A larger pool of random UAs

.gitignore

@@ -26,7 +26,7 @@ resources/content-server/locales.zip
 resources/content-server/mathjax.zip.xz
 resources/content-server/mathjax.version
 resources/mozilla-ca-certs.pem
-resources/common-user-agents.txt
+resources/user-agent-data.json
 icons/icns/*.iconset
 setup/installer/windows/calibre/build.log
 tags

setup/browser_data.py

@@ -0,0 +1,115 @@
+#!/usr/bin/env python2
+# vim:fileencoding=utf-8
+# License: GPLv3 Copyright: 2017, Kovid Goyal <kovid at kovidgoyal.net>
+
+from __future__ import absolute_import, division, print_function, unicode_literals
+
+import os
+import re
+from datetime import datetime
+
+from setup import download_securely
+
+is_ci = os.environ.get('CI', '').lower() == 'true'
+
+
+def filter_ans(ans):
+    return filter(None, (x.strip() for x in ans))
+
+
+def common_user_agents():
+    if is_ci:
+        return [
+            # IE 11 - windows 10
+            'Mozilla/5.0 (Windows NT 10.0; Trident/7.0; rv:11.0) like Gecko',
+            # IE 11 - windows 8.1
+            'Mozilla/5.0 (Windows NT 6.3; Trident/7.0; rv:11.0) like Gecko',
+            # IE 11 - windows 8
+            'Mozilla/5.0 (Windows NT 6.2; Trident/7.0; rv:11.0) like Gecko',
+            # IE 11 - windows 7
+            'Mozilla/5.0 (Windows NT 6.1; Trident/7.0; rv:11.0) like Gecko',
+            # 32bit IE 11 on 64 bit win 10
+            'Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko',
+            # 32bit IE 11 on 64 bit win 8.1
+            'Mozilla/5.0 (Windows NT 6.3; WOW64; Trident/7.0; rv:11.0) like Gecko',
+            # 32bit IE 11 on 64 bit win 7
+            'Mozilla/5.0 (Windows NT 6.1; WOW64; Trident/7.0; rv:11.0) like Gecko',
+        ]
+    print('Getting recent UAs...')
+    raw = download_securely(
+        'https://techblog.willshouse.com/2012/01/03/most-common-user-agents/').decode('utf-8')
+    lines = re.search(
+        r'<textarea.+"get-the-list".+>([^<]+)</textarea>', raw).group(1).splitlines()
+    ans = filter_ans(lines)
+    if not ans:
+        raise ValueError('Failed to download list of common UAs')
+    return ans
+
+
+def firefox_versions():
+    if is_ci:
+        return '51.0 50.0'.split()
+    print('Getting firefox versions...')
+    import html5lib
+    raw = download_securely(
+        'https://www.mozilla.org/en-US/firefox/releases/').decode('utf-8')
+    root = html5lib.parse(raw, treebuilder='lxml', namespaceHTMLElements=False)
+    ol = root.xpath('//div[@id="main-content"]/ol')[0]
+    ol.xpath('descendant::li/strong/a[@href]')
+    ans = filter_ans(ol.xpath('descendant::li/strong/a[@href]/text()'))
+    if not ans:
+        raise ValueError('Failed to download list of firefox versions')
+    return ans
+
+
+def chrome_versions():
+    if is_ci:
+        return []
+    print('Getting chrome versions...')
+    import html5lib
+    raw = download_securely(
+        'https://en.wikipedia.org/wiki/Google_Chrome_version_history').decode('utf-8')
+    root = html5lib.parse(raw, treebuilder='lxml', namespaceHTMLElements=False)
+    table = root.xpath('//*[@id="mw-content-text"]//tbody')[-1]
+    ans = []
+    for tr in table.iterchildren('tr'):
+        cells = tuple(tr.iterchildren('td'))
+        if not cells:
+            continue
+        if not cells[2].text or not cells[2].text.strip():
+            continue
+        s = cells[0].get('style')
+        if '#a0e75a' not in s and 'salmon' not in s:
+            break
+        chrome_version = cells[0].text.strip()
+        ts = datetime.strptime(cells[1].text.strip().split()[
+                               0], '%Y-%m-%d').date().strftime('%Y-%m-%d')
+        try:
+            webkit_version = cells[2].text.strip().split()[1]
+        except IndexError:
+            continue
+        ans.append({'date': ts, 'chrome_version': chrome_version,
+                    'webkit_version': webkit_version})
+    return list(reversed(ans))
+
+
+def all_desktop_platforms(user_agents):
+    ans = set()
+    for ua in user_agents:
+        if 'Mobile/' not in ua and ('Firefox/' in ua or 'Chrome/' in ua):
+            plat = ua.partition('(')[2].partition(')')[0]
+            parts = plat.split(';')
+            if 'Firefox/' in ua:
+                del parts[-1]
+            ans.add(';'.join(parts))
+    return ans
+
+
+def get_data():
+    ans = {
+        'chrome_versions': chrome_versions(),
+        'firefox_versions': firefox_versions(),
+        'common_user_agents': common_user_agents(),
+    }
+    ans['desktop_platforms'] = list(all_desktop_platforms(ans['common_user_agents']))
+    return ans

setup/resources.py

@@ -257,38 +257,14 @@ class CACerts(Command):  # {{{
 
 class RecentUAs(Command):  # {{{
 
-    description = 'Get updated list of recent browser user agents'
-    UA_PATH = os.path.join(Command.RESOURCES, 'common-user-agents.txt')
-
-    def get_list(self):
-        if is_ci:
-            # Dont hammer the server from CI
-            return [
-                 # IE 11 - windows 10
-                 'Mozilla/5.0 (Windows NT 10.0; Trident/7.0; rv:11.0) like Gecko',
-                 # IE 11 - windows 8.1
-                 'Mozilla/5.0 (Windows NT 6.3; Trident/7.0; rv:11.0) like Gecko',
-                 # IE 11 - windows 8
-                 'Mozilla/5.0 (Windows NT 6.2; Trident/7.0; rv:11.0) like Gecko',
-                 # IE 11 - windows 7
-                 'Mozilla/5.0 (Windows NT 6.1; Trident/7.0; rv:11.0) like Gecko',
-                 # 32bit IE 11 on 64 bit win 10
-                 'Mozilla/5.0 (Windows NT 10.0; WOW64; Trident/7.0; rv:11.0) like Gecko',
-                 # 32bit IE 11 on 64 bit win 8.1
-                 'Mozilla/5.0 (Windows NT 6.3; WOW64; Trident/7.0; rv:11.0) like Gecko',
-                 # 32bit IE 11 on 64 bit win 7
-                 'Mozilla/5.0 (Windows NT 6.1; WOW64; Trident/7.0; rv:11.0) like Gecko',
-        ]
-        raw = download_securely('https://techblog.willshouse.com/2012/01/03/most-common-user-agents/').decode('utf-8')
-        lines = re.search(r'<textarea.+"get-the-list".+>([^<]+)</textarea>', raw).group(1).splitlines()
-        return [x.strip() for x in lines if x.strip()]
+    description = 'Get updated list of common browser user agents'
+    UA_PATH = os.path.join(Command.RESOURCES, 'user-agent-data.json')
 
     def run(self, opts):
-        lines = self.get_list()[:10]
-        if not lines:
-            raise RuntimeError('Failed to download list of common user agents')
+        from setup.browser_data import get_data
+        data = get_data()
         with open(self.UA_PATH, 'wb') as f:
-            f.write('\n'.join(lines).encode('ascii'))
+            f.write(json.dumps(data, indent=2))
 # }}}
 
 

src/calibre/__init__.py

@@ -397,12 +397,11 @@ USER_AGENT_MOBILE = 'Mozilla/5.0 (Windows; U; Windows CE 5.1; rv:1.8.1a3) Gecko/
 
 
 def random_user_agent(choose=None, allow_ie=True):
-    try:
-        ua_list = random_user_agent.ua_list
-    except AttributeError:
-        ua_list = random_user_agent.ua_list = P('common-user-agents.txt', data=True, allow_user_override=False).decode('utf-8').splitlines()
+    from calibre.utils.random_ua import common_user_agents
+    ua_list = common_user_agents()
+    ua_list = filter(lambda x: 'Mobile/' not in x, ua_list)
     if not allow_ie:
-        ua_list = filter(lambda x: 'Firefox/' in x or 'Chrome/' in x, ua_list)
+        ua_list = filter(lambda x: 'Trident/' not in x and 'Edge/' not in x, ua_list)
     return random.choice(ua_list) if choose is None else ua_list[choose]
 
 

src/calibre/utils/random_ua.py

@@ -0,0 +1,50 @@
+#!/usr/bin/env python2
+# vim:fileencoding=utf-8
+# License: GPLv3 Copyright: 2017, Kovid Goyal <kovid at kovidgoyal.net>
+
+from __future__ import absolute_import, division, print_function, unicode_literals
+
+import json
+import random
+
+
+def user_agent_data():
+    ans = getattr(user_agent_data, 'ans', None)
+    if ans is None:
+        ans = user_agent_data.ans = json.loads(
+            P('user-agent-data.json', data=True, allow_user_override=False))
+    return ans
+
+
+def common_user_agents():
+    return user_agent_data()['common_user_agents']
+
+
+def random_firefox_version():
+    versions = user_agent_data()['firefox_versions'][:7]
+    return random.choice(versions)
+
+
+def random_desktop_platform():
+    return random.choice(user_agent_data()['desktop_platforms'])
+
+
+def random_firefox_ua():
+    # https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/User-Agent/Firefox
+    return 'Mozilla/5.0 ({p}; rv:{ver}) Gecko/20100101 Firefox/{ver}'.format(
+        p=random_desktop_platform(), ver=random_firefox_version())
+
+
+def random_chrome_version():
+    versions = user_agent_data()['chrome_versions'][:7]
+    return random.choice(versions)
+
+
+def random_chrome_ua():
+    v = random_chrome_version()
+    return 'Mozilla/5.0 ({p}) AppleWebKit/{wv} (KHTML, like Gecko) Chrome/{cv} Safari/{wv}'.format(
+        p=random_desktop_platform(), wv=v['webkit_version'], cv=v['chrome_version'])
+
+
+def random_user_agent():
+    return random.choice((random_chrome_ua, random_firefox_ua))()