Merge pull request #93 from Algolytics/windows-encoding-support

proper windows encoding support on linux machine

docs/usage.rst

@@ -62,7 +62,7 @@ Create new job
 
     job = dq.submit_job(job_config, input_data=input_data)                                         # with data in a variable
 
-    job = dq.submit_job(job_config, input_file='my_file.csv', input_file_encoding='windows-1250')  # with data inside file
+    job = dq.submit_job(job_config, input_file='my_file.csv')                                      # with data inside file
 
     print(job.id)
     print(job.name)
@@ -97,7 +97,7 @@ Create new deduplication job
 	job_config.module_std(address=True, names=True, contact=True)
 	job_config.extend(gus=True, geocode=True, diagnostic=True)
 
-	job = dq.submit_job(job_config, input_data=input_data)  
+	job = dq.submit_job(job_config, input_data=input_data)
 
 	print(job)
 	...
@@ -176,7 +176,7 @@ Analogously for other modules:
     * NIP
     * REGON
 
- 
+
 Check job state
 ---------------
 ::

dq/__init__.py

@@ -5,6 +5,6 @@
 
 __author__ = """Mikołaj Olszewski"""
 __email__ = 'mikolaj.olszewski@algolytics.pl'
-__version__ = '0.4.0'
+__version__ = '0.5.0'
 
 __all__ = ['DQClient', 'JobConfig']

dq/client.py

@@ -24,8 +24,7 @@ def list_jobs(self):
         return self.request_client.get('/jobs')
 
     @from_response(Job)
-    def submit_job(self, config, input_data=None, input_file=None,
-                   input_file_encoding='utf-8'):
+    def submit_job(self, config, input_data=None, input_file=None):
         """ POST 'https://app.dataquality.pl/api/v1/jobs' """
         parts = {
             'config': {
@@ -41,9 +40,7 @@ def submit_job(self, config, input_data=None, input_file=None,
         if input_data:
             parts['file']['content'] = str(input_data)
         else:
-            parts['file']['content'] = open(input_file, 'r',
-                                            encoding=input_file_encoding)
-
+            parts['file']['content'] = open(input_file, 'rb')
         return self.request_client.post_multipart('/jobs', parts=parts)
 
     def job_state(self, job_id):
@@ -61,13 +58,13 @@ def job_report(self, job_id):
         """ GET 'https://app.dataquality.pl/api/v1/jobs/{{job_id}}' """
         return self.request_client.get('/jobs/{}'.format(job_id))
 
-    def job_results(self, job_id, out_file=None, code_page = 'utf-8'):
+    def job_results(self, job_id, out_file=None, encoding='utf-8'):
         """ GET 'https://app.dataquality.pl/api/v1/jobs/{{job_id}}/result' """
-        response = self.request_client.get('/jobs/{}/result'.format(job_id))
+        response = self.request_client.get('/jobs/{}/result'.format(job_id), encoding=encoding)
         if not response.is_ok():
             raise DQError(status=response.status, message=response.content)
         if out_file:
-            with open(out_file, 'w', encoding=code_page) as file:
+            with open(out_file, 'w', encoding=encoding) as file:
                 file.write(response.content)
 
     def delete_job(self, job_id):

dq/request_client.py

@@ -18,8 +18,9 @@ def __init__(self, base_url, token):
     def __make_url(self, path):
         return '{}/{}'.format(self.base_url, path)
 
-    def get(self, path, headers={}):
+    def get(self, path, headers={}, encoding=None):
         response = self.session.get(self.__make_url(path), headers=headers)
+        response.encoding = encoding
         return Response('GET', response.status_code, response.text)
 
     def post(self, path, headers={}):

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.4.0
+current_version = 0.5.0
 commit = True
 tag = True
 

setup.py

@@ -19,7 +19,7 @@
 
 setup(
     name='dq-client',
-    version='0.4.0',
+    version='0.5.0',
     description="Python library which allows to use http://dataquality.pl in easy way.",
     long_description=readme + '\n\n' + history,
     author="Mikołaj Olszewski",