On 05/30/2016 04:10 AM, archus(a)linux.vnet.ibm.com wrote:
> From: Archana Singh <archus(a)linux.vnet.ibm.com>
>
> Wrapper classes/methods for making API calls by reading
> the config file configuration.
>
> Signed-off-by: Archana Singh <archus(a)linux.vnet.ibm.com>
> ---
> tests/fvt/restapilib.py | 738 ++++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 738 insertions(+)
> create mode 100644 tests/fvt/restapilib.py
>
> diff --git a/tests/fvt/restapilib.py b/tests/fvt/restapilib.py
> new file mode 100644
> index 0000000..19634f1
> --- /dev/null
> +++ b/tests/fvt/restapilib.py
> @@ -0,0 +1,738 @@
> +#!/usr/bin/python
> +# -*- coding: utf-8 -*-
You only need to add this header when the source code file has non-ASCII
characters which seems to not be the case.
So you can remove it.
> +#
> +# Project Wok
> +#
> +# Copyright IBM Corp, 2016
> +#
> +# This library is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU Lesser General Public
> +# License as published by the Free Software Foundation; either
> +# version 2.1 of the License, or (at your option) any later version.
> +#
> +# This library is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> +# Lesser General Public License for more details.
> +#
> +# You should have received a copy of the GNU Lesser General Public
> +# License along with this library; if not, write to the Free Software
> +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301USA
> +
> +import ConfigParser
> +import jsonschema
> +import json
> +import requests
> +import logging
> +import os
> +
> +requests.packages.urllib3.disable_warnings()
> +
Why disable warning message? Aren't they important for debug?
> +__all__ = ['APIError', 'APIRequestError', 'APISession', 'SessionParameters']
> +
> +DEFAULT_CONF = os.path.dirname(os.path.abspath(__file__)) + os.sep + 'config'
> +
> +_HTTP_SUCCESS_STATUS_RANGE = range(200, 299) # All of the 2XX status codes
> +
> +
> +class APIError(Exception):
> + """Base class for exceptions raised by this module."""
> +
> + pass
> +
Could you elaborate on the proposal of this class?
Why do not use Exception directly as this class does not add anything to
Exception class?
> +
> +class APISession(object):
> + """Represents an API session with the host Web Services API"""
> +
> + def __init__(self, conffile=DEFAULT_CONF):
> + """ Creates an APISession object for use with the specified Host.
> + """
> + self._session = None
> + self._base_uri = None
> + self.username = None
> + self.passwd = None
I am not comfortable to store password values in the application. Is it
really needed? What are the other possibilities?
> + self.host = None
> + self.port = None
> + # Holds the session from requests if self._session is None.
> + self.session()
> + self.sessionparams = SessionParameters(conffile=conffile)
> + self.logging = self.sessionparams.getLogger()
> + # base URI of connected machine
> + self._base_uri = self.create_base_uri()
> +
> + def session(self):
> + """
> + Create the session object for API Request if no existing session.
> + """
> + if self._session is not None:
> + raise APIError('Already have a session')
> +
> + self._session = requests.Session()
> + return self._session
> +
How is this session different from the Wok session? Why do we need to
handle 2 types of session?
> + def auth(self):
> + """
> + Set authentication to API session.
> + """
> +
> + if self._session is None:
> + self.session()
> +
> + self.username, self.passwd = self.sessionparams.getCredentials()
> +
> + if self.username is None:
> + raise APIError('Username is None')
> + if self.passwd is None:
> + raise APIError('password is None')
> + # Set auth at session level
> + self._session.auth = (self.username, self.passwd)
> +
> + # set password expiration currently it default value
> + # self._session.expires = '1432091741'
> + # For now keeping it False to disable SSL verification
> + self._session.verify = False
> + return
> +
> + def create_base_uri(self):
> + """
> + Create the base URI using host and port.
> + if no host is provided _base_uri is None.
> + if port is None then _base_uri is just host
> + if both is provided then append port to host.
> + """
> +
> + if self._base_uri is not None:
> + return self._base_uri
> +
> + self.host, self.port = self.sessionparams.getUrlParams()
> + if self.host is None:
> + return None
> + if self.port is None:
> + return self.host
> +
> + self._base_uri = 'https://' + self.host + ':' + self.port
> + return self._base_uri
> +
> + def end_session(self):
> + """Ends an API session"""
> +
> + if self._session is None:
> + return None # Silently ignore
> + # The session to be terminated by setting None in session object we
> + # have.
> + self._session = None
> + return
> +
Same I asked about that Session.
> + def request(
> + self,
> + method,
> + uri,
> + body=None,
> + expected_status_values=None,
> + headers=None, ):
> + """
Use one line up to 80 characters.
> + Issue an WS API request and return the response body, if any.
> +
> + \param method
> + The HTTP method to issue (eg. GET, PUT, POST, DELETE).
> +
> + \param uri
> + The URI path and query parameter string for the request.
> +
> + \param body
> + The request body that is input to the request, as a Unicode
> + or String, or None if there is no input body.
> +
> + \param expected_status_values
> + The HTTP status code that is expected on completion of the request.
> + If this parameter is specified, the function raises an exception
> + if the HTTP status from the request was not exactly as specified.
> + Otherwise, it raises an error if the HTTP status is not 2XX.
> +
> + \param headers
> + Request headers for this request, in the form of a Python Dict.
> + Optional. This function automatically augments these headers
> + with the headers needed to specify the API session. Note that if
> + input headers are provided, the supplied Dict object will be
> + modified by this function.
> + """
I like the documentation. But we need to agree on a format to be used
overall.
Do you mind to write a github wiki page with the instructions? (how to
doc parameters and return function, etc)
> +
> + # Start with the headers supplied by caller, if any
> +
> + if headers is not None:
> + hdrs = headers # Use caller's dict, not a copy
> + else:
> + hdrs = dict()
> +
> + # If no session then throw error.
> +
> + if self._session is None:
> + raise APIError('You have no session')
> +
> + resp = self._session.request(method, uri, data=body, headers=hdrs)
> +
> + # If an expected status was specified, check that the status exactly
> + # matches wbat was expected, otherwise check that the status is one of
> + # the success Statuses.
> +
> + if expected_status_values is not None:
> + raise_exc = resp.status_code not in expected_status_values
> + else:
> + raise_exc = resp.status_code not in _HTTP_SUCCESS_STATUS_RANGE
> +
> + # Raise an exceptoin if the result is not as intended.
> +
> + if raise_exc:
> +
> + # If the request fails in some way WEb Services API response
> + # will usually include a standard error response body
> + # in JSON format that includes a more detailed reason
> + # code (and message) for the failure. It provides this data
> + # in JSON format even if the request would return some other
> + # format if the request had been successful.
> + # So if the request has failed, grab that additional info
> + # for use in raising exceptions below.
> +
> + failure_reason = 0
> + failure_code = None
> + # if response.status_code not in _HTTP_SUCCESS_STATUS_RANGE:
> +
> + # The API provides the JSON error response in all usual error
> + # cases. But for certain less common errors this does not occur
> + # because the error is caught higher in the processing stack.
> + # So try to interpret the response as a JSON response body, but
> + # just silently ignore problems if we can't do this.
> +
> + try:
> + error_resp = json.loads(resp.text)
> + failure_reason = error_resp['reason']
> + failure_code = error_resp['code']
> + except (ValueError, KeyError):
> + pass
> +
> + raise APIRequestError(resp.status_code, failure_reason,
> + failure_code)
> +
> + # Return the response
> + return resp
> +
> + def request_octet(
> + self,
> + method,
> + uri,
> + body=None,
> + expected_status_values=None,
> + headers=None,
> + ):
> +
Use one line up to 80 characters. It applies to all the other
functions/methods you created like above.
> + # Start with the request headers the caller supplied if any,
> + # otherwise start with an empty set of headers.
> +
> + if headers is not None:
> + hdrs = headers # Using caller's dict, not a copy
> + else:
> + hdrs = dict()
> +
> + # If a request body was specified, convert it from its assumed Dict
> + # or List form into JSON using the Python JSON library. Also, add in
> + # the required Content-Type HTTP message header to let the API know
> + # that we are supplying input in JSON form. (A Content-Length header
> + # is also needed to specify the body length, but for not not set.)
> +
> + body_json = body
> + if body is not None:
> + body_json = json.dumps(body)
> +
> + # Supply an HTTP Accepts header indicating we accept/expect that any
> + # response body be JSON. The Accepts header is optional, and the API
> + # will defautl to supplying JSON for any operation defined in API V1.1
> + # to return JSON (even if in the future support for other formats would
> + # be added). But its a good idea to specify this header as a safeguard
> + # in case this function were called for an URI that is not capable of
> + # returning JSON, as this function as currently coded would choke on
> + # non-JSON responses.
> +
> + hdrs['Accept'] = 'application/json'
> +
> + # Now issue the request, and retrieve the response object and the
> + # response body if provided.
> +
> + resp = self.request(
> + method,
> + uri,
> + body=body_json,
> + expected_status_values=expected_status_values,
> + headers=hdrs,
> + )
> +
> + return resp
> +
> + def request_json(
> + self,
> + method,
> + uri,
> + body=None,
> + expected_status_values=None,
> + headers=None,
> + ):
Use one line up to 80 characters.
> + """
> + Issue an WS API request that is defined to take JSON input and
> + produce JSON output. (Nearly all WS API requests are like this.)
> +
> + Input and output bodies are specified as Python Dict or List objects,
> + which are converted to/from JSON by this function.
> +
> + \param method
> + The HTTP method to issue (eg. GET, PUT, POST, DELETE).
> +
> + \param uri
> + The URI path and query parameter string for the request.
> +
> + \param body
> + The request body that is input to the request, in the form of a
> + Python Dict or List object. This object is automatically converted
> + to corresponding JSON by this function. Optional.
> +
> + \param expected_status_values
> + The HTTP status code that is expected on completion of the request.
> + See corresponding parameter on request() method for more details.
> +
> + \param headers
> + Request headers for this request, in the form of a Python Dict.
> + Optional. This function automatically augments these headers
> + with the headers needed to specify the JSON content type for input
> + and output, and to specify the API session. Note that if input
> + headers are provided, the supplied Dict object will be modified
> + by this function.
> +
> + """
> +
> + # Start with the request headers the caller supplied if any,
> + # otherwise start with an empty set of headers.
> +
> + if headers is not None:
> + hdrs = headers # Using caller's dict, not a copy
> + else:
> + hdrs = dict()
> +
> + # If a request body was specified, convert it from its assumed Dict
> + # or List form into JSON using the Python JSON library. Also, add in
> + # the required Content-Type HTTP message header to let the API know
> + # that we are supplying input in JSON form. (A Content-Length header
> + # is also needed to specify the body length, but for now have not set.)
> +
> + body_json = None
> + if body is not None:
> + body_json = json.dumps(body)
> + hdrs['Content-Type'] = 'application/json'
> +
> + # Supply an HTTP Accepts header indicating we accept/expect that any
> + # response body be JSON. The Accepts header is optional, and the API
> + # will default to supplying JSON for any operation defined in API V1.1
> + # to return JSON (even if in the future support for other formats would
> + # be added). But its a good idea to specify this header as a safeguard
> + # in case this function were called for an URI that is not capable of
> + # returning JSON, as this function as currently coded would choke on
> + # non-JSON responses.
> +
> + hdrs['Accept'] = 'application/json'
> +
> + # Now issue the request, and retrieve the response object and the
> + # response body if provided.
> +
> + response_body = self.request(
> + method,
> + uri,
> + body=body_json,
> + expected_status_values=expected_status_values,
> + headers=hdrs,
> + )
> +
> + # If a response body was returned, we presume it is JSON and
> + # convert it into a Python dict we will return.
> +
> + resp_json = None
> + if response_body is not None:
> + try:
> + resp_json = response_body.json()
> + except ValueError, err:
> + print 'response_body %s' % response_body
> + print 'response_body headers %s' % response_body.headers
> + raise APIError('Response body expected to be JSON but is'
> + + 'not valid %s', err)
> + return resp_json
> +
> + def request_get_json(
> + self,
> + uri,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a GET request that returns JSON."""
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + return self.request_json('GET', uri,
> + expected_status_values=expected_status_values,
> + headers=headers)
> +
> + def request_put_json(
> + self,
> + uri,
> + body=None,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a GET request that returns JSON."""
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + return self.request_json(
> + 'PUT',
> + uri,
> + body=body,
> + expected_status_values=expected_status_values,
> + headers=headers,
> + )
> +
> + def request_post_json(
> + self,
> + uri,
> + body=None,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a POST request that
> + provides/returns JSON."""
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + return self.request_json(
> + 'POST',
> + uri,
> + body=body,
> + expected_status_values=expected_status_values,
> + headers=headers,
> + )
> +
> + def request_delete_json(
> + self,
> + uri,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a DELETE request that returns JSON."""
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + return self.request_json(
> + 'DELETE',
> + uri,
> + expected_status_values=expected_status_values,
> + headers=headers,
> + )
> +
Those methods request_json_X() have duplicated code. Why do not use only
self.request_json() with the required parameters?
> + def request_get(
> + self,
> + uri,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a GET request with header['Accept']
> + as application/json that returns the response.
> + Does not convert the response to JSON
> + """
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + if headers is None:
> + headers = dict()
> + headers['Accept'] = 'application/json'
> +
> + return self.request('GET', uri,
> + expected_status_values=expected_status_values,
> + headers=headers)
> +
> + def request_put(
> + self,
> + uri,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a PUT request with header['Accept']
> + as application/json that returns the response.
> + Does not convert the response to JSON
> + """
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + if headers is None:
> + headers = dict()
> + headers['Accept'] = 'application/json'
> +
> + return self.request(
> + 'PUT',
> + uri,
> + expected_status_values=expected_status_values,
> + headers=headers,
> + )
> +
> + def request_post(
> + self,
> + uri,
> + body=None,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a POST request with header['Accept']
> + as application/json that returns the response.
> + Does not convert the response to JSON
> + """
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + # If a request body was specified, convert it from its assumed Dict
> + # or List form into JSON using the Python JSON library. Also, add in
> + # the required Content-Type HTTP message header to let the API know
> + # that we are supplying input in JSON form. (A Content-Length header
> + # is also needed to specify the body length, but for now have not set.)
> +
> + body_json = None
> + if body is not None:
> + body_json = json.dumps(body)
> +
> + if headers is None:
> + headers = dict()
> + headers['Content-Type'] = 'application/json'
> + headers['Accept'] = 'application/json'
> +
> + return self.request(
> + 'POST',
> + uri,
> + body=body_json,
> + expected_status_values=expected_status_values,
> + headers=headers,
> + )
> +
> + def request_delete(
> + self,
> + uri,
> + expected_status_values=None,
> + headers=None,
> + ):
> + """Convenience function to do a DELETE request with header['Accept']
> + as application/json that returns the response.
> + Does not convert the response to JSON
> + """
> +
> + if self._base_uri is not None:
> + uri = self._base_uri + uri
> +
> + if headers is None:
> + headers = dict()
> + headers['Accept'] = 'application/json'
> +
> + return self.request(
> + 'DELETE',
> + uri,
> + expected_status_values=expected_status_values,
> + headers=headers,
> + )
> +
> +
Same I commented above.
Those methods request_X() have duplicated code. Why do not use only
self.request_json() with the required parameters?
Having only one method to do a request would be better for maintenance
and testing development.
self.request() and self.request_json() should handle all the methods and
do the right decision on which to do.
So I don't see why specific methods are required.
> +class APIRequestError(APIError):
> + """
> + Raised when an API request ends in error or not as expected.
> +
> + Attributes:
> + \param status
> + HTTP status code from the request
> +
> + \param reason
> + API reason code from the request
> +
> + \param message
> + API diagnostic message from the request
> +
> + \param stack
> + Internal diagnostic info for selected Status 500 errors
> + """
> +
> + def __init__(
> + self,
> + status,
> + reason=None,
> + code=None,
> + ):
Use one line up to 80 characters.
> + self.status = status
> + self.reason = reason
> + self.code = code
> +
> + def __str__(self):
> + if self.code is not None:
> + s = 'Request ended with status %s-%s (%s)' \
> + % (self.status, self.reason, self.code)
> + else:
> + s = 'Request ended with status %s-%s' % (self.status,
> + self.reason)
> + return s
> +
> +
> +class BadJSONResponse(APIError):
> + """exceptions raised for Bad Json Response."""
> +
> + def __init__(
> + self,
> + error,
> + ):
> + self.error = error
> +
> + def __str__(self):
> + s = 'Bad response: %s' % self.error
> + return s
> +
> +
> +class InvalidInput(APIError):
> + """exceptions raised for Invalid Input."""
> +
> + def __init__(
> + self,
> + error,
> + ):
> + self.error = error
> +
> + def __str__(self):
> + s = 'Bad response: %s' % self.error
> + return s
> +
> +
Why do we need specific error exception? Each request, when on error,
will raise a WokException. Why do we need other exceptions types?
> +class SessionParameters(object):
> + """
> + Represents functionalities that could help in getting the
> + session information.
> + """
> +
> + def __init__(
> + self,
> + conffile=DEFAULT_CONF):
Use one line up to 80 characters.
> + """
> + Attributes:
> + \param conffile
> + config file which contains all configuration
> + information with sections
> + """
> + self.conffile = conffile
> + self.username = None
> + self.passwd = None
> + self.host = '127.0.0.1'
> + self.port = '8001'
> + self.params = 'config'
> + self.logfile = 'wok-api-test-suite.log'
> + self.session_section = 'Session'
> + self.logging = logging
> + self.loglevel = 'ERROR'
> + self.LEVELS = {'INFO': self.logging.INFO,
> + 'DEBUG': self.logging.DEBUG,
> + 'WARNING': self.logging.WARNING,
> + 'ERROR': self.logging.ERROR,
> + 'CRITICAL': self.logging.CRITICAL}
> + if self.conffile is None:
> + print 'Configuration file required %s' \
> + % self.conffile
It should raise an exception if it is required to run the tests.
> + else:
> + print 'Reading configuration file %s' % self.conffile
> + self.params = ConfigParser.ConfigParser()
> + print self.params.read(self.conffile)
> + if self.params.has_section(self.session_section):
> + if self.params.has_option(self.session_section, 'logfile'):
> + self.logfile = self.params.get(
> + self.session_section, 'logfile')
> +
> + if self.params.has_option(self.session_section, 'loglevel'):
> + self.loglevel = self.params.get(
> + self.session_section, 'loglevel')
> + else:
> + print "Section %s is not available in the config file " \
> + % self.session_section
> +
Why happens on that case? Default values are assumed?
> + self.logging.basicConfig(format='%(asctime)s %(levelname)s: \
> + %(message)s',
> + filename=self.logfile,
> + level=self.LEVELS[self.loglevel])
> +
> + def getCredentials(self):
> + """
> + Returns user and password details for a session obtained from config
> + file
> + """
> + if self.params.has_section(self.session_section):
> + if self.params.has_option(self.session_section, 'user'):
> + self.username = self.params.get(self.session_section, 'user')
> +
> + if self.params.has_option(self.session_section, 'passwd'):
> + self.passwd = self.params.get(self.session_section, 'passwd')
> + else:
> + print "Section %s is not available in the config file " \
> + % self.session_section
> + raise APIError('Session section not present in config')
> +
> + return self.username, self.passwd
> +
> + def getUrlParams(self):
> + """
> + Returns host and port details for a session obtained from config file
> + """
> + if self.params.has_section(self.session_section):
> + if self.params.has_option(self.session_section, 'host'):
> + self.host = self.params.get(self.session_section, 'host')
> +
> + if self.params.has_option(self.session_section, 'port'):
> + self.port = self.params.get(self.session_section, 'port')
> + else:
> + print "Section %s is not available in the config file " \
> + % self.session_section
> +
> + return self.host, self.port
> +
> + def getLogger(self):
> + return self.logging
> +
> +
> +class Validator(object):
> + """
> + validator class having different validate methods
> + """
> +
> + def __init__(self, logger=None):
> + """
> + :param logger:
> + :return:
> + """
> + if logger is not None:
> + self.logging = logger
> +
> + def validate_json(self, jsn, schma):
> + """
> + validates json against schema using jsonschema.validate library
> +
> + :param jsn: JSON to validate
> + :param schma: Schema against which JSON should get validated
> + :return:
> + """
> + if self.logging is not None:
> + self.logging.info('--> validate_json()')
> + self.logging.debug(
> + 'validate_json(): jsn:%s' % jsn)
> + self.logging.debug(
> + 'validate_json(): schma:%s' % schma)
> + jsonschema.validate(jsn, schma)
> + if self.logging is not None:
> + self.logging.debug(
> + 'validate_json(): jsn is valid')
> + self.logging.info('<-- validate_json()')
