[IMP] test impprove HTTPCase opener
Simplify website crawler using the generic HTTPCase opener
This commit is contained in:
parent
f9e24e1918
commit
78e044b5c7
|
@ -1,5 +1,5 @@
|
|||
# -*- coding: utf-8 -*-
|
||||
import test_converter
|
||||
import test_requests
|
||||
import test_crawl
|
||||
import test_ui
|
||||
import test_views
|
||||
|
|
|
@ -1,35 +0,0 @@
|
|||
# -*- coding: utf-8 -*-
|
||||
import unittest2
|
||||
|
||||
class URLCase(unittest2.TestCase):
|
||||
"""
|
||||
URLCase moved out of test_requests, otherwise discovery attempts to
|
||||
instantiate and run it
|
||||
"""
|
||||
def __init__(self, user, url, source, result):
|
||||
super(URLCase, self).__init__()
|
||||
self.user = user
|
||||
self.url = url
|
||||
self.source = source
|
||||
self.result = result
|
||||
|
||||
@property
|
||||
def username(self):
|
||||
return self.user or "Anonymous Coward"
|
||||
|
||||
def __str__(self):
|
||||
if self.source:
|
||||
return "%s (from %s, as %s)" % (self.url, self.source, self.username)
|
||||
return "%s (as %s)" % (self.url, self.username)
|
||||
|
||||
__repr__ = __str__
|
||||
|
||||
def shortDescription(self):
|
||||
return ""
|
||||
|
||||
def runTest(self):
|
||||
code = self.result.getcode()
|
||||
self.assertIn(
|
||||
code, xrange(200, 300),
|
||||
"Fetching %s as %s returned an error response (%d)" % (
|
||||
self.url, self.username, code))
|
|
@ -0,0 +1,84 @@
|
|||
# -*- coding: utf-8 -*-
|
||||
import logging
|
||||
import urlparse
|
||||
import unittest2
|
||||
import urllib2
|
||||
import time
|
||||
import werkzeug.urls
|
||||
|
||||
import lxml.html
|
||||
|
||||
import openerp
|
||||
from openerp import tools
|
||||
|
||||
_logger = logging.getLogger(__name__)
|
||||
|
||||
class Crawler(openerp.tests.HttpCase):
|
||||
""" Test suite crawling an openerp CMS instance and checking that all
|
||||
internal links lead to a 200 response.
|
||||
|
||||
If a username and a password are provided, authenticates the user before
|
||||
starting the crawl
|
||||
"""
|
||||
|
||||
at_install = False
|
||||
post_install = True
|
||||
|
||||
def crawl(self, url, seen=None, msg=''):
|
||||
if seen == None:
|
||||
seen = set()
|
||||
if url in seen:
|
||||
return seen
|
||||
else:
|
||||
seen.add(url)
|
||||
|
||||
_logger.info("%s %s", msg, url)
|
||||
r = self.url_open(url)
|
||||
code = r.getcode()
|
||||
self.assertIn( code, xrange(200, 300), "%s Fetching %s returned error response (%d)" % (msg, url, code))
|
||||
|
||||
if r.info().gettype() == 'text/html':
|
||||
doc = lxml.html.fromstring(r.read())
|
||||
for link in doc.xpath('//a[@href]'):
|
||||
href = link.get('href')
|
||||
|
||||
parts = urlparse.urlsplit(href)
|
||||
# href with any fragment removed
|
||||
href = urlparse.urlunsplit((
|
||||
parts.scheme,
|
||||
parts.netloc,
|
||||
parts.path,
|
||||
parts.query,
|
||||
''
|
||||
))
|
||||
|
||||
# FIXME: handle relative link (not parts.path.startswith /)
|
||||
if parts.netloc or \
|
||||
not parts.path.startswith('/') or \
|
||||
parts.path == '/web' or\
|
||||
parts.path.startswith('/web/') or \
|
||||
parts.path.startswith('/en_US/') or \
|
||||
(parts.scheme and parts.scheme not in ('http', 'https')):
|
||||
continue
|
||||
|
||||
self.crawl(href, seen, msg)
|
||||
return seen
|
||||
|
||||
|
||||
def test_10_crawl_public(self):
|
||||
t0 = time.time()
|
||||
seen = self.crawl('/', msg='Anonymous Coward')
|
||||
_logger.log(25, "public crawled %s urls in %.2fs", len(seen) ,time.time() - t0)
|
||||
|
||||
def test_20_crawl_demo(self):
|
||||
t0 = time.time()
|
||||
self.authenticate('demo', 'demo')
|
||||
seen = self.crawl('/', msg='demo')
|
||||
_logger.log(25, "demo crawled %s urls in %.2fs", len(seen), time.time() - t0)
|
||||
|
||||
def test_30_crawl_admin(self):
|
||||
t0 = time.time()
|
||||
self.authenticate('admin', 'admin')
|
||||
seen = self.crawl('/', msg='admin')
|
||||
_logger.log(25, "admin crawled %s urls in %.2fs", len(seen), time.time() - t0)
|
||||
|
|
@ -1,155 +0,0 @@
|
|||
# -*- coding: utf-8 -*-
|
||||
import urlparse
|
||||
import unittest2
|
||||
import urllib2
|
||||
import werkzeug.urls
|
||||
|
||||
import lxml.html
|
||||
|
||||
import openerp
|
||||
from openerp import tools
|
||||
|
||||
import cases
|
||||
|
||||
__all__ = ['load_tests', 'CrawlSuite']
|
||||
|
||||
class RedirectHandler(urllib2.HTTPRedirectHandler):
|
||||
"""
|
||||
HTTPRedirectHandler is predicated upon HTTPErrorProcessor being used and
|
||||
works by intercepting 3xy "errors".
|
||||
|
||||
Inherit from it to handle 3xy non-error responses instead, as we're not
|
||||
using the error processor
|
||||
"""
|
||||
|
||||
def http_response(self, request, response):
|
||||
code, msg, hdrs = response.code, response.msg, response.info()
|
||||
|
||||
if 300 <= code < 400:
|
||||
return self.parent.error(
|
||||
'http', request, response, code, msg, hdrs)
|
||||
|
||||
return response
|
||||
|
||||
https_response = http_response
|
||||
|
||||
class CrawlSuite(unittest2.TestSuite):
|
||||
""" Test suite crawling an openerp CMS instance and checking that all
|
||||
internal links lead to a 200 response.
|
||||
|
||||
If a username and a password are provided, authenticates the user before
|
||||
starting the crawl
|
||||
"""
|
||||
|
||||
at_install = False
|
||||
post_install = True
|
||||
|
||||
def __init__(self, user=None, password=None):
|
||||
super(CrawlSuite, self).__init__()
|
||||
|
||||
registry = openerp.registry(tools.config['db_name'])
|
||||
try:
|
||||
# switch registry to test mode, so that requests can be made
|
||||
registry.enter_test_mode()
|
||||
|
||||
self.opener = urllib2.OpenerDirector()
|
||||
self.opener.add_handler(urllib2.UnknownHandler())
|
||||
self.opener.add_handler(urllib2.HTTPHandler())
|
||||
self.opener.add_handler(urllib2.HTTPSHandler())
|
||||
self.opener.add_handler(urllib2.HTTPCookieProcessor())
|
||||
self.opener.add_handler(RedirectHandler())
|
||||
|
||||
self._authenticate(user, password)
|
||||
self.user = user
|
||||
|
||||
finally:
|
||||
registry.leave_test_mode()
|
||||
|
||||
def _request(self, path):
|
||||
return self.opener.open(urlparse.urlunsplit([
|
||||
'http', 'localhost:%s' % tools.config['xmlrpc_port'],
|
||||
path, '', ''
|
||||
]))
|
||||
|
||||
def _authenticate(self, user, password):
|
||||
# force tools.config['db_name'] in user session so opening `/` doesn't
|
||||
# blow up in multidb situations
|
||||
self.opener.open('http://localhost:{port}/web/?db={db}'.format(
|
||||
port=tools.config['xmlrpc_port'],
|
||||
db=werkzeug.urls.url_quote_plus(tools.config['db_name']),
|
||||
))
|
||||
if user is not None:
|
||||
url = 'http://localhost:{port}/login?{query}'.format(
|
||||
port=tools.config['xmlrpc_port'],
|
||||
query=werkzeug.urls.url_encode({
|
||||
'db': tools.config['db_name'],
|
||||
'login': user,
|
||||
'key': password,
|
||||
})
|
||||
)
|
||||
auth = self.opener.open(url)
|
||||
assert auth.getcode() < 400, "Auth failure %d" % auth.getcode()
|
||||
|
||||
def _wrapped_run(self, result, debug=False):
|
||||
registry = openerp.registry(tools.config['db_name'])
|
||||
try:
|
||||
# switch registry to test mode, so that requests can be made
|
||||
registry.enter_test_mode()
|
||||
|
||||
paths = [URL('/')]
|
||||
seen = set(paths)
|
||||
|
||||
while paths:
|
||||
url = paths.pop(0)
|
||||
r = self._request(url.url)
|
||||
url.to_case(self.user, r).run(result)
|
||||
|
||||
if r.info().gettype() != 'text/html':
|
||||
continue
|
||||
|
||||
doc = lxml.html.fromstring(r.read())
|
||||
for link in doc.xpath('//a[@href]'):
|
||||
href = link.get('href')
|
||||
|
||||
parts = urlparse.urlsplit(href)
|
||||
# href with any fragment removed
|
||||
href = urlparse.urlunsplit((
|
||||
parts.scheme,
|
||||
parts.netloc,
|
||||
parts.path,
|
||||
parts.query,
|
||||
''
|
||||
))
|
||||
|
||||
# avoid repeats, even for links we won't crawl no need to
|
||||
# bother splitting them if we've already ignored them
|
||||
# previously
|
||||
if href in seen: continue
|
||||
seen.add(href)
|
||||
|
||||
# FIXME: handle relative link (not parts.path.startswith /)
|
||||
if parts.netloc or \
|
||||
not parts.path.startswith('/') or \
|
||||
parts.path == '/web' or\
|
||||
parts.path.startswith('/web/') or \
|
||||
(parts.scheme and parts.scheme not in ('http', 'https')):
|
||||
continue
|
||||
|
||||
paths.append(URL(href, url.url))
|
||||
|
||||
finally:
|
||||
registry.leave_test_mode()
|
||||
|
||||
class URL(object):
|
||||
def __init__(self, url, source=None):
|
||||
self.url = url
|
||||
self.source = source
|
||||
|
||||
def to_case(self, user, result):
|
||||
return cases.URLCase(user, self.url, self.source, result)
|
||||
|
||||
def load_tests(loader, base, _):
|
||||
base.addTest(CrawlSuite())
|
||||
base.addTest(CrawlSuite('admin', 'admin'))
|
||||
base.addTest(CrawlSuite('demo', 'demo'))
|
||||
return base
|
|
@ -231,12 +231,12 @@ def load_module_graph(cr, graph, status=None, perform_checks=True, skip_modules=
|
|||
registry._init_modules.add(package.name)
|
||||
cr.commit()
|
||||
|
||||
_logger.log(25, "%s modules loaded in %.2fs", len(graph), time.time() - ta0)
|
||||
_logger.log(25, "%s modules loaded in %.2fs", len(graph), time.time() - ta0)
|
||||
|
||||
# The query won't be valid for models created later (i.e. custom model
|
||||
# created after the registry has been loaded), so empty its result.
|
||||
registry.fields_by_model = None
|
||||
|
||||
|
||||
cr.commit()
|
||||
|
||||
return loaded_modules, processed_modules
|
||||
|
|
|
@ -428,7 +428,8 @@ def run_unit_tests(module_name, dbname, position=runs_at_install):
|
|||
tm0 = time.time()
|
||||
_logger.info('%s running tests.', m.__name__)
|
||||
result = unittest2.TextTestRunner(verbosity=2, stream=TestStream(m.__name__)).run(suite)
|
||||
_logger.log(25, "%s tested in %.2fs", m.__name__, time.time() - tm0)
|
||||
if time.time() - tm0 > 5:
|
||||
_logger.log(25, "%s tested in %.2fs", m.__name__, time.time() - tm0)
|
||||
if not result.wasSuccessful():
|
||||
r = False
|
||||
_logger.error("Module %s: %d failures, %d errors", module_name, len(result.failures), len(result.errors))
|
||||
|
|
|
@ -130,6 +130,25 @@ class SingleTransactionCase(BaseCase):
|
|||
cls.cr.rollback()
|
||||
cls.cr.close()
|
||||
|
||||
class RedirectHandler(urllib2.HTTPRedirectHandler):
|
||||
"""
|
||||
HTTPRedirectHandler is predicated upon HTTPErrorProcessor being used and
|
||||
works by intercepting 3xy "errors".
|
||||
|
||||
Inherit from it to handle 3xy non-error responses instead, as we're not
|
||||
using the error processor
|
||||
"""
|
||||
|
||||
def http_response(self, request, response):
|
||||
code, msg, hdrs = response.code, response.msg, response.info()
|
||||
|
||||
if 300 <= code < 400:
|
||||
return self.parent.error(
|
||||
'http', request, response, code, msg, hdrs)
|
||||
|
||||
return response
|
||||
|
||||
https_response = http_response
|
||||
|
||||
class HttpCase(TransactionCase):
|
||||
""" Transactionnal HTTP TestCase with url_open and phantomjs helpers.
|
||||
|
@ -152,6 +171,14 @@ class HttpCase(TransactionCase):
|
|||
self.session.db = DB
|
||||
openerp.http.root.session_store.save(self.session)
|
||||
self.localstorage_path = mkdtemp()
|
||||
# setup an url opener helper
|
||||
self.opener = urllib2.OpenerDirector()
|
||||
self.opener.add_handler(urllib2.UnknownHandler())
|
||||
self.opener.add_handler(urllib2.HTTPHandler())
|
||||
self.opener.add_handler(urllib2.HTTPSHandler())
|
||||
self.opener.add_handler(urllib2.HTTPCookieProcessor())
|
||||
self.opener.add_handler(RedirectHandler())
|
||||
self.opener.addheaders.append(('Cookie', 'session_id=%s' % self.session_id))
|
||||
|
||||
def tearDown(self):
|
||||
rmtree(self.localstorage_path)
|
||||
|
@ -159,11 +186,15 @@ class HttpCase(TransactionCase):
|
|||
super(HttpCase, self).tearDown()
|
||||
|
||||
def url_open(self, url, data=None, timeout=10):
|
||||
opener = urllib2.build_opener()
|
||||
opener.addheaders.append(('Cookie', 'session_id=%s' % self.session_id))
|
||||
if url.startswith('/'):
|
||||
url = "http://localhost:%s%s" % (PORT, url)
|
||||
return opener.open(url, data, timeout)
|
||||
return self.opener.open(url, data, timeout)
|
||||
|
||||
def authenticate(self, user, password):
|
||||
if user is not None:
|
||||
url = '/login?%s' % werkzeug.urls.url_encode({'db': DB,'login': user, 'key': password})
|
||||
auth = self.url_open(url)
|
||||
assert auth.getcode() < 400, "Auth failure %d" % auth.getcode()
|
||||
|
||||
def phantom_poll(self, phantom, timeout):
|
||||
""" Phantomjs Test protocol.
|
||||
|
|
Loading…
Reference in New Issue