Issue 5777 - Update crowdin interface

localeTools.py

 # This Source Code Form is subject to the terms of the Mozilla Public
 # License, v. 2.0. If a copy of the MPL was not distributed with this
 # file, You can obtain one at http://mozilla.org/MPL/2.0/.
 
 import re
 import os
 import sys
 import codecs
 import json
 import urlparse
 import urllib
 import urllib2
+import mimetypes
 from StringIO import StringIO
 from ConfigParser import SafeConfigParser
 from zipfile import ZipFile
 from xml.parsers.expat import ParserCreate, XML_PARAM_ENTITY_PARSING_ALWAYS
 
 langMappingGecko = {
     'bn-BD': 'bn',
     'br': 'br-FR',
     'dsb': 'dsb-DE',
     'fj-FJ': 'fj',
@@ skipping 63 matching lines @@
     'ta',
     'te',
     'th',
     'tr',
     'uk',
     'vi',
     'zh-CN',
     'zh-TW',
 ]
 
-CROWDIN_AP_URL = 'https://api.crowdin.com/api/project/{}/{}'
-
-
-def crowdin_url(project_name, action, key, get={}):

[kzar, 2017/09/28 09:48:29]

IIRC default parameter values in Python are shared, so you generally shouldn't
mutate them or use them when they might be mutated. (Same below.)

[tlucas, 2017/09/28 10:48:40]

As discussed in IRC: You are correct, but the implied problems can't occur since
'key' and 'json' get written explicitly everytime. Won't change unless someone
insists.

-    """Create a valid url for a crowdin endpoint."""
-    url = CROWDIN_AP_URL.format(project_name, action)
-    get['key'] = key
-    get['json'] = 1
-
-    scheme, netloc, path, params, query, fragment = urlparse.urlparse(url)
-
-    query = dict(urlparse.parse_qsl(query))
-    query.update(get)
-
-    return urlparse.urlunparse((
-        scheme, netloc, path, params, urllib.urlencode(query), fragment
-    ))
+CROWDIN_AP_URL = 'https://api.crowdin.com/api/project'
 
 
 def crowdin_request(project_name, action, key, get={}, post_data=None,
                     headers={}, raw=False):
     """Perform a call to crowdin and raise an Exception on failure."""
     request = urllib2.Request(
-        crowdin_url(project_name, action, key, get),
+        '{}/{}/{}?{}'.format(CROWDIN_AP_URL,
+                             urllib.quote(project_name),
+                             urllib.quote(action),
+                             urllib.urlencode(dict(get, key=key, json=1))),
         post_data,
         headers,
     )
 
     try:
         result = urllib2.urlopen(request).read()
     except urllib2.HTTPError as e:
-        err = e.read()
-        raise Exception(

[kzar, 2017/09/28 09:48:29]

Why catch this and raise it again, does it make the output clearer?

[tlucas, 2017/09/28 10:48:40]

It does - e.g. if urllib2 raised an HTTPError(400) this would be the only
output:

  File "/usr/lib/python2.7/urllib2.py", line 556, in http_error_default
    raise HTTPError(req.get_full_url(), code, msg, hdrs, fp)
urllib2.HTTPError: HTTP Error 400: Bad Request

compared to catching and raising again:

Exception: Server indicated that the operation was not successful
{
  "success":false,
  "error":{
    "code":13,
    "message":"Language \"foo\" is invalid or not supported"
  }
}

[kzar, 2017/09/28 11:34:44]

Cool fair enough. Mind making it display the status code too though? Knowing it
was a HTTP 400 error is actually pretty useful.

-            'Server indicated that the operation was not successful\n' + err
-        )
+        raise Exception('Server returned HTTP Error {}:\n{}'.format(e.code,
+                                                                    e.read()))
 
     if not raw:
         return json.loads(result)
 
     return result
 
 
 class OrderedDict(dict):
     def __init__(self):
         self.__order = []
@@ skipping 205 matching lines @@
     if not allowed.issuperset(locales):
         print "Warning, following locales aren't allowed by server: " + ', '.join(locales - allowed)
 
     locales = list(locales & allowed)
     locales.sort()
     params = urllib.urlencode([('languages[]', locale) for locale in locales])
 
     crowdin_request(projectName, 'edit-project', key, post_data=params)
 
 
-def crowdin_body_headers(files):
-    """Create a post body and according headers, which Crowdin can handle."""
+def crowdin_prepare_upload(files):
+    """Create a post body and matching headers, which Crowdin can handle."""
     boundary = '----------ThIs_Is_tHe_bouNdaRY_$'
     body = ''
-    for file, data in files:
-        body += '--%s\r\n' % boundary
-        body += 'Content-Disposition: form-data; name="files[%s]"; filename="%s"\r\n' % (file, file)
-        body += 'Content-Type: application/octet-stream\r\n'
-        body += 'Content-Transfer-Encoding: binary\r\n'
-        body += '\r\n' + data + '\r\n'
-    body += '--%s--\r\n' % boundary
+    for name, data in files:
+        mimetype = mimetypes.guess_type(name)[0]
+        body += (
+            '--{boundary}\r\n'
+            'Content-Disposition: form-data; name="files[{name}]"; '
+            'filename="{name}"\r\n'
+            'Content-Type: {mimetype}; charset=utf-8\r\n'
+            'Content-Transfer-Encoding: binary\r\n'
+            '\r\n{data}\r\n'
+            '--{boundary}--\r\n'
+        ).format(boundary=boundary, name=name, data=data, mimetype=mimetype)
 
     body = body.encode('utf-8')
     return (
         StringIO(body),
         {
-            'Content-Type': 'multipart/form-data; ; charset=utf-8; boundary=' + boundary,

[kzar, 2017/09/28 09:48:29]

This line seems too long, does this code pass linting?

[tlucas, 2017/09/28 10:48:40]

It's about time to get rid of tox.ini 'ignore' entries... Done.

+            'Content-Type': ('multipart/form-data; boundary=' + boundary),
             'Content-Length': len(body)
-        }
+        },
     )
 
 
 def updateTranslationMaster(localeConfig, metadata, dir, projectName, key):
     result = crowdin_request(projectName, 'info', key)
 
     existing = set(map(lambda f: f['name'], result['files']))
     add = []
     update = []
     for file in os.listdir(dir):
@@ skipping 12 matching lines @@
                 newName = file + '.json'
 
             if data:
                 if newName in existing:
                     update.append((newName, data))
                     existing.remove(newName)
                 else:
                     add.append((newName, data))
 
     if len(add):
-        data = {'titles[{}]'.format(name): re.sub(r'\.json', '', name)
-                for name, data in add}
-        data['type'] = 'chrome'
-        data, headers = crowdin_body_headers(add)
-        crowdin_request(projectName, 'add-file', key, post_data=data,
+        query = {'titles[{}]'.format(name): os.path.splitext(name)[0]
+                 for name, _ in add}
+        query['type'] = 'chrome'
+        data, headers = crowdin_prepare_upload(add)
+        crowdin_request(projectName, 'add-file', key, query, post_data=data,
                         headers=headers)
     if len(update):
-        data, headers = crowdin_body_headers(update)
+        data, headers = crowdin_prepare_upload(update)
         crowdin_request(projectName, 'update-file', key, post_data=data,
                         headers=headers)
     for file in existing:
         crowdin_request(projectName, 'delete-file', key, {'file': file})
 
 
 def uploadTranslations(localeConfig, metadata, dir, locale, projectName, key):
     files = []
     for file in os.listdir(dir):
         path = os.path.join(dir, file)
         if os.path.isfile(path):
             if localeConfig['file_format'] == 'chrome-json' and file.endswith('.json'):
                 data = preprocessChromeLocale(path, metadata, False)
                 newName = file
             elif localeConfig['file_format'] == 'chrome-json':
                 fileHandle = codecs.open(path, 'rb', encoding='utf-8')
                 data = json.dumps({file: {'message': fileHandle.read()}})
                 fileHandle.close()
                 newName = file + '.json'
             else:
                 data = toJSON(path)
                 newName = file + '.json'
 
             if data:
                 files.append((newName, data))
     if len(files):
         language = mapLocale(localeConfig['name_format'], locale)
-        data, headers = crowdin_body_headers(files)
+        data, headers = crowdin_prepare_upload(files)
         crowdin_request(projectName, 'upload-translation', key,
                         {'language': language}, post_data=data,
                         headers=headers)
 
 
 def getTranslations(localeConfig, projectName, key):
-    # let crowdin build the project

[kzar, 2017/09/28 09:48:30]

Maybe this should be a docstring for the function instead?

[tlucas, 2017/09/28 10:48:40]

Done.

+    """Download all available translations from crowdin.
+
+    Trigger crowdin to build the available export, wait for crowdin to
+    finish the job and download the generated zip afterwards.
+    """
     crowdin_request(projectName, 'export', key)
 
     result = crowdin_request(projectName, 'download/all.zip', key, raw=True)
     zip = ZipFile(StringIO(result))
     dirs = {}
 
     normalizedDefaultLocale = localeConfig['default_locale']
     if localeConfig['name_format'] == 'ISO-15897':
         normalizedDefaultLocale = normalizedDefaultLocale.replace('_', '-')
     normalizedDefaultLocale = mapLocale(localeConfig['name_format'],
@@ skipping 50 matching lines @@
 
     # Remove any extra files
     for dir, files in dirs.iteritems():
         baseDir = os.path.join(localeConfig['base_path'], dir)
         if not os.path.exists(baseDir):
             continue
         for file in os.listdir(baseDir):
             path = os.path.join(baseDir, file)
             if os.path.isfile(path) and (file.endswith('.json') or file.endswith('.properties') or file.endswith('.dtd')) and not file in files:
                 os.remove(path)