated_1)
# generate word cloud for country 2 terms
wordcloud_2 = WordCloud(
background_color='#DCDCDC',
color_func=color_func2).generate_from_frequencies(
word_cloud_dict_updated_2)
# generate word cloud for common terms
wordcloud_3 = WordCloud(
background_color='#DCDCDC',
color_func=color_func3).generate_from_frequencies(
word_cloud_dict_3)
plt.axis('off')
# image link for word cloud for country 1 terms in static folder
img_link_1 = dir_name + '/static/images/wordcloud_1.png'
# image link for word cloud for country 2 terms in static folder
img_link_2 = dir_name + '/static/images/wordcloud_2.png'
# image link word cloud for country 3 terms in static folder
img_link_3 = dir_name + '/static/images/wordcloud_3.png'
# save word cloud for country 1 terms
plt.imshow(wordcloud_1)
plt.savefig(img_link_1)
img_1 = Image.open(img_link_1)
# call trim()
img_1 = trim(img_1)
img_1.save(img_link_1)
# save word cloud for country 2 terms
plt.imshow(wordcloud_2)
plt.savefig(img_link_2)
img_2 = Image.open(img_link_2)
# call trim()
img_2 = trim(img_2)
img_2.save(img_link_2)
# save word cloud for common terms
plt.imshow(wordcloud_3)
plt.savefig(img_link_3)
img_3 = Image.open(img_link_3)
# call trim()
img_3 = trim(img_3)
img_3.save(img_link_3)
# pass data as context to word_cloud.html
context_3 = {
"country_1_name": country_1_name,
"country_2_name": country_2_name,
"category_name": category_name
}
return render(request, 'word_cloud.html', context_3)
def parse_args(arguments):
prog = 'python tag2wc.py'
description = ('A simple command line interface for tags to wordcloud.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument(
'-t',
'--text',
metavar='file',
type=argparse.FileType(),
default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument(
'-p',
'--image-file',
metavar='file',
dest='imagefile',
type=argparse.FileType('w'),
default='-',
help=
'file the completed PNG image should be written to (default: stdout)')
parser.add_argument(
'-f',
'--font-file',
metavar='path',
dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument('-m',
'--mask-file',
metavar='file',
dest="mask",
type=argparse.FileType(),
help='mask to use for the image form')
parser.add_argument(
'-s',
'--stopwords',
metavar='file',
type=argparse.FileType(),
help=
'specify file of stopwords (containing one word per line) to remove from the given text after parsing'
)
parser.add_argument('-M',
'--colormask',
metavar='[file]',
nargs="*",
help='color mask to use for image coloring')
parser.add_argument('-S',
'--relative_scaling',
type=float,
default=1,
metavar='rs',
help=' scaling of words by frequency (0 - 1)')
parser.add_argument('--margin',
type=int,
default=2,
metavar='size',
help='spacing to leave around words')
parser.add_argument('--max-words',
type=int,
default=500,
dest="max_words",
metavar="mw",
help="define output max words count")
parser.add_argument('--max-font-size',
type=int,
default=80,
dest="max_font_size",
metavar='mfs',
help='define output max font size')
parser.add_argument('--min-font-size',
type=int,
default=2,
dest="min_font_size",
metavar='mfs',
help='define output min font size')
parser.add_argument('-W',
'--width',
type=int,
default=750,
metavar='width',
help='define output image width')
parser.add_argument('-H',
'--height',
type=int,
default=750,
metavar='height',
help='define output image height')
parser.add_argument('--random-state',
type=int,
default=30,
dest="random_state",
metavar='rs',
help='define output random_state')
parser.add_argument(
'-c',
'--color',
metavar='color',
help=
'use given color as coloring for the image - accepts any value from PIL.ImageColor.getcolor'
)
parser.add_argument(
'-b',
'--background',
metavar='color',
default='white',
type=str,
dest='background_color',
help=
'use given color as background color for the image - accepts any value from PIL.ImageColor.getcolor'
)
parser.add_argument(
'--no_collocations',
action='store_true',
help=
'do not add collocations (bigrams) to word cloud (default: add unigrams and bigrams)'
)
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
with args.text:
args.text = txt2dictOrList(args.text.read().decode('utf-8'))
if args.stopwords:
with args.stopwords:
args.stopwords = set(map(str.strip, args.stopwords.readlines()))
print(args.stopwords)
if args.mask:
args.mask = np.array(Image.open(args.mask))
color_func = WC.random_color_func
if args.color:
color_func = WC.get_single_color_func(args.color)
if args.colormask == None:
pass
# print("NONE ")
elif len(args.colormask) == 0:
if args.mask != None:
color_func = WC.ImageColorGenerator(args.mask)
# print "Empty [] SAME AS mask"
elif len(args.colormask) == 1:
cmfile = args.colormask[0]
image = np.array(Image.open(cmfile))
color_func = WC.ImageColorGenerator(image)
# print args.colormask
args.collocations = not args.no_collocations
args.color_func = color_func
return args
a1[i] = 1
else:
a1[i] += 1
bb = sorted(a1.items(), key=lambda x: x[1], reverse=True)
cc = [[x[0].encode('gbk', 'ignore'), str(x[1])] for x in bb]
with open('results_ciping.csv', 'w') as f:
writer = csv.writer(f, lineterminator='\n')
writer.writerows(cc)
# a1 = {u'你好': 13, u'再见': 8, u'滚蛋': 3}
a1 = {}
with open('results_ciping.csv') as f:
tmp = csv.reader(f)
for i in tmp:
if len(i[1]) > 0:
a1[i[0].decode('gbk')] = int(i[1])
for i in a1:
print i + '|' + str(a1[i])
wc = WordCloud(background_color='white',
max_words=1000,
mask=football,
color_func=get_single_color_func('black'),
font_path="FZLTZHUNHJW.TTF",
min_font_size=10,
max_font_size=150,
width=600)
wc.fit_words(a1)
picture_name = 'results.png'
wc.to_file(picture_name)
img = Image.open(picture_name)
img.show()
print u'词云生产完毕'
bar_plot_3 = alt.Chart(df).mark_bar().encode(x="Region",
y="Deaths",
tooltip='Deaths',
color="Region")
bar_plot = bar_plot_1 | bar_plot_2 | bar_plot_3
st.write(
scatter.encode(color=alt.condition(
picked, "Region", alt.value("lightgrey"))).add_selection(picked)
& bar_plot.transform_filter(picked))
# Part III: Word Cloud!
# load interactivity elements
st.cache(suppress_st_warning=True)
st.header('Word Usage in #Covid-19 Tweets (Jan-Mar)')
color_func_twit = wordcloud.get_single_color_func("#00acee")
st.sidebar.write("Choose Word Cloud Options")
remove_eng = st.sidebar.checkbox("Remove English Stop Words")
remove_esp = st.sidebar.checkbox("Remove Spanish Stop Words")
show_chart = st.button('Show Distribution')
slider_ph = st.empty()
value = slider_ph.slider("Choose Max Frequency",
min_value=min_val,
max_value=5 * min_val,
value=2 * min_val,
step=10)
# user text input
custom = st.sidebar.text_input('Add Custom Stopwords (comma separated)')
custom = custom.split(',')
def parse_args(arguments):
# prog = 'python wordcloud_cli.py'
description = ('A simple command line interface for wordcloud module.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument(
'--text', metavar='file', type=FileType(), default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument(
'--regexp', metavar='regexp', default=None, action=RegExpAction,
help='override the regular expression defining what constitutes a word')
parser.add_argument(
'--stopwords', metavar='file', type=FileType(),
help='specify file of stopwords (containing one word per line)'
' to remove from the given text after parsing')
parser.add_argument(
'--imagefile', metavar='file', type=argparse.FileType('wb'),
default='-',
help='file the completed PNG image should be written to'
' (default: stdout)')
parser.add_argument(
'--fontfile', metavar='path', dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument(
'--mask', metavar='file', type=argparse.FileType('rb'),
help='mask to use for the image form')
parser.add_argument(
'--colormask', metavar='file', type=argparse.FileType('rb'),
help='color mask to use for image coloring')
parser.add_argument(
'--contour_width', metavar='width', default=0, type=float,
dest='contour_width',
help='if greater than 0, draw mask contour (default: 0)')
parser.add_argument(
'--contour_color', metavar='color', default='black', type=str,
dest='contour_color',
help='use given color as mask contour color -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--relative_scaling', type=float, default=0,
metavar='rs', help=' scaling of words by frequency (0 - 1)')
parser.add_argument(
'--margin', type=int, default=2,
metavar='width', help='spacing to leave around words')
parser.add_argument(
'--width', type=int, default=400,
metavar='width', help='define output image width')
parser.add_argument(
'--height', type=int, default=200,
metavar='height', help='define output image height')
parser.add_argument(
'--color', metavar='color',
help='use given color as coloring for the image -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--background', metavar='color', default='black', type=str,
dest='background_color',
help='use given color as background color for the image -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--no_collocations', action='store_false', dest='collocations',
help='do not add collocations (bigrams) to word cloud '
'(default: add unigrams and bigrams)')
parser.add_argument(
'--version', action='version',
version='%(prog)s {version}'.format(version=__version__))
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
args = vars(args)
with args.pop('text') as f:
text = f.read()
if args['stopwords']:
with args.pop('stopwords') as f:
args['stopwords'] = set(map(str.strip, f.readlines()))
if args['mask']:
mask = args.pop('mask')
args['mask'] = np.array(Image.open(mask))
color_func = wc.random_color_func
colormask = args.pop('colormask')
color = args.pop('color')
if colormask:
image = np.array(Image.open(colormask))
color_func = wc.ImageColorGenerator(image)
if color:
color_func = wc.get_single_color_func(color)
args['color_func'] = color_func
imagefile = args.pop('imagefile')
return args, text, imagefile
def __init__(self, color_to_words={}, default_color='grey'):
self.color_func_to_words = [(get_single_color_func(color), set(words))
for (color,
words) in color_to_words.items()]
self.default_color_func = get_single_color_func(default_color)
import wordcloud
# 创建词云对象
word = wordcloud.WordCloud(width=500,
height=500,
prefer_horizontal=0.2,
min_font_size=2,
scale=2,
max_words=200,
stopwords=['One', 'night'],
mode='RGBA',
background_color=None)
# 传入需要制作词云的文本
word.generate(
'this is my house, i am leborn james One day,we donlt have to say goodoye,just say good night.'
)
# 将生成的词云保存为图片，保存路径在当前目录的文件夹下
image = word.to_image()
image.save('ss.png')
# 图像颜色生成器
# image_generator = wordcloud.ImageColorGenerator()
# 随机颜色，色相生成器
wordcloud.random_color_func()
# 创建一个颜色函数，该函数返回单个色调和饱和度
color_func1 = wordcloud.get_single_color_func('deepskyblue')
def parse_args(arguments):
# prog = 'python wordcloud_cli.py'
description = ('A simple command line interface for wordcloud module.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument(
'--text', metavar='file', type=FileType(), default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument(
'--regexp', metavar='regexp', default=None, action=RegExpAction,
help='override the regular expression defining what constitutes a word')
parser.add_argument(
'--stopwords', metavar='file', type=FileType(),
help='specify file of stopwords (containing one word per line)'
' to remove from the given text after parsing')
parser.add_argument(
'--imagefile', metavar='file', type=argparse.FileType('wb'),
default='-',
help='file the completed PNG image should be written to'
' (default: stdout)')
parser.add_argument(
'--fontfile', metavar='path', dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument(
'--mask', metavar='file', type=argparse.FileType('rb'),
help='mask to use for the image form')
parser.add_argument(
'--colormask', metavar='file', type=argparse.FileType('rb'),
help='color mask to use for image coloring')
parser.add_argument(
'--contour_width', metavar='width', default=0, type=float,
dest='contour_width',
help='if greater than 0, draw mask contour (default: 0)')
parser.add_argument(
'--contour_color', metavar='color', default='black', type=str,
dest='contour_color',
help='use given color as mask contour color -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--relative_scaling', type=float, default=0,
metavar='rs', help=' scaling of words by frequency (0 - 1)')
parser.add_argument(
'--margin', type=int, default=2,
metavar='width', help='spacing to leave around words')
parser.add_argument(
'--width', type=int, default=400,
metavar='width', help='define output image width')
parser.add_argument(
'--height', type=int, default=200,
metavar='height', help='define output image height')
parser.add_argument(
'--color', metavar='color',
help='use given color as coloring for the image -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--background', metavar='color', default='black', type=str,
dest='background_color',
help='use given color as background color for the image -'
' accepts any value from PIL.ImageColor.getcolor')
parser.add_argument(
'--no_collocations', action='store_false', dest='collocations',
help='do not add collocations (bigrams) to word cloud '
'(default: add unigrams and bigrams)')
parser.add_argument('--version', action='version',
version='%(prog)s {version}'.format(version=__version__))
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
args = vars(args)
with args.pop('text') as f:
text = f.read()
if args['stopwords']:
with args.pop('stopwords') as f:
args['stopwords'] = set(map(str.strip, f.readlines()))
if args['mask']:
mask = args.pop('mask')
args['mask'] = np.array(Image.open(mask))
color_func = wc.random_color_func
colormask = args.pop('colormask')
color = args.pop('color')
if colormask:
image = np.array(Image.open(colormask))
color_func = wc.ImageColorGenerator(image)
if color:
color_func = wc.get_single_color_func(color)
args['color_func'] = color_func
imagefile = args.pop('imagefile')
return args, text, imagefile
def parse_args(arguments):
prog = 'python tag2wc.py'
description = ('A simple command line interface for tags to wordcloud.')
parser = argparse.ArgumentParser(description=description)
parser.add_argument('-t', '--text', metavar='file', type=argparse.FileType(), default='-',
help='specify file of words to build the word cloud (default: stdin)')
parser.add_argument('-p', '--image-file', metavar='file', dest='imagefile', type=argparse.FileType('w'), default='-',
help='file the completed PNG image should be written to (default: stdout)')
parser.add_argument('-f', '--font-file', metavar='path', dest='font_path',
help='path to font file you wish to use (default: DroidSansMono)')
parser.add_argument('-m', '--mask-file', metavar='file', dest="mask", type=argparse.FileType(),
help='mask to use for the image form')
parser.add_argument('-s', '--stopwords', metavar='file', type=argparse.FileType(),
help='specify file of stopwords (containing one word per line) to remove from the given text after parsing')
parser.add_argument('-M', '--colormask', metavar='[file]', nargs="*",
help='color mask to use for image coloring')
parser.add_argument('--relative_scaling', type=float, default=0,
metavar='rs', help=' scaling of words by frequency (0 - 1)')
parser.add_argument('--margin', type=int, default=2,
metavar='size', help='spacing to leave around words')
parser.add_argument('--max-words', type=int, default=500, dest="max_words",
metavar="mw", help="define output max words count")
parser.add_argument('--max-font-size', type=int, default=80, dest="max_font_size",
metavar='mfs', help='define output max font size')
parser.add_argument('--min-font-size', type=int, default=2, dest="min_font_size",
metavar='mfs', help='define output min font size')
parser.add_argument('-W', '--width', type=int, default=750,
metavar='width', help='define output image width')
parser.add_argument('-H', '--height', type=int, default=750,
metavar='height', help='define output image height')
parser.add_argument('--random-state', type=int, default=30, dest="random_state",
metavar='rs', help='define output random_state')
parser.add_argument('-c', '--color', metavar='color',
help='use given color as coloring for the image - accepts any value from PIL.ImageColor.getcolor')
parser.add_argument('-b', '--background', metavar='color', default='white', type=str, dest='background_color',
help='use given color as background color for the image - accepts any value from PIL.ImageColor.getcolor')
parser.add_argument('--no_collocations', action='store_true',
help='do not add collocations (bigrams) to word cloud (default: add unigrams and bigrams)')
args = parser.parse_args(arguments)
if args.colormask and args.color:
raise ValueError('specify either a color mask or a color function')
with args.text:
args.text = txt2dictOrList(args.text.read().decode('utf-8'))
if args.stopwords:
with args.stopwords:
args.stopwords = set(map(str.strip, args.stopwords.readlines()))
if args.mask:
args.mask = np.array(Image.open(args.mask))
color_func = WC.random_color_func
if args.color:
color_func = WC.get_single_color_func(args.color)
if args.colormask == None:
pass
# print("NONE ")
elif len(args.colormask) == 0:
if args.mask != None:
color_func = WC.ImageColorGenerator(args.mask)
# print "Empty [] SAME AS mask"
elif len(args.colormask) == 1:
cmfile = args.colormask[0]
image = np.array(Image.open(cmfile))
color_func = WC.ImageColorGenerator(image)
# print args.colormask
args.collocations = not args.no_collocations
args.color_func = color_func
return args
print("图片打开成功")
#word
wc = WordCloud(font_path=font_path,
mask=mask,
width=1000,
height=1000,
background_color=background_color,
stopwords=STOPWORDS,
max_font_size=int(im.size[1] / 10),
min_font_size=15)
print("WordCloud创建成功")
#generate from dict
from threeKingdoms import counts
wc.generate_from_frequencies(counts)
print("词云生成成功")
#color
color_func = ImageColorGenerator(mask)
color_func2 = get_single_color_func('white')
wc.recolor(color_func=color_func2)
print("染色成功")
#show
# plt.imshow(wc)
# plt.axis('off')
# plt.show()
#save
wc.to_file("cloud_{}.jpg".format(name))
print("保存成功")
