replace browser by session in some places

ADC_function.py

@@ -98,59 +98,113 @@ class TimeoutHTTPAdapter(HTTPAdapter):
             kwargs["timeout"] = self.timeout
         return super().send(request, **kwargs)
 
-def get_html_by_browser(url, cookies: dict = None, ua: str = None, return_type: str = None):
+
+#  with keep-alive feature
+def get_html_session(url:str = None, cookies: dict = None, ua: str = None, return_type: str = None):
+    configProxy = config.getInstance().proxy()
+    session = requests.Session()
+    if isinstance(cookies, dict) and len(cookies):
+        requests.utils.add_dict_to_cookiejar(session.cookies, cookies)
+    retries = Retry(total=configProxy.retry, connect=configProxy.retry, backoff_factor=1, status_forcelist=[429, 500, 502, 503, 504])
+    session.mount("https://", TimeoutHTTPAdapter(max_retries=retries, timeout=configProxy.timeout))
+    session.mount("http://", TimeoutHTTPAdapter(max_retries=retries, timeout=configProxy.timeout))
+    if configProxy.enable:
+        session.verify = config.getInstance().cacert_file()
+        session.proxies = configProxy.proxies()
+    headers = {"User-Agent": ua or G_USER_AGENT}
+    session.headers = headers
+    try:
+        if isinstance(url, str) and len(url):
+            result = session.get(str(url))
+        else: # 空url参数直接返回可重用session对象，无需设置return_type
+            return session
+        if not result.ok:
+            return None
+        if return_type == "object":
+            return result
+        elif return_type == "content":
+            return result.content
+        elif return_type == "session":
+            return result, session
+        else:
+            result.encoding = "utf-8"
+            return result.text
+    except requests.exceptions.ProxyError:
+        print("[-]get_html_session() Proxy error! Please check your Proxy")
+    except Exception as e:
+        print(f"[-]get_html_session() failed. {e}")
+    return None
+
+
+def get_html_by_browser(url:str = None, cookies: dict = None, ua: str = None, return_type: str = None):
     configProxy = config.getInstance().proxy()
     s = requests.Session()
     if isinstance(cookies, dict) and len(cookies):
         requests.utils.add_dict_to_cookiejar(s.cookies, cookies)
-    retries = Retry(connect=configProxy.retry, backoff_factor=1, status_forcelist=[429, 500, 502, 503, 504])
+    retries = Retry(total=configProxy.retry, connect=configProxy.retry, backoff_factor=1, status_forcelist=[429, 500, 502, 503, 504])
     s.mount("https://", TimeoutHTTPAdapter(max_retries=retries, timeout=configProxy.timeout))
     s.mount("http://", TimeoutHTTPAdapter(max_retries=retries, timeout=configProxy.timeout))
     if configProxy.enable:
+        s.verify = config.getInstance().cacert_file()
         s.proxies = configProxy.proxies()
-    browser = mechanicalsoup.StatefulBrowser(user_agent=ua or G_USER_AGENT, session=s)
-    result = browser.open(url)
-    if not result.ok:
-        return ''
-    result.encoding = "utf-8"
-    if return_type == "object":
-        return result
-    elif return_type == "content":
-        return result.content
-    elif return_type == "browser":
-        return result, browser
-    else:
-        return result.text
-
+    try:
+        browser = mechanicalsoup.StatefulBrowser(user_agent=ua or G_USER_AGENT, session=s)
+        if isinstance(url, str) and len(url):
+            result = browser.open(url)
+        else:
+            return browser
+        if not result.ok:
+            return None
+        result.encoding = "utf-8"
+        if return_type == "object":
+            return result
+        elif return_type == "content":
+            return result.content
+        elif return_type == "browser":
+            return result, browser
+        else:
+            return result.text
+    except requests.exceptions.ProxyError:
+        print("[-]get_html_by_browser() Proxy error! Please check your Proxy")
+    except Exception as e:
+        print(f'[-]get_html_by_browser() Failed! {e}')
+    return None
 
 def get_html_by_form(url, form_select: str = None, fields: dict = None, cookies: dict = None, ua: str = None, return_type: str = None):
     configProxy = config.getInstance().proxy()
     s = requests.Session()
     if isinstance(cookies, dict) and len(cookies):
         requests.utils.add_dict_to_cookiejar(s.cookies, cookies)
-    retries = Retry(connect=configProxy.retry, backoff_factor=1, status_forcelist=[429, 500, 502, 503, 504])
+    retries = Retry(total=configProxy.retry, connect=configProxy.retry, backoff_factor=1, status_forcelist=[429, 500, 502, 503, 504])
     s.mount("https://", TimeoutHTTPAdapter(max_retries=retries, timeout=configProxy.timeout))
     s.mount("http://", TimeoutHTTPAdapter(max_retries=retries, timeout=configProxy.timeout))
     if configProxy.enable:
+        s.verify = config.getInstance().cacert_file()
         s.proxies = configProxy.proxies()
-    browser = mechanicalsoup.StatefulBrowser(user_agent=ua or G_USER_AGENT, session=s)
-    result = browser.open(url)
-    if not result.ok:
-        return ''
-    form = browser.select_form() if form_select is None else browser.select_form(form_select)
-    if isinstance(fields, dict):
-        for k, v in fields.items():
-            browser[k] = v
-    response = browser.submit_selected()
-    response.encoding = "utf-8"
-    if return_type == "object":
-        return response
-    elif return_type == "content":
-        return response.content
-    elif return_type == "browser":
-        return response, browser
-    else:
-        return response.text
+    try:
+        browser = mechanicalsoup.StatefulBrowser(user_agent=ua or G_USER_AGENT, session=s)
+        result = browser.open(url)
+        if not result.ok:
+            return None
+        form = browser.select_form() if form_select is None else browser.select_form(form_select)
+        if isinstance(fields, dict):
+            for k, v in fields.items():
+                browser[k] = v
+        response = browser.submit_selected()
+        response.encoding = "utf-8"
+        if return_type == "object":
+            return response
+        elif return_type == "content":
+            return response.content
+        elif return_type == "browser":
+            return response, browser
+        else:
+            return response.text
+    except requests.exceptions.ProxyError:
+        print("[-]get_html_by_form() Proxy error! Please check your Proxy")
+    except Exception as e:
+        print(f'[-]get_html_by_form() Failed! {e}')
+    return None
 
 
 # def get_javlib_cookie() -> [dict, str]:
@@ -645,3 +699,33 @@ def file_not_exist_or_empty(filepath) -> bool:
 # 日语简单检测
 def is_japanese(s) -> bool:
     return bool(re.search(r'[\u3040-\u309F\u30A0-\u30FF\uFF66-\uFF9F]', s, re.UNICODE))
+
+
+if __name__ == "__main__":
+    import sys, timeit
+    from http.client import HTTPConnection
+    s = get_html_session()
+    def benchmark(t, url):
+        print(f"HTTP GET Benchmark times:{t} url:{url}")
+        tm = timeit.timeit(f"_ = session1.get('{url}')",
+            "from __main__ import get_html_session;session1=get_html_session()",
+            number=t)
+        print(f'===={tm:2.5f}s get_html_session() Keep-Alive enable====')
+        tm = timeit.timeit(f"_ = browser1.open('{url}')",
+            "from __main__ import get_html_by_browser;browser1=get_html_by_browser()",
+            number=t)
+        print(f'===={tm:2.5f}s get_html_by_browser() Keep-Alive enable====')
+        tm = timeit.timeit(f"_ = get_html('{url}')",
+            "from __main__ import get_html",
+            number=t)
+        print(f'===={tm:2.5f}s get_html() ====')
+    t = 100
+    #url = "https://www.189.cn/"
+    url = "http://www.chinaunicom.com"
+    HTTPConnection.debuglevel = 1
+    _ = s.get(url)
+    HTTPConnection.debuglevel = 0
+    # Usage: python ./ADC_function.py https://cn.bing.com/
+    if len(sys.argv)>1:
+        url = sys.argv[1]
+    benchmark(t, url)

WebCrawler/__init__.py

@@ -328,6 +328,6 @@ def special_characters_replacement(text) -> str:
                 replace('|', 'ǀ').       # U+01C0 LATIN LETTER DENTAL CLICK @ Basic Multilingual Plane
                 replace('‘', '‘'). # U+02018 LEFT SINGLE QUOTATION MARK
                 replace('’', '’'). # U+02019 RIGHT SINGLE QUOTATION MARK
-                replace('&', '＆').
-                replace('…','…')
+                replace('…','…').
+                replace('&', '＆')
             )

WebCrawler/carib.py

@@ -6,17 +6,16 @@ import re
 from ADC_function import *
 from WebCrawler.storyline import getStoryline
 
+
+G_SITE = 'https://www.caribbeancom.com'
+
+
 def main(number: str) -> json:
     try:
-        # 因演员图片功能还未使用，为提速暂时注释，改为用get_html()
-        #r, browser = get_html_by_browser('https://www.caribbeancom.com/moviepages/'+number+'/index.html',
-        #                return_type='browser')
-        #if not r.ok:
-        #    raise ValueError("page not found")
-        #htmlcode = str(browser.page)
-        htmlbyte = get_html('https://www.caribbeancom.com/moviepages/'+number+'/index.html', return_type='content')
-        htmlcode = htmlbyte.decode('euc-jp')
-        if not htmlcode or '<title>404' in htmlcode or 'class="movie-info section"' not in htmlcode:
+        url = f'{G_SITE}/moviepages/{number}/index.html'
+        result, session = get_html_session(url, return_type='session')
+        htmlcode = result.content.decode('euc-jp')
+        if not result or not htmlcode or '<title>404' in htmlcode or 'class="movie-info section"' not in htmlcode:
             raise ValueError("page not found")
 
         lx = html.fromstring(htmlcode)
@@ -32,13 +31,13 @@ def main(number: str) -> json:
             'actor': get_actor(lx),
             'release': get_release(lx),
             'number': number,
-            'cover': 'https://www.caribbeancom.com/moviepages/' + number + '/images/l_l.jpg',
+            'cover': f'{G_SITE}/moviepages/{number}/images/l_l.jpg',
             'tag': get_tag(lx),
             'extrafanart': get_extrafanart(lx),
             'label': get_series(lx),
             'imagecut': 1,
-#            'actor_photo': get_actor_photo(browser),
-            'website': 'https://www.caribbeancom.com/moviepages/' + number + '/index.html',
+#            'actor_photo': get_actor_photo(lx, session),
+            'website': f'{G_SITE}/moviepages/{number}/index.html',
             'source': 'carib.py',
             'series': get_series(lx),
         }
@@ -101,24 +100,25 @@ def get_series(lx: html.HtmlElement) -> str:
         return ''
 
 def get_runtime(lx: html.HtmlElement) -> str:
-    return str(lx.xpath( "//span[@class='spec-content']/span[@itemprop='duration']/text()")[0]).strip()
+    return str(lx.xpath("//span[@class='spec-content']/span[@itemprop='duration']/text()")[0]).strip()
 
-def get_actor_photo(browser):
-    htmla = browser.page.select('#moviepages > div > div:nth-child(1) > div.movie-info.section > ul > li:nth-child(1) > span.spec-content > a')
+def get_actor_photo(lx, session):
+    htmla = lx.xpath("//*[@id='moviepages']/div[@class='container']/div[@class='inner-container']/div[@class='movie-info section']/ul/li[@class='movie-spec']/span[@class='spec-content']/a[@itemprop='actor']")
+    names = lx.xpath("//*[@id='moviepages']/div[@class='container']/div[@class='inner-container']/div[@class='movie-info section']/ul/li[@class='movie-spec']/span[@class='spec-content']/a[@itemprop='actor']/span[@itemprop='name']/text()")
     t = {}
-    for a in htmla:
-        if a.text.strip() == '他':
+    for i, name in enumerate(names):
+        if name.strip() == '他':
             continue
-        p = {a.text.strip(): a['href']}
+        p = {name.strip(): htmla[i].attrib['href']}
         t.update(p)
     o = {}
     for k, v in t.items():
         if '/search_act/' not in v:
             continue
-        r = browser.open_relative(v)
+        r = session.get(urljoin(G_SITE, v))
         if not r.ok:
             continue
-        html = browser.page.prettify()
+        html = r.text
         pos = html.find('.full-bg')
         if pos<0:
             continue
@@ -126,7 +126,7 @@ def get_actor_photo(browser):
         cssBGjpgs = re.findall(r'background: url\((.+\.jpg)', css, re.I)
         if not cssBGjpgs or not len(cssBGjpgs[0]):
             continue
-        p = {k: urljoin(browser.url, cssBGjpgs[0])}
+        p = {k: urljoin(r.url, cssBGjpgs[0])}
         o.update(p)
     return o
 

WebCrawler/javbus.py

@@ -118,8 +118,15 @@ def main_uncensored(number):
 def main(number):
     try:
         try:
+            url = "https://www." + secrets.choice([
+                'buscdn.fun', 'busdmm.fun', 'busfan.fun', 'busjav.fun',
+                'cdnbus.fun',
+                'dmmbus.fun', 'dmmsee.fun',
+                'fanbus.us',
+                'seedmm.fun',
+                ]) + "/"
             try:
-                htmlcode = get_html('https://www.fanbus.us/' + number)
+                htmlcode = get_html(url + number)
             except:
                 htmlcode = get_html('https://www.javbus.com/' + number)
             if "<title>404 Page Not Found" in htmlcode:

WebCrawler/javdb.py

@@ -4,7 +4,6 @@ import re
 from lxml import etree
 import json
 from ADC_function import *
-from mechanicalsoup.stateful_browser import StatefulBrowser
 from WebCrawler.storyline import getStoryline
 # import io
 # sys.stdout = io.TextIOWrapper(sys.stdout.buffer, errors = 'replace', line_buffering = True)
@@ -30,8 +29,8 @@ def getActor(html):
         idx = idx + 1
     return r
 
-def getaphoto(url,  browser):
-    html_page = browser.open_relative(url).text if isinstance(browser, StatefulBrowser) else get_html(url)
+def getaphoto(url, session):
+    html_page = session.get(url).text if isinstance(session, requests.Session) else get_html(url)
     img_prether = re.compile(r'<span class\=\"avatar\" style\=\"background\-image\: url\((.*?)\)')
     img_url = img_prether.findall(html_page)
     if img_url:
@@ -39,7 +38,7 @@ def getaphoto(url,  browser):
     else:
         return ''
 
-def getActorPhoto(html, javdb_site,  browser): #//*[@id="star_qdt"]/li/a/img
+def getActorPhoto(html, javdb_site, session):
     actorall = html.xpath('//strong[contains(text(),"演員:")]/../span/a[starts-with(@href,"/actors/")]')
     if not actorall:
         return {}
@@ -47,7 +46,7 @@ def getActorPhoto(html, javdb_site,  browser): #//*[@id="star_qdt"]/li/a/img
     actor_photo = {}
     for i in actorall:
         if i.text in a:
-            actor_photo[i.text] = getaphoto(urljoin(f'https://{javdb_site}.com', i.attrib['href']), browser)
+            actor_photo[i.text] = getaphoto(urljoin(f'https://{javdb_site}.com', i.attrib['href']), session)
     return actor_photo
 
 def getStudio(a, html):
@@ -178,15 +177,6 @@ def getDirector(html):
     result1 = str(html.xpath('//strong[contains(text(),"導演")]/../span/text()')).strip(" ['']")
     result2 = str(html.xpath('//strong[contains(text(),"導演")]/../span/a/text()')).strip(" ['']")
     return str(result1 + result2).strip('+').replace("', '", '').replace('"', '')
-def getOutline0(number):  #获取剧情介绍 airav.wiki站点404，函数暂时更名，等无法恢复时删除
-    try:
-        htmlcode = get_html('https://cn.airav.wiki/video/' + number)
-        from WebCrawler.airav import getOutline as airav_getOutline
-        result = airav_getOutline(htmlcode)
-        return result
-    except:
-        pass
-    return ''
 def getOutline(number, title):  #获取剧情介绍 多进程并发查询
     return getStoryline(number,title)
 def getSeries(html):
@@ -224,11 +214,11 @@ def main(number):
             javdb_site = secrets.choice(javdb_sites)
         if debug:
             print(f'[!]javdb:select site {javdb_site}')
-        browser = None
+        session = None
         try:
             javdb_url = 'https://' + javdb_site + '.com/search?q=' + number + '&f=all'
-            res, browser = get_html_by_browser(javdb_url, cookies=javdb_cookies, return_type='browser')
-            if not res.ok:
+            res, session = get_html_session(javdb_url, cookies=javdb_cookies, return_type='session')
+            if not res:
                 raise
             query_result = res.text
         except:
@@ -251,8 +241,9 @@ def main(number):
                     raise ValueError("number not found")
                 correct_url = urls[0]
         try:
-            if isinstance(browser, StatefulBrowser):  # get faster benefit from http keep-alive
-                detail_page = browser.open_relative(correct_url).text
+            if isinstance(session, requests.Session):  # get faster benefit from http keep-alive
+                javdb_detail_url = urljoin(res.url, correct_url)
+                detail_page = session.get(javdb_detail_url).text
             else:
                 javdb_detail_url = 'https://' + javdb_site + '.com' + correct_url
                 detail_page = get_html(javdb_detail_url, cookies=javdb_cookies)
@@ -303,8 +294,8 @@ def main(number):
             'tag': getTag(lx),
             'label': getLabel(lx),
             'year': getYear(detail_page),  # str(re.search('\d{4}',getRelease(a)).group()),
-#            'actor_photo': getActorPhoto(lx, javdb_site,  browser),
-            'website': 'https://javdb.com' + correct_url,
+#            'actor_photo': getActorPhoto(lx, javdb_site,  session),
+            'website': urljoin('https://javdb.com', correct_url),
             'source': 'javdb.py',
             'series': getSeries(lx),
 

WebCrawler/storyline.py

@@ -4,6 +4,7 @@ import re
 import json
 import builtins
 from ADC_function import *
+from lxml.html import fromstring
 from multiprocessing import Pool
 from multiprocessing.dummy import Pool as ThreadPool
 from difflib import SequenceMatcher
@@ -110,24 +111,30 @@ def _getStoryline_mp(site, number, title, debug):
 
 def getStoryline_airav(number, debug):
     try:
-        number_up = number
         site = secrets.choice(('airav.cc','airav4.club'))
         url = f'https://{site}/searchresults.aspx?Search={number}&Type=0'
-        res, browser = get_html_by_browser(url, return_type='browser')
-        if not res.ok:
-            raise ValueError(f"get_html_by_browser('{url}') failed")
-        avs = browser.page.select_one('div.resultcontent > ul > li:nth-child(1) > div')
-        if number_up not in avs.a.h3.text.upper():
+        res, session = get_html_session(url, return_type='session')
+        if not res:
+            raise ValueError(f"get_html_by_session('{url}') failed")
+        lx = fromstring(res.text)
+        urls = lx.xpath('//div[@class="resultcontent"]/ul/li/div/a[@class="ga_click"]/@href')
+        txts = lx.xpath('//div[@class="resultcontent"]/ul/li/div/a[@class="ga_click"]/h3[@class="one_name ga_name"]/text()')
+        detail_url = None
+        for i, txt in enumerate(txts):
+            if re.search(number, txt, re.I):
+                detail_url = urljoin(res.url, urls[i])
+                break
+        if detail_url is None:
             raise ValueError("number not found")
-        detail_url = avs.a['href']
-        res = browser.open_relative(detail_url)
+        res = session.get(detail_url)
         if not res.ok:
-            raise ValueError(f"browser.open_relative('{detail_url}') failed")
-        t = browser.page.select_one('head > title').text
-        airav_number = str(re.findall(r'^\s*\[(.*?)]', t)[0]).upper()
-        if number.upper() != airav_number:
+            raise ValueError(f"session.get('{detail_url}') failed")
+        lx = fromstring(res.text)
+        t = str(lx.xpath('/html/head/title/text()')[0]).strip()
+        airav_number = str(re.findall(r'^\s*\[(.*?)]', t)[0])
+        if not re.search(number, airav_number, re.I):
             raise ValueError(f"page number ->[{airav_number}] not match")
-        desc = browser.page.select_one('li.introduction > span').text.strip()
+        desc = str(lx.xpath('//span[@id="ContentPlaceHolder1_Label2"]/text()')[0]).strip()
         return desc
     except Exception as e:
         if debug:
@@ -140,9 +147,9 @@ def getStoryline_airavwiki(number, debug):
     try:
         kwd = number[:6] if re.match(r'\d{6}[\-_]\d{2,3}', number) else number
         url = f'https://www.airav.wiki/api/video/list?lang=zh-TW&lng=zh-TW&search={kwd}'
-        result, browser = get_html_by_browser(url, return_type='browser')
-        if not result.ok:
-            raise ValueError(f"get_html_by_browser('{url}','{number}') failed")
+        result, session = get_html_session(url, return_type='session')
+        if not result:
+            raise ValueError(f"get_html_session('{url}','{number}') failed")
         j = json.loads(result.content)
         if int(j.get('count')) == 0:
             raise ValueError("number not found")
@@ -150,12 +157,12 @@ def getStoryline_airavwiki(number, debug):
         for r in j["result"]:
             n = r['barcode']
             if re.search(number, n, re.I):
-                link = f'/api/video/barcode/{n}?lng=zh-TW'
+                link = urljoin(result.url, f'/api/video/barcode/{n}?lng=zh-TW')
                 break
         if link is None:
             raise ValueError("number not found")
-        result = browser.open_relative(link)
-        if not result.ok or not re.search(number, browser.url, re.I):
+        result = session.get(link)
+        if not result.ok or not re.search(number, result.url, re.I):
             raise ValueError("detail page not found")
         j = json.loads(result.content)
         if int(j.get('count')) != 1:
@@ -221,7 +228,7 @@ def getStoryline_avno1(number, debug):  #获取剧情介绍 从avno1.cc取得
             form_select='div.wrapper > div.header > div.search > form',
             fields = {'kw' : number},
             return_type = 'browser')
-        if not result.ok:
+        if not result:
             raise ValueError(f"get_html_by_form('{url}','{number}') failed")
         s = browser.page.select('div.type_movie > div > ul > li > div')
         for div in s:
@@ -261,41 +268,45 @@ def getStoryline_amazon(q_title, number, debug):
     if not isinstance(q_title, str) or not len(q_title):
         return None
     try:
-        amazon_cookie, _ = load_cookies('amazon.json')
-        cookie = amazon_cookie if isinstance(amazon_cookie, dict) else None
+        cookie, cookies_filepath = load_cookies('amazon.json')
         url = "https://www.amazon.co.jp/s?k=" + q_title
-        res, browser = get_html_by_browser(url, cookies=cookie, return_type='browser')
-        if not res.ok:
-            raise ValueError("get_html_by_browser() failed")
-        lks = browser.links(r'/black-curtain/save-eligibility/black-curtain')
-        if isinstance(lks, list) and len(lks):
-            browser.follow_link(lks[0])
+        res, session = get_html_session(url, cookies=cookie, return_type='session')
+        if not res:
+            raise ValueError("get_html_session() failed")
+        lx = fromstring(res.text)
+        lks = lx.xpath('//a[contains(@href, "/black-curtain/save-eligibility/black-curtain")]/@href')
+        if len(lks) and lks[0].startswith('/'):
+            res = session.get(urljoin(res.url, lks[0]))
             cookie = None
-        html = etree.fromstring(str(browser.page), etree.HTMLParser())
-        titles = html.xpath("//span[contains(@class,'a-color-base a-text-normal')]/text()")
-        urls = html.xpath("//span[contains(@class,'a-color-base a-text-normal')]/../@href")
+            lx = fromstring(res.text)
+        titles = lx.xpath("//span[contains(@class,'a-color-base a-text-normal')]/text()")
+        urls = lx.xpath("//span[contains(@class,'a-color-base a-text-normal')]/../@href")
         if not len(urls) or len(urls) != len(titles):
             raise ValueError("titles not found")
         idx = amazon_select_one(titles, q_title, number, debug)
         if not isinstance(idx, int) or idx < 0:
             raise ValueError("title and number not found")
-        furl = urls[idx]
-        r = browser.open_relative(furl)
-        if not r.ok:
+        furl = urljoin(res.url, urls[idx])
+        res = session.get(furl)
+        if not res.ok:
             raise ValueError("browser.open_relative()) failed.")
-        lks = browser.links(r'/black-curtain/save-eligibility/black-curtain')
-        if isinstance(lks, list) and len(lks):
-            browser.follow_link(lks[0])
+        lx = fromstring(res.text)
+        lks = lx.xpath('//a[contains(@href, "/black-curtain/save-eligibility/black-curtain")]/@href')
+        if len(lks) and lks[0].startswith('/'):
+            res = session.get(urljoin(res.url, lks[0]))
             cookie = None
-
-        ama_t = browser.page.select_one('#productDescription > p').text.replace('\n',' ').strip()
-        ama_t = re.sub(r'審査番号:\d+', '', ama_t)
+            lx = fromstring(res.text)
+        div = lx.xpath('//*[@id="productDescription"]')[0]
+        ama_t = ' '.join([e.text.strip() for e in div if not re.search('Comment|h3', str(e.tag), re.I) and isinstance(e.text, str)])
+        ama_t = re.sub(r'審査番号:\d+', '', ama_t).strip()
 
         if cookie is None:
-        # 自动创建的cookies文件放在搜索路径表的末端，最低优先级。有amazon.co.jp帐号的用户可以从浏览器导出cookie放在靠前搜索路径
+            # 删除无效cookies，无论是用户创建还是自动创建，以避免持续故障
+            Path(cookies_filepath).unlink(missing_ok=True)
+            # 自动创建的cookies文件放在搜索路径表的末端，最低优先级。有amazon.co.jp帐号的用户可以从浏览器导出cookie放在靠前搜索路径
             ama_save = Path.home() / ".local/share/avdc/amazon.json"
             ama_save.parent.mkdir(parents=True, exist_ok=True)
-            ama_save.write_text(json.dumps(browser.session.cookies.get_dict(), sort_keys=True, indent=4), encoding='utf-8')
+            ama_save.write_text(json.dumps(session.cookies.get_dict(), sort_keys=True, indent=4), encoding='utf-8')
 
         return ama_t