| 12 | # nest_asyncio.apply() |
| 13 | |
| 14 | class getTopSecCom: |
| 15 | def __init__(self, top=None): |
| 16 | self.headers = {"Referer": "http://quote.eastmoney.com/", |
| 17 | "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.141 Safari/537.36"} |
| 18 | self.bk_url = "http://71.push2.eastmoney.com/api/qt/clist/get?cb=jQuery1124034348162124675374_1612595298605&pn=1&pz=85&po=1&np=1&ut=bd1d9ddb04089700cf9c27f6f7426281&fltt=2&invt=2&fid=f62&fs=b:BK0655&fields=f12,f14&_=1612595298611" |
| 19 | self.shares_api = "https://xueqiu.com/S/" |
| 20 | self.top = top |
| 21 | if not os.path.exists("./useful_sec_com_list"): |
| 22 | self.useful_sec_com_list = self.get_sec_com_code() |
| 23 | else: |
| 24 | with open("./useful_sec_com_list", "rb") as fp: |
| 25 | self.useful_sec_com_list = joblib.load(fp) |
| 26 | |
| 27 | def get_sec_com_code(self): |
| 28 | html = rq.get(self.bk_url, headers=self.headers).content.decode("utf-8") |
| 29 | sec_com_list = eval(re.findall("\[(.*?)\]", html)[0]) |
| 30 | useful_sec_com_list = [[i["f12"], i["f14"]] for i in sec_com_list if "ST" not in i["f14"]] |
| 31 | |
| 32 | # 0和3开头的为深证上市股票前缀为sz，6开头的为上证上市股票前缀为sh |
| 33 | for sec_com in useful_sec_com_list: |
| 34 | if sec_com[0][0] == "6": |
| 35 | sec_com[0] = "sh" + sec_com[0] |
| 36 | else: |
| 37 | sec_com[0] = "sz" + sec_com[0] |
| 38 | with open("useful_sec_com_list", "wb") as fp: |
| 39 | joblib.dump(useful_sec_com_list, fp) |
| 40 | return useful_sec_com_list |
| 41 | |
| 42 | async def async_get_shares_details(self, sec_com, url): |
| 43 | async with aiohttp.ClientSession() as session: |
| 44 | async with session.get(url, headers=self.headers) as response: |
| 45 | html = await response.text() |
| 46 | market_value = re.search("<td>总市值：<span>(.*?)亿</span>", html) |
| 47 | if market_value: |
| 48 | return [*sec_com, market_value.groups()[0]] |
| 49 | |
| 50 | async def async_get_all_shares(self): |
| 51 | tasks = [] |
| 52 | for sec_com in self.useful_sec_com_list: |
| 53 | url = self.shares_api + sec_com[0] |
| 54 | tasks.append( |
| 55 | asyncio.create_task( |
| 56 | self.async_get_shares_details(sec_com, url) |
| 57 | ) |
| 58 | ) |
| 59 | done, pendding = await asyncio.wait(tasks) |
| 60 | return [share.result() for share in done if share.result()] |
| 61 | |
| 62 | def get_shares_details(self): |
| 63 | all_shares = [] |
| 64 | for sec_com in self.useful_sec_com_list: |
| 65 | url = self.shares_api + sec_com[0] |
| 66 | response = rq.get(url, headers=self.headers).content.decode("utf-8") |
| 67 | market_value = re.search("<td>总市值：<span>(.*?)亿</span>", response) |
| 68 | if market_value: |
| 69 | all_shares.append([*sec_com, market_value.groups()[0]]) |
| 70 | return all_shares |
| 71 | |