always utf-8, fixes #3209

ox/api.py

@@ -227,6 +227,8 @@ def signin(url):
         url = 'https://%s/api/' % url
     else:
         site = url.split('/')[2]
+    if not url.endswith('/'):
+        url += '/'
     api = API(url)
     update = False
     try:

ox/cache.py

@@ -456,7 +456,47 @@ class RedisCache(KVCache):
         self.backend = redis.from_url(self.url)
 
 
-if cache_path().startswith('fs:'):
+class FallbackCache(KVCache):
+    caches = []
+
+    def __init__(self):
+        fallback = cache_path()
+        for path in fallback.split('|'):
+            os.environ['oxCACHE'] = path
+            if path.startswith('redis:'):
+                store = RedisCache()
+            elif path.startswith('memcache:'):
+                store = MemCache()
+            self.caches.append(store)
+        os.environ['oxCACHE'] = fallback
+
+    def get(self, url, data, headers=None, timeout=-1, value="data"):
+        if timeout == 0:
+            return None
+
+        info_key, data_key = self._keys(url, data, headers)
+        for cache in self.caches:
+            try:
+                info = cache.backend.get(info_key)
+            except:
+                info = None
+            if info:
+                return cache.get(url, data, headers, timeout, value)
+        return None
+
+    def set(self, url, post_data, data, headers):
+        self.caches[0].set(url, post_data, data, headers)
+        for cache in self.caches[1:]:
+            cache.delete(url, post_data, headers)
+
+    def delete(self, url, data=None, headers=None):
+        for cache in self.caches:
+            cache.delete(url, data, headers)
+
+
+if '|' in cache_path():
+    store = FallbackCache()
+elif cache_path().startswith('fs:'):
     store = FileCache()
 elif cache_path().startswith('redis:'):
     store = RedisCache()

ox/web/imdb.py

@@ -187,7 +187,7 @@ class Imdb(SiteParser):
         ], type='int'),
         'keyword': {
             'page': 'keywords',
-            're': '<a href="/keyword/.*?>(.*?)</a>',
+            're': 'data-item-keyword="(.*?)"',
             'type': 'list'
         },
         'language': zebra_list('Language', more=['<a.*?>(.*?)</a>']),
@@ -224,7 +224,7 @@ class Imdb(SiteParser):
         'releasedate': {
             'page': 'releaseinfo',
             're': [
-                '<td class="release_date">(.*?)</td>',
+                '<td class="release-date-item__date".*?>(.*?)</td>',
                 strip_tags,
             ],
             'type': 'list'

ox/web/piratecinema.py

@@ -7,7 +7,7 @@ from ox.net import read_url
 
 def get_poster_url(id):
     url = 'http://piratecinema.org/posters/'
-    html = read_url(url, unicode=True)
+    html = read_url(url).decode('utf-8')
     results = re.compile('src="(.+)" title=".+\((\d{7})\)"').findall(html)
     for result in results:
         if result[1] == id: