#微信好友特征數(shù)據(jù)分析及可視化
# 1.導(dǎo)包操作
from wxpy import *
import re
import jieba
import numpy as np
from scipy.misc import imread
from wordcloud import WordCloud
import matplotlib.pyplot as plt
from pylab import mpl
import pandas as pd

mpl.rcParams['font.sans-serif'] = ['SimHei']
from pyecharts.charts import Map
from pyecharts import options as opts

# 2.登錄操作
bot = Bot()
# 列舉登錄賬號(hào)的所有好友
all_friends = bot.friends()
print(all_friends)

# 獲取登錄賬號(hào)所關(guān)注的所有的公眾號(hào)
all_maps = bot.mps()
print("所有好友列表", all_maps)

# 獲取當(dāng)前登錄賬號(hào)的群聊列表
all_groups = bot.groups()
print("所有群聊列表", all_groups)

# 根據(jù)好友的備注名稱搜索好友
#myfriend = bot.friends().search('許寬')[0]
#print("搜索好友：", myfriend)

# 搜索好友并發(fā)送信息
# bot.friends().search('許寬')[0].send('你好呀')

# 向文件傳輸助手發(fā)送信息
bot.file_helper.send("hello")

# 3.顯示男女比例
sex_dict = {'male': 0, 'female': 0, "no_set": 0}
for friend in all_friends:
  print(friend, friend.sex)
  if friend.sex == 1:
    sex_dict['male'] += 1
  elif friend.sex == 2:
    sex_dict['female'] += 1
  elif friend.sex == 0:
    sex_dict['no_set'] += 1
print(sex_dict)

# 4使用matplotlib可視化
slices = [sex_dict["male"], sex_dict["female"], sex_dict["no_set"]]
activities = ["male", "female", "no_set"]
cols = ["r", "m", "g"]
# startangle：開(kāi)始繪圖的角度，逆時(shí)針旋轉(zhuǎn)
# shadow：陰影
# %1.1f%%:格式化字符串，整數(shù)部分最小1位，小數(shù)點(diǎn)后保留一位，%%：轉(zhuǎn)義字符
plt.pie(slices, labels=activities, colors=cols, startangle=90, shadow=True, autopct='%1.1f%%')
plt.title("微信好友比例圖")
plt.savefig("WeChat_sex.png")

# 統(tǒng)計(jì)登錄賬號(hào)好友的各省人數(shù)
province_dict = {'河北': 0, '山東': 0, '遼寧': 0, '廣西': 0, '吉林': 0,
         '甘肅': 0, '青海': 0, '河南': 0, '江蘇': 0, '湖北': 0,
         '湖南': 0, '江西': 0, '浙江': 0, '廣東': 0, '云南': 0,
         '福建': 0, '臺(tái)灣': 0, '海南': 0, '山西': 0, '四川': 0,
         '陜西': 0, '貴州': 0, '安徽': 0, '北京': 0, '天津': 0,
         '重慶': 0, '上海': 0, '香港': 0, '澳門(mén)': 0, '新疆': 0,
         '內(nèi)蒙古': 0, '西藏': 0, '黑龍江': 0, '寧夏': 0}
# 統(tǒng)計(jì)省份
for friend in all_friends:
  # print(friend.province)
  if friend.province in province_dict.keys():
    province_dict[friend.province] += 1

print("province_dict")
print(province_dict)

# 為了方便數(shù)據(jù)呈現(xiàn)，生成JSON Array格式數(shù)據(jù)
data = []
for key, value in province_dict.items():
  data.append({'name': key, 'value': value}) # 在data列表末尾添加一個(gè)字典元素
print(data)

data_process = pd.DataFrame(data) # 創(chuàng)建數(shù)據(jù)框
data_process.columns = ['city', 'popu']
print(data_process)

map = Map().add("微信好友城市分布圖", [list(z) for z in zip(data_process['city'], data_process['popu'])],
        "china").set_global_opts(
  title_opts=opts.TitleOpts(title="Map-VisualMap（連續(xù)型）"), visualmap_opts=opts.VisualMapOpts(max_=10))
map.render('map.html')


# with...as...語(yǔ)句結(jié)束時(shí)，自動(dòng)調(diào)用f.close()
# a表示：在文件末尾追加
def write_txt_file(path, txt): # 寫(xiě)文件
  with open(path, 'a', encoding='gbk') as f:
    return f.write(txt)


# 每次運(yùn)行程序前，需要?jiǎng)h除上一次的文件
# 默認(rèn)字符編碼為GBK
def read_txt_file(path):
  with open(path, 'r', encoding='gbk') as f:
    return f.read()


# 統(tǒng)計(jì)登錄賬號(hào)好友個(gè)性簽名
for friend in all_friends:
  print(friend, friend.signature)
  # 對(duì)數(shù)據(jù)進(jìn)行清洗，將標(biāo)點(diǎn)符號(hào)等對(duì)詞頻率統(tǒng)計(jì)造成影響的因素剔除
  # [...]:匹配中括號(hào)任意一個(gè)字符
  # r:防止轉(zhuǎn)義
  pattern = re.compile(r'[一-龥](méi)+') # 將正則字符串編譯成正則表達(dá)式對(duì)象，以后在后期的匹配中復(fù)用
  # 對(duì)某一個(gè)簽名進(jìn)行匹配，只匹配中文漢字，結(jié)果是列表
  filterdata = re.findall(pattern, friend.signature)
  print(filterdata)
  write_txt_file('signatures.txt', ''.join(filterdata))

# 讀取文件并輸出。
content = read_txt_file('signatures.txt')
print(content) # 輸出內(nèi)容，僅漢字

# 輸出分詞結(jié)果，結(jié)果為列表
segment = jieba.lcut(content) # 精確模式：不存在冗余數(shù)據(jù)，適合文本分析
print(segment)

# 生成數(shù)據(jù)框且有一列元素
word_df = pd.DataFrame({'segment': segment}) # 字典類(lèi)型
print(word_df)

# index_col=False:第一行不做為索引
# seq=" "：分隔符
# names=['stopword']：列名
# "stopwords.txt":停止詞庫(kù)
stopwords = pd.read_csv("stopwords.txt", index_col=False, sep=" ", names=['stopword'], encoding='gbk')
print(stopwords)

# 查看過(guò)濾停止詞后的數(shù)據(jù)框
word_df = word_df[~word_df.segment.isin(stopwords.stopword)]
print("過(guò)濾后：")
print(word_df)

# 查看分詞的詞頻
# python中的groupby可以看作是基于行或者是基于index的聚合操作
# agg函數(shù)提供基于列的聚合操作，一般與groupby連用
# np.size:numpy庫(kù)中統(tǒng)計(jì)一列中不同值的個(gè)數(shù)
words_stat = word_df.groupby(by=['segment'])['segment'].agg({"計(jì)數(shù)": np.size}) # 警告信息
print(words_stat)

# 根據(jù)計(jì)數(shù)這一列降序排列
words_stat = words_stat.reset_index().sort_values(by=["計(jì)數(shù)"], ascending=False)
print(words_stat)

# 讀入背景圖片
color_mask = imread("black_mask.png")

# 設(shè)置詞云屬性
wordcloud = WordCloud(font_path="Hiragino.ttf", # 設(shè)置字體
           background_color="pink", # 設(shè)置顏色
           max_words=100, # 詞云顯示的最大詞數(shù)
           mask=color_mask, # 設(shè)置背景圖片
           max_font_size=100 # 字體最大值
           )
# 生成詞云字典，獲取詞云最高的前一百詞
word_frequence = {x[0]: x[1] for x in words_stat.head(100).values}
print(word_frequence)

# 繪制詞云圖
wordcloud.generate_from_frequencies(word_frequence)
wordcloud.to_file("wordcloud.png")

# 對(duì)圖像進(jìn)行處理
plt.imshow(wordcloud)
plt.axis("off") # 隱藏坐標(biāo)軸
plt.show()