Merge pull request #2045 from ellisonbg/azurenb

Refactoring notebook managers and adding Azure backed storage.
This commit is contained in:
Brian E. Granger 2012-08-10 17:45:15 -07:00
commit c8e04d991c
5 changed files with 269 additions and 139 deletions

View File

@ -0,0 +1,205 @@
"""A base class notebook manager.
Authors:
* Brian Granger
"""
#-----------------------------------------------------------------------------
# Copyright (C) 2011 The IPython Development Team
#
# Distributed under the terms of the BSD License. The full license is in
# the file COPYING, distributed as part of this software.
#-----------------------------------------------------------------------------
#-----------------------------------------------------------------------------
# Imports
#-----------------------------------------------------------------------------
import os
import uuid
from tornado import web
from IPython.config.configurable import LoggingConfigurable
from IPython.nbformat import current
from IPython.utils.traitlets import List, Dict, Unicode, TraitError
#-----------------------------------------------------------------------------
# Classes
#-----------------------------------------------------------------------------
class BaseNotebookManager(LoggingConfigurable):
# Todo:
# The notebook_dir attribute is used to mean a couple of different things:
# 1. Where the notebooks are stored if FileNotebookManager is used.
# 2. The cwd of the kernel for a project.
# Right now we use this attribute in a number of different places and
# we are going to have to disentagle all of this.
notebook_dir = Unicode(os.getcwdu(), config=True, help="""
The directory to use for notebooks.
""")
def _notebook_dir_changed(self, name, old, new):
"""do a bit of validation of the notebook dir"""
if os.path.exists(new) and not os.path.isdir(new):
raise TraitError("notebook dir %r is not a directory" % new)
if not os.path.exists(new):
self.log.info("Creating notebook dir %s", new)
try:
os.mkdir(new)
except:
raise TraitError("Couldn't create notebook dir %r" % new)
allowed_formats = List([u'json',u'py'])
# Map notebook_ids to notebook names
mapping = Dict()
def load_notebook_names(self):
"""Load the notebook names into memory.
This should be called once immediately after the notebook manager
is created to load the existing notebooks into the mapping in
memory.
"""
self.list_notebooks()
def list_notebooks(self):
"""List all notebooks.
This returns a list of dicts, each of the form::
dict(notebook_id=notebook,name=name)
This list of dicts should be sorted by name::
data = sorted(data, key=lambda item: item['name'])
"""
raise NotImplementedError('must be implemented in a subclass')
def new_notebook_id(self, name):
"""Generate a new notebook_id for a name and store its mapping."""
# TODO: the following will give stable urls for notebooks, but unless
# the notebooks are immediately redirected to their new urls when their
# filemname changes, nasty inconsistencies result. So for now it's
# disabled and instead we use a random uuid4() call. But we leave the
# logic here so that we can later reactivate it, whhen the necessary
# url redirection code is written.
#notebook_id = unicode(uuid.uuid5(uuid.NAMESPACE_URL,
# 'file://'+self.get_path_by_name(name).encode('utf-8')))
notebook_id = unicode(uuid.uuid4())
self.mapping[notebook_id] = name
return notebook_id
def delete_notebook_id(self, notebook_id):
"""Delete a notebook's id in the mapping.
This doesn't delete the actual notebook, only its entry in the mapping.
"""
del self.mapping[notebook_id]
def notebook_exists(self, notebook_id):
"""Does a notebook exist?"""
return notebook_id in self.mapping
def get_notebook(self, notebook_id, format=u'json'):
"""Get the representation of a notebook in format by notebook_id."""
format = unicode(format)
if format not in self.allowed_formats:
raise web.HTTPError(415, u'Invalid notebook format: %s' % format)
last_modified, nb = self.read_notebook_object(notebook_id)
kwargs = {}
if format == 'json':
# don't split lines for sending over the wire, because it
# should match the Python in-memory format.
kwargs['split_lines'] = False
data = current.writes(nb, format, **kwargs)
name = nb.get('name','notebook')
return last_modified, name, data
def read_notebook_object(self, notebook_id):
"""Get the object representation of a notebook by notebook_id."""
raise NotImplementedError('must be implemented in a subclass')
def save_new_notebook(self, data, name=None, format=u'json'):
"""Save a new notebook and return its notebook_id.
If a name is passed in, it overrides any values in the notebook data
and the value in the data is updated to use that value.
"""
if format not in self.allowed_formats:
raise web.HTTPError(415, u'Invalid notebook format: %s' % format)
try:
nb = current.reads(data.decode('utf-8'), format)
except:
raise web.HTTPError(400, u'Invalid JSON data')
if name is None:
try:
name = nb.metadata.name
except AttributeError:
raise web.HTTPError(400, u'Missing notebook name')
nb.metadata.name = name
notebook_id = self.write_notebook_object(nb)
return notebook_id
def save_notebook(self, notebook_id, data, name=None, format=u'json'):
"""Save an existing notebook by notebook_id."""
if format not in self.allowed_formats:
raise web.HTTPError(415, u'Invalid notebook format: %s' % format)
try:
nb = current.reads(data.decode('utf-8'), format)
except:
raise web.HTTPError(400, u'Invalid JSON data')
if name is not None:
nb.metadata.name = name
self.write_notebook_object(nb, notebook_id)
def write_notebook_object(self, nb, notebook_id=None):
"""Write a notebook object and return its notebook_id.
If notebook_id is None, this method should create a new notebook_id.
If notebook_id is not None, this method should check to make sure it
exists and is valid.
"""
raise NotImplementedError('must be implemented in a subclass')
def delete_notebook(self, notebook_id):
"""Delete notebook by notebook_id."""
raise NotImplementedError('must be implemented in a subclass')
def increment_filename(self, name):
"""Increment a filename to make it unique.
This exists for notebook stores that must have unique names. When a notebook
is created or copied this method constructs a unique filename, typically
by appending an integer to the name.
"""
return name
def new_notebook(self):
"""Create a new notebook and return its notebook_id."""
name = self.increment_filename('Untitled')
metadata = current.new_metadata(name=name)
nb = current.new_notebook(metadata=metadata)
notebook_id = self.write_notebook_object(nb)
return notebook_id
def copy_notebook(self, notebook_id):
"""Copy an existing notebook and return its notebook_id."""
last_mod, nb = self.read_notebook_object(notebook_id)
name = nb.metadata.name + '-Copy'
name = self.increment_filename(name)
nb.metadata.name = name
notebook_id = self.write_notebook_object(nb)
return notebook_id
def log_info(self):
self.log.info("Serving notebooks")

View File

@ -6,7 +6,7 @@ Authors:
"""
#-----------------------------------------------------------------------------
# Copyright (C) 2008-2011 The IPython Development Team
# Copyright (C) 2011 The IPython Development Team
#
# Distributed under the terms of the BSD License. The full license is in
# the file COPYING, distributed as part of this software.
@ -19,34 +19,19 @@ Authors:
import datetime
import io
import os
import uuid
import glob
from tornado import web
from IPython.config.configurable import LoggingConfigurable
from .basenbmanager import BaseNotebookManager
from IPython.nbformat import current
from IPython.utils.traitlets import Unicode, List, Dict, Bool, TraitError
from IPython.utils.traitlets import Unicode, Dict, Bool, TraitError
#-----------------------------------------------------------------------------
# Classes
#-----------------------------------------------------------------------------
class NotebookManager(LoggingConfigurable):
notebook_dir = Unicode(os.getcwdu(), config=True, help="""
The directory to use for notebooks.
""")
def _notebook_dir_changed(self, name, old, new):
"""do a bit of validation of the notebook dir"""
if os.path.exists(new) and not os.path.isdir(new):
raise TraitError("notebook dir %r is not a directory" % new)
if not os.path.exists(new):
self.log.info("Creating notebook dir %s", new)
try:
os.mkdir(new)
except:
raise TraitError("Couldn't create notebook dir %r" % new)
class FileNotebookManager(BaseNotebookManager):
save_script = Bool(False, config=True,
help="""Automatically create a Python script when saving the notebook.
@ -59,24 +44,21 @@ class NotebookManager(LoggingConfigurable):
)
filename_ext = Unicode(u'.ipynb')
allowed_formats = List([u'json',u'py'])
# Map notebook_ids to notebook names
mapping = Dict()
# Map notebook names to notebook_ids
rev_mapping = Dict()
def list_notebooks(self):
"""List all notebooks in the notebook dir.
This returns a list of dicts of the form::
dict(notebook_id=notebook,name=name)
"""
def get_notebook_names(self):
"""List all notebook names in the notebook dir."""
names = glob.glob(os.path.join(self.notebook_dir,
'*' + self.filename_ext))
names = [os.path.splitext(os.path.basename(name))[0]
for name in names]
return names
def list_notebooks(self):
"""List all notebooks in the notebook dir."""
names = self.get_notebook_names()
data = []
for name in names:
@ -90,30 +72,20 @@ class NotebookManager(LoggingConfigurable):
def new_notebook_id(self, name):
"""Generate a new notebook_id for a name and store its mappings."""
# TODO: the following will give stable urls for notebooks, but unless
# the notebooks are immediately redirected to their new urls when their
# filemname changes, nasty inconsistencies result. So for now it's
# disabled and instead we use a random uuid4() call. But we leave the
# logic here so that we can later reactivate it, whhen the necessary
# url redirection code is written.
#notebook_id = unicode(uuid.uuid5(uuid.NAMESPACE_URL,
# 'file://'+self.get_path_by_name(name).encode('utf-8')))
notebook_id = unicode(uuid.uuid4())
self.mapping[notebook_id] = name
notebook_id = super(FileNotebookManager, self).new_notebook_id(name)
self.rev_mapping[name] = notebook_id
return notebook_id
def delete_notebook_id(self, notebook_id):
"""Delete a notebook's id only. This doesn't delete the actual notebook."""
"""Delete a notebook's id in the mapping."""
name = self.mapping[notebook_id]
del self.mapping[notebook_id]
super(FileNotebookManager, self).delete_notebook_id(notebook_id)
del self.rev_mapping[name]
def notebook_exists(self, notebook_id):
"""Does a notebook exist?"""
if notebook_id not in self.mapping:
exists = super(FileNotebookManager, self).notebook_exists(notebook_id)
if not exists:
return False
path = self.get_path_by_name(self.mapping[notebook_id])
return os.path.isfile(path)
@ -132,22 +104,7 @@ class NotebookManager(LoggingConfigurable):
path = os.path.join(self.notebook_dir, filename)
return path
def get_notebook(self, notebook_id, format=u'json'):
"""Get the representation of a notebook in format by notebook_id."""
format = unicode(format)
if format not in self.allowed_formats:
raise web.HTTPError(415, u'Invalid notebook format: %s' % format)
last_modified, nb = self.get_notebook_object(notebook_id)
kwargs = {}
if format == 'json':
# don't split lines for sending over the wire, because it
# should match the Python in-memory format.
kwargs['split_lines'] = False
data = current.writes(nb, format, **kwargs)
name = nb.get('name','notebook')
return last_modified, name, data
def get_notebook_object(self, notebook_id):
def read_notebook_object(self, notebook_id):
"""Get the NotebookNode representation of a notebook by notebook_id."""
path = self.find_path(notebook_id)
if not os.path.isfile(path):
@ -165,60 +122,27 @@ class NotebookManager(LoggingConfigurable):
nb.metadata.name = os.path.splitext(os.path.basename(path))[0]
return last_modified, nb
def save_new_notebook(self, data, name=None, format=u'json'):
"""Save a new notebook and return its notebook_id.
If a name is passed in, it overrides any values in the notebook data
and the value in the data is updated to use that value.
"""
if format not in self.allowed_formats:
raise web.HTTPError(415, u'Invalid notebook format: %s' % format)
try:
nb = current.reads(data.decode('utf-8'), format)
except:
raise web.HTTPError(400, u'Invalid JSON data')
if name is None:
try:
name = nb.metadata.name
except AttributeError:
raise web.HTTPError(400, u'Missing notebook name')
nb.metadata.name = name
notebook_id = self.new_notebook_id(name)
self.save_notebook_object(notebook_id, nb)
return notebook_id
def save_notebook(self, notebook_id, data, name=None, format=u'json'):
"""Save an existing notebook by notebook_id."""
if format not in self.allowed_formats:
raise web.HTTPError(415, u'Invalid notebook format: %s' % format)
try:
nb = current.reads(data.decode('utf-8'), format)
except:
raise web.HTTPError(400, u'Invalid JSON data')
if name is not None:
nb.metadata.name = name
self.save_notebook_object(notebook_id, nb)
def save_notebook_object(self, notebook_id, nb):
def write_notebook_object(self, nb, notebook_id=None):
"""Save an existing notebook object by notebook_id."""
if notebook_id not in self.mapping:
raise web.HTTPError(404, u'Notebook does not exist: %s' % notebook_id)
old_name = self.mapping[notebook_id]
try:
new_name = nb.metadata.name
except AttributeError:
raise web.HTTPError(400, u'Missing notebook name')
if notebook_id is None:
notebook_id = self.new_notebook_id(new_name)
if notebook_id not in self.mapping:
raise web.HTTPError(404, u'Notebook does not exist: %s' % notebook_id)
old_name = self.mapping[notebook_id]
path = self.get_path_by_name(new_name)
try:
with open(path,'w') as f:
current.write(nb, f, u'json')
except Exception as e:
raise web.HTTPError(400, u'Unexpected error while saving notebook: %s' % e)
# save .py script as well
if self.save_script:
pypath = os.path.splitext(path)[0] + '.py'
@ -228,6 +152,7 @@ class NotebookManager(LoggingConfigurable):
except Exception as e:
raise web.HTTPError(400, u'Unexpected error while saving notebook as script: %s' % e)
# remove old files if the name changed
if old_name != new_name:
old_path = self.get_path_by_name(old_name)
if os.path.isfile(old_path):
@ -239,6 +164,8 @@ class NotebookManager(LoggingConfigurable):
self.mapping[notebook_id] = new_name
self.rev_mapping[new_name] = notebook_id
del self.rev_mapping[old_name]
return notebook_id
def delete_notebook(self, notebook_id):
"""Delete notebook by notebook_id."""
@ -263,24 +190,7 @@ class NotebookManager(LoggingConfigurable):
break
else:
i = i+1
return path, name
return name
def new_notebook(self):
"""Create a new notebook and return its notebook_id."""
path, name = self.increment_filename('Untitled')
notebook_id = self.new_notebook_id(name)
metadata = current.new_metadata(name=name)
nb = current.new_notebook(metadata=metadata)
with open(path,'w') as f:
current.write(nb, f, u'json')
return notebook_id
def copy_notebook(self, notebook_id):
"""Copy an existing notebook and return its notebook_id."""
last_mod, nb = self.get_notebook_object(notebook_id)
name = nb.metadata.name + '-Copy'
path, name = self.increment_filename(name)
nb.metadata.name = name
notebook_id = self.new_notebook_id(name)
self.save_notebook_object(notebook_id, nb)
return notebook_id
def log_info(self):
self.log.info("Serving notebooks from local directory: %s", self.notebook_dir)

View File

@ -51,7 +51,8 @@ from .handlers import (LoginHandler, LogoutHandler,
MainClusterHandler, ClusterProfileHandler, ClusterActionHandler,
FileFindHandler,
)
from .notebookmanager import NotebookManager
from .basenbmanager import BaseNotebookManager
from .filenbmanager import FileNotebookManager
from .clustermanager import ClusterManager
from IPython.config.application import catch_config_error, boolean_flag
@ -66,7 +67,11 @@ from IPython.zmq.ipkernel import (
aliases as ipkernel_aliases,
IPKernelApp
)
from IPython.utils.traitlets import Dict, Unicode, Integer, List, Enum, Bool
from IPython.utils.importstring import import_item
from IPython.utils.traitlets import (
Dict, Unicode, Integer, List, Enum, Bool,
DottedObjectName
)
from IPython.utils import py3compat
from IPython.utils.path import filefind
@ -215,7 +220,7 @@ flags['read-only'] = (
)
# Add notebook manager flags
flags.update(boolean_flag('script', 'NotebookManager.save_script',
flags.update(boolean_flag('script', 'FileNotebookManager.save_script',
'Auto-save a .py script everytime the .ipynb notebook is saved',
'Do not auto-save .py scripts for every notebook'))
@ -232,7 +237,7 @@ aliases.update({
'port-retries': 'NotebookApp.port_retries',
'keyfile': 'NotebookApp.keyfile',
'certfile': 'NotebookApp.certfile',
'notebook-dir': 'NotebookManager.notebook_dir',
'notebook-dir': 'BaseNotebookManager.notebook_dir',
'browser': 'NotebookApp.browser',
})
@ -260,7 +265,8 @@ class NotebookApp(BaseIPythonApplication):
"""
examples = _examples
classes = IPythonConsoleApp.classes + [MappingKernelManager, NotebookManager]
classes = IPythonConsoleApp.classes + [MappingKernelManager, BaseNotebookManager,
FileNotebookManager]
flags = Dict(flags)
aliases = Dict(aliases)
@ -404,6 +410,10 @@ class NotebookApp(BaseIPythonApplication):
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:
@ -421,7 +431,7 @@ class NotebookApp(BaseIPythonApplication):
else:
self.file_to_run = f
nbdir = os.path.dirname(f)
self.config.NotebookManager.notebook_dir = nbdir
self.config.BaseNotebookManager.notebook_dir = nbdir
def init_configurables(self):
# force Session default to be secure
@ -430,9 +440,10 @@ class NotebookApp(BaseIPythonApplication):
config=self.config, log=self.log, kernel_argv=self.kernel_argv,
connection_dir = self.profile_dir.security_dir,
)
self.notebook_manager = NotebookManager(config=self.config, log=self.log)
self.log.info("Serving notebooks from %s", self.notebook_manager.notebook_dir)
self.notebook_manager.list_notebooks()
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()

View File

@ -7,28 +7,28 @@ from tempfile import NamedTemporaryFile
from IPython.utils.tempdir import TemporaryDirectory
from IPython.utils.traitlets import TraitError
from IPython.frontend.html.notebook.notebookmanager import NotebookManager
from IPython.frontend.html.notebook.filenbmanager import FileNotebookManager
class TestNotebookManager(TestCase):
def test_nb_dir(self):
with TemporaryDirectory() as td:
km = NotebookManager(notebook_dir=td)
self.assertEqual(km.notebook_dir, td)
km = FileNotebookManager(notebook_dir=td)
self.assertEquals(km.notebook_dir, td)
def test_create_nb_dir(self):
with TemporaryDirectory() as td:
nbdir = os.path.join(td, 'notebooks')
km = NotebookManager(notebook_dir=nbdir)
self.assertEqual(km.notebook_dir, nbdir)
km = FileNotebookManager(notebook_dir=nbdir)
self.assertEquals(km.notebook_dir, nbdir)
def test_missing_nb_dir(self):
with TemporaryDirectory() as td:
nbdir = os.path.join(td, 'notebook', 'dir', 'is', 'missing')
self.assertRaises(TraitError, NotebookManager, notebook_dir=nbdir)
self.assertRaises(TraitError, FileNotebookManager, notebook_dir=nbdir)
def test_invalid_nb_dir(self):
with NamedTemporaryFile() as tf:
self.assertRaises(TraitError, NotebookManager, notebook_dir=tf.name)
self.assertRaises(TraitError, FileNotebookManager, notebook_dir=tf.name)

View File

@ -164,6 +164,7 @@ have['oct2py'] = test_for('oct2py')
have['tornado'] = test_for('tornado.version_info', (2,1,0), callback=None)
have['wx'] = test_for('wx')
have['wx.aui'] = test_for('wx.aui')
have['azure'] = test_for('azure')
if os.name == 'nt':
min_zmq = (2,1,7)
@ -303,6 +304,9 @@ def make_exclude():
exclusions.append(ipjoin('extensions', 'rmagic'))
exclusions.append(ipjoin('extensions', 'tests', 'test_rmagic'))
if not have['azure']:
exclusions.append(ipjoin('frontend', 'html', 'notebook', 'azurenbmanager'))
# This is needed for the reg-exp to match on win32 in the ipdoctest plugin.
if sys.platform == 'win32':
exclusions = [s.replace('\\','\\\\') for s in exclusions]