1 # Copyright (C) The Arvados Authors. All rights reserved.
3 # SPDX-License-Identifier: Apache-2.0
5 from __future__ import absolute_import
6 from future import standard_library
7 standard_library.install_aliases()
8 from builtins import range
21 from apiclient import discovery as apiclient_discovery
22 from apiclient import errors as apiclient_errors
28 _logger = logging.getLogger('arvados.api')
30 MAX_IDLE_CONNECTION_DURATION = 30
31 RETRY_DELAY_INITIAL = 2
32 RETRY_DELAY_BACKOFF = 2
35 class OrderedJsonModel(apiclient.model.JsonModel):
36 """Model class for JSON that preserves the contents' order.
38 API clients that care about preserving the order of fields in API
39 server responses can use this model to do so, like this::
41 from arvados.api import OrderedJsonModel
42 client = arvados.api('v1', ..., model=OrderedJsonModel())
45 def deserialize(self, content):
46 # This is a very slightly modified version of the parent class'
47 # implementation. Copyright (c) 2010 Google.
48 content = content.decode('utf-8')
49 body = json.loads(content, object_pairs_hook=collections.OrderedDict)
50 if self._data_wrapper and isinstance(body, dict) and 'data' in body:
55 def _intercept_http_request(self, uri, method="GET", headers={}, **kwargs):
56 if (self.max_request_size and
57 kwargs.get('body') and
58 self.max_request_size < len(kwargs['body'])):
59 raise apiclient_errors.MediaUploadSizeError("Request size %i bytes exceeds published limit of %i bytes" % (len(kwargs['body']), self.max_request_size))
61 if config.get("ARVADOS_EXTERNAL_CLIENT", "") == "true":
62 headers['X-External-Client'] = '1'
64 headers['Authorization'] = 'OAuth2 %s' % self.arvados_api_token
65 if not headers.get('X-Request-Id'):
66 headers['X-Request-Id'] = self._request_id()
68 retryable = method in [
69 'DELETE', 'GET', 'HEAD', 'OPTIONS', 'PUT']
70 retry_count = self._retry_count if retryable else 0
73 time.time() - self._last_request_time > self._max_keepalive_idle):
74 # High probability of failure due to connection atrophy. Make
75 # sure this request [re]opens a new connection by closing and
76 # forgetting all cached connections first.
77 for conn in self.connections.values():
79 self.connections.clear()
81 delay = self._retry_delay_initial
82 for _ in range(retry_count):
83 self._last_request_time = time.time()
85 return self.orig_http_request(uri, method, headers=headers, **kwargs)
86 except http.client.HTTPException:
87 _logger.debug("Retrying API request in %d s after HTTP error",
90 # This is the one case where httplib2 doesn't close the
91 # underlying connection first. Close all open
92 # connections, expecting this object only has the one
93 # connection to the API server. This is safe because
94 # httplib2 reopens connections when needed.
95 _logger.debug("Retrying API request in %d s after socket error",
97 for conn in self.connections.values():
100 delay = delay * self._retry_delay_backoff
102 self._last_request_time = time.time()
103 return self.orig_http_request(uri, method, headers=headers, **kwargs)
105 def _patch_http_request(http, api_token):
106 http.arvados_api_token = api_token
107 http.max_request_size = 0
108 http.orig_http_request = http.request
109 http.request = types.MethodType(_intercept_http_request, http)
110 http._last_request_time = 0
111 http._max_keepalive_idle = MAX_IDLE_CONNECTION_DURATION
112 http._retry_delay_initial = RETRY_DELAY_INITIAL
113 http._retry_delay_backoff = RETRY_DELAY_BACKOFF
114 http._retry_count = RETRY_COUNT
115 http._request_id = util.new_request_id
118 # Monkey patch discovery._cast() so objects and arrays get serialized
119 # with json.dumps() instead of str().
120 _cast_orig = apiclient_discovery._cast
121 def _cast_objects_too(value, schema_type):
123 if (type(value) != type('') and
124 type(value) != type(b'') and
125 (schema_type == 'object' or schema_type == 'array')):
126 return json.dumps(value)
128 return _cast_orig(value, schema_type)
129 apiclient_discovery._cast = _cast_objects_too
131 # Convert apiclient's HttpErrors into our own API error subclass for better
133 # Reassigning apiclient_errors.HttpError is not sufficient because most of the
134 # apiclient submodules import the class into their own namespace.
135 def _new_http_error(cls, *args, **kwargs):
136 return super(apiclient_errors.HttpError, cls).__new__(
137 errors.ApiError, *args, **kwargs)
138 apiclient_errors.HttpError.__new__ = staticmethod(_new_http_error)
140 def http_cache(data_type):
141 homedir = os.environ.get('HOME')
142 if not homedir or len(homedir) == 0:
144 path = homedir + '/.cache/arvados/' + data_type
146 util.mkdir_dash_p(path)
149 return cache.SafeHTTPCache(path, max_age=60*60*24*2)
151 def api(version=None, cache=True, host=None, token=None, insecure=False,
152 request_id=None, timeout=5*60, **kwargs):
153 """Return an apiclient Resources object for an Arvados instance.
156 A string naming the version of the Arvados API to use (for
160 Use a cache (~/.cache/arvados/discovery) for the discovery
164 The Arvados API server host (and optional :port) to connect to.
167 The authentication token to send with each API call.
170 If True, ignore SSL certificate validation errors.
173 A timeout value for http requests.
176 Default X-Request-Id header value for outgoing requests that
177 don't already provide one. If None or omitted, generate a random
178 ID. When retrying failed requests, the same ID is used on all
181 Additional keyword arguments will be passed directly to
182 `apiclient_discovery.build` if a new Resource object is created.
183 If the `discoveryServiceUrl` or `http` keyword arguments are
184 missing, this function will set default values for them, based on
185 the current Arvados configuration settings.
191 _logger.info("Using default API version. " +
192 "Call arvados.api('%s') instead." %
194 if 'discoveryServiceUrl' in kwargs:
196 raise ValueError("both discoveryServiceUrl and host provided")
197 # Here we can't use a token from environment, config file,
198 # etc. Those probably have nothing to do with the host
199 # provided by the caller.
201 raise ValueError("discoveryServiceUrl provided, but token missing")
204 elif not host and not token:
205 return api_from_config(
206 version=version, cache=cache, request_id=request_id, **kwargs)
208 # Caller provided one but not the other
210 raise ValueError("token argument provided, but host missing.")
212 raise ValueError("host argument provided, but token missing.")
215 # Caller wants us to build the discoveryServiceUrl
216 kwargs['discoveryServiceUrl'] = (
217 'https://%s/discovery/v1/apis/{api}/{apiVersion}/rest' % (host,))
219 if 'http' not in kwargs:
220 http_kwargs = {'ca_certs': util.ca_certs_path()}
222 http_kwargs['cache'] = http_cache('discovery')
224 http_kwargs['disable_ssl_certificate_validation'] = True
225 kwargs['http'] = httplib2.Http(**http_kwargs)
227 if kwargs['http'].timeout is None:
228 kwargs['http'].timeout = timeout
230 kwargs['http'] = _patch_http_request(kwargs['http'], token)
232 svc = apiclient_discovery.build('arvados', version, cache_discovery=False, **kwargs)
233 svc.api_token = token
234 svc.insecure = insecure
235 svc.request_id = request_id
236 kwargs['http'].max_request_size = svc._rootDesc.get('maxRequestSize', 0)
237 kwargs['http'].cache = None
238 kwargs['http']._request_id = lambda: svc.request_id or util.new_request_id()
241 def api_from_config(version=None, apiconfig=None, **kwargs):
242 """Return an apiclient Resources object enabling access to an Arvados server
246 A string naming the version of the Arvados REST API to use (for
250 If provided, this should be a dict-like object (must support the get()
251 method) with entries for ARVADOS_API_HOST, ARVADOS_API_TOKEN, and
252 optionally ARVADOS_API_HOST_INSECURE. If not provided, use
253 arvados.config (which gets these parameters from the environment by
256 Other keyword arguments such as `cache` will be passed along `api()`
259 # Load from user configuration or environment
260 if apiconfig is None:
261 apiconfig = config.settings()
263 for x in ['ARVADOS_API_HOST', 'ARVADOS_API_TOKEN']:
264 if x not in apiconfig:
265 raise ValueError("%s is not set. Aborting." % x)
266 host = apiconfig.get('ARVADOS_API_HOST')
267 token = apiconfig.get('ARVADOS_API_TOKEN')
268 insecure = config.flag_is_true('ARVADOS_API_HOST_INSECURE', apiconfig)
270 return api(version=version, host=host, token=token, insecure=insecure, **kwargs)