import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from snownlp import SnowNLP
import datetime
import jieba.analyse
import re
from wordcloud import WordCloud
sns.set_style("darkgrid")

jieba.load_userdict('userdict.txt')

df = pd.read_csv("weibo.csv")

df['publish_time'] = pd.to_datetime(df['publish_time'],format='%Y-%m-%d')

#Sorce:  https://blog.csdn.net/qq_43814415/article/details/119517978
def clean(line):
    rep=['【】','【','】','👍','🤝',
        '🐮','🙏','🇨🇳','👏','❤️','………','🐰','...、、','，，','..','💪','🤓',
         '⚕️','👩','🙃','😇','🍺','🐂','🙌🏻','😂','📖','😭','✧٩(ˊωˋ*)و✧','🦐','？？？？','//','😊','💰','😜','😯',
         '(ღ˘⌣˘ღ)','✧＼٩(눈౪눈)و/／✧','🌎','🍀','🐴',
         '🌻','🌱','🌱','🌻','🙈','(ง•̀_•́)ง！','🉑️','💩',
         '🐎','⊙∀⊙！','🙊','【？','+1','😄','🙁','👇🏻','📚','🙇',
         '🙋','！！！！','🎉','＼(^▽^)／','👌','🆒','🏻',
         '🙉','🎵','🎈','🎊','0371-12345','☕️','🌞','😳','👻','🐶','👄','\U0001f92e\U0001f92e','😔','＋1','🛀','🐸','🐷','➕1',
         '🌚','：：','💉','√','x','！！！','🙅','♂️','💊','👋','o(^o^)o','mei\u2006sha\u2006shi','💉','😪','😱',
         '🤗','关注','……','(((╹д╹;)))','⚠️','Ծ‸Ծ','⛽️','😓','🐵',
         '🙄️','🌕','…','😋','[]','[',']','→_→','💞','😨','"','😁','ฅ۶•ﻌ•♡','😰','🎙️',
         '🤧','😫','(ง•̀_•́)ง','😁','✊','🚬','😤','👻','😣','：','😷','(*^▽^)/★*☆','🐁','🐔','😘','🍋','(✪▽✪)','(❁´ω`❁)','1⃣3⃣','(^_^)／','☀️',
         '🎁','😅','🌹','🏠','→_→','🙂','✨','❄️','•','🌤','💓','🔨','👏','😏','⊙∀⊙！','👍','✌(̿▀̿\u2009̿Ĺ̯̿̿▀̿̿)✌',
         '😊','👆','💤','😘','😊','😴','😉','🌟','♡♪..𝙜𝙤𝙤𝙙𝙣𝙞𝙜𝙝𝙩•͈ᴗ•͈✩‧₊˚','👪','💰','😎','🍀','🛍','🖕🏼','😂','(✪▽✪)','🍋','🍅','👀','♂️','🙋🏻','✌️','🥳','￣￣)σ',
         '😒','😉','🦀','💖','✊','💪','🙄','🎣','🌾','✔️','😡','😌','🔥','❤','🏼','🤭','🌿','丨','✅','🏥','ﾉ','☀','5⃣⏺1⃣0⃣','🚣','🎣','🤯','🌺',
         '🌸','\u200b','\ue627OOTD','收起d'
         ]
    pattern_0=re.compile('#.*?#')#在用户名处匹配话题名称
    pattern_1=re.compile('【.*?】')#在用户名处匹配话题名称
    pattern_2=re.compile('肺炎@([\u4e00-\u9fa5\w\-]+)')#匹配@
    pattern_3=re.compile('@([\u4e00-\u9fa5\w\-]+)')#匹配@
    #肺炎@环球时报
    pattern_4=re.compile(u'[\U00010000-\U0010ffff\uD800-\uDBFF\uDC00-\uDFFF]')#匹配表情
    pattern_7=re.compile('L.*?的微博视频')
    pattern_8=re.compile('（.*?）')
    #pattern_9=re.compile(u"\|[\u4e00-\u9fa5]*\|")#匹配中文

    line=line.replace('O网页链接','')
    line=line.replace('-----','')
    line=line.replace('①','')
    line=line.replace('②','')
    line=line.replace('③','')
    line=line.replace('④','')
    line=line.replace('>>','')
    line=re.sub(pattern_0, '', line,0) #去除话题
    line=re.sub(pattern_1, '', line,0) #去除【】
    line=re.sub(pattern_2, '', line,0) #去除@
    line=re.sub(pattern_3, '', line,0) #去除@
    line=re.sub(pattern_4, '', line,0) #去除表情
    line=re.sub(pattern_7, '', line,0) 
    line=re.sub(pattern_8, '', line,0) 
    line=re.sub(r'\[\S+\]', '', line,0) #去除表情符号
    
    for i in rep:
        line=line.replace(i,'')
    return line

df['content'] = df.apply(lambda x:clean(x['content']),axis = 1)

df['voice'] = df['forward_num']+df['comment_num']+df['like_num']
V = pd.DataFrame()

V['voice'] = df.groupby('publish_time').sum()['voice']
V['num'] = df.groupby('publish_time').count()['voice']

V['voice'].plot(figsize=(14,7))

V['num'].plot(figsize=(14,7))

V.to_csv('安踏微博声量.csv',index= True)

def cal_score(df):
    n = len(df['content'])
    sentiment = np.empty(n)
    for i in range(n):
        text = SnowNLP(df['content'][i])
        sent = text.sentences
        score = list()
        for sen in sent:
            s = SnowNLP(sen)
            score.append(s.sentiments)

        sentiment[i] = np.nanmean(score)
    return sentiment

df['sentiment'] = cal_score(df)

df['sentiment'].describe()

plt.figure(figsize=(12,6))
sns.histplot(df['sentiment'],kde = True)

print('positive:',len(df[df['sentiment'] >= 0.8]))
print('negative:',len(df[df['sentiment'] <= 0.2]))

A = pd.DataFrame()
A['Total'] = df.groupby('publish_time').count()['sentiment']
A['Positive'] = df[df['sentiment']>=0.8].groupby('publish_time').count()['sentiment']
A['Negative'] = df[df['sentiment']<=0.2].groupby('publish_time').count()['sentiment']

A[['Positive','Negative']].plot(figsize=(14,7))

A.to_csv('安踏每日正负情感.csv',index = True)

df[['sentiment']].to_csv('安踏微博情感分析.csv',index = False)

wo = df[(df['publish_time'] >= '2022-02-04')&(df['publish_time'] <= '2022-02-20')].reset_index()
print('positive:',len(wo[wo['sentiment'] >= 0.8]))
print('negative:',len(wo[wo['sentiment'] <= 0.2]))

def Get_Key_Words(df,Num = 10,only_adv = False):
    segments = []
    for index, row in df.iterrows(): 
        content = row['content']
        content = content.replace('中国队','')
        content = content.replace('加油','')
        content = content.replace('冬奥','')
        content = content.replace('奥运','')
        content = content.replace('健儿','')
        if only_adv:
            words = jieba.analyse.extract_tags(content, topK=20, allowPOS=('a','ad','vn','nt'))
        else:
            words = jieba.analyse.extract_tags(content, topK=20, allowPOS=('nr', 'n', 'vn', 'v','nz','nt','a'))
        for word in words:
            segments.append({'word':word, 'count':1})
    dfSeg = pd.DataFrame(segments)
    dfWord = dfSeg.groupby('word').sum()[['count']].sort_values('count',ascending=False).reset_index().head(Num)
    return dfWord

# Source: https://www.bbsmax.com/A/q4zVPxgW5K/

def create_word_cloud(df,num = 15,wid = 4000,hei = 2000):
    frequencies = {}
    for index,row in df.iterrows():
        frequencies[row['word']] = float(row['count'])
    wc = WordCloud(
        font_path="./SimHei.ttf",
        max_words=num,
        width=wid, 
        height=hei,
        margin=2,
        background_color='white',
        prefer_horizontal=1.5
    )
    word_cloud = wc.generate_from_frequencies(frequencies)
    plt.imshow(word_cloud)
    plt.axis("off")
    plt.show()

wo_topic = Get_Key_Words(wo,30)
create_word_cloud(wo_topic)

df2 = df.copy()

df2 = df2[(~df2['content'].str.contains('围观冬奥'))&(~df2['content'].str.contains('冬奥能量'))&(~df2['content'].str.contains('红包'))&
         (~df2['content'].str.contains('品牌时刻'))]

t_1_18 = df2[((df2['publish_time'] == '2022-01-18')|(df2['publish_time'] == '2022-01-21'))&(df2['sentiment'] <= 0.2)].reset_index()

create_word_cloud(Get_Key_Words(t_1_18,30))

df2.groupby('publish_time').count().sort_values('voice',ascending=False)[['voice']].head(6)

t_2_20 = df2[(df2['publish_time'] == '2022-02-20')].reset_index()
t_2_15 = df2[(df2['publish_time'] == '2022-02-15')].reset_index()
t_2_5 = df2[df2['publish_time'] == '2022-02-05'].reset_index()
neg =  wo[wo['sentiment'] <= 0.2].reset_index()
post_mon = df2[(df2['publish_time'] >= '2022-03-01')].reset_index()

create_word_cloud(Get_Key_Words(t_2_20,30))

create_word_cloud(Get_Key_Words(t_2_15,30))

create_word_cloud(Get_Key_Words(t_2_5,30))

create_word_cloud(Get_Key_Words(neg,30))

create_word_cloud(Get_Key_Words(post_mon,30))

erke = pd.read_csv("erke.csv")
erke['publish_time'] = pd.to_datetime(erke['publish_time'],format='%Y-%m-%d')

erke['content'] = erke.apply(lambda x:clean(x['content']),axis = 1)

erke.head(5)

erke['voice'] = erke['forward_num']+erke['comment_num']+erke['like_num']
E = pd.DataFrame()

E['voice'] = erke.groupby('publish_time').sum()['voice']
E['num'] = erke.groupby('publish_time').count()['voice']

E['voice'].plot(figsize=(14,7))

E.to_csv('鸿星尔克微博声量.csv',index = True)

erke['sentiment'] = cal_score(erke)

t = erke[(erke['publish_time']>= '2021-07-21')&(erke['publish_time']<= '2021-07-25')]

tp = erke[erke['publish_time']> '2021-08-01']
pt =  erke[erke['publish_time']< '2021-07-20']

plt.figure(figsize=(12,6))
sns.histplot(t['sentiment'],kde = True)

t[['sentiment']].to_csv('鸿星尔克情感分析.csv',index = False)

print('positive:',len(t[t['sentiment'] >= 0.8]))
print('negative:',len(t[t['sentiment'] <= 0.2]))

t_pos = t[t['sentiment'] >= 0.8]
t_neg = t[t['sentiment'] <= 0.2]

create_word_cloud(Get_Key_Words(t_pos,30,True),15,1000,500)

create_word_cloud(Get_Key_Words(t_neg,30,True),15,1000,500)

create_word_cloud(Get_Key_Words(tp,30))