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
 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',
@@ skipping 65 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={}):
+    """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))

[Vasily Kuznetsov, 2017/09/28 14:35:00]

Sorry that I'm only mentioning this now, but is there any reason to use
`dict(urlparse.parse_qsl(query))` instead of `urlparse.parse_qs(query)`? It
seems that there would only be a difference in cases with multiple values for
the same query parameter and arguably `parse_qs` behaves better in such case (it
doesn't discard all values except for the last, but makes lists of values
instead).

[tlucas, 2017/09/28 15:53:41]

You are right - Done.

+    query.update(get)
+
+    return urlparse.urlunparse((
+        scheme, netloc, path, params, urllib.urlencode(query), fragment
+    ))
+
+
+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),
+        post_data,
+        headers,
+    )
+
+    try:
+        result = urllib2.urlopen(request).read()
+    except urllib2.HTTPError as e:
+        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 = []
 
     def __setitem__(self, key, value):
         self.__order.append(key)
         dict.__setitem__(self, key, value)
 
     def iteritems(self):
@@ skipping 185 matching lines @@
         for match in re.finditer(r'&lang=([\w\-]+)"', firefoxLocales):
             locales.add(mapLocale('BCP-47', match.group(1)))
         langPacks = urllib2.urlopen('https://addons.mozilla.org/en-US/firefox/language-tools/').read()
         for match in re.finditer(r'<tr>.*?</tr>', langPacks, re.S):
             if match.group(0).find('Install Language Pack') >= 0:
                 match2 = re.search(r'lang="([\w\-]+)"', match.group(0))
                 if match2:
                     locales.add(mapLocale('BCP-47', match2.group(1)))
 
     allowed = set()
-    allowedLocales = json.load(urllib2.urlopen(
-        'https://crowdin.com/languages/languages_list?callback='
-    ))
+    allowedLocales = crowdin_request(projectName, 'supported-languages', key)
+
     for locale in allowedLocales:
-        allowed.add(locale['code'])
+        allowed.add(locale['crowdin_code'])
     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])
-    result = urllib2.urlopen('http://api.crowdin.net/api/project/%s/edit-project?key=%s' % (projectName, key), params).read()
-    if result.find('<success') < 0:
-        raise Exception('Server indicated that the operation was not successful\n' + result)
+
+    crowdin_request(projectName, 'edit-project', key, post_data=params)
 
 
-def postFiles(files, url):
+def crowdin_body_headers(files):

[Vasily Kuznetsov, 2017/09/28 14:35:00]

This function name is a bit unclear, probably `crowdin_body_and_headers` would
be better. Or even better and more "in the language of the domain" we could call
it something like `crowdin_prepare_upload`.

[tlucas, 2017/09/28 15:53:40]

Fair enough - Done.

+    """Create a post body and according headers, which Crowdin can handle."""

[Vasily Kuznetsov, 2017/09/28 14:35:00]

I think "matching" instead of "according" would be better here.

[tlucas, 2017/09/28 15:53:40]

Done.

     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
 
     body = body.encode('utf-8')
-    request = urllib2.Request(url, StringIO(body))
-    request.add_header('Content-Type', 'multipart/form-data; boundary=%s' % boundary)
-    request.add_header('Content-Length', len(body))
-    result = urllib2.urlopen(request).read()
-    if result.find('<success') < 0:
-        raise Exception('Server indicated that the operation was not successful\n' + result)
+    return (
+        StringIO(body),
+        {
+            'Content-Type': ('multipart/form-data; ; charset=utf-8; '
+                             'boundary=' + boundary),
+            'Content-Length': len(body)
+        }
+    )
 
 
 def updateTranslationMaster(localeConfig, metadata, dir, projectName, key):
-    result = json.load(urllib2.urlopen('http://api.crowdin.net/api/project/%s/info?key=%s&json=1' % (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):
         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, True)
                 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:
                 if newName in existing:
                     update.append((newName, data))
                     existing.remove(newName)
                 else:
                     add.append((newName, data))
 
     if len(add):
-        titles = urllib.urlencode([('titles[%s]' % name, re.sub(r'\.json', '', name)) for name, data in add])
-        postFiles(add, 'http://api.crowdin.net/api/project/%s/add-file?key=%s&type=chrome&%s' % (projectName, key, titles))
+        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,
+                        headers=headers)
     if len(update):
-        postFiles(update, 'http://api.crowdin.net/api/project/%s/update-file?key=%s' % (projectName, key))
+        data, headers = crowdin_body_headers(update)
+        crowdin_request(projectName, 'update-file', key, post_data=data,
+                        headers=headers)
     for file in existing:
-        result = urllib2.urlopen('http://api.crowdin.net/api/project/%s/delete-file?key=%s&file=%s' % (projectName, key, file)).read()
-        if result.find('<success') < 0:
-            raise Exception('Server indicated that the operation was not successful\n' + result)
+        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):
-        postFiles(files, 'http://api.crowdin.net/api/project/%s/upload-translation?key=%s&language=%s' % (
-            projectName, key, mapLocale(localeConfig['name_format'], locale))
-        )
+        language = mapLocale(localeConfig['name_format'], locale)
+        data, headers = crowdin_body_headers(files)
+        crowdin_request(projectName, 'upload-translation', key,
+                        {'language': language}, post_data=data,
+                        headers=headers)
 
 
 def getTranslations(localeConfig, projectName, key):
-    result = urllib2.urlopen('http://api.crowdin.net/api/project/%s/export?key=%s' % (projectName, key)).read()
-    if result.find('<success') < 0:
-        raise Exception('Server indicated that the operation was not successful\n' + result)
+    """Download all available translations from crowdin.
 
-    result = urllib2.urlopen('http://api.crowdin.net/api/project/%s/download/all.zip?key=%s' % (projectName, key)).read()
+    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'],
                                         normalizedDefaultLocale)
 
     for info in zip.infolist():
@@ skipping 47 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)