##// END OF EJS Templates
explicitly load python mode before IPython mode...
explicitly load python mode before IPython mode ensures IPython mode gets defined at startup, which can sometimes fail because the Python mode was loaded dynamically. closes #3636

File last commit:

r6208:5902725b
r11988:2e7a6314
Show More
nbjson.py
72 lines | 2.0 KiB | text/x-python | PythonLexer
"""Read and write notebooks in JSON format.
Authors:
* Brian Granger
"""
#-----------------------------------------------------------------------------
# Copyright (C) 2008-2011 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 copy
import json
from .nbbase import from_dict
from .rwbase import (
NotebookReader, NotebookWriter, restore_bytes, rejoin_lines, split_lines
)
from IPython.utils import py3compat
#-----------------------------------------------------------------------------
# Code
#-----------------------------------------------------------------------------
class BytesEncoder(json.JSONEncoder):
"""A JSON encoder that accepts b64 (and other *ascii*) bytestrings."""
def default(self, obj):
if isinstance(obj, bytes):
return obj.decode('ascii')
return json.JSONEncoder.default(self, obj)
class JSONReader(NotebookReader):
def reads(self, s, **kwargs):
nb = json.loads(s, **kwargs)
nb = self.to_notebook(nb, **kwargs)
return nb
def to_notebook(self, d, **kwargs):
return restore_bytes(rejoin_lines(from_dict(d)))
class JSONWriter(NotebookWriter):
def writes(self, nb, **kwargs):
kwargs['cls'] = BytesEncoder
kwargs['indent'] = 1
kwargs['sort_keys'] = True
kwargs['separators'] = (',',': ')
if kwargs.pop('split_lines', True):
nb = split_lines(copy.deepcopy(nb))
return py3compat.str_to_unicode(json.dumps(nb, **kwargs), 'utf-8')
_reader = JSONReader()
_writer = JSONWriter()
reads = _reader.reads
read = _reader.read
to_notebook = _reader.to_notebook
write = _writer.write
writes = _writer.writes