def make_cloud(channel, time, myType=None, drawLabels=True, font_path=None):
#get the log file
directory = "logs/" + channel + '/' + time
if myType == None:
file_path = os.path.relpath(directory + '/words.log')
with open(file_path, 'r') as f:
words = f.read().upper()
file_path = os.path.relpath(directory + '/emotes.log')
with open(file_path, 'r') as f:
emotes = " ".join(filter(lambda x:len(x)>3 and x != 'double' and x != 'triple', f.read().split('\n')))
directory = "images/" + channel + '/' + time
if not os.path.exists(directory):
os.makedirs(directory)
print "Generating word cloud... Hold on! (This takes a while if there are a lot of messages)"
scale = 2
w = wordcloud.process_text(words, max_features=1500)
elements = wordcloud.fit_words(w, width=w_words/scale, height=h_words/scale)
wordcloud.draw(elements, os.path.relpath(directory + '/wordcloud.png'),
width=w_words/scale, height=h_words/scale, scale=scale)
print "Word cloud created!"
print "Generating emote cloud..."
w = wordcloud.process_text(emotes, max_features=1500)
elements = wordcloud.fit_words(w, width=w_emotes, height=h_emotes)
wordcloud.draw(elements, os.path.relpath(directory + '/emotecloud.png'),
width=w_emotes, height=h_emotes)
print "Emote cloud created!"
else: #if running the program manually. this is mainly for my debugging purposes.
w_custom = 1100
h_custom = 700
file_path = os.path.relpath(directory + '/'+myType+'.log')
directory = "images/" + channel + '/' + time
if not os.path.exists(directory):
os.makedirs(directory)
with open(file_path, 'r') as f:
data = f.read()
if myType.lower() == 'authors':
data = data.upper()
print "Generating " +myType+ " cloud... Hold on!"
scale = 2
w = wordcloud.process_text(data, max_features=1000)
elements = wordcloud.fit_words(w, width=w_custom/scale,
height=h_custom/scale, font_path=font_path)
wordcloud.draw(elements, os.path.relpath(directory + '/'+myType+'cloud.png'),
width=w_custom/scale, height=h_custom/scale, scale=scale, font_path=font_path)
print myType + " cloud created!"
def make_cloud(self, text): words = wordcloud.process_text(text) elements = wordcloud.fit_words(words, width = 400, height = 400) wordcloud.draw(elements, self.out, width = 400, height = 400, scale = 2) return self.out
def make_cloud(self, text):
words = wordcloud.process_text(text)
elements = wordcloud.fit_words(words, width=400, height=400)
wordcloud.draw(elements, self.out, width=400, height=400, scale=2)
return self.out
def analizer(query, num_topics, dictionary, corpus, alpha):
image_path = "/media/University/UniversityDisc/2-Master/MasterThesis/EjecucionTesis/Desarrollo/PythonProjects/QueryAnalyzer/Models/"
model_path = image_path
alpha=alpha
#lda = models.ldamodel.LdaModel(corpus, num_topics=num_topics, id2word=dictionary, update_every=1, chunksize=50, passes=1)
lda_2 = models.ldamodel.LdaModel(corpus, num_topics=num_topics, id2word=dictionary, alpha=alpha, update_every=1, chunksize=50, passes=1)
dictionary.save(model_path + "tmp_dictionary.dict")
corpora.MmCorpus.serialize(model_path + "tmp_corpus.mm", corpus)
lda_2.save(model_path + 'tmp_model.lda') # same for tfidf, lsi, ...
goals_distribution = ldam.perQueryGoalProportions(query, dictionary, lda_2)
max_goal = ldam.viewPerQueryGoalProportions(goals_distribution)
show_goal = lda_2.show_topic(max_goal)
#print show_goal
new_goal = []
for goal in show_goal:
weight = goal[0]
tag = goal[1]
new_goal.append((tag, weight))
# Compute the position of the words.
elements = wordcloud.fit_words(new_goal, width=100, height=100)
image_path = "/media/University/UniversityDisc/2-Master/MasterThesis/EjecucionTesis/Desarrollo/PythonProjects/QueryAnalyzer/Models/"
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(image_path + 'image.png'), width=100, height=100, scale=2)
lda_model = lda_2
return image_path, lda_model
def makeCloud(self, text, font=None):
if font is None:
font = random.choice(self.fonts)
words, counts = wordcloud.process_text(text, max_features=2000)
elements = wordcloud.fit_words(words, counts, width=self.size,
height=self.size, font_path=font)
wordcloud.draw(elements, self.outFile, width=self.size,
height=self.size, scale=self.scale, font_path=font)
def makeCloud(self, text, font=None):
if font is None:
font = random.choice(self.fonts)
words, counts = wordcloud.process_text(text, max_features=2000)
elements = wordcloud.fit_words(words, counts, width=self.size,
height=self.size, font_path=font)
wordcloud.draw(elements, self.outFile, width=self.size,
height=self.size, scale=self.scale, font_path=font)
def draw_goal(lsi, topic):
other_goal = lsi.show_topic(topic)
new_goal = []
image_path = "/media/University/UniversityDisc/2-Master/MasterThesis/EjecucionTesis/Desarrollo/PythonProjects/QueryAnalyzer/Models/"
for goal in other_goal: weight = goal[0];tag = goal[1];new_goal.append((tag, weight))
elements = wordcloud.fit_words(new_goal, width=100, height=100)
wordcloud.draw(elements, path.join(image_path + 'other_image.png'), width=100, height=100, scale=2)
def make_word_cloud(text, filepath):
import wordcloud #@UnresolvedImport
if isinstance(text, str):
text = wordcloud.process_text(text, max_features=20)
w, h = (400, 400)
text = remove_letters(text)
elements = wordcloud.fit_words(text, width=w, height=h)
wordcloud.draw(elements, filepath, width=w, height=h, scale=1)
return filepath
def makeDoge(self, words, sid):
image = random.choice(self.images)
img = Image.open(image['original'])
width, height = img.size
initialFontSize = int(height * self.config['initialFontSize'])
elements = wordcloud.fit_words(words, width=width, height=height,
font_path=self.config['fontPath'], prefer_horiz=1.0,
initial_font_size=initialFontSize)
imagePath = '{0}-{1}.png'.format(datetime.now().isoformat(), sid)
imagePath = path.join(self.doneDir, imagePath)
self.draw(image, elements, imagePath)
return imagePath
def words_cloud(args):
voc = sio.loadmat(args.vocabulary)[args.voc_key]
folds = [x for x in os.listdir(args.exp) if "fold" in x]
folds.sort()
if args.force_fold >= 0:
folds = [x for x in folds if "%d" % args.force_fold in x]
exp_opts = pickle.load(file(os.sep.join([args.exp, "cmd_opts"]), "rb"))
out_dir = os.sep.join([
args.out,
"u_lambda=%s" % exp_opts.u_lambda,
"w_lambda=%s" % exp_opts.w_lambda
])
if not os.path.exists(out_dir): os.makedirs(out_dir)
logger.info("Number of folds: %d" % len(folds))
for fold in folds:
fold_dir = os.sep.join([args.exp, fold])
epochs = [x for x in os.listdir(fold_dir) if "epoch" in x]
epochs.sort()
if args.force_epoch >= 0:
epochs = [x for x in epochs if "%d" % args.force_epoch in x]
for epoch_name in epochs:
ef_out_dir = os.sep.join([out_dir, fold, epoch_name])
if not os.path.exists(ef_out_dir): os.makedirs(ef_out_dir)
epoch_path = os.sep.join([fold_dir, epoch_name])
epoch = load_compressed(epoch_path)
region_task_words = scored_epoch_words(epoch, voc, args.nwords,
args.force_region)
for r in region_task_words:
for t in region_task_words[r]:
for pn in region_task_words[r][t]:
if args.for_wordle:
outname = "%s_r%dt%d.wordle" % (pn, r, t)
outfile = os.sep.join([ef_out_dir, outname])
f = file(outfile, "w")
for word, score in region_task_words[r][t][pn]:
f.write("%s: %2.5f\n" % (word, score))
f.close()
else:
outname = "%s_r%dt%d.png" % (pn, r, t)
fit_rtw = wordcloud.fit_words(
region_task_words[r][t][pn],
font_path=args.font)
wordcloud.draw(fit_rtw,
os.sep.join([ef_out_dir, outname]),
font_path=args.font)
def ldaGoalDistribution(goals_distribution, max_goal, image_path, lda_n, name):
show_goal = lda_n.show_topic(max_goal)
print show_goal
new_goal = []
for goal in show_goal:
weight = goal[0]
tag = goal[1]
new_goal.append((tag, weight))
# Compute the position of the words.
elements = wordcloud.fit_words(new_goal, width=100, height=100)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(image_path + name), width=100, height=100, scale=2)
def freq_words(string):
print "\n\n\n\t\t\tReading from file"
#tokenize on white spaces
raw_word_list=word_tokenize(string)
#remove stop words
processed_word_list=[word for word in raw_word_list if word not in total_stop_words]
#create an nltk text object
text_obj=nltk.Text(processed_word_list)
print "\n\n\n\t\t\tProcessing"
#Call the frequency distribution method and store the words and corresponding frequencies in a dictionary
fd=FreqDist(text_obj)
#convert the dictionary to a list of tuples conatining key-value pairs
result=fd.items()
#select the 100 most frequent words. If number of words in the result is less than 100, adjust accordingly
if len(result) < 100:
result_length=len(result)
chosen_words=result[: result_length/2]
else:
chosen_words=result[:100]
print "\n\n\n\t\t\tDrawing cloud"
#specify the canvas measurement
elements = wordcloud.fit_words(chosen_words, width=500, height=500)
#draw the cloud
wordcloud.draw(elements, path.join(d, 'frequent_words.png'), width=500, height=500,
scale=2)
print "\n\n\n\t\t\tWord cloud generated in frequent_words.png file"
return
def drawTags(model, lsi, query, dictionary, image_path, tfidf):
print "Init drawTags"
goals_distribution = model.perQueryGoalProportions(query, dictionary, tfidf, lsi)
max_goal = model.viewPerQueryGoalProportions(goals_distribution)
show_goal = lsi.show_topic(max_goal)
print show_goal
new_goal = []
for goal in show_goal:
weight = goal[0]
tag = goal[1]
new_goal.append((tag, weight))
# Compute the position of the words.
elements = wordcloud.fit_words(new_goal, width=100, height=100)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(image_path + 'lsa-image.png'), width=100, height=100, scale=2)
def main():
dictionary = gensim.corpora.Dictionary.load(
"dict_abstracts_corpus_1_cleaned.dict")
corpus = gensim.corpora.MmCorpus("abstracts_corpus_1_cleaned.mm")
lda = gensim.models.LdaModel.load("abstracts_corpus_1_cleaned.lda")
list_of_topics = lda.show_topics(100, formatted=False)
for i in range(len(list_of_topics)):
# Compute the position of the words.
elements = wordcloud.fit_words([(str(l[1]), l[0])
for l in list_of_topics[i]],
font_path="/Library/Fonts/Tahoma.ttf")
# Draw the positioned words to a PNG file.
wordcloud.draw(elements,
"../topic_%i.png" % i,
font_path="/Library/Fonts/Tahoma.ttf")
def words_cloud(args):
voc = sio.loadmat(args.vocabulary)[args.voc_key]
folds = [x for x in os.listdir(args.exp) if "fold" in x]
folds.sort()
if args.force_fold >= 0: folds = [x for x in folds if "%d"%args.force_fold in x]
exp_opts = pickle.load(file(os.sep.join([args.exp,"cmd_opts"]),"rb"))
out_dir = os.sep.join([args.out,"u_lambda=%s"%exp_opts.u_lambda,"w_lambda=%s"%exp_opts.w_lambda])
if not os.path.exists(out_dir): os.makedirs(out_dir)
logger.info("Number of folds: %d"%len(folds))
for fold in folds:
fold_dir = os.sep.join([args.exp,fold])
epochs = [x for x in os.listdir(fold_dir) if "epoch" in x]
epochs.sort()
if args.force_epoch >= 0: epochs = [x for x in epochs if "%d"%args.force_epoch in x]
for epoch_name in epochs:
ef_out_dir = os.sep.join([out_dir,fold,epoch_name])
if not os.path.exists(ef_out_dir): os.makedirs(ef_out_dir)
epoch_path = os.sep.join([fold_dir,epoch_name])
epoch = load_compressed(epoch_path)
region_task_words = scored_epoch_words(epoch,voc,args.nwords,args.force_region)
for r in region_task_words:
for t in region_task_words[r]:
for pn in region_task_words[r][t]:
if args.for_wordle:
outname = "%s_r%dt%d.wordle"%(pn,r,t)
outfile = os.sep.join([ef_out_dir,outname])
f = file(outfile,"w")
for word,score in region_task_words[r][t][pn]:
f.write("%s: %2.5f\n"%(word,score))
f.close()
else:
outname = "%s_r%dt%d.png"%(pn,r,t)
fit_rtw = wordcloud.fit_words(region_task_words[r][t][pn],font_path=args.font)
wordcloud.draw(fit_rtw, os.sep.join([ef_out_dir,outname]),font_path=args.font)
def plot_word_cloud(predata):
#词云图 根据词频
word_df=predata['Comment_Text'].copy()
wordlist=word_df.values.tolist()
wordlist1=','.join(wordlist)
segment=jieba.lcut(wordlist1)
words_df=pd.DataFrame({'segment':segment})
# print(words_df.head())
stopwords=pd.read_csv("C:\\Users\\13174\\Desktop\\my_flask\\flask_01_mysql_Css\\MLmodel\\hit_stopwords.txt",index_col=False,quoting=3,sep="\t",names=['stopword'], encoding='utf-8')#quoting=3全不引用
# stopwords=pd.read_csv("..\MLmodel\hit_stopwords.txt",index_col=False,quoting=3,sep="\t",names=['stopword'], encoding='utf-8')#quoting=3全不引用
# print(stopwords)
word_df1=words_df[~words_df.segment.isin(stopwords.stopword)]
#词频统计
words_stat = word_df1.groupby('segment').agg(计数=pd.NamedAgg(column='segment', aggfunc='size')).reset_index().sort_values(
by='计数', ascending=False)
# print(words_stat.head())
matplotlib.rcParams['figure.figsize'] = (10.0, 5.0)
wordcloud=WordCloud(font_path="simhei.ttf",background_color="white",max_font_size=80) #指定字体类型、字体大小和字体颜色
word_frequence = {x[0]:x[1] for x in words_stat.head(1000).values}
word_frequence_list = []
for key in word_frequence:
temp = (key,word_frequence[key])
word_frequence_list.append(temp)
word_frequence_list1=dict(word_frequence_list)
wordcloud=wordcloud.fit_words(word_frequence_list1)
plt.imshow(wordcloud)
plt.xticks([])#去除x坐标
plt.yticks([])#去除y坐标
# plt.savefig('../static/img/WordCloud.jpg')
plt.savefig('C:\\Users\\13174\\Desktop\\my_flask\\flask_01_mysql_Css\\static\\img\\WordCloud.jpg')
plt.show()
import os
import wordcloud
MODELS_DIR = "models"
final_topics = open(os.path.join(MODELS_DIR, "final_topics.txt"), 'rb')
curr_topic = 0
for line in final_topics:
line = line.strip()[line.rindex(":") + 2:]
scores = [float(x.split("*")[0]) for x in line.split(" + ")]
words = [x.split("*")[1] for x in line.split(" + ")]
freqs = []
for word, score in zip(words, scores):
freqs.append((word, score))
elements = wordcloud.fit_words(freqs, width=120, height=120)
wordcloud.draw(elements, "gs_topic_%d.png" % (curr_topic),
width=120, height=120)
curr_topic += 1
final_topics.close()
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
# print 'Number of arguments:', len(sys.argv), 'arguments.'
# print 'Argument List:', str(sys.argv)
# Experimenting with random seeds
import random
random.seed(42)
if(len(sys.argv) != 5):
print "[ USAGE ]: ", sys.argv[0], " <WordsFile> <OutputFile> <Width> <Height>"
sys.exit()
d = path.dirname(__file__)
# Read the whole text
text = open(path.join(d, sys.argv[1])).read()
# Separate into a list of (word, frequency).
words = wordcloud.process_text(text, max_features=500)
# Compute the position of the words.
elements = wordcloud.fit_words(words, width=int(sys.argv[3]), height=int(sys.argv[4]))
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(d, sys.argv[2]), width=int(sys.argv[3]), height=int(sys.argv[4]), scale=2)
with open(keysFile,'rU') as csvFile:
fileReader = csv.reader(csvFile, delimiter='\t', quotechar='|')
for row in fileReader:
topicKeys[int(row[0])] = row[2].split(' ')
# main work
for source in sources:
print "Processing %s..." %source
topics, weightSum = get_top_topics([source],docWeights,weightThreshold)
text = ""
for topic in topics:
t = " ".join(topicKeys[topic]) + " "
text += t
words = wordcloud.process_text(text)
elements = wordcloud.fit_words(words,font_path=fontPath)
outpath = "/Users/jchan/Desktop/Dropbox/Research/Dissertation/OpenIDEO/Pipeline/Validation/wordclouds/k400t50/%s.png" %source
wordcloud.draw(elements, outpath, font_path=fontPath)
## read in the concept list
#concepts = {}
#with open(conceptFile, 'rU') as csvfile:
# filereader = csv.reader(csvfile, delimiter=',', quotechar='|')
# for row in filereader:
# concepts[row[0]] = row[1]
#
## grab topic-keys -> hash: key = topic, value = list of words
#topicKeys = {}
#with open(keysFile,'rU') as csvFile:
# fileReader = csv.reader(csvFile, delimiter='\t', quotechar='|')
# for row in fileReader:
def generateCloud(text): dir = path.dirname(__file__) words = wordcloud.process_text(text, max_features=1000) elements = wordcloud.fit_words(words, width=1000, height=1000) wordcloud.draw(elements, path.join(dir, 'wordcloud.png'), width=1000, height=1000)
import os
import wordcloud
MODELS_DIR = "."
final_topics = open(os.path.join(MODELS_DIR, "final_topics.txt"), 'rb')
curr_topic = 0
for line in final_topics:
line = line.strip()[line.rindex(":") + 2:]
scores = [float(x.split("*")[0]) for x in line.split(" + ")]
words = [x.split("*")[1] for x in line.split(" + ")]
freqs = []
for word, score in zip(words, scores):
freqs.append((word, score))
elements = wordcloud.fit_words(freqs, width=120, height=120)
wordcloud.draw(elements,
"gs_topic_%d.png" % (curr_topic),
width=120,
height=120)
curr_topic += 1
final_topics.close()
__FILENAME__ = more
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
d = path.dirname(__file__)
# Read the whole text.
text = open(path.join(d, 'alice.txt')).read()
# Separate into a list of (word, frequency).
words = wordcloud.process_text(text, max_features=2000)
# Compute the position of the words.
elements = wordcloud.fit_words(words, width=500, height=500)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(d, 'alice.png'), width=500, height=500,
scale=2)
########NEW FILE########
__FILENAME__ = simple
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
d = path.dirname(__file__)
# Read the whole text.
text = open(path.join(d, 'constitution.txt')).read()
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
d = path.dirname(__file__)
# Read the whole text.
text = open(path.join(d, 'presinaug-addresses.txt')).read()
# Separate into a list of (word, frequency).
words = wordcloud.process_text(text, max_features=10000)
# Compute the position of the words.
elements = wordcloud.fit_words(words, width=900, height=1600)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements,
path.join(d, 'presinaug-wordcloud-1600x900.png'),
width=900,
height=1600,
scale=1)
def wordclouds(x):
d = path.dirname("/Users/MrG/Capstone/")
words = wordcloud.process_text(str(x), max_features = 500)
elements = wordcloud.fit_words(words)
wordcloud.draw(elements, path.join(d,"WC.png"), scale = 5)
return Image(filename='/Users/MrG/Capstone/WC.png', height= 1000, width= 618)
def produceWordCloud(inputText, outputPng): words = wordcloud.process_text(inputText, max_features=400) elements = wordcloud.fit_words(words, width=800, height=500) wordcloud.draw(elements, outputPng, width=800, height=500, scale=2)
# Cluster
km = KMeans(n_clusters=true_k, init='k-means++', max_iter=100, n_init=1, verbose=False)
km.fit(X)
# Create cluster outputs
output_dict = {'cluster': km.labels_, 'values': dataset}
output_df = pd.DataFrame(output_dict)
# Create text files
for i in range(true_k):
print len(output_df[output_df.cluster == i]), round(100*len(output_df[output_df.cluster == i]) / float(len(output_df)), 2)
cluster_text = output_df['values'][output_df.cluster == i].values
temp = "cluster " + str(i) + ".txt"
with open(temp, "w") as outfile:
for j in cluster_text:
outfile.write("%s\n" % j)
# Create wordclouds
for i in range(true_k):
text = open('cluster ' + str(i) + '.txt').read()
# Separate into a list of (word, frequency).
words = wordcloud.process_text(text)
# Compute the position of the words.
elements = wordcloud.fit_words(words, font_path='/Library/Fonts/Arial Black.ttf', width=600, height=300)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, 'cluster ' + str(i) + '.png', font_path="/Library/Fonts/Arial Black.ttf", width=600, height=300)
__FILENAME__ = more
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
d = path.dirname(__file__)
# Read the whole text.
text = open(path.join(d, 'alice.txt')).read()
# Separate into a list of (word, frequency).
words = wordcloud.process_text(text, max_features=2000)
# Compute the position of the words.
elements = wordcloud.fit_words(words, width=500, height=500)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements,
path.join(d, 'alice.png'),
width=500,
height=500,
scale=2)
########NEW FILE########
__FILENAME__ = simple
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
d = path.dirname(__file__)
for doc_top in topics: for ti,_ in doc_top: counts[ti] += 1 # most talked about topics words_max = model.show_topic(counts.argmax(), 50) # least talked about topics words_min = model.show_topic(counts.argmin(), 50) wf_max = [] wlist_max = [] for i,j in words_max: wlist_max.append(j) for i in range(50): wf_max.append((wlist_max[i],counts[i])) wf_min = [] wlist_min = [] for i,j in words_min: wlist_min.append(j) for i in range(50): wf_min.append((wlist_min[i],counts[i+50])) d = path.dirname(__file__) elements_max = wordcloud.fit_words(wf_max) wordcloud.draw(elements_max, path.join(d, 'top50.png'),scale=3) elements_min = wordcloud.fit_words(wf_min) wordcloud.draw(elements_min, path.join(d, 'bottom50.png'),scale=3)
#!/usr/bin/env python2
from os import path
import sys
import wordcloud
d = path.dirname(__file__)
# Read the whole text.
text = open(path.join(d, '4chdata/all.dat')).read()
# Separate into a list of (word, frequency).
words = wordcloud.process_text(text, max_features=1000)
# Compute the position of the words.
elements = wordcloud.fit_words(words, width=1000, height=1000)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(d, str(sys.argv[1])), width=1000, height=1000,
scale=2)
txtfeatWords5 = [lmtzr.lemmatize(word) for word in txtfeatWords4]
# stem using snowball stemmer
txtfeatWords6 = [stemmer.stem(word) for word in txtfeatWords5]
# remove punctuations
txtfeatWords7 = [
word.encode('utf-8').translate(None, string.punctuation)
for word in txtfeatWords6
]
# remove empty strings
txtfeatWords8 = [word for word in txtfeatWords7 if word <> '']
txtfeatWordList[i] = ' '.join(txtfeatWords8)
#pprint('Iteration: %d'% i)
#pprint(txtfeatWordList[i])
pprint(txtfeatWordList)
text = '\n'.join([str(txtfeatWordList[i]) for i in range(num_total)])
#tags = make_tags(get_tag_counts( '\n'.join([ str(txtfeatWordList[i]) for i in range(num_total) ]) ))
#create_tag_image(tags, 'cloud_large.png', size=(1800, 1200), fontname='Lobster')
d = os.path.dirname(__file__)
words = wordcloud.process_text(text)
elements = wordcloud.fit_words(words)
wordcloud.draw(elements, os.path.join(d, 'lemmatized_wordle.png'))
stopwords = pd.read_csv(
"D:\my_documents\competition\government\\stopwords_addition.txt",
encoding='utf8',
index_col=False,
quoting=3,
sep="\t")
segmentDF = segmentDF[~segmentDF.segment.isin(stopwords)]
segStat = segmentDF.groupby(by=["segment"])["segment"].agg({
"计数": numpy.size
}).reset_index().sort_values(["计数"], ascending=False)
segStat.head(100)
#绘画词云
#http://www.lfd.uci.edu/~gohlke/pythonlibs/
wordcloud = WordCloud(font_path='simhei.ttf', background_color="white")
words = segStat.set_index('segment').to_dict()
wordcloud = wordcloud.fit_words(words['计数'])
plt.figure(num=None, figsize=(100, 80), dpi=100, facecolor='w', edgecolor='k')
plt.axis("off")
plt.imshow(wordcloud)
plt.show()
plt.close()
d = path.dirname(__file__)
# String to hold the text from the webpages.
text = "";
# Array of webpages which we'll loop through (from googling Deonte Burton draft).
url_list = ["http://www.draftexpress.com/profile/Deonte-Burton-6487/",
"http://blogs.rgj.com/chrismurray/2014/01/10/nba-scouts-view-nevadas-deonte-burton-as-solid-draft-pick-but-not-a-first-rounder/",
"http://www.nbadraftroom.com/2014/01/deonte-burton.html",
"http://www.nbadraftinsider.com/deonte-burton/",
"http://nbaprospects.blogspot.com/2012/08/scouting-report-deonte-burton-nevada.html",
"http://rushthecourt.net/2014/01/09/a-college-basketball-resolution-for-2014-get-to-know-nevadas-deonte-burton/",
"http://mrsportsblog.wordpress.com/2014/03/05/trust-me-on-this-dynamic-deonte-burton-of-nevada-will-be-making-a-living-in-the-nba/",
"http://www.draftexpress.com/article/NBA-Draft-Prospect-of-the-Week-Deonte-Burton-4392/",
"http://www.nevadawolfpack.com/sports/m-baskbl/spec-rel/021214aad.html"]
# Loop through url items and get the text from each.
for url in url_list:
content = urllib2.urlopen(url)
text += Document(content).summary() + " "
# Separate into a list of word, frequency).
words = wordcloud.process_text(text)
# Compute the position of the words.
elements = wordcloud.fit_words(words)
# Draw the positioned words to a PNG file.
wordcloud.draw(elements, path.join(d, 'db2.png'))
def genWordCloud(filename):
textArray = openTxt(filename)
count = countWords(textArray, 1000)
words = wordcloud.fit_words(count, width=500, height=500)
wordcloud.draw(words, pngPath + os.path.splitext(filename)[0] + '.png', width=500, height=500, scale=2)
return 'Cloud generated for {}'.format(filename)