class NotebookApp(BaseIPythonApplication):
name = 'ipython-notebook'
description = """
The IPython HTML Notebook.
This launches a Tornado based HTML Notebook Server that serves up an
HTML5/Javascript Notebook client.
"""
examples = _examples
classes = IPythonConsoleApp.classes + [
MappingKernelManager, NotebookManager, FileNotebookManager,
NotebookNotary
]
flags = Dict(flags)
aliases = Dict(aliases)
subcommands = dict(list=(NbserverListApp,
NbserverListApp.description.splitlines()[0]), )
kernel_argv = List(Unicode)
def _log_level_default(self):
return logging.INFO
def _log_format_default(self):
"""override default log format to include time"""
return u"%(asctime)s.%(msecs).03d [%(name)s]%(highlevel)s %(message)s"
# create requested profiles by default, if they don't exist:
auto_create = Bool(True)
# file to be opened in the notebook server
file_to_run = Unicode('', config=True)
def _file_to_run_changed(self, name, old, new):
path, base = os.path.split(new)
if path:
self.file_to_run = base
self.notebook_dir = path
# Network related information
allow_origin = Unicode('',
config=True,
help="""Set the Access-Control-Allow-Origin header
Use '*' to allow any origin to access your server.
Takes precedence over allow_origin_pat.
""")
allow_origin_pat = Unicode(
'',
config=True,
help=
"""Use a regular expression for the Access-Control-Allow-Origin header
Requests from an origin matching the expression will get replies with:
Access-Control-Allow-Origin: origin
where `origin` is the origin of the request.
Ignored if allow_origin is set.
""")
allow_credentials = Bool(
False,
config=True,
help="Set the Access-Control-Allow-Credentials: true header")
ip = Unicode('localhost',
config=True,
help="The IP address the notebook server will listen on.")
def _ip_changed(self, name, old, new):
if new == u'*': self.ip = u''
port = Integer(8888,
config=True,
help="The port the notebook server will listen on.")
port_retries = Integer(
50,
config=True,
help=
"The number of additional ports to try if the specified port is not available."
)
certfile = Unicode(
u'',
config=True,
help="""The full path to an SSL/TLS certificate file.""")
keyfile = Unicode(
u'',
config=True,
help="""The full path to a private key file for usage with SSL/TLS.""")
cookie_secret = Bytes(b'',
config=True,
help="""The random bytes used to secure cookies.
By default this is a new random number every time you start the Notebook.
Set it to a value in a config file to enable logins to persist across server sessions.
Note: Cookie secrets should be kept private, do not share config files with
cookie_secret stored in plaintext (you can read the value from a file).
""")
def _cookie_secret_default(self):
return os.urandom(1024)
password = Unicode(u'',
config=True,
help="""Hashed password to use for web authentication.
To generate, type in a python/IPython shell:
from IPython.lib import passwd; passwd()
The string should be of the form type:salt:hashed-password.
""")
open_browser = Bool(True,
config=True,
help="""Whether to open in a browser after starting.
The specific browser used is platform dependent and
determined by the python standard library `webbrowser`
module, unless it is overridden using the --browser
(NotebookApp.browser) configuration option.
""")
browser = Unicode(u'',
config=True,
help="""Specify what command to use to invoke a web
browser when opening the notebook. If not specified, the
default browser will be determined by the `webbrowser`
standard library module, which allows setting of the
BROWSER environment variable to override it.
""")
webapp_settings = Dict(
config=True,
help="Supply overrides for the tornado.web.Application that the "
"IPython notebook uses.")
jinja_environment_options = Dict(
config=True,
help="Supply extra arguments that will be passed to Jinja environment."
)
enable_mathjax = Bool(
True,
config=True,
help="""Whether to enable MathJax for typesetting math/TeX
MathJax is the javascript library IPython uses to render math/LaTeX. It is
very large, so you may want to disable it if you have a slow internet
connection, or for offline use of the notebook.
When disabled, equations etc. will appear as their untransformed TeX source.
""")
def _enable_mathjax_changed(self, name, old, new):
"""set mathjax url to empty if mathjax is disabled"""
if not new:
self.mathjax_url = u''
base_url = Unicode('/',
config=True,
help='''The base URL for the notebook server.
Leading and trailing slashes can be omitted,
and will automatically be added.
''')
def _base_url_changed(self, name, old, new):
if not new.startswith('/'):
self.base_url = '/' + new
elif not new.endswith('/'):
self.base_url = new + '/'
base_project_url = Unicode('/',
config=True,
help="""DEPRECATED use base_url""")
def _base_project_url_changed(self, name, old, new):
self.log.warn("base_project_url is deprecated, use base_url")
self.base_url = new
extra_static_paths = List(
Unicode,
config=True,
help="""Extra paths to search for serving static files.
This allows adding javascript/css to be available from the notebook server machine,
or overriding individual files in the IPython""")
def _extra_static_paths_default(self):
return [os.path.join(self.profile_dir.location, 'static')]
@property
def static_file_path(self):
"""return extra paths + the default location"""
return self.extra_static_paths + [DEFAULT_STATIC_FILES_PATH]
nbextensions_path = List(
Unicode,
config=True,
help=
"""paths for Javascript extensions. By default, this is just IPYTHONDIR/nbextensions"""
)
def _nbextensions_path_default(self):
return [os.path.join(get_ipython_dir(), 'nbextensions')]
mathjax_url = Unicode("", config=True, help="""The url for MathJax.js.""")
def _mathjax_url_default(self):
if not self.enable_mathjax:
return u''
static_url_prefix = self.webapp_settings.get(
"static_url_prefix", url_path_join(self.base_url, "static"))
# try local mathjax, either in nbextensions/mathjax or static/mathjax
for (url_prefix, search_path) in [
(url_path_join(self.base_url,
"nbextensions"), self.nbextensions_path),
(static_url_prefix, self.static_file_path),
]:
self.log.debug("searching for local mathjax in %s", search_path)
try:
mathjax = filefind(os.path.join('mathjax', 'MathJax.js'),
search_path)
except IOError:
continue
else:
url = url_path_join(url_prefix, u"mathjax/MathJax.js")
self.log.info("Serving local MathJax from %s at %s", mathjax,
url)
return url
if os.path.exists("/usr/share/javascript/mathjax/MathJax.js"):
self.log.info("Using system MathJax")
return u"/mathjax/MathJax.js"
# no local mathjax, serve from CDN
url = u"https://cdn.mathjax.org/mathjax/latest/MathJax.js"
self.log.info("Using MathJax from CDN: %s", url)
return url
def _mathjax_url_changed(self, name, old, new):
if new and not self.enable_mathjax:
# enable_mathjax=False overrides mathjax_url
self.mathjax_url = u''
else:
self.log.info("Using MathJax: %s", new)
notebook_manager_class = DottedObjectName(
'IPython.html.services.notebooks.filenbmanager.FileNotebookManager',
config=True,
help='The notebook manager class to use.')
trust_xheaders = Bool(
False,
config=True,
help=
("Whether to trust or not X-Scheme/X-Forwarded-Proto and X-Real-Ip/X-Forwarded-For headers"
"sent by the upstream reverse proxy. Necessary if the proxy handles SSL"
))
info_file = Unicode()
def _info_file_default(self):
info_file = "nbserver-%s.json" % os.getpid()
return os.path.join(self.profile_dir.security_dir, info_file)
notebook_dir = Unicode(
py3compat.getcwd(),
config=True,
help="The directory to use for notebooks and kernels.")
def _notebook_dir_changed(self, name, old, new):
"""Do a bit of validation of the notebook dir."""
if not os.path.isabs(new):
# If we receive a non-absolute path, make it absolute.
self.notebook_dir = os.path.abspath(new)
return
if not os.path.isdir(new):
raise TraitError("No such notebook dir: %r" % new)
# setting App.notebook_dir implies setting notebook and kernel dirs as well
self.config.FileNotebookManager.notebook_dir = new
self.config.MappingKernelManager.root_dir = new
def parse_command_line(self, argv=None):
super(NotebookApp, self).parse_command_line(argv)
if self.extra_args:
arg0 = self.extra_args[0]
f = os.path.abspath(arg0)
self.argv.remove(arg0)
if not os.path.exists(f):
self.log.critical("No such file or directory: %s", f)
self.exit(1)
# Use config here, to ensure that it takes higher priority than
# anything that comes from the profile.
c = Config()
if os.path.isdir(f):
c.NotebookApp.notebook_dir = f
elif os.path.isfile(f):
c.NotebookApp.file_to_run = f
self.update_config(c)
def init_kernel_argv(self):
"""construct the kernel arguments"""
# Scrub frontend-specific flags
self.kernel_argv = swallow_argv(self.argv, notebook_aliases,
notebook_flags)
if any(arg.startswith(u'--pylab') for arg in self.kernel_argv):
self.log.warn('\n '.join([
"Starting all kernels in pylab mode is not recommended,",
"and will be disabled in a future release.",
"Please use the %matplotlib magic to enable matplotlib instead.",
"pylab implies many imports, which can have confusing side effects",
"and harm the reproducibility of your notebooks.",
]))
# Kernel should inherit default config file from frontend
self.kernel_argv.append("--IPKernelApp.parent_appname='%s'" %
self.name)
# Kernel should get *absolute* path to profile directory
self.kernel_argv.extend(["--profile-dir", self.profile_dir.location])
def init_configurables(self):
# force Session default to be secure
default_secure(self.config)
self.kernel_manager = MappingKernelManager(
parent=self,
log=self.log,
kernel_argv=self.kernel_argv,
connection_dir=self.profile_dir.security_dir,
)
kls = import_item(self.notebook_manager_class)
self.notebook_manager = kls(parent=self, log=self.log)
self.session_manager = SessionManager(parent=self, log=self.log)
self.cluster_manager = ClusterManager(parent=self, log=self.log)
self.cluster_manager.update_profiles()
def init_logging(self):
# This prevents double log messages because tornado use a root logger that
# self.log is a child of. The logging module dipatches log messages to a log
# and all of its ancenstors until propagate is set to False.
self.log.propagate = False
# hook up tornado 3's loggers to our app handlers
for name in ('access', 'application', 'general'):
logger = logging.getLogger('tornado.%s' % name)
logger.parent = self.log
logger.setLevel(self.log.level)
def init_webapp(self):
"""initialize tornado webapp and httpserver"""
self.webapp_settings['allow_origin'] = self.allow_origin
if self.allow_origin_pat:
self.webapp_settings['allow_origin_pat'] = re.compile(
self.allow_origin_pat)
self.webapp_settings['allow_credentials'] = self.allow_credentials
self.web_app = NotebookWebApplication(
self, self.kernel_manager, self.notebook_manager,
self.cluster_manager, self.session_manager, self.log,
self.base_url, self.webapp_settings,
self.jinja_environment_options)
if self.certfile:
ssl_options = dict(certfile=self.certfile)
# Disable SSLv3, since its use is discouraged.
ssl_options['ssl_version'] = ssl.PROTOCOL_TLSv1
if self.keyfile:
ssl_options['keyfile'] = self.keyfile
else:
ssl_options = None
self.web_app.password = self.password
self.http_server = httpserver.HTTPServer(self.web_app,
ssl_options=ssl_options,
xheaders=self.trust_xheaders)
if not self.ip:
warning = "WARNING: The notebook server is listening on all IP addresses"
if ssl_options is None:
self.log.critical(warning + " and not using encryption. This "
"is not recommended.")
if not self.password:
self.log.critical(
warning + " and not using authentication. "
"This is highly insecure and not recommended.")
success = None
for port in random_ports(self.port, self.port_retries + 1):
try:
self.http_server.listen(port, self.ip)
except socket.error as e:
if e.errno == errno.EADDRINUSE:
self.log.info(
'The port %i is already in use, trying another random port.'
% port)
continue
elif e.errno in (errno.EACCES,
getattr(errno, 'WSAEACCES', errno.EACCES)):
self.log.warn("Permission to listen on port %i denied" %
port)
continue
else:
raise
else:
self.port = port
success = True
break
if not success:
self.log.critical(
'ERROR: the notebook server could not be started because '
'no available port could be found.')
self.exit(1)
@property
def display_url(self):
ip = self.ip if self.ip else '[all ip addresses on your system]'
return self._url(ip)
@property
def connection_url(self):
ip = self.ip if self.ip else 'localhost'
return self._url(ip)
def _url(self, ip):
proto = 'https' if self.certfile else 'http'
return "%s://%s:%i%s" % (proto, ip, self.port, self.base_url)
def init_signal(self):
if not sys.platform.startswith('win'):
signal.signal(signal.SIGINT, self._handle_sigint)
signal.signal(signal.SIGTERM, self._signal_stop)
if hasattr(signal, 'SIGUSR1'):
# Windows doesn't support SIGUSR1
signal.signal(signal.SIGUSR1, self._signal_info)
if hasattr(signal, 'SIGINFO'):
# only on BSD-based systems
signal.signal(signal.SIGINFO, self._signal_info)
def _handle_sigint(self, sig, frame):
"""SIGINT handler spawns confirmation dialog"""
# register more forceful signal handler for ^C^C case
signal.signal(signal.SIGINT, self._signal_stop)
# request confirmation dialog in bg thread, to avoid
# blocking the App
thread = threading.Thread(target=self._confirm_exit)
thread.daemon = True
thread.start()
def _restore_sigint_handler(self):
"""callback for restoring original SIGINT handler"""
signal.signal(signal.SIGINT, self._handle_sigint)
def _confirm_exit(self):
"""confirm shutdown on ^C
A second ^C, or answering 'y' within 5s will cause shutdown,
otherwise original SIGINT handler will be restored.
This doesn't work on Windows.
"""
# FIXME: remove this delay when pyzmq dependency is >= 2.1.11
time.sleep(0.1)
info = self.log.info
info('interrupted')
print(self.notebook_info())
sys.stdout.write("Shutdown this notebook server (y/[n])? ")
sys.stdout.flush()
r, w, x = select.select([sys.stdin], [], [], 5)
if r:
line = sys.stdin.readline()
if line.lower().startswith('y') and 'n' not in line.lower():
self.log.critical("Shutdown confirmed")
ioloop.IOLoop.instance().stop()
return
else:
print("No answer for 5s:", end=' ')
print("resuming operation...")
# no answer, or answer is no:
# set it back to original SIGINT handler
# use IOLoop.add_callback because signal.signal must be called
# from main thread
ioloop.IOLoop.instance().add_callback(self._restore_sigint_handler)
def _signal_stop(self, sig, frame):
self.log.critical("received signal %s, stopping", sig)
ioloop.IOLoop.instance().stop()
def _signal_info(self, sig, frame):
print(self.notebook_info())
def init_components(self):
pass
@catch_config_error
def initialize(self, argv=None):
super(NotebookApp, self).initialize(argv)
self.init_logging()
self.init_kernel_argv()
self.init_configurables()
self.init_components()
self.init_webapp()
self.init_signal()
def cleanup_kernels(self):
"""Shutdown all kernels.
The kernels will shutdown themselves when this process no longer exists,
but explicit shutdown allows the KernelManagers to cleanup the connection files.
"""
self.log.info('Shutting down kernels')
self.kernel_manager.shutdown_all()
def notebook_info(self):
"Return the current working directory and the server url information"
info = self.notebook_manager.info_string() + "\n"
info += "%d active kernels \n" % len(self.kernel_manager._kernels)
return info + "The IPython Notebook is running at: %s" % self.display_url
def server_info(self):
"""Return a JSONable dict of information about this server."""
return {
'url': self.connection_url,
'hostname': self.ip if self.ip else 'localhost',
'port': self.port,
'secure': bool(self.certfile),
'base_url': self.base_url,
'notebook_dir': os.path.abspath(self.notebook_dir),
}
def write_server_info_file(self):
"""Write the result of server_info() to the JSON file info_file."""
with open(self.info_file, 'w') as f:
json.dump(self.server_info(), f, indent=2)
def remove_server_info_file(self):
"""Remove the nbserver-<pid>.json file created for this server.
Ignores the error raised when the file has already been removed.
"""
try:
os.unlink(self.info_file)
except OSError as e:
if e.errno != errno.ENOENT:
raise
def start(self):
""" Start the IPython Notebook server app, after initialization
This method takes no arguments so all configuration and initialization
must be done prior to calling this method."""
if self.subapp is not None:
return self.subapp.start()
info = self.log.info
for line in self.notebook_info().split("\n"):
info(line)
info(
"Use Control-C to stop this server and shut down all kernels (twice to skip confirmation)."
)
self.write_server_info_file()
if self.open_browser or self.file_to_run:
try:
browser = webbrowser.get(self.browser or None)
except webbrowser.Error as e:
self.log.warn('No web browser found: %s.' % e)
browser = None
if self.file_to_run:
fullpath = os.path.join(self.notebook_dir, self.file_to_run)
if not os.path.exists(fullpath):
self.log.critical("%s does not exist" % fullpath)
self.exit(1)
uri = url_path_join('notebooks', self.file_to_run)
else:
uri = 'tree'
if browser:
b = lambda: browser.open(
url_path_join(self.connection_url, uri), new=2)
threading.Thread(target=b).start()
try:
ioloop.IOLoop.instance().start()
except KeyboardInterrupt:
info("Interrupted...")
finally:
self.cleanup_kernels()
self.remove_server_info_file()
class NotebookApp(BaseIPythonApplication):
name = 'ipython-notebook'
default_config_file_name = 'ipython_notebook_config.py'
description = """
The IPython HTML Notebook.
This launches a Tornado based HTML Notebook Server that serves up an
HTML5/Javascript Notebook client.
"""
examples = _examples
classes = IPythonConsoleApp.classes + [
MappingKernelManager, NotebookManager, FileNotebookManager
]
flags = Dict(flags)
aliases = Dict(aliases)
kernel_argv = List(Unicode)
log_level = Enum(
(0, 10, 20, 30, 40, 50, 'DEBUG', 'INFO', 'WARN', 'ERROR', 'CRITICAL'),
default_value=logging.INFO,
config=True,
help="Set the log level by value or name.")
# create requested profiles by default, if they don't exist:
auto_create = Bool(True)
# file to be opened in the notebook server
file_to_run = Unicode('')
# Network related information.
ip = Unicode(LOCALHOST,
config=True,
help="The IP address the notebook server will listen on.")
def _ip_changed(self, name, old, new):
if new == u'*': self.ip = u''
port = Integer(8888,
config=True,
help="The port the notebook server will listen on.")
port_retries = Integer(
50,
config=True,
help=
"The number of additional ports to try if the specified port is not available."
)
certfile = Unicode(
u'',
config=True,
help="""The full path to an SSL/TLS certificate file.""")
keyfile = Unicode(
u'',
config=True,
help="""The full path to a private key file for usage with SSL/TLS.""")
password = Unicode(u'',
config=True,
help="""Hashed password to use for web authentication.
To generate, type in a python/IPython shell:
from IPython.lib import passwd; passwd()
The string should be of the form type:salt:hashed-password.
""")
open_browser = Bool(True,
config=True,
help="""Whether to open in a browser after starting.
The specific browser used is platform dependent and
determined by the python standard library `webbrowser`
module, unless it is overridden using the --browser
(NotebookApp.browser) configuration option.
""")
browser = Unicode(u'',
config=True,
help="""Specify what command to use to invoke a web
browser when opening the notebook. If not specified, the
default browser will be determined by the `webbrowser`
standard library module, which allows setting of the
BROWSER environment variable to override it.
""")
read_only = Bool(False,
config=True,
help="Whether to prevent editing/execution of notebooks.")
use_less = Bool(False,
config=True,
help="""Wether to use Browser Side less-css parsing
instead of compiled css version in templates that allows
it. This is mainly convenient when working on the less
file to avoid a build step, or if user want to overwrite
some of the less variables without having to recompile
everything.
You will need to install the less.js component in the static directory
either in the source tree or in your profile folder.
""")
webapp_settings = Dict(
config=True,
help="Supply overrides for the tornado.web.Application that the "
"IPython notebook uses.")
enable_mathjax = Bool(
True,
config=True,
help="""Whether to enable MathJax for typesetting math/TeX
MathJax is the javascript library IPython uses to render math/LaTeX. It is
very large, so you may want to disable it if you have a slow internet
connection, or for offline use of the notebook.
When disabled, equations etc. will appear as their untransformed TeX source.
""")
def _enable_mathjax_changed(self, name, old, new):
"""set mathjax url to empty if mathjax is disabled"""
if not new:
self.mathjax_url = u''
base_project_url = Unicode('/',
config=True,
help='''The base URL for the notebook server.
Leading and trailing slashes can be omitted,
and will automatically be added.
''')
def _base_project_url_changed(self, name, old, new):
if not new.startswith('/'):
self.base_project_url = '/' + new
elif not new.endswith('/'):
self.base_project_url = new + '/'
base_kernel_url = Unicode('/',
config=True,
help='''The base URL for the kernel server
Leading and trailing slashes can be omitted,
and will automatically be added.
''')
def _base_kernel_url_changed(self, name, old, new):
if not new.startswith('/'):
self.base_kernel_url = '/' + new
elif not new.endswith('/'):
self.base_kernel_url = new + '/'
websocket_host = Unicode("",
config=True,
help="""The hostname for the websocket server.""")
extra_static_paths = List(
Unicode,
config=True,
help="""Extra paths to search for serving static files.
This allows adding javascript/css to be available from the notebook server machine,
or overriding individual files in the IPython""")
def _extra_static_paths_default(self):
return [os.path.join(self.profile_dir.location, 'static')]
@property
def static_file_path(self):
"""return extra paths + the default location"""
return self.extra_static_paths + [
os.path.join(os.path.dirname(__file__), "static")
]
mathjax_url = Unicode("", config=True, help="""The url for MathJax.js.""")
def _mathjax_url_default(self):
if not self.enable_mathjax:
return u''
static_url_prefix = self.webapp_settings.get("static_url_prefix",
"/static/")
try:
mathjax = filefind(os.path.join('mathjax', 'MathJax.js'),
self.static_file_path)
except IOError:
if self.certfile:
# HTTPS: load from Rackspace CDN, because SSL certificate requires it
base = u"https://c328740.ssl.cf1.rackcdn.com"
else:
base = u"http://cdn.mathjax.org"
url = base + u"/mathjax/latest/MathJax.js"
self.log.info("Using MathJax from CDN: %s", url)
return url
else:
self.log.info("Using local MathJax from %s" % mathjax)
return static_url_prefix + u"mathjax/MathJax.js"
def _mathjax_url_changed(self, name, old, new):
if new and not self.enable_mathjax:
# enable_mathjax=False overrides mathjax_url
self.mathjax_url = u''
else:
self.log.info("Using MathJax: %s", new)
notebook_manager_class = DottedObjectName(
'IPython.frontend.html.notebook.filenbmanager.FileNotebookManager',
config=True,
help='The notebook manager class to use.')
def parse_command_line(self, argv=None):
super(NotebookApp, self).parse_command_line(argv)
if argv is None:
argv = sys.argv[1:]
# Scrub frontend-specific flags
self.kernel_argv = swallow_argv(argv, notebook_aliases, notebook_flags)
# Kernel should inherit default config file from frontend
self.kernel_argv.append("--KernelApp.parent_appname='%s'" % self.name)
if self.extra_args:
f = os.path.abspath(self.extra_args[0])
if os.path.isdir(f):
nbdir = f
else:
self.file_to_run = f
nbdir = os.path.dirname(f)
self.config.NotebookManager.notebook_dir = nbdir
def init_configurables(self):
# force Session default to be secure
default_secure(self.config)
self.kernel_manager = MappingKernelManager(
config=self.config,
log=self.log,
kernel_argv=self.kernel_argv,
connection_dir=self.profile_dir.security_dir,
)
kls = import_item(self.notebook_manager_class)
self.notebook_manager = kls(config=self.config, log=self.log)
self.notebook_manager.log_info()
self.notebook_manager.load_notebook_names()
self.cluster_manager = ClusterManager(config=self.config, log=self.log)
self.cluster_manager.update_profiles()
def init_logging(self):
# This prevents double log messages because tornado use a root logger that
# self.log is a child of. The logging module dipatches log messages to a log
# and all of its ancenstors until propagate is set to False.
self.log.propagate = False
def init_webapp(self):
"""initialize tornado webapp and httpserver"""
self.web_app = NotebookWebApplication(self, self.kernel_manager,
self.notebook_manager,
self.cluster_manager, self.log,
self.base_project_url,
self.webapp_settings)
if self.certfile:
ssl_options = dict(certfile=self.certfile)
if self.keyfile:
ssl_options['keyfile'] = self.keyfile
else:
ssl_options = None
self.web_app.password = self.password
self.http_server = httpserver.HTTPServer(self.web_app,
ssl_options=ssl_options)
if not self.ip:
warning = "WARNING: The notebook server is listening on all IP addresses"
if ssl_options is None:
self.log.critical(warning + " and not using encryption. This"
"is not recommended.")
if not self.password and not self.read_only:
self.log.critical(
warning + "and not using authentication."
"This is highly insecure and not recommended.")
success = None
for port in random_ports(self.port, self.port_retries + 1):
try:
self.http_server.listen(port, self.ip)
except socket.error as e:
if e.errno != errno.EADDRINUSE:
raise
self.log.info(
'The port %i is already in use, trying another random port.'
% port)
else:
self.port = port
success = True
break
if not success:
self.log.critical(
'ERROR: the notebook server could not be started because '
'no available port could be found.')
self.exit(1)
def init_signal(self):
# FIXME: remove this check when pyzmq dependency is >= 2.1.11
# safely extract zmq version info:
try:
zmq_v = zmq.pyzmq_version_info()
except AttributeError:
zmq_v = [int(n) for n in re.findall(r'\d+', zmq.__version__)]
if 'dev' in zmq.__version__:
zmq_v.append(999)
zmq_v = tuple(zmq_v)
if zmq_v >= (2, 1, 9) and not sys.platform.startswith('win'):
# This won't work with 2.1.7 and
# 2.1.9-10 will log ugly 'Interrupted system call' messages,
# but it will work
signal.signal(signal.SIGINT, self._handle_sigint)
signal.signal(signal.SIGTERM, self._signal_stop)
def _handle_sigint(self, sig, frame):
"""SIGINT handler spawns confirmation dialog"""
# register more forceful signal handler for ^C^C case
signal.signal(signal.SIGINT, self._signal_stop)
# request confirmation dialog in bg thread, to avoid
# blocking the App
thread = threading.Thread(target=self._confirm_exit)
thread.daemon = True
thread.start()
def _restore_sigint_handler(self):
"""callback for restoring original SIGINT handler"""
signal.signal(signal.SIGINT, self._handle_sigint)
def _confirm_exit(self):
"""confirm shutdown on ^C
A second ^C, or answering 'y' within 5s will cause shutdown,
otherwise original SIGINT handler will be restored.
This doesn't work on Windows.
"""
# FIXME: remove this delay when pyzmq dependency is >= 2.1.11
time.sleep(0.1)
sys.stdout.write("Shutdown Notebook Server (y/[n])? ")
sys.stdout.flush()
r, w, x = select.select([sys.stdin], [], [], 5)
if r:
line = sys.stdin.readline()
if line.lower().startswith('y'):
self.log.critical("Shutdown confirmed")
ioloop.IOLoop.instance().stop()
return
else:
print "No answer for 5s:",
print "resuming operation..."
# no answer, or answer is no:
# set it back to original SIGINT handler
# use IOLoop.add_callback because signal.signal must be called
# from main thread
ioloop.IOLoop.instance().add_callback(self._restore_sigint_handler)
def _signal_stop(self, sig, frame):
self.log.critical("received signal %s, stopping", sig)
ioloop.IOLoop.instance().stop()
@catch_config_error
def initialize(self, argv=None):
self.init_logging()
super(NotebookApp, self).initialize(argv)
self.init_configurables()
self.init_webapp()
self.init_signal()
def cleanup_kernels(self):
"""Shutdown all kernels.
The kernels will shutdown themselves when this process no longer exists,
but explicit shutdown allows the KernelManagers to cleanup the connection files.
"""
self.log.info('Shutting down kernels')
self.kernel_manager.shutdown_all()
def start(self):
ip = self.ip if self.ip else '[all ip addresses on your system]'
proto = 'https' if self.certfile else 'http'
info = self.log.info
info("The IPython Notebook is running at: %s://%s:%i%s" %
(proto, ip, self.port, self.base_project_url))
info("Use Control-C to stop this server and shut down all kernels.")
if self.open_browser or self.file_to_run:
ip = self.ip or LOCALHOST
try:
browser = webbrowser.get(self.browser or None)
except webbrowser.Error as e:
self.log.warn('No web browser found: %s.' % e)
browser = None
if self.file_to_run:
name, _ = os.path.splitext(os.path.basename(self.file_to_run))
url = self.notebook_manager.rev_mapping.get(name, '')
else:
url = ''
if browser:
b = lambda: browser.open("%s://%s:%i%s%s" % (
proto, ip, self.port, self.base_project_url, url),
new=2)
threading.Thread(target=b).start()
try:
ioloop.IOLoop.instance().start()
except KeyboardInterrupt:
info("Interrupted...")
finally:
self.cleanup_kernels()
class ScriptMagics(Magics):
"""Magics for talking to scripts
This defines a base `%%script` cell magic for running a cell
with a program in a subprocess, and registers a few top-level
magics that call %%script with common interpreters.
"""
script_magics = List(
config=True,
help="""Extra script cell magics to define
This generates simple wrappers of `%%script foo` as `%%foo`.
If you want to add script magics that aren't on your path,
specify them in script_paths
""",
)
def _script_magics_default(self):
"""default to a common list of programs"""
defaults = [
'sh',
'bash',
'perl',
'ruby',
'python',
'python2',
'python3',
'pypy',
]
if os.name == 'nt':
defaults.extend([
'cmd',
])
return defaults
script_paths = Dict(
config=True,
help=
"""Dict mapping short 'ruby' names to full paths, such as '/opt/secret/bin/ruby'
Only necessary for items in script_magics where the default path will not
find the right interpreter.
""")
def __init__(self, shell=None):
super(ScriptMagics, self).__init__(shell=shell)
self._generate_script_magics()
self.job_manager = BackgroundJobManager()
self.bg_processes = []
atexit.register(self.kill_bg_processes)
def __del__(self):
self.kill_bg_processes()
def _generate_script_magics(self):
cell_magics = self.magics['cell']
for name in self.script_magics:
cell_magics[name] = self._make_script_magic(name)
def _make_script_magic(self, name):
"""make a named magic, that calls %%script with a particular program"""
# expand to explicit path if necessary:
script = self.script_paths.get(name, name)
@magic_arguments.magic_arguments()
@script_args
def named_script_magic(line, cell):
# if line, add it as cl-flags
if line:
line = "%s %s" % (script, line)
else:
line = script
return self.shebang(line, cell)
# write a basic docstring:
named_script_magic.__doc__ = \
"""%%{name} script magic
Run cells with {script} in a subprocess.
This is a shortcut for `%%script {script}`
""".format(**locals())
return named_script_magic
@magic_arguments.magic_arguments()
@script_args
@cell_magic("script")
def shebang(self, line, cell):
"""Run a cell via a shell command
The `%%script` line is like the #! line of script,
specifying a program (bash, perl, ruby, etc.) with which to run.
The rest of the cell is run by that program.
Examples
--------
::
In [1]: %%script bash
...: for i in 1 2 3; do
...: echo $i
...: done
1
2
3
"""
argv = arg_split(line, posix=not sys.platform.startswith('win'))
args, cmd = self.shebang.parser.parse_known_args(argv)
try:
p = Popen(cmd, stdout=PIPE, stderr=PIPE, stdin=PIPE)
except OSError as e:
if e.errno == errno.ENOENT:
print("Couldn't find program: %r" % cmd[0])
return
else:
raise
if not cell.endswith('\n'):
cell += '\n'
cell = cell.encode('utf8', 'replace')
if args.bg:
self.bg_processes.append(p)
self._gc_bg_processes()
if args.out:
self.shell.user_ns[args.out] = p.stdout
if args.err:
self.shell.user_ns[args.err] = p.stderr
self.job_manager.new(self._run_script, p, cell, daemon=True)
if args.proc:
self.shell.user_ns[args.proc] = p
return
try:
out, err = p.communicate(cell)
except KeyboardInterrupt:
try:
p.send_signal(signal.SIGINT)
time.sleep(0.1)
if p.poll() is not None:
print("Process is interrupted.")
return
p.terminate()
time.sleep(0.1)
if p.poll() is not None:
print("Process is terminated.")
return
p.kill()
print("Process is killed.")
except OSError:
pass
except Exception as e:
print("Error while terminating subprocess (pid=%i): %s" \
% (p.pid, e))
return
out = py3compat.bytes_to_str(out)
err = py3compat.bytes_to_str(err)
if args.out:
self.shell.user_ns[args.out] = out
else:
sys.stdout.write(out)
sys.stdout.flush()
if args.err:
self.shell.user_ns[args.err] = err
else:
sys.stderr.write(err)
sys.stderr.flush()
def _run_script(self, p, cell):
"""callback for running the script in the background"""
p.stdin.write(cell)
p.stdin.close()
p.wait()
@line_magic("killbgscripts")
def killbgscripts(self, _nouse_=''):
"""Kill all BG processes started by %%script and its family."""
self.kill_bg_processes()
print("All background processes were killed.")
def kill_bg_processes(self):
"""Kill all BG processes which are still running."""
for p in self.bg_processes:
if p.poll() is None:
try:
p.send_signal(signal.SIGINT)
except:
pass
time.sleep(0.1)
for p in self.bg_processes:
if p.poll() is None:
try:
p.terminate()
except:
pass
time.sleep(0.1)
for p in self.bg_processes:
if p.poll() is None:
try:
p.kill()
except:
pass
self._gc_bg_processes()
def _gc_bg_processes(self):
self.bg_processes = [p for p in self.bg_processes if p.poll() is None]
class Kernel(Configurable):
#---------------------------------------------------------------------------
# Kernel interface
#---------------------------------------------------------------------------
# attribute to override with a GUI
eventloop = Any(None)
def _eventloop_changed(self, name, old, new):
"""schedule call to eventloop from IOLoop"""
loop = ioloop.IOLoop.instance()
loop.add_callback(self.enter_eventloop)
session = Instance(Session)
profile_dir = Instance('IPython.core.profiledir.ProfileDir')
shell_streams = List()
control_stream = Instance(ZMQStream)
iopub_socket = Instance(zmq.Socket)
stdin_socket = Instance(zmq.Socket)
log = Instance(logging.Logger)
# identities:
int_id = Integer(-1)
ident = Unicode()
def _ident_default(self):
return unicode_type(uuid.uuid4())
# Private interface
_darwin_app_nap = Bool(
True,
config=True,
help="""Whether to use appnope for compatiblity with OS X App Nap.
Only affects OS X >= 10.9.
""")
# track associations with current request
_allow_stdin = Bool(False)
_parent_header = Dict()
_parent_ident = Any(b'')
# Time to sleep after flushing the stdout/err buffers in each execute
# cycle. While this introduces a hard limit on the minimal latency of the
# execute cycle, it helps prevent output synchronization problems for
# clients.
# Units are in seconds. The minimum zmq latency on local host is probably
# ~150 microseconds, set this to 500us for now. We may need to increase it
# a little if it's not enough after more interactive testing.
_execute_sleep = Float(0.0005, config=True)
# Frequency of the kernel's event loop.
# Units are in seconds, kernel subclasses for GUI toolkits may need to
# adapt to milliseconds.
_poll_interval = Float(0.05, config=True)
# If the shutdown was requested over the network, we leave here the
# necessary reply message so it can be sent by our registered atexit
# handler. This ensures that the reply is only sent to clients truly at
# the end of our shutdown process (which happens after the underlying
# IPython shell's own shutdown).
_shutdown_message = None
# This is a dict of port number that the kernel is listening on. It is set
# by record_ports and used by connect_request.
_recorded_ports = Dict()
# set of aborted msg_ids
aborted = Set()
# Track execution count here. For IPython, we override this to use the
# execution count we store in the shell.
execution_count = 0
def __init__(self, **kwargs):
super(Kernel, self).__init__(**kwargs)
# Build dict of handlers for message types
msg_types = [
'execute_request',
'complete_request',
'inspect_request',
'history_request',
'kernel_info_request',
'connect_request',
'shutdown_request',
'apply_request',
]
self.shell_handlers = {}
for msg_type in msg_types:
self.shell_handlers[msg_type] = getattr(self, msg_type)
control_msg_types = msg_types + ['clear_request', 'abort_request']
self.control_handlers = {}
for msg_type in control_msg_types:
self.control_handlers[msg_type] = getattr(self, msg_type)
def dispatch_control(self, msg):
"""dispatch control requests"""
idents, msg = self.session.feed_identities(msg, copy=False)
try:
msg = self.session.unserialize(msg, content=True, copy=False)
except:
self.log.error("Invalid Control Message", exc_info=True)
return
self.log.debug("Control received: %s", msg)
# Set the parent message for side effects.
self.set_parent(idents, msg)
self._publish_status(u'busy')
header = msg['header']
msg_type = header['msg_type']
handler = self.control_handlers.get(msg_type, None)
if handler is None:
self.log.error("UNKNOWN CONTROL MESSAGE TYPE: %r", msg_type)
else:
try:
handler(self.control_stream, idents, msg)
except Exception:
self.log.error("Exception in control handler:", exc_info=True)
sys.stdout.flush()
sys.stderr.flush()
self._publish_status(u'idle')
def dispatch_shell(self, stream, msg):
"""dispatch shell requests"""
# flush control requests first
if self.control_stream:
self.control_stream.flush()
idents, msg = self.session.feed_identities(msg, copy=False)
try:
msg = self.session.unserialize(msg, content=True, copy=False)
except:
self.log.error("Invalid Message", exc_info=True)
return
# Set the parent message for side effects.
self.set_parent(idents, msg)
self._publish_status(u'busy')
header = msg['header']
msg_id = header['msg_id']
msg_type = msg['header']['msg_type']
# Print some info about this message and leave a '--->' marker, so it's
# easier to trace visually the message chain when debugging. Each
# handler prints its message at the end.
self.log.debug('\n*** MESSAGE TYPE:%s***', msg_type)
self.log.debug(' Content: %s\n --->\n ', msg['content'])
if msg_id in self.aborted:
self.aborted.remove(msg_id)
# is it safe to assume a msg_id will not be resubmitted?
reply_type = msg_type.split('_')[0] + '_reply'
status = {'status': 'aborted'}
md = {'engine': self.ident}
md.update(status)
self.session.send(stream,
reply_type,
metadata=md,
content=status,
parent=msg,
ident=idents)
return
handler = self.shell_handlers.get(msg_type, None)
if handler is None:
self.log.error("UNKNOWN MESSAGE TYPE: %r", msg_type)
else:
# ensure default_int_handler during handler call
sig = signal(SIGINT, default_int_handler)
self.log.debug("%s: %s", msg_type, msg)
try:
handler(stream, idents, msg)
except Exception:
self.log.error("Exception in message handler:", exc_info=True)
finally:
signal(SIGINT, sig)
sys.stdout.flush()
sys.stderr.flush()
self._publish_status(u'idle')
def enter_eventloop(self):
"""enter eventloop"""
self.log.info("entering eventloop %s", self.eventloop)
for stream in self.shell_streams:
# flush any pending replies,
# which may be skipped by entering the eventloop
stream.flush(zmq.POLLOUT)
# restore default_int_handler
signal(SIGINT, default_int_handler)
while self.eventloop is not None:
try:
self.eventloop(self)
except KeyboardInterrupt:
# Ctrl-C shouldn't crash the kernel
self.log.error("KeyboardInterrupt caught in kernel")
continue
else:
# eventloop exited cleanly, this means we should stop (right?)
self.eventloop = None
break
self.log.info("exiting eventloop")
def start(self):
"""register dispatchers for streams"""
if self.control_stream:
self.control_stream.on_recv(self.dispatch_control, copy=False)
def make_dispatcher(stream):
def dispatcher(msg):
return self.dispatch_shell(stream, msg)
return dispatcher
for s in self.shell_streams:
s.on_recv(make_dispatcher(s), copy=False)
# publish idle status
self._publish_status('starting')
def do_one_iteration(self):
"""step eventloop just once"""
if self.control_stream:
self.control_stream.flush()
for stream in self.shell_streams:
# handle at most one request per iteration
stream.flush(zmq.POLLIN, 1)
stream.flush(zmq.POLLOUT)
def record_ports(self, ports):
"""Record the ports that this kernel is using.
The creator of the Kernel instance must call this methods if they
want the :meth:`connect_request` method to return the port numbers.
"""
self._recorded_ports = ports
#---------------------------------------------------------------------------
# Kernel request handlers
#---------------------------------------------------------------------------
def _make_metadata(self, other=None):
"""init metadata dict, for execute/apply_reply"""
new_md = {
'dependencies_met': True,
'engine': self.ident,
'started': datetime.now(),
}
if other:
new_md.update(other)
return new_md
def _publish_execute_input(self, code, parent, execution_count):
"""Publish the code request on the iopub stream."""
self.session.send(self.iopub_socket,
u'execute_input', {
u'code': code,
u'execution_count': execution_count
},
parent=parent,
ident=self._topic('execute_input'))
def _publish_status(self, status, parent=None):
"""send status (busy/idle) on IOPub"""
self.session.send(
self.iopub_socket,
u'status',
{u'execution_state': status},
parent=parent or self._parent_header,
ident=self._topic('status'),
)
def set_parent(self, ident, parent):
"""Set the current parent_header
Side effects (IOPub messages) and replies are associated with
the request that caused them via the parent_header.
The parent identity is used to route input_request messages
on the stdin channel.
"""
self._parent_ident = ident
self._parent_header = parent
def send_response(self,
stream,
msg_or_type,
content=None,
ident=None,
buffers=None,
track=False,
header=None,
metadata=None):
"""Send a response to the message we're currently processing.
This accepts all the parameters of :meth:`IPython.kernel.zmq.session.Session.send`
except ``parent``.
This relies on :meth:`set_parent` having been called for the current
message.
"""
return self.session.send(stream, msg_or_type, content,
self._parent_header, ident, buffers, track,
header, metadata)
def execute_request(self, stream, ident, parent):
"""handle an execute_request"""
try:
content = parent[u'content']
code = py3compat.cast_unicode_py2(content[u'code'])
silent = content[u'silent']
store_history = content.get(u'store_history', not silent)
user_expressions = content.get('user_expressions', {})
allow_stdin = content.get('allow_stdin', False)
except:
self.log.error("Got bad msg: ")
self.log.error("%s", parent)
return
md = self._make_metadata(parent['metadata'])
# Re-broadcast our input for the benefit of listening clients, and
# start computing output
if not silent:
self.execution_count += 1
self._publish_execute_input(code, parent, self.execution_count)
reply_content = self.do_execute(code, silent, store_history,
user_expressions, allow_stdin)
# Flush output before sending the reply.
sys.stdout.flush()
sys.stderr.flush()
# FIXME: on rare occasions, the flush doesn't seem to make it to the
# clients... This seems to mitigate the problem, but we definitely need
# to better understand what's going on.
if self._execute_sleep:
time.sleep(self._execute_sleep)
# Send the reply.
reply_content = json_clean(reply_content)
md['status'] = reply_content['status']
if reply_content['status'] == 'error' and \
reply_content['ename'] == 'UnmetDependency':
md['dependencies_met'] = False
reply_msg = self.session.send(stream,
u'execute_reply',
reply_content,
parent,
metadata=md,
ident=ident)
self.log.debug("%s", reply_msg)
if not silent and reply_msg['content']['status'] == u'error':
self._abort_queues()
def do_execute(self,
code,
silent,
store_history=True,
user_experssions=None,
allow_stdin=False):
"""Execute user code. Must be overridden by subclasses.
"""
raise NotImplementedError
def complete_request(self, stream, ident, parent):
content = parent['content']
code = content['code']
cursor_pos = content['cursor_pos']
matches = self.do_complete(code, cursor_pos)
matches = json_clean(matches)
completion_msg = self.session.send(stream, 'complete_reply', matches,
parent, ident)
self.log.debug("%s", completion_msg)
def do_complete(self, code, cursor_pos):
"""Override in subclasses to find completions.
"""
return {
'matches': [],
'cursor_end': cursor_pos,
'cursor_start': cursor_pos,
'metadata': {},
'status': 'ok'
}
def inspect_request(self, stream, ident, parent):
content = parent['content']
reply_content = self.do_inspect(content['code'], content['cursor_pos'],
content.get('detail_level', 0))
# Before we send this object over, we scrub it for JSON usage
reply_content = json_clean(reply_content)
msg = self.session.send(stream, 'inspect_reply', reply_content, parent,
ident)
self.log.debug("%s", msg)
def do_inspect(self, code, cursor_pos, detail_level=0):
"""Override in subclasses to allow introspection.
"""
return {'status': 'ok', 'data': {}, 'metadata': {}, 'found': False}
def history_request(self, stream, ident, parent):
content = parent['content']
reply_content = self.do_history(**content)
reply_content = json_clean(reply_content)
msg = self.session.send(stream, 'history_reply', reply_content, parent,
ident)
self.log.debug("%s", msg)
def do_history(self,
hist_access_type,
output,
raw,
session=None,
start=None,
stop=None,
n=None,
pattern=None,
unique=False):
"""Override in subclasses to access history.
"""
return {'history': []}
def connect_request(self, stream, ident, parent):
if self._recorded_ports is not None:
content = self._recorded_ports.copy()
else:
content = {}
msg = self.session.send(stream, 'connect_reply', content, parent,
ident)
self.log.debug("%s", msg)
@property
def kernel_info(self):
return {
'protocol_version': release.kernel_protocol_version,
'implementation': self.implementation,
'implementation_version': self.implementation_version,
'language': self.language,
'language_version': self.language_version,
'banner': self.banner,
}
def kernel_info_request(self, stream, ident, parent):
msg = self.session.send(stream, 'kernel_info_reply', self.kernel_info,
parent, ident)
self.log.debug("%s", msg)
def shutdown_request(self, stream, ident, parent):
content = self.do_shutdown(parent['content']['restart'])
self.session.send(stream,
u'shutdown_reply',
content,
parent,
ident=ident)
# same content, but different msg_id for broadcasting on IOPub
self._shutdown_message = self.session.msg(u'shutdown_reply', content,
parent)
self._at_shutdown()
# call sys.exit after a short delay
loop = ioloop.IOLoop.instance()
loop.add_timeout(time.time() + 0.1, loop.stop)
def do_shutdown(self, restart):
"""Override in subclasses to do things when the frontend shuts down the
kernel.
"""
return {'status': 'ok', 'restart': restart}
#---------------------------------------------------------------------------
# Engine methods
#---------------------------------------------------------------------------
def apply_request(self, stream, ident, parent):
try:
content = parent[u'content']
bufs = parent[u'buffers']
msg_id = parent['header']['msg_id']
except:
self.log.error("Got bad msg: %s", parent, exc_info=True)
return
md = self._make_metadata(parent['metadata'])
reply_content, result_buf = self.do_apply(content, bufs, msg_id, md)
# put 'ok'/'error' status in header, for scheduler introspection:
md['status'] = reply_content['status']
# flush i/o
sys.stdout.flush()
sys.stderr.flush()
self.session.send(stream,
u'apply_reply',
reply_content,
parent=parent,
ident=ident,
buffers=result_buf,
metadata=md)
def do_apply(self, content, bufs, msg_id, reply_metadata):
"""Override in subclasses to support the IPython parallel framework.
"""
raise NotImplementedError
#---------------------------------------------------------------------------
# Control messages
#---------------------------------------------------------------------------
def abort_request(self, stream, ident, parent):
"""abort a specifig msg by id"""
msg_ids = parent['content'].get('msg_ids', None)
if isinstance(msg_ids, string_types):
msg_ids = [msg_ids]
if not msg_ids:
self.abort_queues()
for mid in msg_ids:
self.aborted.add(str(mid))
content = dict(status='ok')
reply_msg = self.session.send(stream,
'abort_reply',
content=content,
parent=parent,
ident=ident)
self.log.debug("%s", reply_msg)
def clear_request(self, stream, idents, parent):
"""Clear our namespace."""
content = self.do_clear()
self.session.send(stream,
'clear_reply',
ident=idents,
parent=parent,
content=content)
def do_clear(self):
"""Override in subclasses to clear the namespace
This is only required for IPython.parallel.
"""
raise NotImplementedError
#---------------------------------------------------------------------------
# Protected interface
#---------------------------------------------------------------------------
def _topic(self, topic):
"""prefixed topic for IOPub messages"""
if self.int_id >= 0:
base = "engine.%i" % self.int_id
else:
base = "kernel.%s" % self.ident
return py3compat.cast_bytes("%s.%s" % (base, topic))
def _abort_queues(self):
for stream in self.shell_streams:
if stream:
self._abort_queue(stream)
def _abort_queue(self, stream):
poller = zmq.Poller()
poller.register(stream.socket, zmq.POLLIN)
while True:
idents, msg = self.session.recv(stream, zmq.NOBLOCK, content=True)
if msg is None:
return
self.log.info("Aborting:")
self.log.info("%s", msg)
msg_type = msg['header']['msg_type']
reply_type = msg_type.split('_')[0] + '_reply'
status = {'status': 'aborted'}
md = {'engine': self.ident}
md.update(status)
reply_msg = self.session.send(stream,
reply_type,
metadata=md,
content=status,
parent=msg,
ident=idents)
self.log.debug("%s", reply_msg)
# We need to wait a bit for requests to come in. This can probably
# be set shorter for true asynchronous clients.
poller.poll(50)
def _no_raw_input(self):
"""Raise StdinNotImplentedError if active frontend doesn't support
stdin."""
raise StdinNotImplementedError("raw_input was called, but this "
"frontend does not support stdin.")
def getpass(self, prompt=''):
"""Forward getpass to frontends
Raises
------
StdinNotImplentedError if active frontend doesn't support stdin.
"""
if not self._allow_stdin:
raise StdinNotImplementedError(
"getpass was called, but this frontend does not support input requests."
)
return self._input_request(
prompt,
self._parent_ident,
self._parent_header,
password=True,
)
def raw_input(self, prompt=''):
"""Forward raw_input to frontends
Raises
------
StdinNotImplentedError if active frontend doesn't support stdin.
"""
if not self._allow_stdin:
raise StdinNotImplementedError(
"raw_input was called, but this frontend does not support input requests."
)
return self._input_request(
prompt,
self._parent_ident,
self._parent_header,
password=False,
)
def _input_request(self, prompt, ident, parent, password=False):
# Flush output before making the request.
sys.stderr.flush()
sys.stdout.flush()
# flush the stdin socket, to purge stale replies
while True:
try:
self.stdin_socket.recv_multipart(zmq.NOBLOCK)
except zmq.ZMQError as e:
if e.errno == zmq.EAGAIN:
break
else:
raise
# Send the input request.
content = json_clean(dict(prompt=prompt, password=password))
self.session.send(self.stdin_socket,
u'input_request',
content,
parent,
ident=ident)
# Await a response.
while True:
try:
ident, reply = self.session.recv(self.stdin_socket, 0)
except Exception:
self.log.warn("Invalid Message:", exc_info=True)
except KeyboardInterrupt:
# re-raise KeyboardInterrupt, to truncate traceback
raise KeyboardInterrupt
else:
break
try:
value = py3compat.unicode_to_str(reply['content']['value'])
except:
self.log.error("Bad input_reply: %s", parent)
value = ''
if value == '\x04':
# EOF
raise EOFError
return value
def _at_shutdown(self):
"""Actions taken at shutdown by the kernel, called by python's atexit.
"""
# io.rprint("Kernel at_shutdown") # dbg
if self._shutdown_message is not None:
self.session.send(self.iopub_socket,
self._shutdown_message,
ident=self._topic('shutdown'))
self.log.debug("%s", self._shutdown_message)
[s.flush(zmq.POLLOUT) for s in self.shell_streams]
class IPythonConsoleApp(Configurable):
name = 'ipython-console-mixin'
default_config_file_name = 'ipython_config.py'
description = """
The IPython Mixin Console.
This class contains the common portions of console client (QtConsole,
ZMQ-based terminal console, etc). It is not a full console, in that
launched terminal subprocesses will not be able to accept input.
The Console using this mixing supports various extra features beyond
the single-process Terminal IPython shell, such as connecting to
existing kernel, via:
ipython <appname> --existing
as well as tunnel via SSH
"""
classes = classes
flags = Dict(flags)
aliases = Dict(aliases)
kernel_manager_class = BlockingKernelManager
kernel_argv = List(Unicode)
# frontend flags&aliases to be stripped when building kernel_argv
frontend_flags = Any(app_flags)
frontend_aliases = Any(app_aliases)
# create requested profiles by default, if they don't exist:
auto_create = CBool(True)
# connection info:
ip = Unicode(LOCALHOST,
config=True,
help="""Set the kernel\'s IP address [default localhost].
If the IP address is something other than localhost, then
Consoles on other machines will be able to connect
to the Kernel, so be careful!""")
sshserver = Unicode(
'',
config=True,
help="""The SSH server to use to connect to the kernel.""")
sshkey = Unicode(
'',
config=True,
help="""Path to the ssh key to use for logging in to the ssh server."""
)
hb_port = Int(0,
config=True,
help="set the heartbeat port [default: random]")
shell_port = Int(0,
config=True,
help="set the shell (ROUTER) port [default: random]")
iopub_port = Int(0,
config=True,
help="set the iopub (PUB) port [default: random]")
stdin_port = Int(0,
config=True,
help="set the stdin (DEALER) port [default: random]")
connection_file = Unicode(
'',
config=True,
help=
"""JSON file in which to store connection info [default: kernel-<pid>.json]
This file will contain the IP, ports, and authentication key needed to connect
clients to this kernel. By default, this file will be created in the security-dir
of the current profile, but can be specified by absolute path.
""")
def _connection_file_default(self):
return 'kernel-%i.json' % os.getpid()
existing = CUnicode('',
config=True,
help="""Connect to an already running kernel""")
confirm_exit = CBool(
True,
config=True,
help="""
Set to display confirmation dialog on exit. You can always use 'exit' or 'quit',
to force a direct exit without any confirmation.""",
)
def build_kernel_argv(self, argv=None):
"""build argv to be passed to kernel subprocess"""
if argv is None:
argv = sys.argv[1:]
self.kernel_argv = swallow_argv(argv, self.frontend_aliases,
self.frontend_flags)
# kernel should inherit default config file from frontend
self.kernel_argv.append("--KernelApp.parent_appname='%s'" % self.name)
def init_connection_file(self):
"""find the connection file, and load the info if found.
The current working directory and the current profile's security
directory will be searched for the file if it is not given by
absolute path.
When attempting to connect to an existing kernel and the `--existing`
argument does not match an existing file, it will be interpreted as a
fileglob, and the matching file in the current profile's security dir
with the latest access time will be used.
After this method is called, self.connection_file contains the *full path*
to the connection file, never just its name.
"""
if self.existing:
try:
cf = find_connection_file(self.existing)
except Exception:
self.log.critical(
"Could not find existing kernel connection file %s",
self.existing)
self.exit(1)
self.log.info("Connecting to existing kernel: %s" % cf)
self.connection_file = cf
else:
# not existing, check if we are going to write the file
# and ensure that self.connection_file is a full path, not just the shortname
try:
cf = find_connection_file(self.connection_file)
except Exception:
# file might not exist
if self.connection_file == os.path.basename(
self.connection_file):
# just shortname, put it in security dir
cf = os.path.join(self.profile_dir.security_dir,
self.connection_file)
else:
cf = self.connection_file
self.connection_file = cf
# should load_connection_file only be used for existing?
# as it is now, this allows reusing ports if an existing
# file is requested
try:
self.load_connection_file()
except Exception:
self.log.error("Failed to load connection file: %r",
self.connection_file,
exc_info=True)
self.exit(1)
def load_connection_file(self):
"""load ip/port/hmac config from JSON connection file"""
# this is identical to KernelApp.load_connection_file
# perhaps it can be centralized somewhere?
try:
fname = filefind(self.connection_file,
['.', self.profile_dir.security_dir])
except IOError:
self.log.debug("Connection File not found: %s",
self.connection_file)
return
self.log.debug("Loading connection file %s", fname)
with open(fname) as f:
s = f.read()
cfg = json.loads(s)
if self.ip == LOCALHOST and 'ip' in cfg:
# not overridden by config or cl_args
self.ip = cfg['ip']
for channel in ('hb', 'shell', 'iopub', 'stdin'):
name = channel + '_port'
if getattr(self, name) == 0 and name in cfg:
# not overridden by config or cl_args
setattr(self, name, cfg[name])
if 'key' in cfg:
self.config.Session.key = str_to_bytes(cfg['key'])
def init_ssh(self):
"""set up ssh tunnels, if needed."""
if not self.sshserver and not self.sshkey:
return
if self.sshkey and not self.sshserver:
# specifying just the key implies that we are connecting directly
self.sshserver = self.ip
self.ip = LOCALHOST
# build connection dict for tunnels:
info = dict(ip=self.ip,
shell_port=self.shell_port,
iopub_port=self.iopub_port,
stdin_port=self.stdin_port,
hb_port=self.hb_port)
self.log.info("Forwarding connections to %s via %s" %
(self.ip, self.sshserver))
# tunnels return a new set of ports, which will be on localhost:
self.ip = LOCALHOST
try:
newports = tunnel_to_kernel(info, self.sshserver, self.sshkey)
except:
# even catch KeyboardInterrupt
self.log.error("Could not setup tunnels", exc_info=True)
self.exit(1)
self.shell_port, self.iopub_port, self.stdin_port, self.hb_port = newports
cf = self.connection_file
base, ext = os.path.splitext(cf)
base = os.path.basename(base)
self.connection_file = os.path.basename(base) + '-ssh' + ext
self.log.critical("To connect another client via this tunnel, use:")
self.log.critical("--existing %s" % self.connection_file)
def _new_connection_file(self):
cf = ''
while not cf:
# we don't need a 128b id to distinguish kernels, use more readable
# 48b node segment (12 hex chars). Users running more than 32k simultaneous
# kernels can subclass.
ident = str(uuid.uuid4()).split('-')[-1]
cf = os.path.join(self.profile_dir.security_dir,
'kernel-%s.json' % ident)
# only keep if it's actually new. Protect against unlikely collision
# in 48b random search space
cf = cf if not os.path.exists(cf) else ''
return cf
def init_kernel_manager(self):
# Don't let Qt or ZMQ swallow KeyboardInterupts.
signal.signal(signal.SIGINT, signal.SIG_DFL)
# Create a KernelManager and start a kernel.
self.kernel_manager = self.kernel_manager_class(
ip=self.ip,
shell_port=self.shell_port,
iopub_port=self.iopub_port,
stdin_port=self.stdin_port,
hb_port=self.hb_port,
connection_file=self.connection_file,
config=self.config,
)
# start the kernel
if not self.existing:
self.kernel_manager.start_kernel(extra_arguments=self.kernel_argv)
elif self.sshserver:
# ssh, write new connection file
self.kernel_manager.write_connection_file()
atexit.register(self.kernel_manager.cleanup_connection_file)
self.kernel_manager.start_channels()
def initialize(self, argv=None):
"""
Classes which mix this class in should call:
IPythonConsoleApp.initialize(self,argv)
"""
self.init_connection_file()
default_secure(self.config)
self.init_ssh()
self.init_kernel_manager()
class ControllerFactory(HubFactory):
"""Configurable for setting up a Hub and Schedulers."""
usethreads = Bool(False, config=True)
# internal
children = List()
mq_class = CStr('zmq.devices.ProcessMonitoredQueue')
def _usethreads_changed(self, name, old, new):
self.mq_class = 'zmq.devices.%sMonitoredQueue' % ('Thread' if new else
'Process')
def __init__(self, **kwargs):
super(ControllerFactory, self).__init__(**kwargs)
self.subconstructors.append(self.construct_schedulers)
def start(self):
super(ControllerFactory, self).start()
child_procs = []
for child in self.children:
child.start()
if isinstance(child, ProcessMonitoredQueue):
child_procs.append(child.launcher)
elif isinstance(child, Process):
child_procs.append(child)
if child_procs:
signal_children(child_procs)
def construct_schedulers(self):
children = self.children
mq = import_item(self.mq_class)
# maybe_inproc = 'inproc://monitor' if self.usethreads else self.monitor_url
# IOPub relay (in a Process)
q = mq(zmq.PUB, zmq.SUB, zmq.PUB, 'N/A', 'iopub')
q.bind_in(self.client_info['iopub'])
q.bind_out(self.engine_info['iopub'])
q.setsockopt_out(zmq.SUBSCRIBE, '')
q.connect_mon(self.monitor_url)
q.daemon = True
children.append(q)
# Multiplexer Queue (in a Process)
q = mq(zmq.XREP, zmq.XREP, zmq.PUB, 'in', 'out')
q.bind_in(self.client_info['mux'])
q.setsockopt_in(zmq.IDENTITY, 'mux')
q.bind_out(self.engine_info['mux'])
q.connect_mon(self.monitor_url)
q.daemon = True
children.append(q)
# Control Queue (in a Process)
q = mq(zmq.XREP, zmq.XREP, zmq.PUB, 'incontrol', 'outcontrol')
q.bind_in(self.client_info['control'])
q.setsockopt_in(zmq.IDENTITY, 'control')
q.bind_out(self.engine_info['control'])
q.connect_mon(self.monitor_url)
q.daemon = True
children.append(q)
# Task Queue (in a Process)
if self.scheme == 'pure':
self.log.warn("task::using pure XREQ Task scheduler")
q = mq(zmq.XREP, zmq.XREQ, zmq.PUB, 'intask', 'outtask')
q.bind_in(self.client_info['task'][1])
q.setsockopt_in(zmq.IDENTITY, 'task')
q.bind_out(self.engine_info['task'])
q.connect_mon(self.monitor_url)
q.daemon = True
children.append(q)
elif self.scheme == 'none':
self.log.warn("task::using no Task scheduler")
else:
self.log.info("task::using Python %s Task scheduler" % self.scheme)
sargs = (self.client_info['task'][1], self.engine_info['task'],
self.monitor_url, self.client_info['notification'])
kwargs = dict(scheme=self.scheme,
logname=self.log.name,
loglevel=self.log.level,
config=dict(self.config))
q = Process(target=launch_scheduler, args=sargs, kwargs=kwargs)
q.daemon = True
children.append(q)
class TaskScheduler(SessionFactory):
"""Python TaskScheduler object.
This is the simplest object that supports msg_id based
DAG dependencies. *Only* task msg_ids are checked, not
msg_ids of jobs submitted via the MUX queue.
"""
hwm = Integer(1,
config=True,
help="""specify the High Water Mark (HWM) for the downstream
socket in the Task scheduler. This is the maximum number
of allowed outstanding tasks on each engine.
The default (1) means that only one task can be outstanding on each
engine. Setting TaskScheduler.hwm=0 means there is no limit, and the
engines continue to be assigned tasks while they are working,
effectively hiding network latency behind computation, but can result
in an imbalance of work when submitting many heterogenous tasks all at
once. Any positive value greater than one is a compromise between the
two.
""")
scheme_name = Enum(
('leastload', 'pure', 'lru', 'plainrandom', 'weighted', 'twobin'),
'leastload',
config=True,
allow_none=False,
help="""select the task scheduler scheme [default: Python LRU]
Options are: 'pure', 'lru', 'plainrandom', 'weighted', 'twobin','leastload'"""
)
def _scheme_name_changed(self, old, new):
self.log.debug("Using scheme %r" % new)
self.scheme = globals()[new]
# input arguments:
scheme = Instance(FunctionType) # function for determining the destination
def _scheme_default(self):
return leastload
client_stream = Instance(zmqstream.ZMQStream) # client-facing stream
engine_stream = Instance(zmqstream.ZMQStream) # engine-facing stream
notifier_stream = Instance(zmqstream.ZMQStream) # hub-facing sub stream
mon_stream = Instance(zmqstream.ZMQStream) # hub-facing pub stream
# internals:
graph = Dict() # dict by msg_id of [ msg_ids that depend on key ]
retries = Dict() # dict by msg_id of retries remaining (non-neg ints)
# waiting = List() # list of msg_ids ready to run, but haven't due to HWM
depending = Dict() # dict by msg_id of Jobs
pending = Dict() # dict by engine_uuid of submitted tasks
completed = Dict() # dict by engine_uuid of completed tasks
failed = Dict() # dict by engine_uuid of failed tasks
destinations = Dict(
) # dict by msg_id of engine_uuids where jobs ran (reverse of completed+failed)
clients = Dict() # dict by msg_id for who submitted the task
targets = List() # list of target IDENTs
loads = List() # list of engine loads
# full = Set() # set of IDENTs that have HWM outstanding tasks
all_completed = Set() # set of all completed tasks
all_failed = Set() # set of all failed tasks
all_done = Set() # set of all finished tasks=union(completed,failed)
all_ids = Set() # set of all submitted task IDs
auditor = Instance('zmq.eventloop.ioloop.PeriodicCallback')
ident = CBytes() # ZMQ identity. This should just be self.session.session
# but ensure Bytes
def _ident_default(self):
return self.session.bsession
def start(self):
self.engine_stream.on_recv(self.dispatch_result, copy=False)
self.client_stream.on_recv(self.dispatch_submission, copy=False)
self._notification_handlers = dict(
registration_notification=self._register_engine,
unregistration_notification=self._unregister_engine)
self.notifier_stream.on_recv(self.dispatch_notification)
self.auditor = ioloop.PeriodicCallback(self.audit_timeouts, 2e3,
self.loop) # 1 Hz
self.auditor.start()
self.log.info("Scheduler started [%s]" % self.scheme_name)
def resume_receiving(self):
"""Resume accepting jobs."""
self.client_stream.on_recv(self.dispatch_submission, copy=False)
def stop_receiving(self):
"""Stop accepting jobs while there are no engines.
Leave them in the ZMQ queue."""
self.client_stream.on_recv(None)
#-----------------------------------------------------------------------
# [Un]Registration Handling
#-----------------------------------------------------------------------
def dispatch_notification(self, msg):
"""dispatch register/unregister events."""
try:
idents, msg = self.session.feed_identities(msg)
except ValueError:
self.log.warn("task::Invalid Message: %r", msg)
return
try:
msg = self.session.unserialize(msg)
except ValueError:
self.log.warn("task::Unauthorized message from: %r" % idents)
return
msg_type = msg['header']['msg_type']
handler = self._notification_handlers.get(msg_type, None)
if handler is None:
self.log.error("Unhandled message type: %r" % msg_type)
else:
try:
handler(asbytes(msg['content']['queue']))
except Exception:
self.log.error("task::Invalid notification msg: %r",
msg,
exc_info=True)
def _register_engine(self, uid):
"""New engine with ident `uid` became available."""
# head of the line:
self.targets.insert(0, uid)
self.loads.insert(0, 0)
# initialize sets
self.completed[uid] = set()
self.failed[uid] = set()
self.pending[uid] = {}
# rescan the graph:
self.update_graph(None)
def _unregister_engine(self, uid):
"""Existing engine with ident `uid` became unavailable."""
if len(self.targets) == 1:
# this was our only engine
pass
# handle any potentially finished tasks:
self.engine_stream.flush()
# don't pop destinations, because they might be used later
# map(self.destinations.pop, self.completed.pop(uid))
# map(self.destinations.pop, self.failed.pop(uid))
# prevent this engine from receiving work
idx = self.targets.index(uid)
self.targets.pop(idx)
self.loads.pop(idx)
# wait 5 seconds before cleaning up pending jobs, since the results might
# still be incoming
if self.pending[uid]:
dc = ioloop.DelayedCallback(
lambda: self.handle_stranded_tasks(uid), 5000, self.loop)
dc.start()
else:
self.completed.pop(uid)
self.failed.pop(uid)
def handle_stranded_tasks(self, engine):
"""Deal with jobs resident in an engine that died."""
lost = self.pending[engine]
for msg_id in lost.keys():
if msg_id not in self.pending[engine]:
# prevent double-handling of messages
continue
raw_msg = lost[msg_id][0]
idents, msg = self.session.feed_identities(raw_msg, copy=False)
parent = self.session.unpack(msg[1].bytes)
idents = [engine, idents[0]]
# build fake error reply
try:
raise error.EngineError(
"Engine %r died while running task %r" % (engine, msg_id))
except:
content = error.wrap_exception()
# build fake header
header = dict(
status='error',
engine=engine,
date=datetime.now(),
)
msg = self.session.msg('apply_reply',
content,
parent=parent,
subheader=header)
raw_reply = map(zmq.Message,
self.session.serialize(msg, ident=idents))
# and dispatch it
self.dispatch_result(raw_reply)
# finally scrub completed/failed lists
self.completed.pop(engine)
self.failed.pop(engine)
#-----------------------------------------------------------------------
# Job Submission
#-----------------------------------------------------------------------
def dispatch_submission(self, raw_msg):
"""Dispatch job submission to appropriate handlers."""
# ensure targets up to date:
self.notifier_stream.flush()
try:
idents, msg = self.session.feed_identities(raw_msg, copy=False)
msg = self.session.unserialize(msg, content=False, copy=False)
except Exception:
self.log.error("task::Invaid task msg: %r" % raw_msg,
exc_info=True)
return
# send to monitor
self.mon_stream.send_multipart([b'intask'] + raw_msg, copy=False)
header = msg['header']
msg_id = header['msg_id']
self.all_ids.add(msg_id)
# get targets as a set of bytes objects
# from a list of unicode objects
targets = header.get('targets', [])
targets = map(asbytes, targets)
targets = set(targets)
retries = header.get('retries', 0)
self.retries[msg_id] = retries
# time dependencies
after = header.get('after', None)
if after:
after = Dependency(after)
if after.all:
if after.success:
after = Dependency(
after.difference(self.all_completed),
success=after.success,
failure=after.failure,
all=after.all,
)
if after.failure:
after = Dependency(
after.difference(self.all_failed),
success=after.success,
failure=after.failure,
all=after.all,
)
if after.check(self.all_completed, self.all_failed):
# recast as empty set, if `after` already met,
# to prevent unnecessary set comparisons
after = MET
else:
after = MET
# location dependencies
follow = Dependency(header.get('follow', []))
# turn timeouts into datetime objects:
timeout = header.get('timeout', None)
if timeout:
# cast to float, because jsonlib returns floats as decimal.Decimal,
# which timedelta does not accept
timeout = datetime.now() + timedelta(0, float(timeout), 0)
job = Job(
msg_id=msg_id,
raw_msg=raw_msg,
idents=idents,
msg=msg,
header=header,
targets=targets,
after=after,
follow=follow,
timeout=timeout,
)
# validate and reduce dependencies:
for dep in after, follow:
if not dep: # empty dependency
continue
# check valid:
if msg_id in dep or dep.difference(self.all_ids):
self.depending[msg_id] = job
return self.fail_unreachable(msg_id, error.InvalidDependency)
# check if unreachable:
if dep.unreachable(self.all_completed, self.all_failed):
self.depending[msg_id] = job
return self.fail_unreachable(msg_id)
if after.check(self.all_completed, self.all_failed):
# time deps already met, try to run
if not self.maybe_run(job):
# can't run yet
if msg_id not in self.all_failed:
# could have failed as unreachable
self.save_unmet(job)
else:
self.save_unmet(job)
def audit_timeouts(self):
"""Audit all waiting tasks for expired timeouts."""
now = datetime.now()
for msg_id in self.depending.keys():
# must recheck, in case one failure cascaded to another:
if msg_id in self.depending:
job = self.depending[msg_id]
if job.timeout and job.timeout < now:
self.fail_unreachable(msg_id, error.TaskTimeout)
def fail_unreachable(self, msg_id, why=error.ImpossibleDependency):
"""a task has become unreachable, send a reply with an ImpossibleDependency
error."""
if msg_id not in self.depending:
self.log.error("msg %r already failed!", msg_id)
return
job = self.depending.pop(msg_id)
for mid in job.dependents:
if mid in self.graph:
self.graph[mid].remove(msg_id)
try:
raise why()
except:
content = error.wrap_exception()
self.all_done.add(msg_id)
self.all_failed.add(msg_id)
msg = self.session.send(self.client_stream,
'apply_reply',
content,
parent=job.header,
ident=job.idents)
self.session.send(self.mon_stream,
msg,
ident=[b'outtask'] + job.idents)
self.update_graph(msg_id, success=False)
def maybe_run(self, job):
"""check location dependencies, and run if they are met."""
msg_id = job.msg_id
self.log.debug("Attempting to assign task %s", msg_id)
if not self.targets:
# no engines, definitely can't run
return False
if job.follow or job.targets or job.blacklist or self.hwm:
# we need a can_run filter
def can_run(idx):
# check hwm
if self.hwm and self.loads[idx] == self.hwm:
return False
target = self.targets[idx]
# check blacklist
if target in job.blacklist:
return False
# check targets
if job.targets and target not in job.targets:
return False
# check follow
return job.follow.check(self.completed[target],
self.failed[target])
indices = filter(can_run, range(len(self.targets)))
if not indices:
# couldn't run
if job.follow.all:
# check follow for impossibility
dests = set()
relevant = set()
if job.follow.success:
relevant = self.all_completed
if job.follow.failure:
relevant = relevant.union(self.all_failed)
for m in job.follow.intersection(relevant):
dests.add(self.destinations[m])
if len(dests) > 1:
self.depending[msg_id] = job
self.fail_unreachable(msg_id)
return False
if job.targets:
# check blacklist+targets for impossibility
job.targets.difference_update(blacklist)
if not job.targets or not job.targets.intersection(
self.targets):
self.depending[msg_id] = job
self.fail_unreachable(msg_id)
return False
return False
else:
indices = None
self.submit_task(job, indices)
return True
def save_unmet(self, job):
"""Save a message for later submission when its dependencies are met."""
msg_id = job.msg_id
self.depending[msg_id] = job
# track the ids in follow or after, but not those already finished
for dep_id in job.after.union(job.follow).difference(self.all_done):
if dep_id not in self.graph:
self.graph[dep_id] = set()
self.graph[dep_id].add(msg_id)
def submit_task(self, job, indices=None):
"""Submit a task to any of a subset of our targets."""
if indices:
loads = [self.loads[i] for i in indices]
else:
loads = self.loads
idx = self.scheme(loads)
if indices:
idx = indices[idx]
target = self.targets[idx]
# print (target, map(str, msg[:3]))
# send job to the engine
self.engine_stream.send(target, flags=zmq.SNDMORE, copy=False)
self.engine_stream.send_multipart(job.raw_msg, copy=False)
# update load
self.add_job(idx)
self.pending[target][job.msg_id] = job
# notify Hub
content = dict(msg_id=job.msg_id, engine_id=target.decode('ascii'))
self.session.send(self.mon_stream,
'task_destination',
content=content,
ident=[b'tracktask', self.ident])
#-----------------------------------------------------------------------
# Result Handling
#-----------------------------------------------------------------------
def dispatch_result(self, raw_msg):
"""dispatch method for result replies"""
try:
idents, msg = self.session.feed_identities(raw_msg, copy=False)
msg = self.session.unserialize(msg, content=False, copy=False)
engine = idents[0]
try:
idx = self.targets.index(engine)
except ValueError:
pass # skip load-update for dead engines
else:
self.finish_job(idx)
except Exception:
self.log.error("task::Invaid result: %r", raw_msg, exc_info=True)
return
header = msg['header']
parent = msg['parent_header']
if header.get('dependencies_met', True):
success = (header['status'] == 'ok')
msg_id = parent['msg_id']
retries = self.retries[msg_id]
if not success and retries > 0:
# failed
self.retries[msg_id] = retries - 1
self.handle_unmet_dependency(idents, parent)
else:
del self.retries[msg_id]
# relay to client and update graph
self.handle_result(idents, parent, raw_msg, success)
# send to Hub monitor
self.mon_stream.send_multipart([b'outtask'] + raw_msg,
copy=False)
else:
self.handle_unmet_dependency(idents, parent)
def handle_result(self, idents, parent, raw_msg, success=True):
"""handle a real task result, either success or failure"""
# first, relay result to client
engine = idents[0]
client = idents[1]
# swap_ids for XREP-XREP mirror
raw_msg[:2] = [client, engine]
# print (map(str, raw_msg[:4]))
self.client_stream.send_multipart(raw_msg, copy=False)
# now, update our data structures
msg_id = parent['msg_id']
self.pending[engine].pop(msg_id)
if success:
self.completed[engine].add(msg_id)
self.all_completed.add(msg_id)
else:
self.failed[engine].add(msg_id)
self.all_failed.add(msg_id)
self.all_done.add(msg_id)
self.destinations[msg_id] = engine
self.update_graph(msg_id, success)
def handle_unmet_dependency(self, idents, parent):
"""handle an unmet dependency"""
engine = idents[0]
msg_id = parent['msg_id']
job = self.pending[engine].pop(msg_id)
job.blacklist.add(engine)
if job.blacklist == job.targets:
self.depending[msg_id] = job
self.fail_unreachable(msg_id)
elif not self.maybe_run(job):
# resubmit failed
if msg_id not in self.all_failed:
# put it back in our dependency tree
self.save_unmet(job)
if self.hwm:
try:
idx = self.targets.index(engine)
except ValueError:
pass # skip load-update for dead engines
else:
if self.loads[idx] == self.hwm - 1:
self.update_graph(None)
def update_graph(self, dep_id=None, success=True):
"""dep_id just finished. Update our dependency
graph and submit any jobs that just became runable.
Called with dep_id=None to update entire graph for hwm, but without finishing
a task.
"""
# print ("\n\n***********")
# pprint (dep_id)
# pprint (self.graph)
# pprint (self.depending)
# pprint (self.all_completed)
# pprint (self.all_failed)
# print ("\n\n***********\n\n")
# update any jobs that depended on the dependency
jobs = self.graph.pop(dep_id, [])
# recheck *all* jobs if
# a) we have HWM and an engine just become no longer full
# or b) dep_id was given as None
if dep_id is None or self.hwm and any(
[load == self.hwm - 1 for load in self.loads]):
jobs = self.depending.keys()
for msg_id in sorted(
jobs, key=lambda msg_id: self.depending[msg_id].timestamp):
job = self.depending[msg_id]
if job.after.unreachable(self.all_completed, self.all_failed)\
or job.follow.unreachable(self.all_completed, self.all_failed):
self.fail_unreachable(msg_id)
elif job.after.check(self.all_completed,
self.all_failed): # time deps met, maybe run
if self.maybe_run(job):
self.depending.pop(msg_id)
for mid in job.dependents:
if mid in self.graph:
self.graph[mid].remove(msg_id)
#----------------------------------------------------------------------
# methods to be overridden by subclasses
#----------------------------------------------------------------------
def add_job(self, idx):
"""Called after self.targets[idx] just got the job with header.
Override with subclasses. The default ordering is simple LRU.
The default loads are the number of outstanding jobs."""
self.loads[idx] += 1
for lis in (self.targets, self.loads):
lis.append(lis.pop(idx))
def finish_job(self, idx):
"""Called after self.targets[idx] just finished a job.
Override with subclasses."""
self.loads[idx] -= 1
class HubAuth(BaseAuth):
"""Jupyter hub authenticator."""
graders = List([],
config=True,
help="List of JupyterHub user names allowed to grade.")
proxy_address = Unicode(
config=True, help="Address of the configurable-http-proxy server.")
def _proxy_address_default(self):
return self._ip
proxy_port = Int(8001,
config=True,
help="Port of the configurable-http-proxy server.")
hub_address = Unicode(config=True, help="Address of the hub server.")
def _hub_address_default(self):
return self._ip
hub_port = Int(8000, config=True, help="Port of the hub server.")
hubapi_address = Unicode(config=True, help="Address of the hubapi server.")
def _hubapi_address_default(self):
return self._ip
hubapi_port = Int(8081, config=True, help="Port of the hubapi server.")
hubapi_cookie = Unicode("jupyter-hub-token",
config=True,
help="Name of the cookie used by JupyterHub")
notebook_url_prefix = Unicode(None,
config=True,
allow_none=True,
help="""
Relative path of the formgrader with respect to the hub's user base
directory. No trailing slash. i.e. "Documents" or "Documents/notebooks". """
)
def _notebook_url_prefix_changed(self, name, old, new):
self.notebook_url_prefix = new.strip('/')
hub_base_url = Unicode(config=True, help="Base URL of the hub server.")
def _hub_base_url_default(self):
return 'http://{}:{}'.format(self.hub_address, self.hub_port)
generate_hubapi_token = Bool(False,
config=True,
help="""Use `jupyterhub token` as a default
for HubAuth.hubapi_token instead of $JPY_API_TOKEN.""")
hub_db = Unicode(config=True,
help="""Path to JupyterHub's database. Only
manditory if `generate_hubapi_token` is True.""")
hubapi_token = Unicode(config=True,
help="""JupyterHub API auth token.
Generated by running `jupyterhub token`. If not explicitly set,
nbgrader will use $JPY_API_TOKEN as the API token.""")
def _hubapi_token_default(self):
if self.generate_hubapi_token:
return check_output(
['jupyterhub', 'token',
'--db={}'.format(self.hub_db)]).decode('utf-8').strip()
else:
return os.environ.get('JPY_API_TOKEN', '')
proxy_token = Unicode(config=True,
help="""JupyterHub configurable proxy
auth token. If not explicitly set, nbgrader will use
$CONFIGPROXY_AUTH_TOKEN as the API token.""")
def _proxy_token_default(self):
return os.environ.get('CONFIGPROXY_AUTH_TOKEN', '')
remap_url = Unicode(config=True,
help="""Suffix appened to
`HubAuth.hub_base_url` to form the full URL to the formgrade server. By
default this is '/hub/{{NbGraderConfig.course_id}}'. Change this if you
plan on running more than one formgrade server behind one JupyterHub
instance.""")
def _remap_url_default(self):
return '/hub/nbgrader/' + self.parent.course_id
def _remap_url_changed(self, name, old, new):
self.remap_url = new.rstrip('/')
def __init__(self, *args, **kwargs):
super(HubAuth, self).__init__(*args, **kwargs)
# Create base URLs for the hub and proxy.
self._hubapi_base_url = 'http://{}:{}'.format(self.hubapi_address,
self.hubapi_port)
self._proxy_base_url = 'http://{}:{}'.format(self.proxy_address,
self.proxy_port)
# Register self as a route of the configurable-http-proxy and then
# update the base_url to point to the new path.
response = self._proxy_request(
'/api/routes' + self.remap_url,
method='POST',
body={'target': 'http://{}:{}'.format(self._ip, self._port)})
if response.status_code != 201:
raise Exception(
'Error while trying to add JupyterHub route. {}: {}'.format(
response.status_code, response.text))
self._base_url = self.hub_base_url + self.remap_url
# Redirect all formgrade request to the correct API method.
self._app.register_blueprint(blueprint,
static_url_path=self.remap_url +
'/static',
url_prefix=self.remap_url,
url_defaults={'name': 'hub'})
def authenticate(self):
"""Authenticate a request.
Returns a boolean or flask redirect."""
# If auth cookie doesn't exist, redirect to the login page with
# next set to redirect back to the this page.
if 'jupyter-hub-token' not in request.cookies:
return redirect(self.hub_base_url + '/hub/login?next=' +
self.remap_url)
cookie = request.cookies[self.hubapi_cookie]
# Check with the Hub to see if the auth cookie is valid.
response = self._hubapi_request('/hub/api/authorizations/cookie/' +
self.hubapi_cookie + '/' + cookie)
if response.status_code == 200:
# Auth information recieved.
data = response.json()
if 'name' in data:
user = data['name']
# Check if the user name is registered as a grader.
if user in self.graders:
self._user = user
return True
else:
self.log.warn(
'Unauthorized user "%s" attempted to access the formgrader.'
% user)
else:
self.log.warn(
'Malformed response from the JupyterHub auth API.')
abort(
500,
"Failed to check authorization, malformed response from Hub auth."
)
elif response.status_code == 403:
self.log.error(
"I don't have permission to verify cookies, my auth token may have expired: [%i] %s",
response.status_code, response.reason)
abort(
500,
"Permission failure checking authorization, I may need to be restarted"
)
elif response.status_code >= 500:
self.log.error("Upstream failure verifying auth token: [%i] %s",
response.status_code, response.reason)
abort(502, "Failed to check authorization (upstream problem)")
elif response.status_code >= 400:
self.log.warn("Failed to check authorization: [%i] %s",
response.status_code, response.reason)
abort(500, "Failed to check authorization")
else:
# Auth invalid, reauthenticate.
return redirect(self.hub_base_url + '/hub/login?next=' +
self.remap_url)
return False
def notebook_server_exists(self):
"""Does the notebook server exist?"""
return True
def get_notebook_url(self, relative_path):
"""Gets the notebook's url."""
if self.notebook_url_prefix is not None:
relative_path = self.notebook_url_prefix + '/' + relative_path
return self.hub_base_url + "/user/{}/notebooks/{}".format(
self._user, relative_path)
def _hubapi_request(self, *args, **kwargs):
return self._request('hubapi', *args, **kwargs)
def _proxy_request(self, *args, **kwargs):
return self._request('proxy', *args, **kwargs)
def _request(self, service, relative_path, method='GET', body=None):
base_url = getattr(self, '_%s_base_url' % service)
token = getattr(self, '%s_token' % service)
data = body
if isinstance(data, (dict, )):
data = json.dumps(data)
return requests.request(method,
base_url + relative_path,
headers={'Authorization': 'token %s' % token},
data=data)
