Enable automated testing by mastering system dependencies. Use when reality is simply not good enough.
See the The documenation website <http://http://stubo-app.readthedocs.org/en/latest/>_ to view
user documentation.
Stub-O-Matic is offered under GPLv3, see LICENSE for more details.
Stub-O-Matic is made available by OpenCredo <http://opencredo.com>_
and a team of contributors.
(Linux Red Hat Enterprise, CentOS, Fedora, or Amazon Linux)
Redis Server Installation::
$ sudo yum install redis-server
$ sudo service redis-server start
Mongo server Installation::
$ sudo yum install mongodb-org
$ sudo service mongod start | stop
dependencies::
$ sudo yum install python27-virtualenv libxslt libxslt-devel libxml2 libxml2-devel texlive-full
(Linux Ubuntu)
Redis Server Installation::
$ sudo apt-get install -y redis-server
$ sudo service redis-server start
Mongo server Installation::
$ sudo apt-get install -y mongodb-org
$ sudo service mongod start | stop
dependencies::
$ sudo apt-get install -y python-dev python-pip wget python2.7-dev libxml2 libxml2-dev libxslt1-dev
Stubo::
$ virtualenv --no-site-packages env
$ source ./env/bin/activate
(env) $ git clone git@github.com:Stub-O-Matic/stubo-app.git
(env) $ cd stubo_app
(env) $ python setup.py develop
Perform the following::
$ sudo service mongod start
$ sudo service redis-server start
Active your virtual python env:
$ source ./env/bin/activate
You can start Stubo from your env root dir:
$ cd env
Note you need to create a log dir under the dir you call stubo from:
(first run only)
$ mkdir log
$ create_tracker_collection
$ stubo
Logging is in env/log/stubo.log
Default config is picked up from ./etc/dev.ini
Alternatively you can run from another location:
(env) $ stubo -c path_to_config
You can run the test suite by running the following command (in the project root: stubo-app)::
(env) $ nosetests stubo
Use the -x flag to stop the tests at the first error::
(env) $ nosetests stubo -x
Or to run a specific test:
(env) $ nosetests stubo.tests.integration.test_api:TestAcceptance.test_exec_first_commands
Use the --with-coverage flag to display a coverage report after
running the tests, this will show you which files / lines have not
been executed when the tests ran::
(env) $ nosetests stubo --with-coverage --cover-package=stubo
Or to run a specific test:
(env) $ nosetests stubo.tests.integration.test_api:TestAcceptance.test_exec_first_commands
and code analysis
(env) $ pylint stubo
If you want to call functions to debug in a python shell and they have a dependency on redis or mongo you must initialise them first
from stubo.utils import start_redis
slave, master = start_redis({})
slave
<redis.client.Redis object at 0x10168e650>
from stubo.utils import init_mongo
dbclient = init_mongo()
dbclient
Database(MongoClient('localhost', 27017), u'stubodb')
Now you can make calls that use redis
>>> from stubo.cache import Cache
>>> cache = Cache(host='localhost')
>>> cache.get_session('first', 'first_1')
{u'status': u'dormant', u'system_date': u'2015-02-24', u'scenario': u'localhost:first', u'last_used': u'2015-02-24 11:19:21', u'scenario_id': u'54ec5e3981875908f911a71b', u'session': u'first_1'}
& mongodb
from stubo.model.db import Scenario
scenario = Scenario()
stubs = list(scenario.get_stubs('localhost:first'))
import pprint
pp = pprint.PrettyPrinter(indent=4)
pp.pprint(stubs[0])
{ u'_id': ObjectId('537c8f1cac5f7303ad704d85'),
u'scenario': u'localhost:first',
u'stub': { u'recorded': u'2014-05-21',
u'request': { u'bodyPatterns': [ { u'contains': [ u'get my stub\n']}],
u'method': u'POST'},
u'response': { u'body': u'Hello {{1+1}} World\n',
u'delayPolicy': u'slow',
u'status': 200}}}
from stubo.model.db import Tracker
tracker = Tracker()
list(tracker.find_tracker_data({'function' : 'get/response'}, 0, 1))
[{u'function': u'get/response', u'scenario': u'conv', u'start_time': datetime.datetime(2013, 10, 29, 10, 38, 16, 534000, tzinfo=<bson.tz_util.FixedOffset object at 0x1015b6910>), u'return_code': 200, u'session': u'mary', u'duration_ms': 5, u'stubo_response': u'a2 response\n', u'_id': ObjectId('526f90180cfb4403fc27c0fa')}]
Load performance test data (over 10,000 stubs in 100 scenarios) with
stubo/api/exec/cmds?cmdfile=/static/cmds/perf/perf_setup.commands
this will take some time to run and leaves 100 sessions in playback mode.
optionally demonstrate that the load was successfull with:
stubo/api/exec/cmds?cmdfile=/static/cmds/perf/perf_getresponse.commands
this gets one response from each session.
When finished end sessions and delete stubs with:
stubo/api/exec/cmds?cmdfile=/static/cmds/perf/perf_teardown.commands
Certain Metrics are extracted from the track record and sent via UDP to a statsd server which aggregates the stats every 10 seconds before sending to a graphite carbon server via TCP.
The location of the statsd server and metrics prefix to use is specified in the stubo ini file
statsd.host = localhost
statsd.prefix = stubo
An analytics dashboard is provided to see that stats. Graphite connection details are specified in the stubo ini file::
graphite.host = http://your-stubo-graphite.com/
# auth to connect to graphite server if required
graphite.user = <user>
graphite.password = <passwd>
Documentation is generated with Sphinx. The doc sources are located under /docs.
prereq for generating PDF:
$ yum install texlive
To generate docs
(env) $ python setup.py build_sphinx --build-dir=./stubo/static/docs
To generate PDF
$ cd docs
$ make latexpdf
$ cp _build/latex/Stub-O-Matic.pdf ../stubo/static/docs
Stubo profiling is available via two different profilers yappi and plop.
yappi function stats
name: name of the function being profiled
#n: is the total callcount of the function.
tsub: total spent time in the function minus the total time spent in the other functions called from this function.
ttot: total time spent in the function.
tavg: is same as ttot/ccnt. Average total time.
Plop is a stack-sampling profiler for Python. Profile collection can be turned on and off in a live process with minimal performance impact.
Both profilers are enabled via an HTTP inteface in stubo.
yappi - generate stats for 3 mins, output is saved to a csv file:
_profile?interval=180
plop - generate stats for 3 mins, output is saved by default to /tmp/plop.out:
_profile2?interval=180&output=/tmp/plop.out
To view plop results in a bubble graph:
(env) $ python -m plop.viewer --datadir=/tmp
View via the plop app at http://localhost:8888