示例#1
0
def Eps_crab(industry_new):

    # 要抓取的網址
    url = 'https://goodinfo.tw/StockInfo/StockList.asp?MARKET_CAT=全部&INDUSTRY_CAT=' + industry_new + '&SHEET=年獲利能力_近N年一覽&SHEET2=EPS(元)&RPT_TIME=最新資料'
    """
    #user_agent = UserAgent()
    
    headers = {
        "Authorization": "Bearer " + "YkrXjA4k7pswPML2wkdNxgcRhqSKPcrBysvLmIClsvd",
        "Content-Type": "application/x-www-form-urlencoded",
        "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9", 
        "Accept-Encoding": "gzip, deflate, br", 
        "Accept-Language": "zh-TW,zh;q=0.9", 
        #"Host": "goodinfo.tw/StockInfo/index.asp",  #目標網站 
        "Sec-Fetch-Dest": "document", 
        "Sec-Fetch-Mode": "navigate", 
        "Sec-Fetch-Site": "none", 
        "Upgrade-Insecure-Requests": "1", 
        #隨機設定 使用者代理(User-Agent)
        #"User-Agent":user_agent.random,
        "User-Agent": "Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36", #使用者代理
        "Referer": "https://www.google.com/"
        
    }
    """

    headers = stockmd.get_headers()

    #請求網站
    list_req = requests.post(url, headers=headers)

    #將整個網站的程式碼爬下來
    soup = BeautifulSoup(list_req.content, "html.parser")

    stock_result = soup.find_all("a", {
        "class": "link_black",
        "style": "cursor:pointer;"
    })

    #取出近3年EPS欄位名稱
    stock_name_list = []
    for i in range(0, 19):
        name = stock_result[i].text
        if i >= 15:
            stock_name_list.append(name)

    stockEps = soup.find_all(id=re.compile('^row'))

    stock_num_list = []
    #計算倒數前4個table
    stockEps_reciprocal1_table = []
    stockEps_reciprocal2_table = []
    stockEps_reciprocal3_table = []
    #stockEps_reciprocal4_table=[]
    #stockEpsYr_list=[stockEps_reciprocal1_table,stockEps_reciprocal2_table,stockEps_reciprocal3_table,stockEps_reciprocal4_table]
    stockEpsYr_list = [
        stockEps_reciprocal1_table, stockEps_reciprocal2_table,
        stockEps_reciprocal3_table
    ]

    #取出近3年的EPS數值
    table = -1
    for stocklist in stockEpsYr_list:
        for num in range(0, len(stockEps)):
            result = stockEps[num].find_all('nobr')[table].text
            stocklist.append(result)
        table = table - 1

    #取出代號
    for num in range(0, len(stockEps)):
        result = stockEps[num].find_all('nobr')[0].text
        stock_num_list.append(result)

    #欄位合併
    # stockresult_list={"代號":stock_num_list,stock_name_list[0]:stockEps_reciprocal4_table,stock_name_list[1]:stockEps_reciprocal3_table,stock_name_list[2]:stockEps_reciprocal2_table
    #                  ,stock_name_list[3]:stockEps_reciprocal1_table}
    stockresult_list = {
        "代號": stock_num_list,
        stock_name_list[0]: stockEps_reciprocal3_table,
        stock_name_list[1]: stockEps_reciprocal2_table,
        stock_name_list[2]: stockEps_reciprocal1_table
    }
    df_stock = pd.DataFrame(stockresult_list)

    #去除空白
    df_stock = df_stock.replace(r'^\s*$', np.nan, regex=True)

    #判斷第3行欄位空值數量 只取3年的EPS
    """
    isnanum=df_stock[df_stock.columns[3]].isna().sum()
    if isnanum>10:
        #空值填補至2020Q3 (新調整的部分)
        df_stock.loc[df_stock[df_stock.columns[4]].isnull(),df_stock.columns[4]]=df_stock[df_stock.columns[3]]
        df_stock=df_stock.drop([df_stock.columns[3]],axis=1)
    else:
        df_stock=df_stock.drop([df_stock.columns[1]],axis=1)
    """

    for i in df_stock.iloc[:, 1:]:
        df_stock[i] = df_stock[i].astype(str).astype(float)

    df_stock_result = df_stock.loc[(df_stock[df_stock.columns[1]] >= 1.5)
                                   & (df_stock[df_stock.columns[2]] >= 1.5) &
                                   (df_stock[df_stock.columns[3]] >= 1.5)]

    df_stock_num_result = df_stock_result["代號"].tolist()

    result_stock = stockmd.stock_result(headers, df_stock_num_result)
    """
    result_stock=""

    for i in df_stock_num_result:
        url1='http://jsjustweb.jihsun.com.tw/z/zc/zca/zca_'+i+'.djhtm'
        #請求網站
        list_req1 = requests.post(url1, headers = headers)
        #將整個網站的程式碼爬下來
        soup1 = BeautifulSoup(list_req1.content, "html.parser")
        td=soup1.find_all('td')[4]
        a=td.text
        a=a.lstrip()
        a=a.split(" ")
        a=a[0]
        a=a.split("(")
        a=a[0]
        a=a.replace("*","")
        #print(i+a+' ')
        url2='https://tw.stock.yahoo.com/q/q?s='+i
        #請求網站
        list_req2 = requests.post(url2, headers = headers)
        #將整個網站的程式碼爬下來
        soup2 = BeautifulSoup(list_req2.content, "html.parser")
        b=soup2.find_all('td')[5].text
        result=i+a+' '+'目前價格:'+b+'元'
        result_stock+=result+'\n'
        #print(result_stock)
        #result_stock_list.append(result)
    """
    #print(industry_new+'相關類股其近三年EPS>=1.5,適合購買的股票為:'+'\n'+result_stock)

    result = '相關類股其近三年EPS>=1.5,適合購買的股票為:' + '\n' + result_stock

    #params = {"message": industry_new+'相關類股其近三年EPS>=1.5,適合購買的股票為:'+'\n'+result_stock}
    #r = requests.post("https://notify-api.line.me/api/notify",
    #headers=headers, params=params)

    return result
示例#2
0
def Ni_crab(industry_new):

    # 要抓取的網址
    url = 'https://goodinfo.tw/StockInfo/StockList.asp?MARKET_CAT=全部&INDUSTRY_CAT=' + industry_new + '&SHEET=年獲利能力_近N年一覽&SHEET2=稅後淨利年成長率(%25)&RPT_TIME=最新資料'
    """
    #user_agent = UserAgent()
    
    headers = {
        
        "Content-Type": "application/x-www-form-urlencoded",
        "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9", 
        "Accept-Encoding": "gzip, deflate, br", 
        "Accept-Language": "zh-TW,zh;q=0.9", 
        #"Host": "goodinfo.tw/StockInfo/index.asp",  #目標網站 
        "Sec-Fetch-Dest": "document", 
        "Sec-Fetch-Mode": "navigate", 
        "Sec-Fetch-Site": "none", 
        "Upgrade-Insecure-Requests": "1", 
        #隨機設定 使用者代理(User-Agent)
        #"User-Agent":user_agent.random,
        "User-Agent": "Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36", #使用者代理
        "Referer": "https://www.google.com/"
        
    }
    """

    headers = stockmd.get_headers()

    #請求網站
    list_req = requests.post(url, headers=headers)

    #result=str(list_req)
    #return result

    #將整個網站的程式碼爬下來
    soup = BeautifulSoup(list_req.content, "html.parser")

    stockNi = soup.find_all(id=re.compile('^row'))

    #取出股票代號及Ni結果

    stock_num_list = []
    stock_Ni_list = []

    for Ni in range(0, len(stockNi)):
        stocknum = stockNi[Ni].find_all('nobr')[0].text
        resultNi = stockNi[Ni].find_all('nobr')[-1].text
        stock_num_list.append(stocknum)
        stock_Ni_list.append(resultNi)

    #欄位合併
    stockresult_list = {"代號": stock_num_list, "稅後淨利年成長率(%)": stock_Ni_list}
    df_stock = pd.DataFrame(stockresult_list)

    df_stock = df_stock[df_stock['稅後淨利年成長率(%)'].str.contains("\+")]

    df_stock_num_result = df_stock["代號"].tolist()

    result_stock = stockmd.stock_result(headers, df_stock_num_result)
    """
    result_stock=""
    
    for i in df_stock_num_result:
        url1='http://jsjustweb.jihsun.com.tw/z/zc/zca/zca_'+i+'.djhtm'
        #請求網站
        list_req1 = requests.post(url1, headers = headers)
        #將整個網站的程式碼爬下來
        soup1 = BeautifulSoup(list_req1.content, "html.parser")
        td=soup1.find_all('td')[4]
        a=td.text
        a=a.lstrip()
        a=a.split(" ")
        a=a[0]
        a=a.split("(")
        a=a[0]
        a=a.replace("*","")
        #print(i+a+' ')
        url2='https://tw.stock.yahoo.com/q/q?s='+i
        #請求網站
        list_req2 = requests.post(url2, headers = headers)
        #將整個網站的程式碼爬下來
        soup2 = BeautifulSoup(list_req2.content, "html.parser")
        b=soup2.find_all('td')[5].text
        result=i+a+' '+'目前價格:'+b+'元'
        result_stock+=result+'\n'
        #print(result_stock)
        #result_stock_list.append(result)
    """

    #print(industry_new+'相關類股其今年累計稅後淨利年度成長率優於去年者,適合購買的股票為:'+'\n'+result_stock)

    result = '相關類股其今年累計稅後淨利年度成長率優於去年者,適合購買的股票為:' + '\n' + result_stock

    #params = {"message": industry_new+'相關類股其今年累計稅後淨利年度成長率優於去年者,適合購買的股票為:'+'\n'+result_stock}
    #r = requests.post("https://notify-api.line.me/api/notify",
    #headers=headers, params=params)

    return result
示例#3
0
def PER_crab(industry_new):

    # 要抓取的網址
    url = 'https://goodinfo.tw/StockInfo/StockList.asp?MARKET_CAT=全部&INDUSTRY_CAT=' + industry_new + '&SHEET=交易狀況&SHEET2=日&RPT_TIME=最新資料'
    """
    #user_agent = UserAgent()
    
    headers = {
            
        "Content-Type": "application/x-www-form-urlencoded",
        "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9", 
        "Accept-Encoding": "gzip, deflate, br", 
        "Accept-Language": "zh-TW,zh;q=0.9", 
        #"Host": "goodinfo.tw/StockInfo/index.asp",  #目標網站 
        "Sec-Fetch-Dest": "document", 
        "Sec-Fetch-Mode": "navigate", 
        "Sec-Fetch-Site": "none", 
        "Upgrade-Insecure-Requests": "1", 
        #隨機設定 使用者代理(User-Agent)
        #"User-Agent":user_agent.random,
        "User-Agent": "Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/85.0.4183.102 Safari/537.36", #使用者代理
        "Referer": "https://www.google.com/"
        
    }
    """

    headers = stockmd.get_headers()

    #請求網站
    list_req = requests.post(url, headers=headers)

    #result=str(list_req)
    #return result

    #將整個網站的程式碼爬下來
    soup = BeautifulSoup(list_req.content, "html.parser")

    #取text數字(股票代碼) 先存成list 再用迴圈取出來
    stock_mun = soup.find_all(class_="link_black",
                              target="_blank",
                              text=re.compile('\d{4}'),
                              href=re.compile("StockDetail"))

    stock_mun_list = []

    for num in stock_mun:
        #print(num.text)
        stock_mun_list.append(num.text)
        #result=str(num.text)

    #去除重複的股票代碼
    stock_mun_list = np.unique(stock_mun_list).tolist()

    stock = []

    for num in stock_mun_list:
        #print(num)
        try:
            url_PBR = 'http://jsjustweb.jihsun.com.tw/z/zc/zca/zca_' + num + '.djhtm'
            #請求網站
            list_req_PBR = requests.post(url_PBR, headers=headers)
            #將整個網站的程式碼爬下來
            soup_PBR = BeautifulSoup(list_req_PBR.text, "html.parser")
            tr = soup_PBR.find_all('tr')[6]
            td = tr.find_all('td')[1]
            PBR = td.text
            PBR = PBR.replace(",", "")
            if PBR == 'N/A':
                #print("空值")
                pass
                #continue
                #break
            else:
                tr_PBR = soup_PBR.find_all('tr')[7]
                td_PBR = tr_PBR.find_all('td')[1]
                avrPBR = td_PBR.text
                PBR = float(PBR)
                avrPBR = avrPBR.replace(",", "")
                avrPBR = float(avrPBR)

                if PBR <= 15:

                    if PBR < avrPBR:
                        #print('本益比小')
                        stock.append(num)
                    else:
                        #print('本益比大')
                        pass
                else:
                    #print('本益比大')
                    pass
        except IndexError:
            pass

    #result_stock_list=[]

    result_stock = stockmd.stock_result(headers, stock)
    """
    result_stock=''
    
    for i in stock:
        url1='http://jsjustweb.jihsun.com.tw/z/zc/zca/zca_'+i+'.djhtm'
        #請求網站
        list_req1 = requests.post(url1, headers = headers)
        #將整個網站的程式碼爬下來
        soup1 = BeautifulSoup(list_req1.content, "html.parser")
        td=soup1.find_all('td')[4]
        a=td.text
        a=a.lstrip()
        a=a.split(" ")
        a=a[0]
        a=a.split("(")
        a=a[0]
        a=a.replace("*","")
        #print(i+a+' ')
        url2='https://tw.stock.yahoo.com/q/q?s='+i
        #請求網站
        list_req2 = requests.post(url2, headers = headers)
        #將整個網站的程式碼爬下來
        soup2 = BeautifulSoup(list_req2.content, "html.parser")
        b=soup2.find_all('td')[5].text
        result=i+a+' '+'目前價格:'+b+'元'
        result_stock+=result+'\n'
        #print(result_stock)
        #result_stock_list.append(result)
     """

    result = '相關類股其本益比<=15適合購買的股票為:' + '\n' + result_stock
    #params = {"message": '半導體業相關類股其本益比較小適合購買的股票為:'+'\n'+result_stock}
    #r = requests.post("https://notify-api.line.me/api/notify",
    #headers=headers2, params=params)

    return result
示例#4
0
from linebot.models import (
    MessageEvent,
    TextSendMessage,
    TemplateSendMessage,
    ButtonsTemplate,
    MessageTemplateAction,
    PostbackEvent,
    PostbackAction,
    CarouselColumn,
    CarouselTemplate,
    PostbackTemplateAction
)



headers=stockmd.get_headers()



#global 定義全域變數 交換變數
def save_industry(industry):
    global industry_new
    if industry !="":
        industry_new=industry
        return industry_new
    else:
        return industry_new
    

def data_list(name,token):
    if name !="":