def app(self):
return Blocker(upstream_app,
domains=[{
"domain": "giraffe.com",
"template": "disallow_access.html.jinja2",
"status": 400
}, {
"domain": "birds.com",
"template": "could_not_process.html.jinja2",
"status": 451
}])
def test_it_ignores_invalid_lines_in_blocklist(self, tmp_path):
blocklist_path = str(tmp_path / "test-blocklist.txt")
blocklist_content = """
timewaster.com blocked
invalid-line
foo bar baz
# This is a comment
"""
_write_file(blocklist_path, blocklist_content)
app = Blocker(upstream_app, blocklist_path)
client = Client(app, Response)
resp = client.get("/timewaster.com")
assert "cannot be annotated" in resp.data
def test_it_reads_blocklist_from_file(self, file_open, tmp_path):
blocklist_path = str(tmp_path / "test-blocklist.txt")
_write_file(blocklist_path, "timewaster.com blocked")
app = Blocker(upstream_app, blocklist_path)
client = Client(app, Response)
# The blocklist should be fetched when the app is instantiated.
file_open.assert_called_with(blocklist_path)
# Fetch a site that is blocked in the custom blocklist.
resp = client.get("/timewaster.com")
assert "cannot be annotated" in resp.data
# Fetch a site that is not blocked in the custom blocklist,
resp = client.get("/youtube.com")
assert UPSTREAM_CONTENT in resp.data
def test_it_rereads_blocklist_if_mtime_changes(self, client, file_open,
tmp_path):
blocklist_path = str(tmp_path / "test-blocklist.txt")
_write_file(blocklist_path, "", mtime=1000)
app = Blocker(upstream_app, blocklist_path)
client = Client(app, Response)
# An initial request should not re-read the blocklist file,
# as the mtime is unchanged.
file_open.reset_mock()
resp = client.get("/timewaster.com")
file_open.assert_not_called()
assert UPSTREAM_CONTENT in resp.data
# Simulate a change in content and mtime of the blocklist file, which
# should cause it to be re-read on the next request.
_write_file(blocklist_path, "timewaster.com blocked", mtime=2000)
resp = client.get("/timewaster.com")
file_open.assert_called_with(blocklist_path)
assert "cannot be annotated" in resp.data
def app(environ, start_response):
embed_url = os.environ.get("H_EMBED_URL", "https://hypothes.is/embed.js")
template_params = environ.get("pywb.template_params", {})
template_params["h_embed_url"] = embed_url
environ["pywb.template_params"] = template_params
return pywb.apps.wayback.application(environ, start_response)
application = RequestHeaderSanitiser(app)
application = ResponseHeaderSanitiser(application)
application = Blocker(
application,
checkmate_host=os.environ["CHECKMATE_URL"],
api_key=os.environ["CHECKMATE_API_KEY"],
)
application = UserAgentDecorator(application, "Hypothesis-Via")
application = ConfigExtractor(application)
application = wsgi.DispatcherMiddleware(
application,
{
"/favicon.ico": static.Cling("static/favicon.ico"),
"/robots.txt": static.Cling("static/robots.txt"),
"/static": static.Cling("static/"),
"/static/__pywb": static.Cling(resource_filename("pywb", "static/")),
"/static/__shared/viewer/web/viewer.html": redirect_old_viewer,
"/h": redirect_strip_matched_path,
"/_status": status_endpoint,
},
def blocker(self, application):
return Blocker(application,
checkmate_host="http://checkmate.example.com")
return redirect(path, code=301)
def app(environ, start_response):
embed_url = os.environ.get("H_EMBED_URL", "https://hypothes.is/embed.js")
template_params = environ.get("pywb.template_params", {})
template_params["h_embed_url"] = embed_url
environ["pywb.template_params"] = template_params
return pywb.apps.wayback.application(environ, start_response)
application = RequestHeaderSanitiser(app)
application = ResponseHeaderSanitiser(application)
application = Blocker(application, checkmate_host=os.environ["CHECKMATE_URL"])
application = UserAgentDecorator(application, "Hypothesis-Via")
application = ConfigExtractor(application)
application = wsgi.DispatcherMiddleware(
application,
{
"/favicon.ico": static.Cling("static/favicon.ico"),
"/robots.txt": static.Cling("static/robots.txt"),
"/static": static.Cling("static/"),
"/static/__pywb": static.Cling(resource_filename("pywb", "static/")),
"/static/__shared/viewer/web/viewer.html": redirect_old_viewer,
"/h": redirect_strip_matched_path,
},
)
application = newrelic.agent.WSGIApplicationWrapper(application, name="proxy")
template_params = environ.get("pywb.template_params", {})
template_params["h_embed_url"] = embed_url
environ["pywb.template_params"] = template_params
return pywb.apps.wayback.application(environ, start_response)
blocker_kwargs = {}
blocklist_path = os.environ.get("BLOCKLIST_PATH")
if blocklist_path:
blocker_kwargs["blocklist_path"] = blocklist_path
application = RequestHeaderSanitiser(app)
application = ResponseHeaderSanitiser(application)
application = Blocker(application, **blocker_kwargs)
application = UserAgentDecorator(application, "Hypothesis-Via")
application = ConfigExtractor(application)
application = wsgi.DispatcherMiddleware(
application,
{
"/favicon.ico": static.Cling("static/favicon.ico"),
"/robots.txt": static.Cling("static/robots.txt"),
"/static": static.Cling("static/"),
"/static/__pywb": static.Cling(resource_filename("pywb", "static/")),
"/static/__shared/viewer/web/viewer.html": redirect_old_viewer,
"/h": redirect_strip_matched_path,
},
)
application = newrelic.agent.WSGIApplicationWrapper(application, name="proxy")
def app(self):
return Blocker(upstream_app)
self.application = application
def __call__(self, environ, start_response):
host_to_replace = "www.facebook.com"
host_replacing = "m.facebook.com"
start = environ["PATH_INFO"].find(host_to_replace)
if start == 8 or start == 9:
path = environ["PATH_INFO"]
path = path[:start] + host_replacing + \
path[start + len(host_to_replace):]
start_response('302 Found', [('Location', path)])
return ['1']
return self.application(environ, start_response)
application = RequestHeaderSanitiser(app)
application = ResponseHeaderSanitiser(application)
application = Blocker(application)
application = UserAgentDecorator(application, 'Hypothesis-Via')
application = FacebookRedirector(application)
application = ConfigExtractor(application)
application = wsgi.DispatcherMiddleware(
application, {
'/favicon.ico': static.Cling('static/favicon.ico'),
'/robots.txt': static.Cling('static/robots.txt'),
'/static': static.Cling('static/'),
'/static/__pywb': static.Cling(resource_filename('pywb', 'static/')),
'/static/__shared/viewer/web/viewer.html': redirect_old_viewer,
'/h': redirect_strip_matched_path,
})
def blocker(self, application):
return Blocker(
application,
checkmate_host="http://checkmate.example.com",
api_key="test_api_key",
)
def app(self):
return Blocker(upstream_app,
prefixes=['/giraffe', '/birds/duck'],
template='your eyes are protected')