From a66b964ecc5f56a21caa91efe493102949b85eef Mon Sep 17 00:00:00 2001
From: Nikolay Kim <fafhrd91@gmail.com>
Date: Mon, 7 Jul 2014 16:14:01 -0700
Subject: [PATCH] add chardet support to ClientResponse.text() and
 ClientResponse.json() #105

---
 aiohttp/client.py    | 29 ++++++++++++++++++++---------
 setup.py             |  2 +-
 tests/test_client.py | 34 ++++++++++++++++++++++++++++++++++
 3 files changed, 55 insertions(+), 10 deletions(-)

diff --git a/aiohttp/client.py b/aiohttp/client.py
index 28d4503a03f..84e612c67d5 100644
--- a/aiohttp/client.py
+++ b/aiohttp/client.py
@@ -15,6 +15,10 @@
 import urllib.parse
 import weakref
 import warnings
+try:
+    import chardet
+except ImportError:  # pragma: no cover
+    chardet = None
 
 import aiohttp
 from . import helpers, streams
@@ -701,17 +705,26 @@ def read_and_close(self, decode=False):
         )
         return (yield from self.read(decode))
 
+    def _get_encoding(self, encoding):
+        ctype = self.headers.get('CONTENT-TYPE', '').lower()
+        mtype, stype, _, params = helpers.parse_mimetype(ctype)
+
+        if not encoding:
+            encoding = params.get('charset')
+            if not encoding and chardet:
+                encoding = chardet.detect(self._content)['encoding']
+            if not encoding:  # pragma: no cover
+                encoding = 'utf-8'
+
+        return encoding
+
     @asyncio.coroutine
     def text(self, encoding=None):
         """Read response payload and decode."""
         if self._content is None:
             yield from self.read()
 
-        ctype = self.headers.get('CONTENT-TYPE', '').lower()
-        mtype, stype, _, params = helpers.parse_mimetype(ctype)
-
-        encoding = encoding or params.get('charset', 'utf-8')
-        return self._content.decode(encoding)
+        return self._content.decode(self._get_encoding(encoding))
 
     @asyncio.coroutine
     def json(self, *, encoding=None, loads=json.loads):
@@ -720,16 +733,14 @@ def json(self, *, encoding=None, loads=json.loads):
             yield from self.read()
 
         ctype = self.headers.get('CONTENT-TYPE', '').lower()
-        mtype, stype, _, params = helpers.parse_mimetype(ctype)
-        if not (mtype == 'application' or stype == 'json'):
+        if 'json' not in ctype:
             client_log.warning(
                 'Attempt to decode JSON with unexpected mimetype: %s', ctype)
 
         if not self._content.strip():
             return None
 
-        encoding = encoding or params.get('charset', 'utf-8')
-        return loads(self._content.decode(encoding))
+        return loads(self._content.decode(self._get_encoding(encoding)))
 
 
 class HttpClient:
diff --git a/setup.py b/setup.py
index 1b4daf195fa..8b119637e42 100644
--- a/setup.py
+++ b/setup.py
@@ -18,7 +18,7 @@
 else:
     install_requires = ['asyncio']
 
-tests_require = install_requires + ['nose', 'gunicorn']
+tests_require = install_requires + ['nose', 'gunicorn', 'chardet']
 
 
 def read(f):
diff --git a/tests/test_client.py b/tests/test_client.py
index 00b41dc3004..7ffec8c3ce2 100644
--- a/tests/test_client.py
+++ b/tests/test_client.py
@@ -173,6 +173,23 @@ def second_call(*args, **kwargs):
         self.assertEqual(res, '{"тест": "пройден"}')
         self.assertTrue(self.response.close.called)
 
+    def test_text_detect_encoding(self):
+        def side_effect(*args, **kwargs):
+            def second_call(*args, **kwargs):
+                raise aiohttp.EofStream
+            fut = asyncio.Future(loop=self.loop)
+            fut.set_result('{"тест": "пройден"}'.encode('cp1251'))
+            content.read.side_effect = second_call
+            return fut
+        self.response.headers = {'CONTENT-TYPE': 'application/json'}
+        content = self.response.content = unittest.mock.Mock()
+        content.read.side_effect = side_effect
+        self.response.close = unittest.mock.Mock()
+
+        res = self.loop.run_until_complete(self.response.text())
+        self.assertEqual(res, '{"тест": "пройден"}')
+        self.assertTrue(self.response.close.called)
+
     def test_json(self):
         def side_effect(*args, **kwargs):
             def second_call(*args, **kwargs):
@@ -230,6 +247,23 @@ def second_call(*args, **kwargs):
         self.assertEqual(res, {'тест': 'пройден'})
         self.assertTrue(self.response.close.called)
 
+    def test_json_detect_encoding(self):
+        def side_effect(*args, **kwargs):
+            def second_call(*args, **kwargs):
+                raise aiohttp.EofStream
+            fut = asyncio.Future(loop=self.loop)
+            fut.set_result('{"тест": "пройден"}'.encode('cp1251'))
+            content.read.side_effect = second_call
+            return fut
+        self.response.headers = {'CONTENT-TYPE': 'application/json'}
+        content = self.response.content = unittest.mock.Mock()
+        content.read.side_effect = side_effect
+        self.response.close = unittest.mock.Mock()
+
+        res = self.loop.run_until_complete(self.response.json())
+        self.assertEqual(res, {'тест': 'пройден'})
+        self.assertTrue(self.response.close.called)
+
     def test_override_flow_control(self):
         class MyResponse(ClientResponse):
             flow_control_class = aiohttp.FlowControlDataQueue