基于百度AI的微博内容情感分析

基于百度AI的微博内容情感分析

在使用代码之前，需要先去百度AI官网注册一个账号，获取秘钥。

from aip import AipNlp
import pandas as pd
import numpy as np
import time
import random

""" 你的 APPID AK SK """
#需要注册百度AI账号
APP_ID = ''
API_KEY = ''
SECRET_KEY = ''
client = AipNlp(APP_ID, API_KEY, SECRET_KEY)


def is_chinese(uchar):
    """判断一个unicode是否是汉字"""
    if uchar >= u'\u4e00' and uchar <= u'\u9fa5':
        return True
    if uchar in ['“','”','，',':','。','（','）','！','？','【','】','《','》',';']:
        return True
    else:
        return False

def format_str(content):
    content_str = ''
    for i in content:
        if is_chinese(i):
            content_str = content_str+i
    return content_str

datass = pd.read_csv('result.csv')
# In[]
datass['text'] = datass['text'].fillna('0')#用0填补缺失值
datass['text'] = datass['text'].map(lambda x:str(x))#把所有类型都转化为字符串
# In[]
ls3 = []
ls4 = []
ls5 = []
ls6 = []
ls11 = []
ls12 = []
for i in range(len(datass)):
    ls11.append(datass['mid'].iloc[i])
    ls12.append(datass['text'].iloc[i])
    #链接百度AI
    try:
        datas = client.sentimentClassify(datass['text'].iloc[i][:1000])
    except Exception:
        datas = client.sentimentClassify(str(format_str(datass['text'].iloc[i][:1000])))
    try:
        data = datas['items'][0]
    except Exception:
        if datas['error_code'] == 18:
            print('速度太快了！！！')
            time.sleep(3)
            datas = client.sentimentClassify(str(format_str(datass['text'].iloc[i][:1000])))
        else:
            pass
    
    try:
        data = datas['items'][0]
    except Exception:
        ls3.append(np.nan)
        ls4.append(np.nan)
        ls5.append(np.nan)
        ls6.append(np.nan)
    else:
        print(i,datas)
        try:   
            df1 = data['positive_prob']
        except Exception:
            ls3.append(np.nan)
        else:
            ls3.append(df1)
        
        try:
            df2 = data['confidence']
        except Exception:
            ls4.append(np.nan)
        else:
            ls4.append(df2)
            
        try:
            df3 = data['negative_prob']
        except Exception:
            ls5.append(np.nan)
        else:
            ls5.append(df3)
        
        try:
            df4 = data['sentiment']
        except Exception:
            ls6.append(np.nan)
        else:
            ls6.append(df4)

    time.sleep(random.uniform(1,2))
    
# In[]
datass['positive_prob'] = ls3
datass['confidence'] = ls4
datass['negative_prob'] = ls5
datass['sentiment'] = ls6
datass.to_csv('情感分析.csv',index = None)
# In[]
results = pd.DataFrame()
results['id'] = ls11
results['text'] = ls12
results['positive_prob'] = ls3
results['confidence'] = ls4
results['negative_prob'] = ls5
results['sentiment'] = ls6
#results.to_excel('E:/all.xlsx',index = None)